Comparative Economics of Rice Production in Pakistan: A Price risk Analysis

1 European Journal of Economics, Finance And Administrative Sciences - Issue 6 (2006)

EUROPEAN JOURNAL OF ECONOMICS, FINANCE AND ADMINISTRATIVE SCIENCES

ISSN: 1450-2275

Issue 6 October, 2006


EUROPEAN JOURNAL OF ECONOMICS, FINANCE AND ADMINISTRATIVE SCIENCES http://www.eurojournals.com/EJEFAS.htm Editor-In-Chief Adrian M. Steinberg, Wissenschaftlicher Forscher Editorial Advisory Board Bansi Sawhney, University of Baltimore Jwyang Jiawen Yang, The George Washington University Zhihong Shi, State University of New York Zeljko Bogetic, The World Bank Jatin Pancholi, Middlesex University Christos Giannikos, Columbia University Hector Lozada, Seton Hall University Jan Dutta, Rutgers University Chiaku Chukwuogor-Ndu, Eastern Connecticut State University Neil Reid, University of Toledo John Mylonakis, Hellenic Open University (Tutor) M. Femi Ayadi, University of Houston-Clear Lake Emmanuel Anoruo, Coppin State University H. Young Baek, Nova Southeastern University Jean-Luc Grosso, University of South Carolina Sumter Richard Omotoye, Virginia State University Mahdi Hadi, Kuwait University Jean-Luc Grosso, University of South Carolina Ali Argun Karacabey, Ankara University Felix Ayadi, Texas Southern University Bansi Sawhney, University of Baltimore David Wang, Hsuan Chuang University Cornelis A. Los, Kazakh-British Technical University Leo V. Ryan, DePaul University Richard J. Hunter, Seton Hall University Said Elnashaie, Auburn University Panayiotis Tahinakis, University of Macedonia Mukhopadhyay Bappaditya, Management Development Institute M. Carmen Guisan, University of Santiago de Compostela Subrata Chowdhury, University of Rhode Island Teresa Smith, University of South Carolina Wassim Shahin, Lebanese American University Mete Feridun, Cyprus International University Teresa Smith, University of South Carolina Sumter Ranjit Biswas, Philadelphia University Katerina Lyroudi, University of Macedonia Maria Elena Garcia-Ruiz, University of Cantabria Zulkarnain Muhamad Sori, University Putra Malaysia Indexing / Abstracting European Journal of Social Sciences is indexed in Scopus, Elsevier Bibliographic Databases, EMBASE, Ulrich, DOAJ, Cabell, Compendex, GEOBASE, and Mosby.


Aims and Scope The European Journal of Scientific Research is a quarterly, peer-reviewed international research journal that addresses both applied and theoretical issues. The scope of the journal encompasses research articles, original research reports, reviews, short communications and scientific commentaries in the fields of economics, finance and administrative sciences. Editorial Policies 1) The journal realizes the meaning of fast publication to researchers, particularly to those working in competitive & dynamic fields. Hence, it offers an exceptionally fast publication schedule including prompt peer-review by the experts in the field and immediate publication upon acceptance. It is the major editorial policy to review the submitted articles as fast as possible and promptly include them in the forthcoming issues should they pass the evaluation process. 2) All research and reviews published in the journal have been fully peer-reviewed by two, and in some cases, three internal or external reviewers. Unless they are out of scope for the journal, or are of an unacceptably low standard of presentation, submitted articles will be sent to peer reviewers. They will generally be reviewed by two experts with the aim of reaching a first decision within a two-month period. Suggested reviewers will be considered alongside potential reviewers identified by their publication record or recommended by Editorial Board members. Reviewers are asked whether the manuscript is scientifically sound and coherent, how interesting it is and whether the quality of the writing is acceptable. Where possible, the final decision is made on the basis that the peer reviewers are in accordance with one another, or that at least there is no strong dissenting view. 3) In cases where there is strong disagreement either among peer reviewers or between the authors and peer reviewers, advice is sought from a member of the journal's Editorial Board. The journal allows a maximum of three revisions of any manuscripts. The ultimate responsibility for any decision lies with the Editor-in-Chief. Reviewers are also asked to indicate which articles they consider to be especially interesting or significant. These articles may be given greater prominence and greater external publicity. 4) Any manuscript submitted to the journals must not already have been published in another journal or be under consideration by any other journal. Manuscripts that are derived from papers presented at conferences can be submitted even if they have been published as part of the conference proceedings in a peer reviewed journal. Authors are required to ensure that no material submitted as part of a manuscript infringes existing copyrights, or the rights of a third party. Contributing authors retain copyright to their work. 5) The journal makes all published original research immediately accessible through www.EuroJournals.com without subscription charges or registration barriers. Through its open access policy, the journal is committed permanently to maintaining this policy. This process is streamlined thanks to a user-friendly, web-based system for submission and for referees to view manuscripts and return their reviews. The journal does not have page charges, color figure charges or submission fees. However, there is an article-processing and publication fee payable only if the article is accepted for publication. Submissions All papers are subjected to a blind peer review process. Manuscripts are invited from academicians, research students, and scientists for publication consideration. The journal welcomes submissions in all areas related to science. Each manuscript must include a 200 word abstract. Authors should list their contact information on a separate paper. Electronic submissions are acceptable. The journal publishes both applied and conceptual research. Articles for consideration are to be directed to the editor through [email protected]. In the subject line of your e-mail please write "EJEFAS submission"


Articles are accepted in MS-Word or pdf formats Contributors should adhere to the format of the journal. All correspondence should be directed to the editor There is no submission fee Publication fee for each accepted article is $150 USD

European Journal of Economics, Finance and Administrative Sciences is published in the United States of America at Lulu Press, Inc (Morrisville, North Carolina) by EuroJournals, Inc.


European Journal of Economics, Finance and Administrative Sciences Issue 6 October, 2006 Contents The Determinants of Strategic Alliance Sustainment Following an Alliance Change 6-13 Craig R. Erwin Trading Systems, Foreign Direct Investment and Economic Growth: Evidence from Asean Countries 14-28 Kevin Odulukwe. Onwuka and Ahmad Zubaidi Baharumshah Day-of-the-Week Effect, Volatility, and Linkages to Macro-Economic Indicators: Evidence from North American Financial Markets, 1997-2004 29-37 Chiaku Chukwuogor-Ndu Governance, Taxation and Fiscal Policy in Nigeria 38-51 Olu Okotoni The Economic and Business Driving Forces of the Russian Market Environment and Main Decision Makers Geo-Economic World Perspectives 52-60 Evangelos Karafotakis and John Mylonakis Proposed Models for Integrating Marketing and R&D Departments in High Tech Markets: Literature Review Approach 61-76 Thomas A. Fotiadis Is There A Link Between Strategic Human Resource Management Practices and Organizational Outcomes? A Study of The Northern Cyprus Manufacturing Industry 77-88 Serife Zihni Eyüpoglu The Effect of Relative Earnings Performance on Firms' Accrual Decisions: Evidence from France Ramzi Benkraiem 89-102 An Application of Casetti's Expansion Method to a Variable Coefficient Regression Model of Electricity Demand: Simulation Results of Alternative Estimation Methods 103-113 Roger L. Burford and Susan M. L. Zee Comparative Economics of Rice Production in Pakistan: A Price Risk Analysis 114-123 Mohammad F. Hussain, Sofia Anwar and Zakir Hussain The Predictability of Amman Stock Exchange Performance: A Univariate Autoregressive Integrated Moving Average (ARIMA) Model 124-139 Mohammad Al-Shiab Managerial Leadership Values across Cultures 140-149 Osarumwense Iguisi

European Journal of Economics, Finance and Administrative Sciences ISSN 1450-2887 Issue 6 (2006) © EuroJournals, Inc. 2006 http://www.eurojournalsn.com

The Determinants of Strategic Alliance Sustainment Following an Alliance Change

Craig R. Erwin

Eastern Connecticut State University 83 Windham Street, Willimantic, CT 06226 Phone: 860-465-4632 Fax: 860-465-4469

E-mail: [email protected]

Abstract Strategic alliances tend to have difficulty weathering changes and frequently dissolve prematurely. This paper examines the factors that best predict alliance sustainment or dissolution following an alliance change. All Arizona high-tech, manufacturers with between ten and 100 employees were surveyed to test seven hypotheses derived from a theoretical model using Student’s t Distribution. Regression analysis was used to test the model after determining that multicollinearity was not excessive and after applying Bonferroni corrections as needed to reduce the likelihood of Type I error. Regression analysis identified factors with the greatest effect on sustainment as the extent to which a change is planned jointly and the impact of the change. Keywords: Strategic alliance, alliance change, alliance sustainment, alliance continuity, alliance dissolution, joint venture, strategic partnership

I. Introduction Strategic alliances are susceptible to changes and frequently dissolve prematurely (Das and Teng, 2000). Previous research has focused primarily on identifying determinants of changes in alliances involving large, international firms. This study addresses a gap in alliance literature (McGee, 1994); we know little about alliances involving small firms, including alliance evolution and the outcomes of changes. In the study it is found that dyadic alliances involving small high-tech firms are more likely to be sustained following a change that is jointly planned by alliance partners or that is small in magnitude. This has implications for the way managers structure, monitor and maintain alliances. II. Literature Review Although strategic alliances may benefit partners in numerous ways, outcomes are not always positive, and costs may be quite high (Pearce, 1997). A strategic alliance is a voluntary, cooperative, inter-firm arrangement that enables organizations to collaborate to achieve mutual goals over time (Das and Teng, 2000). At worst, a failed alliance may result in the failure of one or both alliance partners. In less severe cases, an alliance may hinder a firm's performance or restrict it from pursuing lucrative opportunities.

Frequently short-lived and often failing to meet their objectives, alliances are plagued by instability; they frequently undergo major unplanned changes or premature dissolution (Das and Teng,


2000). Although some changes are necessary if relationships are to grow, such changes can damage alliances and firms, especially when firms are highly interdependent.

Scholars have often acknowledged the importance of decision-making in alliances (Van de Ven and Walker, 1984). Pearce (1997) found that management decision-making difficulties due to factionalism and organizational differences may lead to alliance dissolution. In contrast, joint decision-making has been identified as a determinant of alliance success and sustainment (Meyer, Alvarez, and Blasick, 1994). Saxton (1997) discovered a relationship between shared decision-making and alliance performance while Anderson and Narus (1990) found that decision-making affects the success of an alliance. Finally, in examining entrepreneurial alliances Larson (1992) observed that successful alliances are characterized by joint planning. Thus, although effective joint decision-making may be uncommon and difficult to accomplish, it appears to positively affect alliance continuity as well as success.

Hypothesis 1. The greater the inter-firm interaction prior to an alliance change the greater the likelihood that the alliance will be sustained following a change. Hypothesis 2. The greater the extent to which partners make alliance decisions jointly prior to an alliance change, the greater the likelihood that the alliance will be sustained following a change. Hypothesis 3. The greater the extent to which an alliance change is planned, the greater the likelihood that the alliance will be sustained following a change. Hypothesis 4. The greater the extent to which an alliance change is jointly planned, the greater the likelihood that the alliance will be sustained following a change.

The literature contains both theoretical (Pfeffer and Salancik, 1978) and empirical support (Inkpen and Beamish, 1997) for the importance of dependence in sustaining alliances.

Hypothesis 5. The greater the extent to which alliance partners are interdependent, the greater the likelihood that the alliance will be sustained following an alliance change. Scholars have also frequently identified trust as a factor in alliance sustainment and success

(Young-Ybarra and Wiersema, 1999). Some researchers have found that trust is a particularly important factor in the success of small and/or high-tech alliances (Meyer, Alvarez, and Blasick, 1994).

Hypothesis 6. The greater the level of inter-firm trust, the greater the likelihood that the alliance will be sustained following an alliance change. Alliance instability is defined as a major change of an alliance (Inkpen and Beamish, 1997) because

it is assumed that major changes (i.e., mergers, restructurings) are frequently deleterious to alliances. As such, a minor change would be expected to be less harmful to an alliance than a major change and more likely to result in sustainment.

Hypothesis 7. The less the impact of an alliance change on partners, the greater the likelihood that the alliance will be sustained following the change.

III. Data and Methodology This study attempts to test a theoretical model on alliance change (see Figure 1). The model predicts whether a strategic alliance is sustained or dissolved following an alliance change. It contains predictors that are either characteristics of the alliance or of the change.

The study consisted of a survey that made use of a short mailed questionnaire. The questionnaire asked respondents to select any alliance change other than formation or termination. The study population, selected from the Arizona Manufacturers Directory (2000), consisted of all public and private Arizona businesses with between ten and one hundred employees that manufacture high-tech products. The study questionnaire was mailed to the top executive of each of the 196 firms in the study population. 73 respondents returned the questionnaire indicating that it did not apply to their firm because they had no alliance changes to report on. This left 123 firms in the population with alliance changes. 52 respondents returned a completed questionnaire. The questionnaire response rate of 42.3%


was determined by dividing the number of completed questionnaires by the number of firms with alliance changes. Since seven of the 52 completed questionnaires were unusable for various reasons, data from the remaining 45 questionnaires were analyzed.

Figure 1: Model 1 – A Model of Change in Strategic Alliances Error!

PRIOR TO AT THE TIME OF OUTCOMES CHANGE CHANGE OF CHANGE

The items that made up the multi-item measures were factor analyzed and a four factor solution

resulted. The factors were trust, joint decision making, dependence of partner A on partner B, and dependence of partner B on partner A. Next, reliability tests were conducted on each factor and items for each factor except trust were deleted to increase reliabilities.

To measure joint decision making, two scales were combined; a scale used by Heide and Miner (1992) on shared problem solving and a scale on frequency of interaction used by Parkhe (1993). Cronbach’s alpha for the resulting measure was .61. To measure trust, Young-Ybarra and Wiersema's (1999) multi-item trust measure was slightly modified (α =.79). To measure interdependence, two multi-item measures used by Ganesan (1994) were modified. Cronbach’s alpha was .83 for Ganesan’s first measure and .66 for Ganesan’s other measure. The dependent variable was whether an alliance was sustained or terminated. Control variables included partner nationality, alliance age and lack of agreement to terminate the alliance at a future date (no planned dissolution).

Questionnaire data were coded and converted to an electronic format. Missing data and respondent errors were resolved by contacting respondents when possible. Data entry and coding errors were also identified and fixed. IV. Empirical Results Respondents reported on a wide variety of alliances and alliance changes. The most common types of alliances were the fifteen long-term sourcing arrangements and the ten joint marketing arrangements. Although most alliance partners were American, eight respondents reported on alliances with foreign firms.

• Frequency of Interaction

• Joint Decision Making

• Trust • Interdependence

• Planned/Unplanned Change

• Unilaterally/Jointly Planned Change

• Impact on Partner Firms

Alliance Outcome • Sustainment • Dissolution


Twenty changes were reported as mostly planned, while twenty-four were mostly unplanned and unilateral. Although respondents reported on a wide variety of alliance changes, the most frequent changes were mergers or acquisitions of one partner by a 3rd party. Slightly more than half of the alliances still existed a year after the change.

In addition to characteristics of alliances and alliance partners, and changes and their effects, other information was gleaned from questionnaire responses and conversations with respondents. It was found that firms in the study population commonly use strategic alliances. Of the 196 firms in the study population, 52, or 26.5%, returned a completed questionnaire, indicating that their firm had at least one strategic alliance.

Forty-five of the 125 population members that responded to the questionnaire reported that a change had taken place. The most frequent change reported was a change in ownership of one partner. Nearly as frequent were increases and decreases in alliance scope.

After questionnaire responses had been converted to an electronic format and errors and omissions resolved, bivariate correlation coefficients for all variables were computed (shown in Table 1). The independent variables that had significant bivariate correlations with the dependent variable, alliance sustainment (Kendall’s tau-b coefficients), prior to Bonferroni corrections, were the extent to which the change is planned jointly (.580**) and alliance age (.287*). Bonferroni corrections (.05/66) were applied because, with 66 correlation coefficients in the table, chances were that at least three correlations would be statistically significant by chance (i.e., .05 =1/20). Bonferroni corrections reduce the likelihood of Type I error by making statistical tests as stringent as necessary to account for the number of variables being tested. No correlations were significant after Bonferroni corrections were made.

Collinearity diagnostics were used to determine if multicollinearity would be likely to cause problems during regression analysis. Variance inflation factors (VIFs), statistics used to check multicollinearity (Neter, Kutner, Nachtsheim, and Wasserman, 1996) were also computed because the existence of excessive multicollinearity can artificially inflate R2 values, leading to incorrect conclusions about how well independent variables account for the variation in the dependent variable. The highest VIF among all independent

variables was 5.59 and the mean VIF for the data was 3.52, not high enough to indicate multicollinearity problems (Neter et al.).

Model 1 was able to explain most of the variation in the dependent variable for the alliances in the data set (R2 = .803). However, adjusted R2 was only .368 and the F ratio for the model was only 1.848 (p = .258). Regression results for Model 1 are shown in Table 2.

Model 1 was analyzed to determine which factors would best be retained in a more parsimonious model (Model 2) to be tested later. It seemed likely that some of the independent variables could be removed from the model without sacrificing much explanatory power. Indicators were removed from the model, using hierarchical regression to determine which predictors from Model 1 to include in a revised model. The resulting model, Model 2, contained only two predictors. It was highly significant (F = 14.208, p ≤ .001), and able to account for much of the variation in the dependent variable. Model 2 predictors were the extent to which the change is planned jointly (if planned) and the impact of the change. The regression results are shown in Table 2. R2 for Model 2 was .670 (adjusted R2 = .623).

Hypothesis 1 predicted that frequency of interfirm interaction prior to an alliance change would positively affect alliance sustainment following the change. Hypothesis 2 predicted that joint decision making prior to the change would positively affect sustainment. Neither hypothesis was supported. Since the two items used to measure frequency of interaction failed to load on a common factor during factor analysis, one item was discarded. The other item loaded on the same factor as one of the items used to measure joint decision making. These two items were used to make up the joint decision making factor. Joint decision making, was modestly and negatively correlated to alliance sustainment (-.130). In addition, it was not a significant predictor in the Model 1 global regression equation (t = -


.038, p = .971). Since little evidence of an association between joint decision making and alliance sustainment was found, joint decision making was not retained as an indicator in Model 2.

Table 1: Model 1 Means, Standard Deviations, and Correlationsa

(Kendall's tau-b Correlation Coefficients)

Variable Mean s.d. 1 2 3 4 5 6 7 8 9 10 11 ____________________________________________________________________________________________________________________________ 1. Nationality 1.18 .39 2. No Planned 1.82 .39 -.06

Dissolution 3. Alliance Age 4.21 4.98 .10 -.02 4. Joint Decision 4.99 1.27 .10 .12 -.23

Making 5. Unplanned/ 1.53 .50 .08 -.26 .06 -.10

Planned 6. Unilateral/Joint 3.32 1.98 .12 -.06 -.10 .14 .12 7. Responsible for 5.04 1.84 -.16 -.04 .10 .10 -.24 -.32

Change 8. Trust 4.41 1.26 -.14 -.28 .00 -.13 .06 -.19 .03 9. Dependence of A 4.74 1.56 .15 -.14 .14 .01 -.03 .20 .35 .036

on B 10 Dependence of 4.68 1.24 -.22 -.01 -.14 -.02 -.09 -.30 -.06 .017 -.02

B on A 11. Change Impact 4.49 1.49 .02 -.25 -.08 .19 .11 .31 .01 -.12 .01 -.08 12. Sustainment 1.15 1.20 .04 -.13 .29 -.13 .24 .58 -.16 .21 .10 -.18 .11 ____________________________________________________________________________________________________________________________ a N = 45. Note – None of the correlation coefficients are significant because I applied a Bonferroni correction (.05/66), since there are 66 correlations in the table, to keep the significance level at .05 for the entire study. Several correlations appeared to be significant without the Bonferroni correction.

Hypothesis 3 predicted that planned changes would be more likely than unplanned changes to result in alliance sustainment. Hypothesis 3 was not supported (t = -.081, p = .938). Kendall’s tau-b correlation coefficient was .242, which, although

not significant, indicates that the extent to which an alliance change is planned is associated with alliance sustainment. However, since the variable explained very little of the variation in sustainment, the extent to which an alliance change is planned was not retained in Model 2.

Hypothesis 4 predicted that jointly planned changes would be more likely than unilaterally planned changes to result in alliance sustainment. Hypothesis 4 was supported. In Model 1, the best predictor of alliance sustainment was joint planning of the change (t = 3.474, p = .018 prior to Bonferroni correction). Kendall’s correlation coefficient was .580, which was significant at p < .01, prior to Bonferroni correction. In Model 2, joint planning was a highly significant predictor of sustainment (t = 5.279, p ≤ .001).

Hypothesis 5 predicted that interdependence would positively affect alliance sustainment following a change. However, no support was found for Hypothesis 5. Dependence of the respondent’s firm on the partner was not a significant predictor in the Model 1 global regression equation (t = -.316, p = .764). Nor was dependence of the partner on the respondent’s firm a significant predictor (t = -.057, p = .956). Kendall’s tau-b correlation coefficients for each variable (.084 and -.184, respectively) also failed to provide evidence of a link between interdependence and alliance sustainment. As a result, interdependence was not retained as a predictor in Model 2.

Hypothesis 6 predicted that interfirm trust would affect alliance sustainment following a change. However, Model 1 regression analysis failed to support Hypothesis 7 (t = 1.380, p = .226). Trust explained very little of the variation in the dependent variable. Hence, even though a positive (although


not significant) association was found between trust and alliance sustainment (Kendall’s tau-b coefficient was .205), trust was not retained in Model 2.

Table 2: Regression Results for Models 1 and 2: Determinants of Alliance Sustainmenta ______________________________________________________________________ Variable Model 1 Model 2 ______________________________________________________________________ Constant -2.272 1.322 Control Variables

Nationality -.162 No Planned Dissolution 1.050 Alliance Age .100 Independent Variables Joint Decision Making -.0133 Unplanned/Planned .103 Unilateral/Joint .797 .561** Responsible for change .114 Trust .568 Dependence of A on B -.140 Dependence of B on A -.0339 Change Impact -.314 -.322* R2 .803 .670 Adjusted R2 .368 .623 F 1.848 14.208** ______________________________________________________________________ DV = Alliance Sustainment a Unstandardized coefficients are reported. N = 45 alliance changes. † p<.10, one-tailed test * p<.05, one-tailed test ** p<.01, one-tailed test

Note - Bonferroni corrections were applied to all independent variables in Model 1. Hypothesis 7 predicted that the impact of a change would affect alliance sustainment. Hypothesis 7

was supported. Although the impact of the change was not a significant predictor in Model 1 (t = -1.756, p = .139 prior to Bonferroni correction), it was the second best indicator of alliance sustainment. However, it was not highly correlated with alliance sustainment (Kendall’s tau-b = .109). Impact of the change was included in Model 2 where it was a significant predictor (t = -2.862, p = .013). V. Conclusion The results of this study have helped to advance the knowledge of alliances, especially alliances involving small firms. The results suggest that the most important factors affecting alliance sustainment after a change are the extent to which a change is planned jointly and the impact of the change. The results also suggest that alliance changes are more likely to result in sustainment if they are jointly planned and executed and affect partners moderately. The study provides less compelling


evidence that two other predictors, interfirm trust and the extent to which an alliance change is planned, may also be related to sustainment.

Overall, this study makes several contributions to alliance literature. First, it confirms that alliance changes are not uncommon. Second, it provides evidence that scholars’ attention to alliance instability has not been misguided. It appears that alliance changes are important; scholars have been wise to focus on major, unplanned changes. A key assumption made by scholars (Inkpen and Beamish, 1997) is that alliance instability must be unplanned from the perspective of one or more partners. This study furnished support for the assumption prevalent in alliance change literature that unplanned changes are deleterious to alliances.

This study also provides evidence that another key assumption is valid; the assumption that instability consists of major alliance changes. It was found that the greater the impact of a change, the greater the likelihood that an alliance will dissolve. Thus, major changes, those with greater impact, tend to be associated with dissolution and minor changes tended to be associated with sustainment. Hence, this study’s results provide support for the assumption, common in alliance change literature (Inkpen and Beamish, 1997), that major changes are detrimental to alliances.

This study suggests that a change that has a high impact on an alliance may threaten its survival even if the change is largely beneficial whereas minor changes, even if they result in problems, may be less threatening to an alliance. Even under the best conditions it is difficult to manage and sustain an alliance and impossible to ensure it is successful. Hence, managers need to carefully plan and implement large changes to prevent disastrous consequences. They may be wise to break large changes up into more manageable pieces to reduce implementation risks. Further, managers may choose to forgo certain opportunities (mergers and acquisitions) that satisfy cost/benefit analysis criteria, but may result in unknown negative consequences.

It is not surprising that this research suggests that alliances are more likely to survive changes that have a minor impact on alliance partners than those with a major impact. However, even a change with a positive impact on alliance partners may lead to dissolution if that change has a major impact on the firms, due to the difficulty that partners may have in responding to the change and coping with the effects of the change and the tensions that may arise between partners as a result. A greedy partner may want more than its share of the benefits of a change, leading it to act opportunistically.

More research on alliance change is needed. There is much to be learned about different types and magnitudes of alliance changes and their effects. As alliances proliferate, unless more can be learned about how to sustain them through tumultuous times, many may continue to be costly, short-lived disappointments. References 1] Anderson, J. and Narus, J. (1990). A model of distribution firm and manufacturing firm

working partnerships. Journal of Marketing, 54 (1): 42-59. 2] Arizona Manufacturers Directory. (2000). Evanston, IL: Manufacturers' News, Inc., 3] Das, T. K., and Teng, B. S. (2000). Instabilities of strategic alliances: An internal tensions

perspective. Organization Science, 11 (1): 77-101. 4] Ganesan, S. (1994). Determinants of long-term orientation in buyer-seller relationships. Journal

of Marketing, 58 (2): 1-19. 5] Heide, J., and Miner, A. (1992). The shadow of the future: Effects of anticipated interaction and

frequency of contact on buyer-seller cooperation. Academy of Management Journal, 35 (2): 265-291.

6] Inkpen, A. C., and Beamish, P. W. (1997). Knowledge, bargaining power, and the instability of international joint ventures. Academy of Management Review, 22 (1): 177-202.

7] Larson, A. (1992). Network Dyads in entrepreneurial settings: A study of the governance of exchange relationships. Administrative Science Quarterly, 37: 76-104.


8] McGee, J. (1994). Cooperative strategy and new venture performance: The role of managerial experience. In W. Bygrove, S. Birley, N. Churchill, E. Gatewood, F. Hoy, R. Kelley, and W. Wetzel, Jr. (Eds.), Frontiers of Entrepreneurship Research. Proceedings of 14th Annual Entrepreneurship Research Conference: 445-459. Wellesley, MA: Babson College.

9] Meyer, G., Alvarez, S., and Blasick, J. (1994). Benefits of technology based strategic alliances: An entrepreneurial perspective. In W. Bygrove, S. Birley, N. Churchill, E.

10] Gatewood, F. Hoy, R. Kelley, and W. Wetzel, Jr. (Eds.), Frontiers of Entrepreneurship Research. Proceedings of 17th Annual Entrepreneurship Research Conference: 629-642. Wellesley, MA: Babson College.

11] Neter, J., Kutner, M., Nachtsheim, C., and Wasserman, W. (1996). Applied Linear Statistical Models. Chicago: Richard D. Irwin, Inc.

12] Parkhe, A. (1993). Partner nationality and the structure-performance relationship in strategic alliances. Organization Science, 4 (2): 301-324.

13] Pearce, R. (1997). Toward understanding joint venture performance and survival: A bargaining and influence approach to transaction cost theory. Academy of Management Review, 22 (1): 203-225.

14] Pfeffer, J. and Salancik, G. R. (1978). The External Control of Organizations: A Resource Dependence Perspective. New York: Harper and Row.

15] Saxton, T. (1997). The effects of partner and relationship characteristics on alliance outcomes. Academy of Management Journal, 40 (2): 443-461.

16] Van de Ven, A., & Walker, G. (1984). The dynamics of interorganizational coordination. 17] Administrative Science Quarterly, 29, 598-621. 18] Young-Ybarra, C., and Wiersma, M. (1999). Strategic flexibility in information technology

alliances: The influence of transaction costs economics and social exchange theory. Organization Science, 10 (4): 439-459.


Trading Systems, Foreign Direct Investment and Economic Growth: Evidence from Asean Countries

Kevin Odulukwe. Onwuka

Department of Economics, Faculty of Economics and Administrative Sciences Yaşar Universitesi, Izmir, Turkey

Tel. + 90 232 461 4111; Fax: +90 232 4614121 [email protected]

Ahmad Zubaidi Baharumshah

Muzafar Shah Habibullah Department of Economics, Faculty of Economics and Management

Universiti Putra Malaysia, 43400, UPM, Serdang, Selangor, Malaysia

Abstract Over a decade the regional trade relations are growing in number and scope giving rise to trade policy dilemma among the developing nations. This paper presents empirical evidence on the effect of the regional trading system along with multilateral trading system and foreign direct investment (FDI) on the economic growth of ASEAN-5 countries (namely Malaysia, Singapore, Thailand, Indonesia and the Philippines) over the period 1976 to 2002. Three measures are constructed for each trading system and fixed effects estimation technique is applied to estimate the parameters of the model. The results reveal that multilateral trading system leads to faster growth rate than the regional trading system. Most of its measures are positive and are significant. The effect of regional trading system on growth rate is rather mixed. With the exception of RTA, the measures of regional trading system are negative giving inconclusive result on direction of the effect of regional trade bloc on economic growth. The effect of FDI on growth rate is more pronounced when the country is liberalising broadly. This implies that regional groupings can be utilized as a second best alternative to improve economic growth in ASEAN economies. Keywords: Trading systems, foreign direct investments and growth JEL Classification: F15, F21, O40

I. Introduction The resurgence of regionalism in the past decades vis-à-vis multilateral trading system creates trade policy dilemma among the developing nations. That is whether joining to multilateral or regional trading system would lead to improved economic growth and facilitate the inflow of the inward foreign direct investment. Trade is the engine of growth, no doubt, and apart from the comparative advantages, theories suggest additional gains from trade arising through economies of scale, exposure to competition and diffusion of knowledge. But which of the trading path will benefit the developing nations still remains a puzzling question among policymakers and academics. Secondly the nature of the relationship between regional trade agreements and economic growth is very much unclear, especially among the so-called south-south regional trade agreements (RTAs). This is the controversy


over few decades. The regionalism, which is seen as free trade areas and / or the custom unions, has been growing over the past decades especially in the 1980s and 1990s in Western Europe, North America, Africa and Asia.

The regional focus emerged as multilateralism weakened. This is attributed to declining U.S. hegemony in the world economy, rising pressures for increased protection in the major industrial countries, and fundamental changes in global political and security issues since the collapse of communism in the former Soviet Union and Eastern Europe. The growth of regional trade arrangements is also attributed to the delay in the Uruguay Round negotiations, which were not concluded until 1993, and this encourages a number of countries to explore bilateral approaches to expanding their economic relations, particularly through the formation of regional trading arrangements (DeRosa, 1995). This resurgence of regionalism in world trade is perceived to continue and will influence the nature and evolution of world economy. For example the uncompromising stance of European Union on CAP (Common Agriculture Policy) will fortify the resurgence of regionalism. The recent growth of interest in regional economic arrangements also came from apparent sharp shift in trade policy by the United States since the early 1980s. Prior to this time the United States placed overriding priority on the global trade regime. It refused to participate in regional initiatives and looked askance at regional initiatives by others. However, since then United States has simultaneously pursued global liberalisation and a number of regional initiatives. It has negotiated bilateral free trade arrangement with Israel and Canada, converted the later into North America free trade area (NAFTA) with addition of Mexico. It has accepted the steady evolution and creation of single economic market in Western Europe and promoted far-reaching free trade agreements in the Asia Pacific region via APEC.

The shift in trade policy by a leader of global liberalisation leads to rethinking about the relationship between global trading system and growth process. This apparent shift has created trade policy choice dilemma in most developing economies between regional and multilateral trading system. That is, would an economy perform better via regional initiatives or multilateral initiatives? Perceived to threaten the global trading system and the objective of achieving economic gains from broad multilateral trade liberalisation, regional arrangements, however, grow tremendously. About 250 RTAs are notified to World Trade Organisation (WTO) by the end of 2002, as against 130 before 1995. The number of RTAs is expected to reach 300 by 2005 (WTO, 2002). Hence the link between the regional trade arrangements and growth process requires urgent examination empirically. Hence the focal objective of this study is to examine the relationship between the regional trade arrangements, multilateral trade liberalisation and the economic growth in ASEAN-5 economies. In achieving this we extend the earlier work and take into account the shocks that have rocked the growth process in these economies in the model. We also construct regional trading system measures similar to multilateral trade measures to capture the effect of regional trade arrangement on economic growth. We explore the effects of interactions between the FDI and human capital, exports, and some measures of trading systems.

The remainder of this paper is organised as follows. The next takes a look at possible theoretical link between regional trade arrangements and economic growth and empirical evidence. Section three presents the model used to investigate the link between regional trading system and growth in panel framework. Sources of data are given in section four, while section five presents the results of the empirical investigations. Section six concludes the paper. II. Theoretical and Empirical Evidence May be, the best way for an economy to prosper is from the ground up (that is from the community and the neighbourhood) or belonging to smaller group of countries where rules or laws are easy to enforce. Regional trade arrangements provide such a community or neighbourhood. Thus regional trade agreements can be linked to growth from the community. Regional trade arrangements are conceived


with the notion that by lowering barriers on trade between member-countries trade can be created leading to improved economic performance and welfare. However theory and empirical literatures do suggest that countries that pursue multilateral trade policies grow faster than the economies that follow regional trade agreements because the total regional trade may be smaller in volume. The reason for this strong bias in favour of multilateral trade liberalisation is partly based on the conclusions of wide range of empirical studies, which claimed that outward-oriented economies consistently have higher growth rate than inward-oriented countries. Output growth or an outstanding economic performance in most East Asian countries has been linked to multilateral trade liberalisation as well as foreign direct investment. And this suggests that a significant progress in the East Asian countries is due to reducing tariff barriers and dismantling of non-tariff barriers.

The early empirical work started in 1950s with Viner’s seminal work (Viner, 1950). In a simple partial equilibrium model under perfect competition Viner (1950) demonstrates that RTA may increase the level of trade between members at the expense of less efficient domestic producers (trade creation) but also of more efficient third countries (trade diversion). The net effect of RTA on trade thus depends on the relative size of these two effects. Apart from trade creation and diversion, dynamic effects are incorporated into the static approach to regional integration. These dynamic effects are competition and scale effects. These dynamic effects of regional integration are often used to justify the formation of such regional trading arrangements. Both the European project and NAFTA have been justified on economies of scale that not only allowed RTA members to increase their intra-regional exports but also their trade with the rest of the world. The Viner’s framework presents ambivalent conclusion that RTAs can enhance or reduce welfare of member states. Several authors have attempted to clarify empirically these ambiguous effects of RTAs but fail to solve the puzzle (World Bank, 2000; Yeats, 1998; Schiff, 1997; Park, 1995). These studies arrived at similar conclusions that RTAs, between the developing countries are more likely to generate trade diversion, especially when external tariffs are high than the RTAs involving the developed and developing countries. On the other hand, studies like Elbadawi (1997), Evans (1998), Lewis et al. (1999) and Flores (1997) have come up with empirical evidence that regional integration can generate positive net effects necessary to trigger the much-needed strategic complementarities. A more recent study Cernat (2001) assesses the effect of RTAs on the bilateral trade and finds that some intra- regional dummy variables are negative and some positive and significant. Still the impact of RTAs on trade, thus welfare remains unclear.

An attempt is made by Vamvakidis (1999) to reconcile theory of regional trade arrangements with empirical evidence on some macroeconomic variables such as growth rate and FDI based on time- series set data of 17 regional groupings over the period 1950-1992. He finds the economies that liberalised broadly to grow faster and have higher investment shares in both the short and long run compared to economies that participate in regional trade arrangements. Although his study was beset with aggregation problem, he arrives at similar conclusion that RTAs between developing nations are more of trade diverting than trade creation. The results cannot not be used to make general conclusion on the effects of the regional trading blocs. More specific study is required which can take into account in the model the shocks in these economies. The present study addresses these problems by using ASEAN free trade area (AFTA) and incorporates into the model Asian financial crisis. III. Modelling Trading Systems, FDI and Economic Growth in a Panel Framework In modelling the impact of trading systems on economic growth, we divide the trading system into two - multilateral and regional trading systems. The multilateral trading system is a trade involving many, if not all nations, while the regional trading system can be taken as regional trade blocs that involve countries within the same geography, and agree to trade among themselves by reducing trade barriers. Most studies on the determinants of economic growth either use growth accounting framework, which is based on an aggregate production function expressed in growth rates (Vamvakidis, 1999 and Harrison, 1996). Recent empirical studies rely on endogenous growth model that assume constant or


increasing returns to scale in capital as in Romer, 1990; Grossman and Helpman, 19991 and Borensztein et al. 1998. Following the work of Levine and Renelt (1992), which searched for a set of robust variables to model growth and the theoretical contributions of Romer (1990) to the new growth theory literature, a degree of convergence on most appropriate empirical specification has occurred. The core explanatory variables for economic growth identified in these studies include investment, initial per capita GDP and human capital and FDI. This study extends this model to include exports, wage rate and the trading system measures. The separate inclusion of multilateral and regional trading variables is based on the observation and historical empirical evidence that turbulent world trading system has a significant impact on terms of trade and hence economic growth of developing countries. Also included in the model are measures of the Asian financial crisis to test robustness of the results. The basic specification for the model is therefore:

tititititi

tititiiti

PFTDSEXWAGFDIEDEMPyg

,,8,7,6,5

,4,3,276,10, lnεββββ

βββββ+++++

++++= (1)

where tig , is the real GDP per capita growth of the country i , 76,iy is the real GDP per capita in 1976,

tiEMP, is the working population growth ( number of labour force employed), tiED , is the ratio of educational expenditure to GDP (representing human capital), tiFDI , is the ratio of FDI inflow to GDP, tiWAG , is the growth of expected wage rate, tiEX , is the export growth, tiTDS , is the vector of trading systems and tiPF , is the financial crisis dummy and ti,ε is the stochastic error term satisfying

the assumption ),0( 2σIIDuit = . The approach here is much in keeping with the work of Barro (1991), Easterly and Levin (1997), Greenaway et al. (2002) and Li and Liu (2005).

Three alternative indexes of multilateral trading system and regional trade arrangement are constructed and used to investigate whether the evidence supports the view that with other things being equal following multilateral trading system leads to faster economic growth rates than the regional trade blocs or regional trading arrangements offer better growth prospects to developing nations in lieu of shocks in the world markets. In constructing the multilateral trade measures we utilize imports plus exports over gross domestic product (OPEN), the average tariff (policy measure) that is calculated by dividing customs revenue by the value of imports (AVTA) and the Dollar’s real exchange rate distortion (DSTORT) that measures trade restrictions. For regional trading system, similar measures were constructed using intra-ASEAN trade over GDP (ROPEN) and Singapore dollar as a benchmark to construct regional exchange rate distortion index (RDSTORT) and a dummy variable (RTA) is employed which takes the value of zero before 1977 and one thereafter capture the effect of participation in regional trade blocs. The measures help us understand the form of relationship between multilateral trade or regional trade and growth. The shock such as financial crisis is measured by a dummy variable, which takes the value of one in 1997 and 1998 or otherwise zero. Also explored are the interactions between FDI and trading system measures, human capital and exports to examine their effects on economic growth are contingent on each other.

The use of trade share as a measure of openness has been criticized as not being a policy tools but an outcome and is obviously endogenous. But nonetheless it helps us to measure the outcomes of trading system effect on economic growth rate. The tools at the disposal of governments are tariff and non-tariff barriers and exchange rate and not export and imports levels (Rodrik, 2000) and are exogenous policy measure. They cannot represent all but a fraction of restrictions inherent in developing countries. We use different indexes because nature of commercial policy. Tariffs, quotas, licenses, prohibitions and exchange rate controls can affect international trade differently and hence growth rate. Attempts to construct a single indicator of trade restrictive index may be futile and will tend to generate disagreements and controversies among economists (Edwards, 1998). This means that


for research on the relationship between the trade policy and some macroeconomic variables to be persuasive, its results have to be robust to the way in which openness is measured.

In estimating Eq (1), we employ the fixed effects estimation technique. The ASEAN sample permits the use of annual data instead of averages over time, as is often done in the cross-country empirical literature. This study recognizes the severe restriction of fixed-effects estimation on the parameters, but this allows us to exploit the cross-country variability in the data to learn about the growth process and implications of multilateral and regional trading systems. These restrictions permit the direct identification of the factors affecting the steady path of growth either in the multilateral or regional trading systems. There are reasons to believe in common long-run coefficients for the ASEAN countries given that they have access to common technologies markets and have intensive-trade and foreign direct investment, all factors lead to similar long-run production functions parameters. However, we assume that the speed of convergence to the steady-states would be the same across countries, but there is no reason to be so, because of policies underlining growth process in each ASEAN country and growth may differ due to population growth rate. Thus according to Nagayasu (1998) pooling the data not only increases the statistical power to accept or reject the null hypothesis without increasing the possibility of a structural shifts but also helps us draw a general conclusion that applies to a broad group of countries. IV. Data A data set of five ASEAN countries is used to analyse the robustness of the relationships between multilateral trading system, regional trading system and growth, over the period 1976 to 2002. Exports, imports, GDP and exchange rate data are collected from the International Financial Statistics of IMF. Trade exposure indices are constructed, that measure how the ASEAN countries are being represented in the international and regional markets or competitiveness of ASEAN economies in the global economy. These indices indicate how representation in either international or regional affects the growth rate and trade policy or restrictions. Human capital is proxied by educational expenditure (ED). Educational expenditure and employment level are taken from Asian Development Bank, Key indicators database. V. Empirical Results Panel Unit root test Testing for stationarity has become a conventional in econometrics analysis involving panel data as it is in time series. The Levin and Lin (1993) (LL), Levin, Lin and Chu (LLC) (2002) and Im, Peraran and Shin (1997) (IPS) tests are the most widely used methods for panel data unit root tests in the literature. For this study we rely on IPS procedure to test the stationarity of our data series, as it performs better than (LL) and (LLC) because it relaxes the identical assumptions of LL test, which requires all countries within each group to share a common average speed of adjustment to steady state equilibrium and estimates the ADF test equation for each individual. Also the IPS test is easy to use because the Tables of the critical values are made available in the same paper (Im et al., 1997). The IPS method pools N separate independent Augmented Dickey-Fuller (ADF) regressions:

tiktktikittiiitti uXXXXXX ,,,11,, )()()( +−∆+−+=−∆ −−−− ∑γρδ (2)

This procedure allows for heterogeneity in ρ by testing the null hypothesis, 0=iρ for all i , against the alternative hypothesis, 0<iρ for at least one i . The limiting distribution for their t-statistics is given as:


)1,0()(2

NtN

ADF

ADFADF →−

σ

µ (3)

where the moments ADFµ and 2ADFσ are obtained from Monte Carlo simulations and ADFt is the

average estimated ADF t-statistics from the sample. The results of the stationarity test are reported in Table 1 and it shows that all the variables are stationary at least 1% level of significance. Table 1: Panel Unit root Test

Variables IPS-Statistics-W Prob GDP -7.207 0.0000 FDI -7.754 0.0000 EMP -11.522 0.0000 ED -2.559 0.0052 WAG -8.568 0.0000 EX -5.298 0.0000 MER -3.660 0.0001 RER -2.097 0.0180 AVTA (T) -4.196 0.0000 GDP= GDP growth rate, FDI = growth of Foreign direct investment, ED= the growth rate human capital, EX = the growth rate of export, EMP = growth rate of employment level, WAG = wage rate, MER = exchange rate distortion index based on US dollar, AVTA = average tariff (tariff revenue/imports)

Correlation Across Trading System Measures The investigation starts with a simple Spearman’s correlation test on the three measures of multilateral trading system (OPEN, DSTORT and AVTA) and regional trade arrangements (ROPEN, RDSTORT and RTA) and it is observed that the relationship between these variables is low (Table 2). The relationship between average tariffs (AVTA) and trade share (OPEN) is negative. This indicates, according to the theory, that trade barriers are fairly effective in reducing trade in the international market. Thus, it is safe to conclude that trade barriers may have negative effects on growth through reducing the size of the external sector of a country. The relationship between trade share and exchange rate distortion is positive and low, indicating that the higher the exchange rate distortion the higher the trade share. There is also no indication of multicollinearity among the measures of trading systems constructed for the analysis; as such they can be included in the same model or independently. Table 2: Spearman correlation coefficients for measures of trading systems

Multilateral trading system measures OPEN DSTORT AVTA OPEN 1.0000 DSTORT 0.1578 1.0000 AVTA -0.6000 0.0689 1.0000

Regional trading system measures ROPEN RDSTORT RTA OPEN 1.0000 DSTORT 0.1055 1.0000 RTA 0.4739 0.0268 1.0000


Multilateral Trading System and Growth Rate In analysing the impact of the multilateral trading system on cross-country growth rates we use fixed effects estimation technique. In the estimation, we assume that the form of heteroskedasticity is unknown. To deal with this problem of heteroskedastictiy together with serial correlation, we allow the cross-section weights and white heteroskedasticity consistent covariance in the fixed effect estimation. This implies that each pool will have an unrestricted intercept and that each pool equation is down weighted by an estimate of the cross-section residual standard deviation. The results of this investigation are reported in Table (3).

In column (1) we report the results of the core variables – FDI, initial GDP, exports, employment level, human capital and wage rate. The impact of FDI is negative and significant. Other variables namely exports, human capital, employment level and wage rate carry positive signs and are significant at least at 5% level, except employment level. Column (2) reports the results of regression that includes the three measures of multilateral trading system and Asian financial crisis. Columns 3 through 8, report the results of interactions between FDI and human capital, trade share and exports.

The three measures carry positive signs and are statistically significant, at least at 1% level, except the average tariff (AVTA). The coefficient of OPEN is 0.0434 (column 2) implying that a 10% increase in trade shares would increase the average growth rate of GDP by 0.434% per year. This also suggests that ASEAN countries will grow on average by 0.434% a year after joining multilateral trading system. The result is consistent with a number of empirical studies, (see Harrison 1996; Vamvakidis, 1999; Yanikkaya, 2003) that used the trade share as a measure of trade reform. One of the important criticisms of trade share as a measure of trade orientation is that it is an outcome of trade policy reform and not a trade policy per se. Its effect on growth will likely depend on the effect of trade policy reform on exports and imports. Another criticism, as noted by some studies (see Yanikkaya, 2003) is the likelihood of the reverse causation between growth and trade share. In this case fast growth may cause higher trade shares. And it does not necessarily mean that the direction of causation goes from higher trade shares to higher growth rates.

The coefficient of average tariff is positive and not significant. However, it is significant when foreign direct investment interacted with trade share (FDOP) (see column 8 of Table 3). This result contradicts the long established theory of the late 20th century, that average tariff is negatively correlated with the growth rate (Harrison, 1996 and Edwards, 1998) but supports the hypothesis of new growth theory, the theory of strategic trade policy, the infant-industries arguments and development economics that trade restrictions can promote growth in certain countries. On the empirical side it is consistent with the findings of most recent studies. For example, Bairoch, 1972; Capie, 1983; O’Rourke, 2000 and Yanikkaya, 2003 find a positive relationship between the average tariff and economic growth. These studies argue that protectionist countries grow faster in the 19th century, not slower as economists have believed in the late 20th century. The view in the 19th century is that tariff (protection) enhances economic performance, while the 20th century economists see it as having deleterious effect on growth.

Does tariff-growth positive correlation in ASEAN economies emanate from import substitution? As Irwin (2002) noted several individual countries’ experiences in the late nineteenth century are not consistent with the view that import substitution promoted growth. For example, Argentina and Canada are high tariff countries that grew not because of protectionist trade policies but because of capital import that helped to stimulate export-led-growth in the agricultural staple products. ASEAN economies grow because of expansion in agricultural exports not principally because of import tariff revenue. If tariff causes the reallocation of productive resources to the goods in which ASEAN countries have comparative advantage from goods in which they have no advantage, then tariffs are likely to affect growth positively. The tariff structure in ASEAN countries was designed to move resources to sectors that have relatively higher positive externalities for the whole economy. As remarked by Rodriguez and Rodrik (2001) there has been a tendency to overstate the pure tariff effect on growth and whether some of the 20th century results would be unequivocal if they control for


omitted trade policies. It should be noted however that “it is difficult to isolate the effect of trade policies alone, since other policies may change at the same time and liberalism typically comes as a package. Thus, countries that liberalized their trade also liberalized their domestic factor markets, liberalized their domestic commodity markets and set up better property-right enforcement” (Clemens and Williamson, 2002).

Table 3: GDP growth rate and multilateral trading system Independent variables

Dependent variable: GDP growth rate

1 2 3 4 5 6 7 8 IGDP 0.0343

[0.751] -0.0190*** [-2.784]

-0.0220*** [-2.914]

-0.0186** [-2.307]

-0.0203*** [-2.770]

-0.0219*** [-2.934]

-0.0080 [-0.664]

0.0229* [1.798]

FDI -0.0479*** [-2.960]

0.0316*** [3.618]

0.0318*** [3.624]

0.0332*** [3.342]

0.0274*** [3.205]

EDEX 0.0960** [2.300]

0.0014 [0.096]

0.0104 [0.607]

0.011 [0.073]

0.0051 [0.319]

0.0004 [0.027]

0.0003 [0.027]

EMP 0.4900 [1.148]

0.4193*** [3.893]

0.4994*** [4.468]

0.5003*** [4.391]

0.4236*** [3.913]

0.4736*** [5.330]

0.4529*** [4.114]

0.5385*** [4.503]

WAG 0.4187** [2.707]

0.2470*** [7.069]

0.2491*** [7.118]

0.2506*** [7.127]

0.2447*** [7.177]

0.2772*** [7.749]

0.2391*** [7.236]

0.2271*** [6.638]

EX 0.6298*** [5.045]

0.2519*** [4.384]

0.2756*** [4.608]

0.2958*** [5.060]

0.2361*** [3.044]

0.2539*** [4.318]

0.2711*** [4.922]

OPEN 0.0434*** [4.539]

0.0465*** [4.748]

0.0450*** [4.691]

0.0439*** [4.663]

0.0397*** [4.267]

0.0355*** [3284]

MER 0.6663*** [6.576]

0.6838*** [6.650]

0.6753*** [6.619]

0.6605*** [6.310]

0.6101*** [5.954]

0.6709*** [6.785]

0.6636*** [7.055]

AVTA 0.0067 [0.397]

0.0093 [0.560]

0.0097 [0.593]

0.0032 [0.182]

-0.0125 [-0.649]

0.0070 [0.417]

0.0348** [2.460]

PF -0.0426*** [-3.791]

-0.0345** [-2.566]

-0.0359*** [-2.944]

-0.0436*** [-3.468]

-0.0664*** [-6.666]

-0.0348*** [-2.675]

-0.0190 [-1.322]

FDED 0.0128* [1.799]

0.0106** [2.276]

FDEX 0.9553 [0.579]

5.5584*** [5.077]

FDOP 0.1993 [1.447]

0.5844*** [4.718]

2R 0.44 0.644 0.636 0.638 0.640 0.618 0.636 0.624 Log Likelihood

45.56 99.13 99.54 99.14 99.18 94.75 99.35 97.89

DW Statistics

2.13 2.14 2.14 2.16 2.14 2.10 2.14 2.14

F-Value 5.623 [0.0193]

21.82 [0.0000]

23.16 [0.0000]

23.14 [0.0000]

21.95 [0.0000]

23.02 [0.0000]

17.49 [0.0000]

32.94 [0.0000]

Observations 135 110 110 110 110 110 110 110 ***, **, and * denote 1%, 5% and 10% level of significance respectively. Figures in brackets are t values. GDP= GDP growth rate, FDI = growth of Foreign direct investment, EDEX = the growth rate human capital, IGDP = initial GDP 1975, EX = the growth rate of export, EMP = growth rate of employment level, WAG = wage rate, OPEN = imports+ exports/ GDP, MER = exchange rate distortion index based on US dollar, AVTA = average tariff (tariff revenue/imports), FDED = interaction between FDI and human capital, FDOP = the interaction between FDI and OPEN and FDEX = the interaction between FDI and Export (EX), D = dummy variable used to capture the effect of commodity and Asia Financial crises.


Exchange rate distortion index measures the extent of trade restrictions in the economies under investigation. Trade restrictions represent policy response to real or perceived market imperfections or at other extreme are mechanisms for rent-extraction; they will work differently from natural or geographical barriers to trade and other exogenous determinants. The coefficient of exchange rate distortion is positive and significant at 1% level (column 2 of Table 2) implying a stable exchange rate in ASEAN economies. Thus a 10% increase in exchange rate stability leads the ASEAN-5 economies to grow on average by 6.7% per year. The result, however, contradicts Dollar’s (1992) findings in 95 developing countries and Ghura and Grennes’ (1993) where the impact of exchanges rate distortion on growth was found to be negative and significant but supports the argument of Michaely, et al. (1991) which emphasized the fundamental role of exchange rate stability in liberalization episodes. They argue that countries with a more volatile real exchange rate experienced poorer overall performance than those nations that have managed to maintain a more stable real exchange rate. The result indicates that ASEAN economies follow exchange rate polices that are more consistent with price stability and anti-exports bias. The result also stresses the importance of nominal exchange rate devaluation to eliminate real exchange rate overvaluation, which most ASEAN countries pursued and implemented during the liberalization periods. Devaluation is a package that demands dismantling of trade, capital and exchange controls. These measures are channelled towards increasing exports and improving the business environment for both domestic and foreign direct investments. It can be seen that exchange rate is one of the policy instruments that can be manipulated to achieve a desired level of economic growth. Exchange rate stability is a sine qua non attraction to investments as volatility in exchange rate deters investors, affects economy by limiting amount of capital inflows, discourages savings and less opportunity to create more jobs and thus decrease in productivity and GDP.

The coefficients of FDI is negative and significant (column 1) but turns positive and significant when trading system and Asian financial crisis are controlled for (column 2). Thus a 10% percent increase in FDI inflow will increase the growth by 0.32% a year after joining the multilateral trading system. The result is a strong indication of liberalisation of government policies towards FDI in ASEAN-5, which has, according to neoclassical growth model, ability to increasing the volume of investments. Also the result indicates evidence of transmission of technology on business practices and management techniques and or spill over effect into ASEAN-5 economies, as endogenous growth models would allow us to understand, leading to faster growth. The FDI has independent effect on the economic growth as it is still positive and significant when interacted with human capital, exports and trade share measure. Thus FDI-led-growth hypothesis upholds in ASEAN economies investigated.

The coefficient of the initial GDP is positive in column (1) but becomes negative and significant when we control for trading system measures and Asian financial crisis (column 2). The result does not change in other regressions. This suggests that ASEAN-5 countries are converging towards the long run equilibrium income. This result is consistent with convergence hypothesis, which states that economies with similar structures tend to converge to similar steady-state.

Employment level is positive and significant. Employment contributes to economic growth through two main factors – income tax generation and reduction in social security payment. As employment levels rises so is the income tax revenue and reduction in payment on social security, leading to faster GDP growth. The employment level rises as countries move to liberalise their economies. A 10% increase in employment level will lead the economies to grow by 4.2 percent a year after joining the multilateral trading system.

The positive relationship between wage rate and growth rates indicates that ASEAN economies have begun to enter into the new stage of industrialization according to Lewis theory of dual economy model of industrialization. In this stage the surplus labour, which induced FDI, decreases (i.e. demand for labour outweighs supply of labour) and pushes up wage and thus the overall per capita share of GDP. Growth is not led by investments of corporations at this stage but by private consumption. The profits of the multinational corporations that can be efficiently invested go into the hand of consumers in form of increased wage. This gives the consumers more money to consume goods and services.


The coefficient of human capital proxied by educational expenditure is not robust. It is positive and not significant. This result casts doubt on the traditional role given to human capital in development process as separate factor of production, which is to increase the country’s absorptive capacity for new innovations and catch-up. Lack of clear role of human capital in the growth process of sampled ASEAN countries, according to Benhabid and Spiegel (1994), can be attributed to specification and data source problem as well as bias encountered when regressing per capita income growth on accumulated factor of production. Intuitively, one possible explanation is that as knowledge is not acquired in a day and is an ongoing process its returns on education will take time manifest. From human capital perspective it is seen that human capital development is not growing at the same rate as the economic growth or educational expenditure does not match growth in ASEAN-5 economies. Vamvakidis (1999) found a similar results using secondary school enrolment as proxy for human capital. He attributed the results to poor quality of the proxy for human capital compared to those used by Bassanini, Scarpetta and Hemings (2001)3 and de la Fuente and Domenech (2000). The results on human capital have some important policy implications. Insofar as sustainable economic growth is the cardinal objective of ASEAN countries labour policy needs be flexible to accommodate the shortfall in human capital in the region. Too, appropriate educational policy needs to be put in place to address the accumulation of human capital in the region.

The coefficient of export (EX) is positive and significant, confirming the export-led-growth hypothesis in ASEAN economies. The Asian financial crisis carries a negative sign and is significant at 1% level, implying that the Asian crisis derails the path of growth in the economies under investigation. A 10% increase in shocks will lead to decrease in growth rate by 0.43% a year. This implies that shocks that affect the physical structures are likely to disrupt economic growth process.

It is clear from the preceding discussion that FDI, EX, EDEX and OPEN are positively correlated with growth. Consequently one would expect the interaction between FDI and EX, EDEX and OPEN to have positive effect on growth. This might be so because FDI is attracted to foreign locations due to trade reforms, or exports and or abundance of absorptive capacity in the host country. In column 3 through column 8 we test this hypothesis. In all the coefficients of interacting variables are positive and are highly significant especially when the interacting variables are excluded from the regression (columns 4, 6 and 8). The results show FDI, exports and trade reform have independent effect on growth. However exports promotion, human capital and trade reforms enhance the role of FDI in economic growth. Regional Trading System and Growth Rate Unlike in the multilateral trading system, the measures of regional trading system give mixed results. The estimates of the effect of regional trading system on economic growth are reported in Table 4. The coefficients of regional trade share (ROPEN) and exchange rate distortion (RDSTORT) carry negative signs and are significant at 1% level (column 1 of Table 4). This implies that regional trade liberalisation may not be a good pathway to improve economic growth. A 10% increase in regional trade share and exchange rate distortion will lead to a fall in growth rate by 1.4% and 3.1% respectively after joining regional trade blocs. However participation in the regional trade arrangements (RTA) fosters economic growth. The coefficient of RTA is positive and significant (column 1). This indicates that ASEAN economies will grow by 0.46% percent a year after joining the regional trade arrangement. This result does not agree with the findings of Vamvakidis (1999), who finds that participation in regional trade arrangements adversely affects economic growth based on a set of 18 RTAs with a total of 109 member countries over a sample period 1950-1992.

The adverse impact of regional trade share and exchange rate distortion on growth shows that intra-ASEAN trade is small compared to their trade with the world. With the exception of Singapore, the

3 They used average years of schooling for the population between 25-64 years and found high return to education, as coefficients (0.04 to 0.071) are positive and significant One extra year of schooling would lead to an average increase in steady state output per capita by about 4-7 per cent


economies of ASEAN members are not complementary. They engage in similar industrial sectors. Most of their primary commodities and manufactured goods export compete in the same world markets. In such situation exchange rate policy becomes the main useful tool in ASEAN economies to make their goods competitive in international markets. Table 4: GDP growth rate and Regional trading arrangement

Independent variables

Dependent variable: GDP growth rate

1 2 3 4 5 6 7 IGDP -0.0647*

[-1.886] -0.0512 [-1.453]

-0.0586* [-1.720]

-0.0913*** [-2.917]

-0.1150*** [-3.684]

-0.0759** [-2.414]

-0.0753** [-2.580]

FDI -0.0081 [-0.653]

-0.0083 [-0.643]

-0.0138 [-1.165]

-0.0080 [-0.663]

EDEX 0.0071 [0.394]

0.0334 [1.263]

0.0210 [1.208]

0.0190 [1.260]

0.0104 [0.590]

0.0081 [0.591]

EMP 0.1861 [1.618]

0.2302** [1.993]

0.1882 [1.611]

0.2008 [1.7353]*

0.1604 [1.579]

0.1919* [1.702]

0.1818 [1.634]

WAG 0.3675*** [5.275]

0.3323*** [4.982]

0.3536*** [5.375]

0.3088*** [4.531]

0.3426*** [4.697]

0.3551*** [5.298]

0.3571*** [5.699]

EX 0.1196*** [4.906]

0.1350*** [5.148]

0.1323*** [5.169]

0.1005*** [4.359]

0.1152*** [4.808]

0.1190*** [4.966]

ROPEN -0.1398*** [-3.432]

-0.1786*** [-3.873]

-0.1682*** [-4.012]

-0.0717 [-1.522]

-0.0895* [-1.757]

-0.0206 [-0.317]

RER -0.3087*** [-3.906]

-0.3140*** [-3.723]

-0.3132*** [-4.067]

-0.3279*** [-4.586]

-0.3421*** [-5.988]

-0.3057*** [-3.873]

-0.3091*** [-4.102]

RTA 0.0460** [2.232]

0.0596*** [3.096]

0.0583*** [3.175]

0.1620*** [6.035]

0.1653*** [6.647]

0.0450** [2.248]

0.0490*** [3.212]

PF -0.0937*** [-7.226]

-0.0830*** [-6.236]

-0.0856*** [-6.897]

-0.0979*** [-8.418]

-0.1038*** [-8.915]

-0.0762*** [-5.639]

-0.0740*** [-6.024]

FDED 0.0361*** [2.635]

0.0219*** [4.888]

FDEX 0.3533*** [5.0017]

0.3543*** [5.654]

FDROP -0.6366*** [-2.750]

-0.6792*** [-4.871]

2R 0.433 0.444 0.449 0.467 0.458 0.441 0.454

Log Likelihood

61.78 62.97 62.24 65.22 62.45 62.75 62.52

DW Statistics 2.11 2.10 2.081 2.09 1.984 2.109 2.122

F-value 11.49 [0.0000]

10.42 [0.0000]

12.98 [0.0000]

16.28 [0.0000]

19.71 [0.0000]

11.11 [0.0000]

12.74 [0.0000]

No of Obs 108 108 108 108 108 108 108

***, **, and * denote 1%, 5% and 10% level of significance respectively. Figures in brackets are t values. GDP= GDP growth rate, FDI = growth of Foreign direct investment, EDEX = the growth rate human capital, IGDP = initial GDP 1975, REX = the growth of regional export, EMP = growth of employment level, WAG = wage rate, ROPEN = imports+ exports/ GDP within the region, RER = regional exchange rate distortion index based on Singapore dollar, RTA = participation in regional trade arrangement, FDED = interaction between FDI and human capital, FDROP = the interaction between FDI and ROPEN and FDREX = the interaction between FDI and regional export (REX), D = dummy variable used to capture the effect of commodity and Asia Financial crises.


In regional trading system FDI, employment level (EMP) and human capital (EDEX), though their coefficients carry positive signs, except FDI, do not have any effect on the growth rate. Wage rate (WAG) and regional export (REX) have similar effect on growth as in multilateral trading system. Its coefficients carry the expected sign and are statistically significant at 1% level. The coefficient of initial GDP (IGDP) is negative after joining regional trade arrangement, implying that convergence of ASEAN economies in the long run is possible. They will converge to similar steady-state and income inequality will disappear over time. The effect of commodity and financial crises on growth rate in regional trading system does not differ from its effect on growth after joining multilateral trading system. Its coefficient is negative and significant at 1% level, indicating the following regional trade initiatives a 10% increase in shocks in the ASEAN economies will reduce growth rate by 0.94% a year.

We also test the effect of interaction between FDI and regional trading system measures. The results of this exercise are reported in columns 2 to 7. The coefficient of the interaction terms, FDED and FDREX, are positive and significant at least at 5% level except FDROP which is negative. The analysis shows that the interacting variables have independent effect on economic growth under regional trading system.

From the ongoing discussion, the measures of trading system used in the growth equation cannot be said to have covered all the restrictions in ASEAN economies under investigation. There are other forms of restrictions that relate to health and environmental regulations that affect international trade and thus growth. These are not included in the model. The measures of the trading system have some limitations. The trade share (dependency ratio) is a measure that is not related to policy; a country can distort trade heavily and still have high trade dependency ratio. Thus it is largely endogenous and outcome, not a trade policy per se. Many cross-country studies on openness-growth nexus used it because its data is readily available. It could be better used to measure the impact of trade policy reforms. Average import tariff and exchange rate controls are trade related policies and are tools at the disposal of the governments. Average tariff has some flaws. It may underweight high tariff rates where the corresponding import levels tend to be low and would not cover both tariff and non-tariff barriers that operate in the economy. According to Edwards (1998) an attempt to construct a single index of trade-orientation may lead to disagreements and generate controversies among economists. For this reason and for research on the relationship between trade policy and growth to be persuasive, its results have to be robust to the way in which openness is measured. VI. Conclusion and Policy Implications This paper has presented a cross-country evidence of the effect multilateral and regional trading systems on cross-country growth rate using fixed effect estimation technique. It is found that ASEAN-5 economies grow faster after joining multilateral trading system. The effect of regional trading system is rather mixed. While the variable that captures participation in regional trade arrangements is positive, other regional trading measures are negative leading to inconclusive results. Multilateral trading system provides a good environment for investments, regional trading system is characterised by competition among the ASEAN countries. Thus the exchange rate policy in ASEAN economies is engineered towards creating incentives for FDI after joining the multilateral trading system. The FDI-led-growth hypothesis is evident in ASEAN-5 countries and can be fully realised if there is enough human capital and the country is liberalising.

This study can say with caution that tariff hardly constitute the most serious trade restrictions in ASEAN economies but they help to indicate the trend in ASEAN trade policy. Tariff averages do not reflect the overall protection levels and if it does, one expects to find a negative relationship between tariff and growth rates, controlling for other determinants of growth rate. Thus tariff measures are imperfect indicator of a country’s trade policy orientation and may not always reflect protectionist policies. The tariff structures in ASEAN economies are not protective tariff but rather reallocation tools of resources to sectors that have positive externalities and or where ASEAN countries have


comparative advantage. In this case it is allowed to lower the long run cost curve as in the infant-industry case or to foster industrialization, as in those modern growth models where industry is the carrier of technological change and capital deepening. Tariffs do necessarily change the domestic economic structure in a way that promotes or impede growth (Irwin, 2002) and might influence growth by influencing access to international credit market (Lane, 2001).

As shown by the results of this study, regional integration might assume economic importance in the future. For ASEAN to realize the full benefit of integration, its economies should be complementary rather than competitive. Competition leaves no room for deep integration. As new technologies increase the scope of the markets beyond the boundaries of individual members and ASEAN to realize the full gains of integration, the existing governance structure needs to change (i.e. members should be willing to sacrifice some powers and be more prosperous); the trade policies of members should be harmonised. Selective intervention, if well executed could contribute to growth as shown by the result in Table 3, but it requires capable bureaucracy and political will, as it may lead to negative impacts on resource allocation, productivity and state corporate governance. References 1] Asian Development Bank, Key indicators, various issues 2] Bairoch, P. (1972) Free trade and European economic development in the 19th century.

European Economic Review, vol. 3: 211-245. 3] Barro, R. J. (1991) Economic growth in a cross section of countries. Quarterly Journal of

Economics, vol. 106 (2): 407-433. 4] Bassanini, A., S. Scrapetta and Hemmings, P. (2001) Economic Growth: The role of policies

and institutions, Panel data evidence from OECD countries, Economic Department Working paper No 283, Organization for economic Co-operation and Development (OECD), France, PP. 1-63.

5] Benhabib, J. and Spiegel, M. (1994) The Roles of human capital in economic development: evidence from aggregate cross-country data. Journal of Monetary Economics, vol. 34: 143-173.

6] Borensztein, E., De Gregorio, J. and Lee, J-W (1998) How does foreign direct investment affect economic growth? Journal of International Economics, vol. 45: 115-135.

7] Capie, F. (1983) Tariff protection and Economic performance in the Nineteenth century in J. Black and L.A. Winters (eds) Policy and Performance in International Trade, London Macmillan

8] Cernat, L. (2001) Assessing Regional trade arrangements: Are South-South RTAs more trade diverting? Policy issues in international trade and commodities, Study series No. 16, United Nations Conference on Trade and Development (UNCTAD), Geneva, Switzerland

9] Clemens, M. A. and Williamson, J. G. (2002) Why did the tariff-growth correlation reverse after 1950? NBER Working Paper 9181, http://www.nber.org/papers/w9181

10] [10] de la Fuente, A. and Domenech, R. (2000) Human capita in growth regressions, how much difference does data quality make? CSIC Campue de la Universidad Autonomede Barcelona, Mimeo.

11] DeRosa, D.A. (1995) Regional Trading Arrangements among Developing countries: The ASEAN Example. Research Report, International Food Policy Research Institute, No. 103.

12] Dollar, D. (1992) Outward-oriented developing countries do grow more rapidly: evidence from LDCs, 1976-1985. Journal of Development and Cultural Change, vol.40 (3): 523-544.

13] Edwards, S. (1998) Openness, Productivity and Growth: What do We Really Know? Economic Journal, vol. 108: 383-398.

14] Easterly, W. and Levine, R. (1997) Africa’s growth tragedy: Policies and Ethnic divisions. Quarterly Journal of Economics, 112: 1203 – 1250.


15] Elbadwi, I. (1997) The impact of regional trade and monetary schemes on intra-Sub-Saharan Africa trade in Ademola, O, I. Elbadawi and P. Collier, eds., Regional Integration and Trade Liberalisation in Sub-Saharan Africa, Macmillan, Houndmills, Basingstoke, London.

16] Evans, D. (1998) Options for regional integration in Southern Africa, IDS working Paper 94, Sussex: Institute of Development Studies.

17] Flores, R. Jr. (1997) The gains from MERCOSUL: A general equilibrium, Imperfect competition evaluation, Journal of Policy Modeling, 19 (1): 1-18.

18] Ghura, D. and T. J. Grennes (1993) The real exchange rate and macroeconomic performance in Sub-Sahara Africa. Journal of Development Economics, vol. 42 (1): 155-174.

19] Greenaway, D. Morgan, C. W. and Wright, P.W. (2002) Trade Liberalisation and Growth in Developing Countries. Journal of Development Economics, 67: 855-865

20] Grossman, G. and E. Helpman (1991) Innovation and Growth in Global Economy, MIT Press Cambridge, MA.

21] Harrison, A. (1996) Openness and growth. A time series cross-country analysis for developing countries. Journal of Development Economics, vol. 48: 419-447.

22] Im, K. S., Pesaran, M. H and Shin, Y. (1997) Testing for Unit Roots in Heterogeneous Panels. Discussion paper, Department of Applied Economics. University of Cambridge.

23] International Financial Statistics of IMF, various issues 24] Irwin, D. A. (2002) Interpreting the tariff-growth correlation of the late nineteenth century.

NBER Working Paper 25] Lane, P.R. (2001) International trade and economic convergence: the credit channel. Oxford

Economic Paper, vol. 53:221-240 26] Levine, R and Renelt, D. (1992) A Sensitivity Analysis of Cross Country Growth Regression,

American Economic Review, 82(4), 942-63. 27] Levin, A. and Lin, C.F. (1993) Unit Root Tests in Panel Data: New Results. Discussion Paper,

Department of Economics, University of California, San Diego, No. 93-56, p. 35 28] Levin, A., Lin, C.F. and Chu, C. J. (2002) Unit Root Tests in Panel Data: Asymptotic and

Finite-Sample Properties. Journal of Econometrics, 108: 1-24. 29] Li, Xiaoying and Liu, Xiaming (2005) Foreign Direct Investment and Economic Growth: An

Increasingly Endogenous Relationship. World Development, 33(3): 393-407. 30] Lewis, J. D., S. Robinson and Thierfelder, K. (1999) After the negotiations: Assessing the

impact of Free Trade Agreements in Southern Africa, TMD Discussion Paper. Washington: International Food Policy Research Institute.

31] O’ Rourke, K. H. (2000) Tariffs and Growth in late 19th century. Economic Journal, vol. 110:456-483.

32] Park, J, H. (1995) The new regionalism and third world development. Journal of Developing Societies, XI (1): 21-35.

33] Michaely, M. M., Papageorgiou, D. and Choksi, A. (1991) Liberalizing Foreign Trade, Vol.7 Lessons of experiences in the developing World. Oxford and Cambridge, M.A. Basil Blackwell.

34] Nagayasu, J. (1998) Does the long-run PPP hypothesis hold for African evidence from Panel cointegration. IMF Working Paper, WP/98/123

35] Rodrik, D. (2000) Comments on “Trade, Growth and Poverty” by D. Dollar and A Kraay, http://ksghome.harvard.edu/drodrik.academic.ksg/Rodrik

36] [36] Rodriguez, F. and D. Rodrik (2001). “Trade policy and Economic growth: A Skeptic’s guide to the cross national evidence” in Ben Bernanke and Kenneth S. Rogoff (eds) Macroeconomics Annual, 2000, Cambridge, M.A. MIT Press for NBER

37] Romer, P. M (1990) Endogenous technological change. Journal of Political Economy, vol. 98: S71-S102


38] Schiff, M. (1997) Small is beautiful: Preferential trade agreements and the impact of country size, market share and smuggling, Journal of Economic Integration, 12:359-87.

39] Vamvakidis, A. (1999) Regional trade agreement and broad liberalization: which path leads to faster growth? IMF Staff Papers, Vol. 46 (1): 42-68.

40] Viner, J. (1950) The custom Union Issue, New York Caregie Endowment for International Peace.

41] World Bank, (2000) Trade blocs, New York: Oxford University Press. 42] WTO (2002) WTO Annual Report Chapter 3: Overview of Developments in the International

Trading Environment, P. 39. 43] Yanikkaya, H. (2003) Trade Openness and economic growth: a cross-country empirical

investigation. Journal of Development Economics, 72: 57-89. 44] Yeats, A. (1998). What can be expected from African regional trade arrangements? Some

empirical evidence, World Bank mimeo.


Day-of-the-Week Effect, Volatility, and Linkages to Macro-Economic Indicators: Evidence from North American

Financial Markets, 1997-2004

Chiaku Chukwuogor-Ndu Eastern Connecticut State University

Abstract This paper examines the financial markets’ trends such as the annual returns, daily returns and volatility of returns in three North America markets namely S&P/TSX, Canada: IPC GRAL (MXSE), Mexico and S&P 500, US for the period 1997-2004. In the face of recurrent economic crisis in different parts of the world, their contagion effects, and the recession in the US, there is need to evaluate the impact of negative economic developments on the financial markets’ returns and volatility in these countries. A set of parametric and non-parametric tests is used to test the equality of mean returns and standard deviations of the returns across the-days-of-the-week.

I. Introduction This paper examines the financial market’s trends such as the annual returns, daily returns and volatility of returns in three North America markets namely S&P/TSX, Canada; IPC GRAL (MXSE), Mexico and S&P 500, US for the period 1997-2004. In the face of recurrent economic crisis in different parts of the world and their contagion effects, the recession in the US and the present globalized system there is need to determine the current situation with respect to patterns in annual stock returns, daily returns and their volatility for a period. The period 1997 to 2004 is particularly interesting for North America because it encompasses period of economic booms and recessions, financial crises in Asia, Europe and Latin America. A set of parametric and non-parametric tests is used to test the equality of mean returns and standard deviations of the returns across the-days-of-the-week. II. Literature Review The presence of anomalies in returns of common stocks has intrigued researchers since the last century challenging the appropriateness of the Capital Asset Pricing model (CAPM) and the whole theory of market efficiency.

Day-of-the-week-effect in stock returns in the US Market has been documented by a large number of studies. For instance, in the US stock market the mean Monday stock return has been found to be negative or significantly lower than the non-Monday return. Many studies have shown that in addition, mean stock return on Fridays is significantly high relative to other days. See for example Cross (1973), French (1980), Jaff, Westerfield and Ma (1989), Gibbons and Hess (1981), Lakonishok and Levi (1982), Rogalski (1984), Keim and Stambaugh (1984), Harris (1986), Lakonishok and Smidt (1988), Mehdian and Perry (2001), among others document the Monday effect or other daily anomalies in the


US stock market. Jaff and Westerfield (1985a, b) found a negative Monday effect in Canada and the UK but a negative Tuesday effect in Japan and Australia. Condoyanni et al. (1987) confirms these findings on the Japanese and Australian markets. Kato (1990) also finds that the Tuesday return is negative and Wednesday and Saturday returns are strongly positive in Japan. Jaff, Westerfield, and Ma (1989) drew attention to this phenomenon when they provided international evidence. III. Data and Methodology We use the daily closing values of three selected Latin American stock market indices from January 2nd, 1997 to December 31st 2004 to determine the daily returns, day-of-the-week effect and volatility of returns. We further use the average annual returns to depict the annual trends in stock market returns. The financial markets studied are S&P/TSX, Canada; IPC GRAL (MXSE), Mexico and S&P 500, US.

The daily stock returns for these North American stock indices are calculated as follows:

ln(Pt/Pt-1) *100 (1)

Where Pt is the stock index at date t. Except for the returns on Monday, any returns that are preceded by a holiday were excluded. This exclusion was done in previous studies to avoid speculation that observed day-of-the-week-effect could be partially due to these non-trading days. To determine the nature of the volatility of returns, the distributions of daily returns are analyzed using such measures as variance, standard deviations, kurtosis, skewness and coefficient of variation. The results were substantiated by parametric and non-parametric tests.

Since the result of the normality test indicates that the distributions of the returns are non normal, we use the non-parametric test, the Kruskal-Wallis to check for the results on equality on mean returns. The Kruskal-Wallis statistic is as follows:

( ) ( )131

121

2

+−+ ∑

=

nnR

NN

k

j j

J (2)

Where: k = number of samples; nj = number of values in jth sample; N = ∑nj =total number of values; Rj = sum of ranks in the sample when N values are ranked together (the statistic is approximately Chi-square distributed degrees of freedom equal to k-1).

To test for the equality of variance across the days of the week, we employ the Bartlett’s homogeneity test.. The test criterion is as follows (Snedecor and Cochran, 1970)

⎟⎟⎠

⎞⎜⎜⎝

⎛−= ∑

−2

2

lnln issavM (3)

Where, a = the number of samples; v = degree of freedom; 2−

s = asi /2∑ 2js = estimate of the σ2 from sample ; then, the quantity M/C is distributed approximately as a Chi

square distribution with degrees of freedom equal to (a-1). The above test is for the case when all groups have the same degrees of freedom. When the degrees

of freedom differ, as with samples of unequal sizes, the test criterion is as follows as follows:

( ) ( )∑∑ −=−

22

lnln iii svsVM (4)

( )[ ] ⎟⎟⎠

⎞⎜⎜⎝

⎛−

⎭⎬⎫

⎩⎨⎧

−+= ∑ ∑ ii vva

C 1113

11 (5)


where ( ) ∑∑−

iii vsvs /22

2is is an estimate of the σ2 from sample I, a = the number of samples, vi = the degree of freedom of

samples i. The quantity M/C is distributed approximately as a Chi-square with degrees of freedom equal to (a-1). In our case, as we have five weekdays in a week, degrees of freedom are four.

However, as Bartlett’s test of homogeneity of variance is sensitive to non normality in stock return distribution, the Levene’s (1960) test is also employed to check the results on equality of variance. In measuring the variation within a class, Levene’s test uses the average of the absolute deviations instead of the mean square of deviations. This avoidance of squaring makes the test criterion much less sensitive to non-normal distributions (Snedecor and Cochran, 1976). The Levene’s statistic is as follows:

( ) ( ) ( )( ) ⎥

⎦

⎤⎢⎣

⎡−−

⎥⎦

⎤⎢⎣

⎡−−= ∑ ∑∑

= = = 1/..

1 1 1

2.

2. J

JNxDDDDnFJ

j

J

j

n

ijijjj

j

(6)

where ,. jijij MRD −= Rij is the return for week I and weekday j for j =1, 2,…., J and J =5 if the last trading day of the week is a Friday. IV. Empirical Findings A. Annual Returns 1997-2004 During the period 1997 -2004, the MXSE Index, Mexico achieved a phenomenal growth of by 284.35 percent. For the same period, the Canadian S&P/TSX Composite Index grew by 56.16 percent and the US S&P 500 grew by 50.47 percent. The MXSE Index, Mexico displayed great volatility in annual movements during the period. The highest gain of 80.06 was achieved in 1999. Other significant increases were 55.59 percent in 1997 and 65.14 percent in 2004. The declines of -24.28 percent in 1998 and -20.73 in 2000 helped to accentuate the volatility in the index movement already mentioned. The Canadian S&P/TSX Composite Index and the US S&P 500 also achieved significant gains during the period. For example in US S&P 500 gained 29.76 percent in 1998 while the Canadian S&P/TSX Composite Index gained 30.22. However the declines in these indexes overshadowed the gains. For three years in a row the US S&P 500 achieved the following declines -10.14 percent in 2000, -13.04 percent in 2001 and-23.37 in 2002. These declines coincide with the period of US recession. The Canadian S&P/TSX Composite Index declined two years in 2001 by 13.11 percent and in 2002 by 14.03 percent. A regression analysis using the closing index values as dependent and independent variables indicates a high SE Coefficient of 709.0, R-Sq of 60.6% and a P-Value of 0.000 between the Canadian S&P/TSX Composite Index movements and MXSE Index, Mexico. The regression analysis did not indicate any significant relationship between the Canadian stock index movement and the US stock index movement nor between the US stock index movement and the Canadian stock index movement. The annual closing index values, the period percentage changes, and the annual percentage changes are contained in Table 1. Figure 1 illustrates the annual changes in the index annual closing prices. B. Daily Returns Analysis The mean returns for Canada were negative on Tuesday and Wednesday while Mexico showed a negative Monday return and the US a negative Friday return. This observation is consistent with earlier documented evidence regarding Mexican stock market returns but at variance with earlier documented evidence regarding the US and Canadian stock markets returns. In longer period studies of the US S&P 500, a negative Monday return has been consistently observed. This further underscores the problems associated with reliance on research results based on data that encapsulates. Possible reasons for these different results may be the difference in the period of data covered, possible increased efficiency of


the US stock market, and increased action of market participants to take advantage of earlier observed daily anomalies in stock returns. The average daily returns of the three indexes are contained on Tab-2. Table 1: Annual Closing Index Values and Period Percentage Changes

Index/Country/Annual Percentage changes

1996 1997 1998 1999 2000 2001 2002 2003 2004 Period Percentage change in closing values

S&P/TSX Comp Index, Canada

5,921 6,671 6,430 8,374 8,848 7,688 6,610 8,209 9,247 56.16

Annual Percentage changes

12.66 -3.60 30.22 5.67 -13.11 -14.03 24.19 12.64

MXSE IPC GRAL Index, Mexico

3,361 5,229 3,960 7,130 5,652 6,372 6127 7822 12,918

284.35


55.59 -24.28 80.06 -20.73 12.74 -3.85 27.67 65.14

S & P 500, US 741 948 1230 1470 1320 1148 880 975 1115 50.47


27.88 29.76 19.53 -10.14 -13.04 -23.37 10.76 14.37

Figure1. Annual Percentage Changes in Index Closing Prices

-40

-20

0

20

40

60

80

100

1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

Years

Ann

ual P

erce

ntag

e C

hang

e

S&P/TSXComp.Index, Canada

MXSE IPC GRALIndex, Mexico

S&P 500, US


Table 2: Average Daily Returns of the NAFM for the Period January 2nd 1997- December 31st, 2004. Name of Index/ Country Monday Tuesday Wednesday Thursday Friday

S&P/TSX, Canada 0.0430

-0.0092

-0.0219

0.00130

0.0382

MXSE, Mexico -0.0284

0.0642

0.0722

0.0258

0.0061

S&P 500, US 0.0163

0.0276

0.0050

0.0058

-0.0025

The highest mean return and standard deviation occurred for Canada on Monday and Wednesday respectively; for Mexico on Wednesday and Thursday respectively; and for the US on Tuesday and Monday respectively. The implication of this observation is that investors can take advantage of these observed possible higher returns to dispose of their stock and take advantage of these anomalies. For example, liquation of $1,000,000.00 in the highest return days for the three markets will yield the investor, $72, 200.00 in the Mexican market, $29, 500.00 in the US market and $43, 000.00 in the Canadian markets. It is obvious that the highest return in nominal terms is from the Mexican market. For an investor who is not a resident of any of these countries, it might be necessary to factor in the possible exchange rate and tax implications on funds remittance. Conversely investors can also make significant gains by purchasing on the days with lowest returns which signifies that stock prices are generally down. Whatever the case may be , this observation is a further challenge to the appropriateness of the capital market theory. Table 3 contains a summary of maximum/minimum returns/standard deviations of the NAFM for the Period January 2nd 1997- December 31st, 2004. Table 3: Summary of Maximum/Minimum Returns/Standard Deviations of the NAFM for the Period January 2nd 1997- December 31st, 2004.

Name of Index/ Country

Maximum return/Std. Dev.

Day of Occurrence

Minimum Return/ Std. Dev.

Day of Occurrence

S&P/TSX, Canada 0.043* 0.4903**

Monday Wednesday

-0.0219* 0.4452**

Wednesday Thursday

MXSE, Mexico 0.0722* 0.7706**

Wednesday/ Thursday

-0.0284* 0.6577**

Monday/Friday

S&P 500, US 0.0276* 0.0295**

Tuesday/ Monday

-0.00252* 0.0252**

Friday/ Wednesday

* Mean returns ** Standard deviation of returns

To test the day of the week effect using the Kruskal-Wallis test, the following null and alternate hypotheses are tested for each market.

Ho: There is no difference in the returns across the days of the week; H1: There a difference in the returns across the days of the week. If the null hypothesis is rejected, this means that there is a day-of-the-week effect. According to the result of the Kruskal-Wallis test, there is evidence of the day of the week effect in

Canada and Mexico but not in the US. This finding is also at variance with past documented regarding the US. For example Cross (1973), French (1980) and Jaff, Westerfield and Ma (1989), Gibbons and Hess (1981), Lakonishok and Levi (1982), Rogalski (1984), Keim and Stambaugh (1984), Harris


(1986), Lakonishok and Smidt (1988), Mehdian and Perry (2001), among others document the Monday effect or other daily anomalies in the US stock market. V. Volatility of Returns The daily returns were mostly skewed to the left for Canada and Mexico and mostly skewed to the right for the US. There was generally reduced variation and kurtosis. The S&P 500 had the lowest standard deviations of returns and the MXSE, Mexico had the highest standard deviations of returns. The highest standard deviations of 0.4903 for Canada occurred on Wednesday, 0.7706 occurred on Thursday and the highest standard deviation of 0295 for the US occurred on Monday. There was generally higher kurtosis for the Mexican and Canadian markets and lower kurtosis for the US market except on Monday. Even though the MXSE, Mexico had the highest standard deviations for all the days of the week, S&P/TSX, Canada and S&P 500, US showed the highest coefficients of variation of 34202.3 on Thursday for Canada coinciding with the day of lowest standard deviation and -21058.52 on Friday for the US coinciding with the day of lowest return. The highest daily coefficient of variation for the Mexican MXSE index was 10868.64 which occurred on Friday and this case coinciding with the day of lowest standard deviation. The levels of returns and standard deviations are responsible for these observations. Table 4 contains the standard deviation of returns and the analysis is supported by the basic statistics of the three indexes contained in the Appendix. Table 4: Standard deviations of Returns of the NAFM for the Period January 2nd 1997- December 31st, 2004.

The results of the Levene’s (1960) test of the equality standard deviations of the returns at the 5 percent confidence level could not reject the Null Hypothesis that mean returns are equal across the days of the week for all the North American markets. The results of the Levene’s test are contained on Table 5. Table 5: Results of Test of Normality, Equality of Means/Variance across Day-of-the-Week Effect for the Period January 2nd, 1997-December 31st, 2004.

Name of Index/Country

Kruskal-Wallis W Test for Normality Levene’s Test

Bartlett's Test

Chi-square P Value Statistics R Statistics P Value Statistics P Value

S&P/TSX Canada

12.97 .011 .4694 0.972 0.27 0.897 4.97 0.291

MXSE, Mexico

3.17 0.530 0.7227 0.966 1.17 0.323 9.84 0.043

S&P 500, US

1.68 0.794 0.5403 0.988 0.39 0.819 6.44 0.169

Name of Index/ Country Monday Tuesday Wednesday Thursday Friday

S&P/TSX, Canada 0.4547

0.4789

0.4903

0.4452

0.4743

MXSE, Mexico 0.7344

0.7169

0.7272

0.7706

0.6577

S&P 500, US 0.0295

0.0273

0.0252

0.0265

0.0265


VI. Conclusion We find that the MXSE Index, Mexico achieved a phenomenal growth of by 284.35 percent, the Canadian S&P/TSX Composite Index grew by 56.16 percent and the US S&P 500 grew by 50.47 percent. The MXSE Index, Mexico displayed great volatility in annual movements during the period. For three years in a row the, US S&P 500 suffered the following declines -10.14 percent in 2000, -13.04 percent in 2001 and

-23.37 in 2002. These declines coincide with the period of US recession. The result of a regression analysis using the closing index values as dependent and independent variables indicates a high SE Coefficient of 709.0, R-Sq of 60.6% between the Canadian S&P/TSX Composite Index and The MXSE Index, Mexico. This relationship showed a high significant P-value.

The mean returns for Canada were negative on Tuesday and Wednesday while Mexico showed a negative Monday return and the US a negative Friday return. This observation is consistent with earlier documented evidence regarding Mexican stock market returns but at variance with earlier documented evidence regarding the US and Canadian stock markets returns.

According to the result of the Kruskal-Wallis test, there is evidence of the day of the week effect in Canada and Mexico but not in the US. This finding is also at variance with past documented regarding the US. For example Cross (1973), French (1980) and Jaff, Westerfield and Ma (1989), Gibbons and Hess (1981), Lakonishok and Levi (1982), Rogalski (1984), Keim and Stambaugh (1984), Harris (1986), Lakonishok and Smidt (1988), Mehdian and Perry (2001), among others document the Monday effect or other daily anomalies in the US stock market.

The daily returns were mostly skewed to the left for Canada and Mexico and mostly skewed to the right for the US. There was generally reduced variation and kurtosis. There was generally higher kurtosis for the Mexican and Canadian markets and lower kurtosis for the US market except on Monday. Even though the MXSE, Mexico had the highest standard deviations for all the days of the week, S&P/TSX, Canada and S&P 500, US showed the highest coefficients of variation of 34202.3 on Thursday for Canada coinciding with the day of lowest standard deviation and -21058.52 on Friday for the US coinciding with the day of lowest return. The results of the Levene’s (1960) test of the equality standard deviations of the returns at the 5 percent confidence level could not reject the null Hypothesis that mean returns are equal across the days of the week for all the North American markets

VII. References 1] 1.F. Cross, (1973) The behavior of stock price on Fridays and Mondays, Financial Analyst

Journal, 29, 67-69. 2] 2.k. French, (1980) Stock returns and the week-end effect, Journal of Financial Economics, 8,

55-70. 3] 3.R.S. Gibbons, and P. Hess, (1981), Day of the Week Effects and Asset Return, Journal of

Business, 54, 579-96. 4] 4.L. Harris, (1986), A transaction Data Study of Weekly and Intra daily Patterns in Stock

Returns, Journal of Financial Economics, 16, 99-117. 5] 5.J. Jaff, and R. Westerfield (1985a), The week-end effect in common stock return: the

international evidence, Journal of Finance, 40, 433-54. 6] 6. , (1985b), Patterns in Japanese common stock returns: day of the week

and turn of the year effect, Journal of Financial Quantitative Analysis, 20, 261-72 7] 7.J. Jaff, R.L. Westerfield, and C. Ma (1989), A twist on the Monday effect in stock prices:

Evidence from the U.S. and foreign stock markets, Journal of Banking and Finance, 13, 641-50. 8] 8.K. Kato, (1990), Weekly patterns in Japanese stock returns, Management Science, 36, 1031-

43. 9] 9.J. Lakonishok, and M. Levi, (1982), Week-end effects on stock returns: a note, Journal of

Finance, 37, 883-89.


10] 10. and S. Smidt, (1988), Are seasonal anomalies real? A ninety-year perspective, Review of Financial Studies, 1, 403-25.

11] 11.S. Mehdian, S. and M. Perry, (2001), The reversal of the Monday effect: new evidence from US equity markets, Journal of Business Finance and Accounting, 28, 1043-1066.

12] 12.R. Rogalski, (1984), New findings regarding day of the week returns over trading and non-trading period, Journal of Finance, 39, 1603-14.

VIII. Appendix Table 1: Daily Returns: Basic Statistics, Canada S&P/TSX

Description Monday Tuesday Wednesday Thursday Friday

Mean 0.0430 -0.00917 -0.0219 0.00130 0.0382

Median 0.0909 0.00470 -0.0254 0.0316 0.0584

Maximum 1.4917 1.7772 1.6998 2.0340 1.7726

Minimum -2.7680 -1.5953 -3.6766 -2.7056 -2.8734

Standard Dev. 0.4547 0.4789 0.4903 0.4452 0.4743

Skewness -1.07 0.06 -0.84 -0.49 -0.91

Kurtosis 4.49 2.41 8.15 4.70 6.14

Variance 0.2068 0.2293 0.2404 0.1982 0.2250

Coefficient of variation 1058.48 -5223.22 -2243.01 34202.3 1241.22

Observation 376 407 411 410 403

Table 2: Daily Returns: Basic Statistics, Mexico (MXSE)

Description Monday Tuesday Wednesday Thursday Friday

Mean -0.0284 0.0642 0.0722 0.0258 0.00605

Median 0.0000 0.0295 0.0620 0.0493 0.0478

Maximum 2.5786 5.2783 3.7746 2.7565 3.2543

Minimum -6.2167 -2.5424 -2.2224 -4.4910 -3.5905

Standard Dev. 0.7344 0.7169 0.7272 0.7706 0.6577

Skewness -1.53 1.71 0.44 -0.59 -0.42

Kurtosis 13.72 11.93 2.64 4.19 4.13

Variance 0.5393 0.5140 0.5289 0.5938 0.4326

Coefficient of variation -2582.11 1116.04 1007.24 2985.89 10868.64

Observations 381 390 390 383 378


Table 3: Daily Returns: Basic Statistics, US (S&P 500) Description Monday Tuesday Wednesday Thursday Friday

Mean 0.0163 0.0276 0.00572 0.00495 -0.00252

Median 0.0284 0.0258 0.0169 0.00863 0.00925

Maximum 2.2873 2.1671 2.4204 2.0211 1.6640

Minimum -3.0895 -1.8424 -1.5006 -1.6992 -2.6077

Standard Dev. 0.0295 0.0273 0.0252 0.0265 0.0265

Skewness -0.92 0.30 0.50 0.24 -0.35

Kurtosis 5.03 1.69 2.02 1.19 1.63

Variance 0.3311 0.3079 0.2630 0.2817 0.2809

Coefficient of variation

3519.56 2008.11 8967.43 10722.8 -21058.52

Observations 381 412 413 402 401


Governance, Taxation and Fiscal Policy in Nigeria

Dr. Olu Okotoni∗ Department of Public Administration

Obafemi Awolowo University Ile-Ife, Nigeria

E-mail: [email protected] [email protected]

Abstract

The paper examines governance, taxation and fiscal policy in Nigeria. It identifies that fiscal policy remains one of the greatest threats to political stability in Nigeria because the colonial architects of Nigeria laid the foundation on fiscal convenience, rather than on mutual trust, cultural and historical affinity. The article also establishes that governance and taxation exist side by side since taxes are needed to pay government's bills, but citizens see taxes as means of depriving them part of their income and therefore constitute potential source(s) of conflicts. It is further noted that the incursion of the military into Nigeria's politics since 1966 has aggravated the governance crises rather than addressed them and that while successive military administrations have succeeded in breaking up the federal structure into more sub-national units, they have failed in all other areas of governance – transparency and accountability, economic development, fiscal policy and democratisation. The paper also notes that Nigeria has not only failed to evolve a successful fiscal policy, but its fiscal federalism from 1914 to date portends a bleak scenario since most efforts have been geared toward sharing rather than mobilizing of resources, leading to a monolithic economy. The paper recommends that proper accountability, transparency and honesty are necessary for sustaining democratic governance in Nigeria and that all existing and potential sources of revenue be maximally explored. The article concludes that an effective management of the symbiotic relationship between governance and taxation is capable of producing good fiscal policy for the country.

I. Preamble Governance carries a variety of definitions. The purpose sometimes determines the nature, scope and perspective of its definition. The World Bank (1992:1) for example views governance from power perspective, defining it as "the manner in which power is exercised in the management of a country's economic and social resources for development.” Mohammed Fofana (1997) who takes an economic approach sees governance as "the process by which a society makes decision about the production and distribution of scarce resources." To William Zartman (1997:1) governance is conflict management.

∗ Dr. Olu Okotoni is a Senior Lecturer in the Department of Public Administration, Obafemi Awolowo University, Ile-Ife, Nigeria. He is the currently the Acting Head of Department and the immediate past vice dean of the Faculty of Administration.


Landell-Mills and Serageldin (1991:304)1 define governance as "how people are ruled, how the affairs of a state are administered and regulated". Goran Hyden (1992) equates governance with regime, which he defines as "the conscious management of regime structures with a view to enhancing the legitimacy of the public realms" (1992:7). Rosenau (1992:5) who views governance from a political angle argues that governance is not synonymous with government, stating that "governance is a system of rule that works only if it is accepted by the majority (or at least, by the most powerful of those it affects, whereas governments can function even in the face of widespread opposition to their policies." Nazrul Islam and Om Prakash Mathur (1995:3) also view governance from political perspective state that "governance broadly refers to the system of government concentrating on effective and accountable institution, democratic principles and electoral processes, representative and responsible structures of government in order to ensure an open and legitimate relationship between the civil society and the state." Trevor Gordon-Somers’ (1997:133) attempt appears like a summary:

The legitimate exercise of political, economic and administrative authority in the management of a country's affairs, at all levels. It comprises the complex mechanisms, processes and institutions through which citizens and groups articulate their interests, exercise their legal rights and obligations and mediate their differences. Management of development, in terms of policy formulation, resource allocation and balancing of economic interests, is crucial". … Sound governance is therefore participatory, transparent, accountable, effective, equitable and promotes the rule of law.

From the various definitions, governance is viewed within the narrow context of state; whereas the

term governance covers a wider range such as civil society, non-governmental organizations, community affairs etc.

Governance and taxation exist side by side. Taxation which is an integral part of governance constitutes a potential source of conflict. Taxes are needed to pay government's bills, but citizens see taxes as means of depriving them part of their income and resources. The Peasants Revolt in the fourteenth century was caused by the introduction of a poll tax; “and throughout history, unfair or seemingly unfair systems of taxation have been at the heart of many such conflicts.” (Nightingale, 2001). One of the fundamental causes of the American War for Independence was the rebellion of British colonists against taxes that were perceived as inequitable.2 (Pechman,1985). In Nigeria, property tax is seen as unjust and in fact, culturally unacceptable as it seems to penalize people for owning property. Thus, this tax which has the greatest potential for local government revenue generation cannot be fully tapped. In many instances, tax payers cannot justify the continuous payment of some certain rates and taxes, as there are no corresponding services from government.

Arbitrary increase in taxation triggers resistance from tax payers. This was the case of the 1968 civil disturbances in the Western State of Nigeria (now Oyo, Ogun, Ondo, Osun and Ekiti states), occasioned by what was considered as an outrageous increase of tax rates in the state. As a result, a group of farmers known as Agbekoya (meaning Farmers against Oppression) revolted against the arbitrary increase in taxes and levies. This resulted in a bloody clash that lasted for several months before the revolt was quelled. The following extracts from the Drum Magazine (June 1969) provides an insight into the story:

1Cited from Nazrul Islam and Om Prakash Mathur (1995) "Urban Governance in Asia" in Urban Governance (Regional paper presented at the Second Urban Forum: November 27-29, 1995 at the UN Office, Nairobi, pg 1. 2For example, a "protest Congress gathered in 1765 in reaction to the 1763 Stamp Tax for raising revenue among the colonies. The Congress asserted such fundamental principles as the entitlements of colonists to the 'inherent rights and liberties' of British subjects. Henry J. Merry (1986) The Constitutional System: The Group Character of the Elected Institutions New York, Praeger Publishers, page 20.


The fracas began in Ibadan towards the beginning of December 1968 when a mammoth crowd of tax agitators marched towards Mapo hall singing a war song: Oke mefa l'ao san. Oke mefa l'ao san. Bi o ba gba kumo, a o gbo'ri bibe. Oke mefa l'ao san. This means in English: "We are paying only thirty shillings. If this cannot be achieved by the application of cudgel, it will be by the cutting off of certain heads. We are paying only thirty shillings.

At the time of this praetorian threat of carnage and murder, income tax (poll tax) in the state had risen from #1.17.6d to #3 including a state development fund of 7s.6d. In addition, the agitators were expected to pay 10 shillings national reconstruction fund, an average of 30 shillings water rate, all adding to #9 (Nine pounds) in certain parts of the state.

The military governor, Brigadier Adebayo, was forced to succumb to some of their demands. He

immediately announced a cut of five shillings from every item under taxation and also warned civil servants in the state "not to make the people pay for services which are not yet available."

There is the problem of perception. Many citizens believe that governance is about sharing rather than mobilizing resources. They expect government to provide all social services and facilities free of charge. The era of oil boom in Nigeria strengthened this idea that government have enough resources to provide free services and cater for government's expenditure. Thus, taxes and rates are seen as unnecessary burdens from government. Poor sharing of national resources and bad fiscal policy also constitute potential sources of conflict. This is the case of the oil producing areas in Nigeria (especially the Ogonis) that are aggrieved over the sharing and distribution of mineral resources extracted from their land. Inter-governmental transfers of fiscal resources among the various levels of government and jurisdictional tax powers constitute another problem in governance. In summary, governance, taxation, fiscal policy and conflict are interrelated. A poor management of one constitutes a potential source of crisis. The dynamics of their interrelationship is the pre-occupation of this paper. II. Research Problem Fiscal policy remains one of the greatest threats to political stability in Nigeria. Fiscal policy and taxation are central issues in Nigeria’s governance arena since 1914. In the first instance, one of the principal reasons why the Northern and the Southern Protectorates were amalgamated was to make up from the fiscal surplus of the south for the rather perpetual fiscal deficit of the north. Peter Ekeh (1997:40) explained, "Frugal British administration were not pleased that although Southern Nigeria could pay for its own administration, Northern Nigeria's administrative costs were subsidized from London. It was for the sake of achieving economic balance that the amalgamation of the two separate colonies of Southern and Northern Nigeria was effected in 1914". The colonial architects of Nigeria laid the foundation on fiscal convenience, rather than on mutual trust, cultural and historical affinity. To Ogundowole (1994:ix),

The amalgamation of the peoples of Nigeria into a federation devoid of nationality formations is in itself denationalization par excellence, i.e. taking away the very souls of the peoples of the nationalities that constitute Nigeria, while the federation represents nothing more than the amalgamation of the lifeless ghosts of the nationalities.

Okoth-Ogendo (1996:53) corroborates this point, "the state in Africa at independence was not a

constitutional state", but a constituted state; "one erected on pillars that were not fashioned out of past experiences or future aspirations"; and argues that "as constituted, the machinery of the state was not part of the shared experience of the African people." Since Nigeria’s foundation was laid on fiscal considerations, any thing that touches on fiscal issues becomes very crucial to its existence. This explains why Nigeria’s fiscal policy has received more attention and public debate than any federal problems since Independence. Other federal matters (such as the creation of states and local governments, federal character, population, census, boundaries, intergovernmental relations,


constitutional jurisdictions, inter-group relations, ethnicity, elections etc) revolve around this. The federal system which was adopted in 1954 gave birth to acrimony, inter-ethnic and inter-group conflicts. A civil war was fought from 1967 to 1970 around the subject of what should constitute an acceptable social, political and economic order.

The incursion of the military into Nigerian politics since 1966 did not help matter. Successive military administrations have aggravated the failures of the Nigerian state rather than address them. The military has ruled the country for more than three decades out of forty-two years of independence in 1960. Under military regimes, federal government arrogated to itself "extraordinary powers which do not constitutionally belong to it". (Ekeh, 1997:7). These include "the seizure of funds and appropriation of assets that should belong to local governments, states, and corporate institutions established by the constitution and legislative provisions." (ibid:7). It is against this background the paper discusses the problems associated with governance, taxation and fiscal policies in Nigeria. III. Objectives of the Study The overall objective of this study is to examine and document some of the most successful strategies for tackling the problems associated with taxation and fiscal policies in Nigeria. Specifically, the study sets out to:

examine the role of taxation and fiscal policy in Nigerian governance; determine the extent the fiscal crisis has affected the Nigeria's federation; examine and document the strategies employed in tackling one and two above; proffer solutions on how to mitigate the problems of fiscal crisis and taxation on governance in

Nigeria, and highlight policy challenges for the next millennium IV. Theoretical and Intellectual Discourse The subject of taxation has received considerable intellectual and theoretical attention in literature. Taxation is one of the most volatile subjects in governance both in developing and developed countries. It is, therefore, no surprise that some of the most familiar quotations in history have been comments by famous men about taxes (Pechman, 1985:1). Benjamin Franklin once remarked, "in this world, nothing is certain but death and taxes." (ibid:1). Chief Justice John Marshall in the United States of America was quoted as saying: "the power to tax involves the power to destroy." The above quotations show clearly the volatility of taxation.

Tax and taxation are used synonymously in this paper. Tax refers to a "compulsory levy by a public authority for which nothing is received directly in return." (James and Nobes, 1992:266). To Kath Nightingale (2001) “a tax is compulsory contribution, imposed by government, and while tax payers may receive nothing identifiable in return for their contribution, they nevertheless have the benefit of living in a relatively educated, healthy and safe society.” Nightingale explains that taxation is “part of the price to be paid for an organized society.” She identified six reasons for taxation: provision of public goods, redistribution of income and wealth, promotion of social and economic welfare, economic stability, harmonization and regulation.

Fiscal policy refers to "government's use of taxation and public expenditure to influence the aggregate level of economic activity." (James and Nobes, 1992:266). According to Jhingan (1996:348) fiscal policy means "the use of taxation, public borrowing, and public expenditure by government for purposes of 'stabilization' or 'development'." Fiscal policy also involves government's measures to control and monitor public finance with a view of achieving some economic goals. The role of fiscal policy varies significantly within various economies. For instance, the role of fiscal policy in advanced economies is to stabilise the rate of growth, whereas in "the context of an underdeveloped economy, the role of fiscal policy is to accelerate the rate of capital formation." (ibid:348). Jhingan adds, "fiscal policy plays a dynamic role in underdeveloped countries" - as its extensive use is indispensable for


economic development. The objectives of fiscal policy are to (i) increase the rate of investment; (ii) encourage socially optimal investment; (iii) increase employment opportunities; (iv) promote economic stability in the face of international instability; (v) counteract inflation; and (vi) increase and redistribute national income. (Jhingan, 1996) The objectives of fiscal policy are quite laudable and are capable of yielding tremendous results if judiciously pursued by any government, but feasible in a stable political arena. The absence of political stability in Nigeria since independence in 1960 has contributed to its inability to embark on fiscal policies that could lead to economic development.

Joseph Schumpeter (1996) addresses the crisis of the tax state. He argues that "the fiscal history of a people is above all an essential part of its general history" and that taxes do not only help to create the state, but also help to form it. Schumpeter's argument appears relevant to Nigeria's political history where fiscal finance forms a dominant part. Nobert EIias (1996:345-6) focuses on the state’s monopoly of violence and explains the inter-relationship between the use of physical violence and taxation. He comments,

The society of what we call the modern age is characterized, above all in the West, by a certain level of monopolization. Free use of military weapons is denied the individual and reserved to a central authority of whatever kind, and likewise the taxation of the property or income of individuals is concentrated in the hands of a central social authority. The financial means thus flowing into this central authority maintain its monopoly of military force, while this in turn maintains the monopoly of taxation.

Many African States, including Nigeria, have abused this monopoly of violence and taxation. In Nigeria, the military has used its weapons not only to harass, but to extort money from innocent and defenceless citizens. In December 1968, "a combined team of police and army opened fire on (tax) agitators" in Western State. At the end of the demonstrations, "it was estimated that more than fifty people had been killed in different parts of the state with several more wounded." (Drum, 1969). This type of state violence has replayed itself times without number in the country.

State monopoly can be viewed from another angle. The federal government enjoys some certain monopoly of taxing powers as contained in the constitution.1 It also enjoys monopoly over the armed forces (the army, the air force and navy), the police and other para-military bodies.Human (1995:20) provides explanation for the State’s monopoly of violence:

States justify the violence they use to this end by saying that they cannot do the good that they do for their societies if they are not given the right of ultimate recourse to violence against those who impede, resist or impair their efforts. To forestall a situation where, in every case, individuals or groups of people resist their actions, states have to employ violent means; states establish their control over society by demanding a monopoly over the means of violence. ... Hence the force used by the state is sometimes referred to as 'legitimate'. Violence used by any one else in society is seen as 'illegitimate.'

Human’s argument appears reasonable and logical provided it is not turned to oppression of

citizens. Some scholars have dealt with specific aspects of public finance in Nigeria. Adedeji (1969) reviews Nigeria’s fiscal policy from 1946 to 1969; Mbanefoh and Anyanwu (1990) deal with the constitutional role of revenue mobilization and allocation among the various tiers of government; and Adesola (1989) examines the machinery for collection and disbursement of public funds in Nigeria with a prescription for the aborted Third Republic. A great deal of work has also been devoted to tax administration in Nigeria. They include (Adesola, 1986; Okele, 1986; Omopariola and Nassar, 1986; Oyelere and Oyewole, 1986; Oribabor, 1986). Others have limited their scope to local government

1See item 39 of the Exclusive Legislative List of the 1999 Constitution of the Federal Republic of Nigeria.


finance. In this category, are Adedeji and Rowland (1971) and Bello-Imam (1989). The intellectual and theoretical discourse has shown that the subject of taxation and fiscal policy has received substantial attention; although not from governance perspective, which is the focus of this paper. V. Nigerian Governance: Historical and Geographical Profile Nigerian governance environment can best be understood within its political, historical and geographical profile. The amalgamation of the Southern and Northern Protectorates and the Lagos Colony gave birth to Nigeria from a motley of nationalities and ethnic groups on January 1, 1914 . In 1954, the country adopted a federal system and attained Independence on October 1960 with Westminster Parliamentary system of government. The federal system has transmuted from a two-tiered federal structure, comprising three unequal regions to a three-tiered federal system of 36 states, one federal capital territory (FCT) and 774 local governments. Nigeria is 924,000 square kilometres with two distinct climatic zones - the tropical south and the hot, dry north. Three main cultural groups dominate, situated in the southwest (Yorubaland), the southeast (Igboland) and the north (Hausaland/Borno). In all, there are over 300 different ethnic groups with as many languages and over 1,000 dialects. Nigeria’s population was 88.9 million in 1991 and estimated as 129.9 million in 2001 by the World Bank. Fiscal Policy and Taxation in Nigeria This section is an overview of federal revenue in Nigeria. Up till late 1960s, custom and excise duties dominated the federal revenue. Table one shows that customs and excise duties topped with 68.1% in 1959/60 and 1969/70; followed by direct tax (9.4%); mining (8.1%); and interests and payments 6.1%. The remaining revenues were not significant as seen in the table. From mid 1970s, oil rents became dominant in Nigeria's economy. It increased from 26.3% in 1970 to 77.4% in 1975 and 81% in 1980. Oil revenue started to decline from early 1980s. In 1981 oil revenue reduced sharply to 64.4%. (CBN, 1997). Despite this, as a monolithic economy, oil revenue remains dominant. In 2001 proceeds from oil accounted for 76.5% of the total federal revenue.

The period preceding 1950 was characterized by a strong central government in fiscal matters. As late as 1951, the federal government took almost 100% of the total revenue that accrued to the nation's accounts with almost no share for the regional governments. This trend changed in 1954 when the recommendations of Hicks-Philipson Commission of 1951 which gave regional governments considerable fiscal powers were implemented. The 1951 Constitution embraced the recommendations and for the first time, gave the regional governments some measure of autonomy on revenues and tax jurisdictions. Regions were to have shares from all revenues accruing to the centre from direct and indirect taxes. Consequently, in 1954, regional governments had a share of 28.6% of the total federal revenue, while the federal government retained 71.4%. In 1959, the revenue share of regional governments slightly dropped to 21.8%, while that of the federal government witnessed an upward review (78.2%). This was occasioned by the fiscal revision commission of 1958 headed by Jeremy Raisman. The Commission emphasised the principle of population rather than that of derivation, which was earlier emphasized by the Chick's Commission. The regional share of federal revenues witnessed a slight rise in 1965 and 1975 to 23.6% and 26.4% respectively, while that of the federal government fell to 76.4% and 73.6% in 1965 and 1975 respectively.

Between 1961 and 1969, certain political developments occurred which necessitated adjustments in the existing federal fiscal arrangements. These included the loss of Southern Cameroon in 1961, the creation of the Midwest Region in 1963, and the Military Decree No. 15 of 27th May, 1967 which balkanised the four regions into 12 states. The Binn's recommendations which formed the basis for the working of federal finance until 1967 emphasised the principle of fiscal needs. To redress the problems that emanated from this, Dina Commission was set up in July 1968.1 The committee submitted its

1Dina Commission is significant because members of the previous commissions (1954, 1958 and 1964) were foreigners drawn from the United Kingdom or Australia; the 1968 commission consisted of Nigerians (See Awa 1976:70).


report in February 1969 and pointed out the problem of great imbalance in economic development among the various states of the federation, and accordingly made strong recommendations that fiscal needs should form the basis of revenue sharing among the various states of the federation. Revenue sharing in the country has not been without trauma, confusion, strife and conflicts. Several ad hoc commissions and committees were set up by government to address the problems. There were the following ad hoc Fiscal Review Commissions between 1946 and 1969:

Philipson Commission (1946) Hicks Philipson (1951) Chicks Commission (1953) Raisman Commission (1959) Binns Commission (1964) Dina Committee (1968)

The major achievements of these early fiscal policies were summarised by Danjuma1 (1992) as follows:

the establishment of regional/state government autonomy over certain revenue sources (e.g. personal income tax);

the establishment of federal government exclusive control of some revenue sources (e.g. armed forces income tax);

the creation of a Distributable Pool Account into which other revenues (including import and export taxes, mining rents and royalties, etc) were paid and which was subsequently distributed between the federal, and region/state governments;

the development of revenue allocation principles, such as derivation, population, even development, etc on the basis of which funds in the Distributable Pool Account were shared among regions/states.

Between 1977 and 1984 a more comprehensive federal fiscal arrangements were put in place. Some of the fiscal arrangements during this period include:

Aboyade Committee (1977) Okigbo Commission (1980) Allocation of Revenue Act (1981) Allocation of Revenue Amendment Decree (1984)

As part of the transition programme to the Second Republic, General Olusegun Obasanjo Military administration appointed the Aboyade Technical Committee on Revenue Allocation in 1977 to review the inter-governmental tax jurisdiction and revenue allocation arrangement in order to facilitate better efficiency in the working of fiscal federalism. The Committee recommended that all federally-collected revenue (except the personal income tax of the Armed Forces, External Affairs Officers and Federal Capital Territory) be consolidated into one account, which would be shared by the federal, states and local governments, using the following percentages:

Federal Government - 57 State Government - 30 Local Government - 10 Special Grants - 03 Total - 100

In addition to the 10% share for local governments, each state was to contribute 10% of its total

revenue to the share of its constituent local governments. This was subsequently reviewed to 10% of state's internally generated revenue. Most states of the federation never honoured this arrangement. The special grants account was earmarked to deal with problems requiring special provision, such as oil

1 Lt. General Theophilus Y. Danjuma (Rtd) was the Chairman, National Revenue Mobilisation, Allocation and Fiscal Commission.


pollution, general national ecological degradation, national emergencies and disasters. Aboyade's report was later jettisoned as it was regarded to be too technical for practical implementation. (Danjuma, 1992). Okigbo Committee was set up by President Shehu Shagari on November 21, 1979, barely a month of assuming office on October 1, 1979. The Committee which put into consideration the twin objectives of equity and efficiency recommended that the federal account be shared, using the following percentages:

Federal Government - 53 State Government - 30

Local Government - 10 Special Fund to be distributed as follows: Initial development of FCT - 2.5 Mineral producing areas - 2.0 Ecological and other disaster - 1.5 Total - 100 The Government White Paper on Okigbo's Committee Report amended the recommendations and

came out with the following percentages: Federal Government (55%); State Government (30%); Local Government (8%); Development of Federal Capital Territory (2.5%); Mineral Producing States (Derivation) (2%); Development of Mineral producing areas (1.5%); General Ecological Problems (1%). The Revenue Act was enacted in 1981 and it upheld most of the recommendations of Government White Paper, except in few areas. The State had an upward review from 30% to 30.5%; and also Local Government from 8% to 10%; while the development of Federal Capital Territory was totally scraped. An Amendment was made in 1984 which mainly affected the State with an upward review from 30.5% to 32.5%.

The federal fiscal arrangement in the country did not witness any major change until in 1988 when "The National Revenue Mobilization, Allocation and Fiscal Commission" was established by Decree No. 49 of 1989 by General Ibrahim Babangida, as part of the transition programme to civilian rule. The commission which was headed by Lt. General T.Y. Danjuma (Rtd) drew most of its members from the academics. At the inauguration on September 6, 1988, the commission had nine members.1 Later, the composition of membership was changed to have a chairman and one member from each state of the federation as a representative. Unlike in the past when ad hoc committees/commissions were set up to review fiscal arrangements, the Commission which was more of a permanent institution was charged with the following responsibilities:

systematic design and effective mobilisation of all sources of public sector revenues; periodic review of the revenue allocation principles and formulae such that would minimise

short-term political pressure; prescription and application of revenue allocation formulae after due approval by the federal

government for the purpose of sharing the Federation Account between the federal, state and local governments;

monitoring the accruals and disbursement of revenue from the Federal Account, the States Joint Account, the Local Government Joint Account, the various Special Purposes Accounts and such other Accounts that may from time to time be established or designated by the Commission with the approval of the federal government;

ensure full compliance with established revenue sharing arrangements as well as full public accountability for all funds so allocated to various governments and/or agencies involved in the disposition of the Federation Account;

liaise with the National Planning Commission and similar statutory bodies in the orderly fiscal development of each other tier of government;

1The members included: Lt. General T.Y. Danjuma ((Rtd) - Chairman, Dr. A.S. Abam - Secretary, Professor M.O. Kayode, Professor Chike Obi, Professor G.I. Osayinwese, Professor U. Damachi, Mrs. O. Olakunrin, Alhaji A. Abdullahi, Alhaji Umaru Mutallab.


collaborate with all layers of government as well as their ministries, departments, agencies, and extra-ministerial units in the prompt, regular, and faithful production of public financial statistics;

determination of the remuneration which it may deem appropriate for political office holders such as members of the executive and legislative branches of government outside the Consolidated Account;

commissioning, undertaking or sponsoring studies, analyses and deliberations on subjects which may bear directly or impinge significantly on the policy and operation domains of federal fiscal system and inter-governmental financial relations;

making whatsoever general or specific recommendations as the Commission may consider necessary for more effective mobilisation, collection, allocation and distribution of federal, state and local government revenues, as well as providing guidelines for their efficient implementation; and

submitting regular and timely annual reports to the federal government on its general activities over and beyond its specific recommendations or ad hoc submissions on particular subjects, with such annual reports also incorporating the Commission's audited accounts.

The Commission came up with its first major recommendations on a new allocation formula which was approved by the Armed Forces Ruling Council (AFRC) and became operational with effect from January 1990. See Table One. Appraising the commission, it has not performed creditably in mobilising resources which was its primary assignment. Its major pre-occupation was revenue sharing which led its reneging on its other functions.

The last major fiscal review was in 1992, when the Federal Government increased the share of the Local Government from Federation Account from 15% to 20%, and reduced that the States from 30% to 25%. The arrangement was occasioned by the transferring of the management and funding of primary education to local government with effect from January 1991. In June 1992, a new revenue allocation formula was announced that further reduced the states' share from 25% to 24% but doubled the general ecology fund and the development of mineral producing areas fund from 1% to 2% and 1.5% to 3% respectively. Table Four shows actual federal fiscal transfers to state and local governments in Nigeria between 1976 and 2001. The fiscal transfers to state governments during the period under review have not been consistent. It ranges from 7.8% (the lowest) in 1996 and 30.5% (the highest) in 1982. On the other hand, federal actual transfers to local governments have increased progressively from 1.7% 1976 to 16% in 1990 and 20% in 1998. The regime of General Babangida administration committed huge financial resources to local government in the country. From 1991, actual federal transfers to local governments began to witness a downward trend - from 16% in 1990 to 3.5% in 1997.

The experience of fiscal policies in Nigeria portends not only a dangerous trend, but seems threatening and disappointing, since the emphasis has always been on revenue sharing, rather than resource mobilization. The most threatening aspect of Nigerian fiscal federalism is its dependence syndrome on oil rents. State and local governments across the country depend largely on federal financial transfers. Their internal revenues have continued to dwindle unabated. Statistics shows that federal transfers to local governments constitute over 90% of their total revenues. Taxation has virtually been neglected in the country. Value-Added Tax (VAT) remains the only active tax apart from Pay as you earn (PAYE) deducted from workers. VAT was introduced in 1994. Its contribution to federal revenue was 3.6% in 1994, with progressive increase to 4.6% in 1995; 5.9% in 1996 and 7.9% in 1999, but decreased to 3.0% in Year 2000 and slightly moved to 4.1% in Year 2001. A viable tax like the property rate is neglected for lack of political will to put appropriate legislations in place for local governments. Existing tax laws in the country are rather obsolete and cannot face the challenges of the 21st century except they are reviewed.


VI. Conclusions and Conjectures Nigerian fiscal federalism from 1914 to date portends a bleak scenario. Most efforts have been geared toward sharing rather than mobilizing of resources. The rents from oil have been wrongly appropriated, looted and squandered by successive governments coupled with lack of transparency and accountability from leaders in the three tiers of government. Public enterprises and corporations are badly managed. Services such as electricity, telecommunication, postal, water, health and education are epileptic and erratic despite large subventions from government. The economy remains a monolithic one, resulting from lack of investment of the oil monies in other sectors (such as agriculture, iron and steel industry, technology etc), which are capable of yielding revenues for the country. Consequently, all tiers of government have consistently depended on oil rents since the 1970s.

The incursion of the military into Nigeria's politics since 1966 has aggravated the governance crises rather than addressed them. While successive military administrations have succeeded in breaking up the federal structure into more sub-national units, they have failed in all other areas of governance – transparency and accountability, economic development, fiscal policy and democratisation. The 1979, 1989, 19951 and 1999 Constitutions designed by the Military clearly reflected its bias for centralization. The constitutional status of state and local governments is made subordinate to the federal government in many respects. Their taxing powers are not only limited but are encroached upon by the federal government at will. The federal exclusive legislative list of the 1999 Constitution contains 68 items, whereas the concurrent legislative list (federal and state legislative powers) has only 30 items. Local government is conspicuously absent from the list. Local government taxing powers are determined by the State House of Assembly. During military regimes, local governments are placed under the Office of the Chief of General Staff (CGS) at the federal level and supervised by the Governor's Office or Deputy Governor's Office at the state level. A review of the Constitution is required to reflect a true federal system as practised in advanced countries such as the United States and Canada.

The general misconception of governance by citizens to mean sharing and looting of national resources complicates the matter. Revenues from oil are seen as national cake meant to be shared. A new orientation of governance is required for both rulers and the ruled. Similarly, a re-definition of Nigeria federal system is a sine-qua-non as one way of addressing the governance crises and fiscal problems. Some have advocated a weak centre and strong units as a means of curbing the excesses of the federal government. Taxing powers for States and local governments with appropriate legislative backings are equally necessary to facilitate their autonomy within the federation. For effective federal system, the various units comprising the federation must be allowed to control some of their affairs in their own way with their own resources. (Awa, 1976:64). The National Revenue Mobilization, Allocation and Fiscal Commission needs to be resuscitated and centres Its activities on revenue mobilization rather than distribution of fiscal resources. Composition of memberships of the Commission should reflect all stake-holders including communities where resources are extracted. Proper accountability, transparency and honesty are necessary for sustaining democratic governance in Nigeria. It is reiterated that all existing and potential sources of revenue must be maximally explored and managed. In conclusion, an effective management of the symbiotic relationship between governance and taxation is capable of producing good fiscal policy.

1The 1995 Constitution designed by late General Sani Abacha military administration has been jettisoned. That of 1989 was not allowed to operate before the Third Republic was aborted in 1993.


Appendixes Table 1: Average Contribution to Federal Revenue 1959/60-1969/70

Source % 1. Customs and Excise 68.1 2. Direct Taxes 9.4 3. Licences and internal revenue 0.8 4. Mining 8.1 5. Earnings and sales 0.9 6. Rent of Government Property 0.3 7. Interests and Payments 6.1% 8. Reimbursements 0.9 9. Miscellaneous 3.4

Source: F.S. Idachaba "Statistical Evidence on Tax Revenue Instability in Nigeria in Quarterly Journal of Administration Volume IX, Number Three, April 1975. Table 2: Federal Revenue: 1970-2001

Year 1970 1975 1980 1985 1990 1995 1998 2001 Oil Revenue 26.3 77.4 81.0 72.6 73.3 70.5 62.4 76.5 Non-oil Revenue 73.7 22.6 19.0 27.4 26.7 29.5 37.6 33.5 Company Income Tax 7.2 4.7 3.8 6.7 3.0 4.8 7.2 3.1 Customs & Excise 58.3 13.8 11.9 13.7 8.8 8.1 12.4 7.6 Value Added Tax - - - - 8.1 7.9 4.1 Privatisation Proceeds - - - - - - - 3.5 FG Indep Rev* 8.1 4.0 3.2 6.2 1.8 4.4 2.5 1.2 Education Tax - - - - - - - 0.7 Others** - - - - 13.1 7.6 7.5 1.1

Sources: 1. Central Bank of Nigeria (1997) Statistical Bulletin Volume 8, Number 1, June 1997 2. Central Bank of Nigeria (1998, 2001) Annual Reports and Statements of Accounts * Federal Government Independent revenue comprises revenue from interest payments, rents on government

properties, personal income tax of Armed Forces, Police, External Affairs and federal Capital residents, GSM. ** Others: This item includes drawn-down from Fertiliser Reserves, Customs Levies, Subvention/Grants, and

Sterilised Oil Windfall Proceeds and Grants. Table 3: Vertical Allocation: Commission's

Beneficiaries Commission's Recommendations (in %)

Govt Approval in %)

Federal Government 47 50 State Government 30 30 Local Government 15 15 Special Funds: 8 5 a. FCT (FA) 1 1

• Stabilization (FA) 0.5 0.5 • Savings (FA) 2 0 • Derivation (MR) 2 1 • Development of oil producing areas (MR) 1.5 1.5 • Development of Non-oil producing areas (NOMR) 0.5 0 • General Ecology (FA) 0.5 1

Notes: FA = Federal Account MR = Mineral Revenue NOMR = Non-Oil Mineral Revenue


Source: T.Y. Danjuma "Revenue Sharing and the Political Economy of Federalism" (Paper delivered at the National Conference on Federalism and Nation Building: The Challenges of the Twenty First Century, organized by the National Council on Intergovernmental Relations, 14th -18th December, 1992 at Sheraton Hotel Towers, Abuja), pages 20-21. Table 4: Federal Transfers to State and Local Governments in Nigeria from Federation Account, 1976 - 1998

Fiscal Year State Govts (Naira Million)

% of Federal Revenue

Local Govts (Naira Million)

% of Federal Revenue

1976 n.a. n.a. 100.0 1.7 1977 n.a. n.a. 250.0 4.2 1978 1,37.1 24.0 150.0 2.2 1979 2,642.4 24.2 261.2 2.4 1980 3,776 23.8 352.6 2.2 1981 3,825.6 25.9 1,085.0 7.3 1982 3,095.3 30.5 1,018.7 8.0 1983 E 2,538.8 23.4 996.8 9.1 1984 2,799 25.1 1,061.5 9.5 1985 3,260.8 22.3 1,327.5 8.5 1986 2,843.8 23.1 1,166.9 9.5 1987 6,197.1 24.6 2,117.8 8.4 1988 8,181.3 29.9 2,727.1 10.1 1989 11,502.1 22.8 3,399.3 10.0 1990 13,509.7 20.2 7,780.0 16.0 1991 19,742 19.5 10,199.0 13.0 1992 24,497.3 12.8 15,720.0 11.3 1993 29,363.5 15.2 18,316.4 9.5 1994 29,017.5 14.3 17,321.3 8.5 1995 38,385.2 8.3 17,983.4 3.9 1996 40,619.1 7.8 21,590.6 4.1 1997 50,902.5 8.7 20,443.3 3.5 1998 61,759.5 24.0 51,466.2 20.0 1999 108,214.8 18.6 90,179.2 15.5 2000 248,561.7 19.6 207,146.6 16.4

Source: Central Bank of Nigeria Annual Report and Statement of Accounts (1977-2000). Select Bibliography 1] Adedeji, Adebayo (1969) Nigerian Federal Finance: Its Development, Problems and Prospects

London, Hutchinson Educational 2] Adedeji, Adebayo and L. Rowland (1972) Local Government Finance in Nigeria Ile-Ife,

University of Ife Press 3] Adesola, S.M. (1986) "Evolving a Rational tax structure for Nigeria" The Quarterly Journal of

Administration Volume XX, Nos. One and Two, pp 7-28. 4] Aladejare, E.A. (1986) "The Role of Accounting in Tax Administration" 5] The Quarterly Journal of Administration Volume XX, Nos. One and Two, pp 43-66 6] Aluko, S.A. (1970) "Nigerian Federal Finance - A General Review" The Quarterly Journal of

Administration Volume IV, No. Two, pp 77-82. 7] Anderson, & J.L. Hilley, Financing State and Local Governments Fourth Edition, Washington,

D.C., The Brooking Institutions, 1986. 8] Ashford, Douglase (1982) British Dogmatism and French Pragmatism: Central-Local Policy

Making in the Welfare State, London, George Allen and Unwin.


9] Awa, E.O. (1964) Federal Government in Nigeria, Berkeley and Los Angeles, University of California Press.

10] Awa, E.O. (1976) Issues in Federalism Benin City, Ethiope Publishing House 11] Bello-Imam, I.B. ed (1990) Local Government Finance in Nigeria Ibadan, NISER 12] Central Bank of Nigeria Annual Report and Statement of Accounts (1977-2001). 13] Concord, J. ed (1975) The Failure of the State: On the Distribution of Political and Economic

Power in Europe, Totowa, N.J. Rowman and Little Field. 14] Danjuma, T.Y. "Revenue Sharing and the Political Economy of Federalism" (Paper delivered at

the National Conference on Federalism and Nation Building: The Challenges of the Twenty First Century, organized by the National Council on Intergovernmental Relations, 14th -18th December, 1992 at Sheraton Hotel Towers, Abuja).

15] Due, J.F. and A.F. Friedlaender (1973) Government Finance: Economics of the Public Sector Homewood, Richard D. Irwin Inc.

16] Ekeh, Peter P. ed (1997) Nigerian Federalism New York The Association of Nigerian Scholars for Dialogue.

17] Elias, Nobert (1996) The Civilizing Process: The History of Manners and State Formation and Civilization Oxford, Blackwell.

18] Erero, E.J. and M.O. Okotoni (1998) "Decentralization Programs in Africa: The Nigerian Component", Report presented to the World Bank October, 1998.

19] Federal Republic of Nigeria (1976) Guidelines for Local Government Reform Government Printer, Kaduna.

20] Federal Republic of Nigeria (1979) The Constitution of the Federal Republic of Nigeria 21] Federal Republic of Nigeria (1999) The Constitution of the Federal Republic of Nigeria 22] Federal Republic of Nigeria (1987) Report of the Political Bureau Federal Government Printer,

Lagos. 23] Fofana, M. (1997) Is there any organic Relationship Between Good Governance and

Development" in African Institute for Democracy, April 1997, pp 142-160. 24] Human, Piet (1995) Managing Towards Self-Reliance: Effectiveness of Organizations in Africa

Cape Town, Phoenix Publishing. 25] Hyden, Goran (1992) "Governance and the Study of Politics" in Goran Hyden and Michael

Bratton eds Governance and Politics in Africa Boulder and London Lynne Reinner Publishers, pp 1-26.

26] Idachaba, F.S. (1975) "Statiscal Evidence on Tax Revenue Instability in Nigeria in Quarterly Journal of Administration Volume IX, Number Three, April 1975.

27] Islam, Nazrul and Om Prakash Mathur (1995) "Urban Governance in Asia" in Urban Governance (Regional paper presented at the Second Urban Forum: November 27-29, 1995 at the UN Office, Nairobi.

28] James, S. and Christopher Nobes (1992) The Economics of Taxation Hemel Hempstead, Prentice Hall.

29] Jhingan, M.L. (1996) The Economics of Development and Planning Delhi Konark Publishers PVT Ltd.

30] Mbanefoh, G.F, and J.C. Anyanwu "Revenue Allocation" (1990) The Quarterly Journal of Administration Volume XXIV, No. Three, 166-178.

31] Nightingale, K. (2001) Taxation: Theory and Practice London, Prentice Hall. 32] Nwabueze, B.O. (1964) Constitutional Law of the Nigerian Republic, London, Butterworths. 33] Ogundowole, Kolawole (1994) Colonial Amalgam: Federalism and the National Question: A

Philosophical Examination Lagos, Pumark Nigeria Ltd. 34] Okele, J.B. "Aspects of the Nigerian Tax Laws that require Improvement or are unsettled with

regard to their Interpretation and application" (1986) The Quarterly Journal of Administration Volume XX, Nos. One and Two, pp 29-41.


35] Okoth-Ogendo, H.W.O. (1996) "Constitutionalism without Constitutions" in C.M. Zoetbout et al eds (1996) Constitutionalism in Africa: A Quest for Autochthonous Principles Rotterdam, International Association of Constitutional Law, pp 49-61.

36] Okotoni, Matthew (1993) "Problem of Internally Generated Revenue for Local Governments" (Paper presented at the National Workshop on Revenue Allocation and Generation for Local Government; at the Nigeria Institute of International Affairs, Lagos, January 27-28, 1993).

37] Okotoni, Olu (1996) "Local Government and National Development in Nigeria The Nigerian Journal of Local Government Studies Vol. 6 pp 44-59

38] Olaloku, F. Akin (1979) "Nigerian Federal Finances: Issues and Choices" in A.B. Akinyemi et al (eds) Readings in Federalism Lagos NIIA pp 109-125

39] Olowu, Dele; C.A. Ajayi, Dotun Popoola, M.O. Okotoni, S.R. Akinola (1994) Property Taxation and Nigerian Local Government, sponsored by IDRC, Canada.

40] Olowu, Dele & Olu Okotoni (1996) "The Informal Sector in Nigeria: Some Analytical and Developmental Issues" in E.U. Olisadebe and Olu Ajakaiye Conceptual and Methodological Framework for Informal Sector Research in Nigeria Lagos, Ibadan, Central Bank of Nigeria and Nigeria Institute of Social and Economic Research, pp 27-39.

41] Oribabor, P.E. "Improving Staff Training in Tax Administration: A Diagnostic Approach to Assessment of Training Needs" The Quarterly Journal of Administration Volume XX, Nos. One and Two, pp 95-104.

42] Oyelere, B.A. and P.O. Oyewole (1986) "Computer Application to Tax Administration" The Quarterly Journal of Administration Volume XX, Nos. One and Two, pp 81-94.

43] Pechman, J.A (1985) The Promise of Tax Reform Englewood Cliffs, Prentice-Hall Inc. 44] Rosenau, J.N. (1992) "Governance, Order, and Change in World Politics" in J.N Rosenau and

Ernest-Otto Czempiel eds Governance with Government: Order and Changes in World Politics New York, Port Chester Cambridge University Press, pp 1-29.

45] Schumpeter, J.A. (1996) "The Crisis of the Tax State in Richard Swedberg ed Economic Sociology Cheltenham, pp 5-38.

46] Stolper, W.F. (1970) "Some Considerations concerning the Allocation of Fiscal Resources" The Quarterly Journal of Administration Vol. IV, No. 2, pp 83-91.

47] Trevor, G. "Statement on Good Governance" in African Institute for Democracy, April 1997, pp 131-136.

48] Weber, Max (1978) Economy and Society: An Outline of Interpretive Sociology Berkeley, University of California Press.

49] Wunsch, J.S. and Dele Olowu (1995) ed The Failure of the Centralized State San Francisco, Institute of Contemporary Studies.

50] Zartman, I. William ed (1997) Governance as Conflict Management Washington, D.C. Brookings Institution Press.


The Economic and Business Driving Forces of the Russian Market Environment and Main Decision Makers

Geo-Economic World Perspectives

Evangelos Karafotakis 119 Ipocratous, Spata

190 04, Athens, Greece, John Mylonakis 10 Nikiforou, Glyfada, 166 75, Athens, Greece

John Mylonakis

10 Nikiforou str., Glyfada, 166 75 Athens, Greece

Abstract

The Russian economic and business environment constitutes a special pattern, which does not derive from the country’s political authoritarian tradition but, mainly, from its being met with reserve by powerful Europeans and the USA. The management of the Russian economy is currently carried out by two strong groups: that of the Russian President and the oligarchs. The purpose of this paper is to describe the immergence and struggle of power between the main groups for controlling the Russian economy and managing natural resources, as well as, to show the business geo-economic priorities of the present market leaders. Of great economic interest are, also, the interconnection of politicians and executive market people, the necessity for protectionism of authority and the crucial role of the middle class.

Keywords: Russian economy, oligarchs, protectionism, oil markets and industry, geo-economic developments, autonomist activity, globalisation

JEL Classification: P26, P27, P28, N44, N45, N55

I. Introduction This study aims at revealing the dynamics of the Russian economy in terms of its political background, the ambition of its middle class to dominate, as well as, the existing weave problems in parliamentary democratically-governed countries.

Russian economy constitutes a special pattern which does not derive from the country’s authoritarian political tradition, but mainly from its being met with reserve by powerful Europeans and the USA. This reserved behavior causes great trouble in Russian leaders, thus reversing every effort of achieving substantial convergence. This intense situation particularly manifests itself in the relationship between NATO and Russia, especially after the North Atlantic Alliance and European Union enlargements, as well as, governmental changes in Ukraine and Georgia.


II. The Economic Conditions and the Managerial Control of the Russian Market The management of the Russian economy is currently carried out by two teams. The first refers to the extremely powerful team of President Putin, the so-called “Siloviki”, roughly translated as “the strongmen”. The Siloviki team draws its economic power from state monopolies, such as Gasprom, RAO UES and Rosneft. The second refers to the oligarchs who remained after the derogation of Gusinsky and Berezovsky, the expatriation of Abramovich and the imprisonment of Hodorkovski.

The continuous conflict between the above two teams for the management of the Russian economy is less intense today, especially after the long-term imprisonment of Hodorkovski. Yet, it will not eliminate and there is a chance that it will grow after President Putin has abandoned presidency due to the constitutional facts of his service.

The Siloviki team currently controls the main ministries, such as the Ministries of Foreign Affairs, Defense, Internal Affairs, Economy, as well as state monopolies and the majority of the country’s government houses. Its authority has grown after the dismissal of Prime Minister Kasyanov in 2004 and his being replaced with Fradkov. However, the team has not managed to create authority basis in the capital market or in banks and it has small influence on industry, apart from its institutional and executive influence through state mechanisms.

If the authority of the Siloviki is considered to be concentrated in the public sector, then the authority of the oligarchs in the private economic sector, and especially in industry and services, should be regarded as greatly concentrated and possibly centralised. According to a specialised market research carried out in 2003 by World Bank in 1,700 large enterprises across 45 sector of the Russian economy, the situation exhibits strong evidence of economic power centralisation. These 45 sectors represented 40% of total employment and, more specifically in industry they represented 77% of total sales. Table 1: Indicative control evidence of the Russian market No Sector Oligarchs share of

sales (%) Centralisation percentage (%)

1 Oil 72 59 2 Ferrous Metallurgy 78 66 3 Automotives 71 71 4 Mechanics 12 12 5 Non-ferrous Metals 92 95 6 Paper 30 41 7 Milk 18 23 8 Aluminium 80 90 9 Building Materials 6 32 10 Pharmaceuticals 17 37 11 Fertilisers 46 66 12 Pipe Manufacturing 55 85 13 Furniture 3 23 14 Extraction of Ferrous Metals 73 59

Source: Guriev-Rachinski 2004

Based on the above tables, the dynamics and influence of the “oligarchs” are apparent across the Russian economy. According to Guriev and Rachinski, the complexus of the enterprises that are controlled by either the Oligarchs or by the most dynamic representatives of the newly formulated middle class is spreading horizontically and vertically, thus perfectly interconnecting production to profit distribution and utilisation.


III. The Formation of A New Middle Class The new middle class of Russia, in its vast majority, has originated from the once nomenclature and from various regional elements which emerged in the upper classes during the period of the so-called “predatory capitalism" (i.e. the 1992-1996 period). Table 2: Some of the “Oligarchs” No Sector Oligarchs Controlling Enterprises 1 Oil Abramovich,

Alekperov, Bogdanov, Fridman

Sibneft, Lukoil, Surgutneftegaz, TNK-BP

2 Ferrous Metallurgy Mordashov, Rashnikov

Severstal, Magnitogorsk

3 Automotives Kadanikov Avtovaz 4 Mechanics Bendukize United Machinery 5 Non-ferrous Metals Mahmudov, Potanin UGMK, Inteross 6 Paper Zingarevich Ilim Pulp Ent. 7 Milk Dubinin, Plastinin WimmBillDann 8 Aluminium Deripaska Rusal, Base element 9 Building Materials - - 10 Pharmaceuticals Popov-Pumpiansky MDM 11 Fertilisers Popov-Pumpiansky MDM 12 Pipe Manufacturing Popov-Pumpiansky MDM 13 Furniture - - 14 Extraction of Ferrous

Metals Potanin, Ivanishvili Norilsk, Metalloinvest

Source: Guriev-Rachinski 2004 Table 3: Social stratification of Russia (estimations) No.

Stratificational distribution 1994, in % 2004, in %

1 Middle class 2 2,5 2 Middle layers 15 21 Suburbanites (out of the above) 5 7 petit bourgeois (out of the above) 10 14 3 Workers 70 64,5 4 Rural layers 13 12 Totals 100 100

Sources: Goskomstat, T.Malevoi

The modern formulation of the middle class, including the “siloviki”, is characterised by a certain peculiarity in Russia, considering the fragile balance between political authority and economy. As shown in Table 3, the middle class grows with time, as is also the case with middle layers, especially after 2000. Economical circumstances provide the city’s middle layers with the possibility to develop through the services sector, provided that a larger part of the income becomes available for improving living conditions compared to the first transition decade.

What creates impression is the fact that, in the course of 10 years, society changes have been noted, which neither are probably nor accompanied by an understanding of the role of each social layer


(A.Ryabov 2005). It is seriously doubted whether Russian citizens comprehend equally with other Europeans the dynamics of globalisation and the role of their country in this process. In reality, socially rising Russian citizens regard their country as a unique place for business development creation and believe in the country’s recovery at a global level (T.Malevoi 2003). This fact in turn undermines the country’s spreading processes and favors protectionist attitude (see nationalism), which is mostly observed in lower social layers.

In any case, the formation of a new middle class always involves risk in the sense that modern middle classes create their notions based on personal profit and not on long-term planning. This fact derives from the way middle classes are formulated and from their insecurities regarding their place in the global market. A typical example would be the disagreement that took place four years ago between the “oligarchs” Deripaska and Hodorkovski regarding Russia’s joining the World Trade Organisation. The former was against due to his fear of international competition, while the latter was in favor for reasons of attracting foreign capital. The lack of a common position in the new middle class regarding the country’s internal and external economic policy, combined with past susceptibility on the issue of wealth tactics, creates space for the imposing of “siloviki” beliefs, thus resulting in the prevailing of intrigue. IV. Interconnection Between Politicians and Executive Market People Until 2000 the interconnection between politicians and powerful people engaged in the market and in the economy constituted a rule, which had been placed by the people of President Yeltsin (the so-called "Family"), in order to take over industry and services and lead to the growing of production. This strategy resulted in the creation of the team of the “oligarchs” who, anticipating their rapid enrichment, lead to a state of protectionism and to the creation of powerful oligopolic clusters. The politicians’ financing practice aiming at strictly serving their own interests, as well as the switching from market to politics may be also described through the cases of Hodorkovski and Abramovich. The former, despite the remarks of people of the Siloviki environment continued, even after the strengthening of Putin, to finance the opposition and even the Communist Group and Zirinovski, in order to enjoy the protection of Duma (the house of the Russian Parliament). This practice, not only was disapproved by those close to the President, but also led him to prison at the very crucial moment of his negotiations for the merging of Yukos and Sibneft to create the largest oligopoly in oil history. Abramovich, on the other hand, managed to become elected as the Governor of Yakutsk, in order to enjoy protection of the upper House and to cover his businesses in this area.

The end of the « oligarchs » also signalled the limitation, though not the abolition, of the relationship between politicians and market people. Today, this relationship is only developing under the direct supervision and possibly the permission of the “siloviki”, who control the country’s executive and judicial authorities. V. The Necessity for the Protectionism of Authority The existence of safety mechanisms around authority during Putin’s presidency has been widely criticised by the international community. Yet, why such a large number of employees, presidency associates, ministry and state companies’ executives originate from the KGB, the GPU and the Army? It is estimated that over 70% of the presidency, the state companies’ and main ministries’ executive members have previously served and have been distinguished within the aforementioned governmental services (The Saint Petersburg Times 20/1/2004, Novaya gazeta 14/8/2003).

A rather plain and rational explanation lies in the fact that the “siloviki” have been the only team of state executives that was organised, disciplined and adequately prepared for the change. They were the only people who were able to prevail in the 1989-1997 chaotic period and to impose their presence on President Yeltsin, due to the needs of the presidency of the time. This team not only managed to evolve


but was also able to promote the current President Putin to the highest governmental post without extreme social or other types of disturbances. It was this success that made the team popular among the people, who were disappointed by the corruption of the 1989-2000 periods. Naturally these people were promoted in key posts with the help of their Team’s fellow member Vladimir Putin after his election to the country’s highest position.

Modern functions of the "siloviki", in reality, ensure that the President’s choices are securely met and that he is protected against all kinds of danger, provided that the main objectives of the team are met, i.e. the re-incorporation of the country in the international power lobby. Post-soviet authority in Russia is in need of safety due to the complexity and the unavoidable concentration of authority, as a result of the vast area the country covers and its being multicultural (The televised address to the nation by President Vladimir Putin, September 4, 2004). Gorbachev’s and Yeltsin’s fall has apparently has provided a lesson to President Putin on administration security and maintaining balance.

The extraordinary protectionism of authority creates obstacles in a more liberal administration regime as it is not a part of it. It is possible that this effort to create a leading party with Putin in the centre constitutes a part of a more general “siloviki” dynamic process, aiming at their long-term presence in the main administrative bodies of the country. VI. The Role and Dynamics of Energy According to BP-Amoco’s year book, Russia controls 4.6% of the known oil reserves, while together with the rest of the Eurasian countries, they control 6.4%. In total, Middle-East countries control 65.3% of global reserves while the Organisation of Petroleum Exporting Countries (OPEC) controls 77.8% of reserves. Russia and Eurasian countries control 11% of global oil production against 9.2% controlled by Europe and 9.8% controlled by the USA. Overall, Middle East controls 31% of global oil production and the Organisation of Petroleum Exporting Countries (OPEC) controls 41%. When it comes to the distribution of global spending, the EU holds 18%, North America 30.4% and Russia 5%. Europe is the largest oil importer internationally, absorbing 27.3% of global crude oil exports and 24% of its finished products, followed by the USA whose corresponding values are 27% and 23% (2003 data). Russia together with the rest of the Eurasian countries control 8.6% of global crude oil exports and 15% of oil products, while the Middle East overall controls 50.3% of crude oil exports and 23.4% of oil products’ exports. From the above information, it may be concluded that, among other countries, Russia manages to withhold the added value of crude oil processing, contrary to Middle East countries where a significant percentage of added value is lost. Moreover, it is evident that The USA and Europe are the world’s largest oil consumers, a fact that results in them being dependent on OPEC and, to a certain extent, on Russia and Eurasian countries. The dependence rate of the USA, however, is only slight compared to that of Europe, given the power and dynamics of oil multinational enterprises of the Middle East. In practice, the international oil market is controlled by American oil groups that also control to a great extent the OPEC itself.

In this market need to strengthen their position the Russian oil industries, which are deprived of resources and financing, in order to exploit new sources and to reduce their operating costs. Indeed, if we consider the increasing tendency of oil consuming, estimated to double in Europe by 2015 and to become seven times greater in Asia, the energy game may be easily comprehended (ΒΡ-Amoco). American multinational enterprises try, and will try even harder, to restrict Eurasia, with the aim of acquiring East market control, mainly China, while they do not intend to extricate themselves from the Middle East. It is quite remarkable that the International Energy Agency (IEA) estimates Chinese consuming to reach 10% of global consuming by 2020, maintaining the same production representing only the 3% of global production. This means that China will need to find the remaining 7%. Respectively, Russian oil industries will try to penetrate the European market, thus replacing the American multinational enterprises while, at the same time, they will try to control China’s oil


supplies. In the following years, the most dangerous games will unfold in the above two fields of Eurasia/ Far East and the Middle East bringing international peace and safety at stake.

In terms of natural gas, its global consuming is distributed among North America by 32% or the

USA by 27%, Europe by 18%, the Commonwealth of Independent States (CIS)1 by 23% and Asian countries by 12% (BP-Amoco). The International Monetary Fund estimates that, by 2020, North America consuming will fall down to 22%, transitional countries’ consuming will fall down to 20%, the European share will maintain its levels, while the Asian share will increase by four times. More particularly, the Middle East countries’ share is expected to triple. Based on year 2000 data on production, the largest producing countries at a global level are Russia that controls 22.5% of global production and the USA with almost the same percentage. Considering that the USA are far from Asia, the only large producer is Russia and possibly the Middle East, which currently holds only 8.7% of global production while owning 35% of international reserves. The only country controlling the 15.3% of international reserves is Iran. It is worth mentioning that the most abundant reserves can be found in Russia, which controls 32.1% of international reserves. Therefore, among the countries close to Asia, whose consuming will grow by four times by 2020, the largest natural gas producers are Russia and Iran. Within this context, it is self-evident that multinational companies become interested in settling their positions not only in the Middle East but also in Eurasia.

Based on the above and on President Putin’s professional and theoretical background, he may easily consider that Russia and Eurasia will eventually become the victims of globalisation, in order for international companies to enjoy the country's energy reserves. For this reason, he declared war on the “oligarchs” and protected authority with the help of the “siloviki”, by concentrating state power on energy, having full control over natural gas through Gasprom, controlling oil transportation through Rosneft and obtaining full control of the power company, even though it is typically presided by Anatoli Chubais (М а r t а О l к о т 2003). A rather unknown fact is that President Putin himself compiled in 1997 a dissertation on the exploitation of Russia’s energy resources at the St. Petersburg University. In his dissertation, he supported everything he now implements in his country’s energy policy. In brief, President Putin mentioned in his dissertation that his country’s energy sector was the ticket to a globalised economy, as well as to Russia's recovery and placement among the top forces, provided that control is exercised by the government to the benefit of national interests and not by the “oligarchs” (Putin В. 1999).

Today, it may be said that the energy sector has been (substantially and not typically) placed under the authority of the state, especially after the result of the Yukos case and the interception of the connection between Yukos-Sibneft and American multinational enterprises, as was set in the Bush-Putin dialogue in 2002. In this very context, a lot of new oil transportation pathway designs were cancelled. Today’s focus is the increasing incorporation of Gasprom in the EU, the restructuring of the oil sector and the ensuring of Russia's share in the international oil market, as well as a place in the Chinese and Indian energy markets (D. Orlov 2005). In order for these Russian authority plans to be implemented, certain requirements have to be met: above all, funds must be acquired for the growth of fixed investment in all aspects of the Russian oil industry (up and downstream). This may currently seem as extremely difficult but can be achieved by 2020. VII. The Central Asia Issue Russia faces three different forces in Central Asia that interconnect, resulting in a casual negative cumulative balance. These forces are:

• The increasing desire for western penetration in the area (Afghanistan and Kyrgyzstan American bases),

1 Covers the territory of the former Soviet Union less the Baltic States of Estonia, Latvia and Lithuania.


• The will of the Chinese to penetrate the energy resources of the area • Muslim radicals

Facing these forces, Russia followed two strategic moves: The first refers to the Shanghai

Cooperation Organisation and the second to the Collective Security Pact. The Shanghai Cooperation Organisation (SCO) was established in 2001 and consists of Russia, China, Kazakhstan, Kyrgyzstan, Tajikistan and Uzbekistan. Through the SCO, the Russian leadership attempted to ensure Chinese and Russian mutual interests in Central Asia. The Collective Security Pact was established in 2002 and includes Russia, Armenia, Kazakhstan, Kyrgyzstan, Belarus and Tajikistan, while Georgia, Azerbaijan and Uzbekistan have withdrawn. Through the Collective Security pact, the Russian leadership attempted to ensure its defense against Muslim danger and, at the same time, to maintain its influence on the participating countries. If the Collective Security Pact was functioning as planned, then Moscow would be able to control over a vast area from Caucasus to the Baltic Sea and from the Poland borders up to the Altai Mountains. This area would be much larger and of greater geo-economic importance than the area controlled by NATO. Moscow would immediately transform into a key place for Eurasian and Far Eastern geo-economic developments.

However, none of these pacts functioned as was planned by their founders, mainly due to China’s cautiousness and disbelief on the part of Central Asia. Chinese, on their part, considered that, in this way, they would reinforce arms system supplies from Russia. The rest of the countries believed that they would obtain the necessary partners for the administration of their countries and the management of their energy resources. The non-functioning of these agreements had a negative effect on the construction of new energy conduits between Russia and China and gave free scope to the Fergana valley Muslims, resulting in the recent Uzbekistan crisis. Furthermore, the summit conducted by President Putin and his Chinese counterpart Jiang Zemin in December 2002, aiming at the discovery of cooperation development measures on oil and energy matters failed to do so. The Russian offer of part-financing an oil conduit between the two countries remained unfulfilled. According to Putin, this conduit would alter the total division of labor in Eurasia. Russian analysts that referred to this failure, believe that the Russian attempt to approach the West and the USA played a significant role, as it created a sense of inhibition in China and India, which was about to join the Shanghai Cooperation Organisation (Die Zeit/23.06.2005).

Today, Moscow continues to face the above mentioned forces without substantial developments. The recent decision of the Uzbekistan leader Karimov on the immediate retirement of the American air base should not be regarded as a significant development, even though it is a sudden action, approved by Moscow. In any case, President Putin himself characterises the foreign policy of Central Asian countries as a primary goal of Russia, a fact that proves the great importance rendered to this area on the part of Moscow (Putin: Speech at the Security Council Meeting 2005). VIII. The ‘Soft Underbelly’ Issues The Russian “soft underbelly” is perceived as the Caucasian area between the Black Sea and the Caspian Sea. This area faces three main problems:

• The autonomist activity of the Caucasian populations of Turkmen • The control of the Black Sea • The separation of interest zones in the Caspian Sea

Regarding the first, the situation is critical because the problem of exporting autonomistic activity

from Chechnya to the neighboring democracies, such as Ingushetia, has not yet been resolved. Governmental changes in Georgia and the fate of Ossetia and Abkhazia intensify the problem because Georgia, in its effort to annex these areas, is possible to support Chechen autonomists with unknown consequences, just like in the past. An extensive autonomistic Caucasian crisis will be the worst thing


for Moscow because there is a large possibility that such crisis would be exported towards other Russian regions where the autonomistic movement smolders (e.g. Tatarstan). The inability to subject the Chechens, especially after Beslan massacre, is apparent and this fact does not go unnoticed in Eurasia, where Russia strongly desires to play the role of the great guardian.

The control of the Black Sea should also be considered as relevant to the above issue and is compulsory for Russia due to the existence of the oil conduit in Novorossiysk. In the absence of other conduits, this one in Novorossiysk has not ceased to be of great importance not only for economic reasons but also for its geopolitical significance. Losing control of the Caucasian area may create huge problems in the Novorossiysk role and operation. In addition, a possible Ukraine and Georgia entry in NATO (requested by both of the two countries), limits Russian dynamics in the Black Sea, as all other countries in the area will be part of the same defense bloc. In such an extreme scenario, Moscow will play a very small role in sea transports from the Channels (Dardanellia/Instanbul) and, therefore, there is the possibility of blackmail, since a decrease in oil transportations through the Channels is likely to lead Russia to an extensive crisis due to a smaller oil income.

Today, Moscow does not seem to have altered its methods in the Caucasian area and has maintained its tough tactics imposed by President Putin. This fact should be regarded as a weakness and a precursor of a different strategy and policy. The alternative Moscow method greatly depends on the management of Caspian operational zones, a difficult problem that has many constituents. Yet, what is the actual value of the Caspian Sea? Is it its oil reserves or its geopolitical position? In 2002 the manager of the Caspian Energy Politics Section of the State Department Stephen Mann and the president of the ENI Italian Oil Industry Gros-Pietro stated that the Caspian Sea has much less reserves than anticipated in the 90s. Their statements reflect reality and it is probable that the actual value of the Caspian Sea is its huge amount of natural gas reserves of Turkmenistan, which are operated by Gasprom based on specific terms for the country’s leader’s (Niyazov) income and security. Russian’s failure to separate the Caspian zones, accompanied by governmental changes in Azerbaijan (ostracism of the Aliyev family) and an American presence in Iran would endanger the second source of natural gas in Russia. It is not accidental that for many years Moscow has been trying, by diplomatic means, to define the Caspian operating zones, by giving way to pressure and Iran requests and, at the same time, it has been supporting Tehran on various issues which are critical to the rest of the world (see nuclear energy industry). On the other hand, a Russian defeat in the Caspian Sea would cause chain reactions in the Caucasian area, resulting in the authority of Russia’s “soft underbelly” being transferred to other western forces. This transfer of authority has been aborted with great effort from the part of Moscow during the last two centuries. IX. Conclusions Russian has substantial natural resources in order to finance its development and world power, to the extent the world oil prices will continue to be controlled by oligopoly forces and maintained in high prices. In our days, the power is shared by ‘siloviski’, who control, directly or indirectly, all natural resources. Also, the Siloviki team controls the main ministries, such as the Ministries of Foreign Affairs, Defense, Internal Affairs, Economy, as well as, state monopolies and the majority of the country’s government houses.

The new middle class of Russia, in its vast majority, has originated from the once nomenclature and from various regional elements which emerged in the upper classes during the period of the so-called “predatory capitalism".

Russia is faced by severe problems in different fronts, a significant part of which has not emerged yet and therefore has not created additional notable disputes.


References 1] Andrei Ryabov, (2005), “Originality Instead of Modernization” (in Russian), Moscow Centre

Carnegie, pp.12-20 2] Annual Presidential Address to Duma, May 16, 2003 3] BP-Amoco, (2002), “Statistical Review of World Energy”, pp. 120-153 4] Die Zeit/23.06.2005, p. 9 5] Dmitry Orlov, (2005), “Russia's Turn To Make a Move”, Moscow News No 49 6] Goskomstat, (2001), “Social conditions and living standards of the population” (in Russian),

pp. 314-330 7] Guriev-Rachinski, (20040, “Corporate control and corporate governance at a company”,

December, Moscow, pp. 25-58 8] International Energy Agency, (2003), “Oil market report”, pp. 4-16 9] Marta Olkot, (2005), “Putin and the oil policy in Russia Working papers” (in Russian), 1,

Moscow Centre Carnegie, pp. 6-13 10] Novaya gazeta, 14/8/2003, p.7 11] President Putin, (2005), Speech at the Security Council Meeting 12] Tatiana Malevoi, (2003), “Middle classes in Russia” (in Russian), Moscow Centre Carnegie,

pp.12-85 13] The Saint Petersburg Times, 20/1/2004, p. 5 14] Vladimir Putin, (1999), “Mineral resources in development strategy of the Russian economy”

(in Russian), Zapiski/Gori Saint Petersburg, pp. 51-68 15] World Bank: From Transition to Development, (2004), “A country Economic Memorandum for

Russian Federation”, April, pp. 78-115.


Proposed Models for Integrating Marketing and R&D Departments in High Tech Markets:

Literature Review Approach

Thomas A. Fotiadis Thomas Fotiadis

Tel. ++306944630505 49, Karakasi street, 54644, Thessaloniki, Greece

Email: [email protected]

Abstract Biographical notes: Thomas A. Fotiadis has a Ph.D in Marketing. His doctoral studies were concluded at the University of Macedonia, Social and Economical Sciences, Thessalonica, Greece. He has a Masters’ degree in Business Administration, which he also obtained from the University of Macedonia as a scholarship student, and he has concluded his Masters’ degree (2nd) in Business Computing early in 2005 at the Aristotle University of Thessaloniki. He has been teaching Marketing and Economic courses in the TEI of Serres, Katerini and Thessaloniki for the last six years .He has also been lecturing at the University of Macedonia, . His research field focuses on empirical research on High Tech Marketing and Management of Information Systems. His research work is published in Books, Journals, and he has taken part in several International Conferences. For the last two years, he works as a Lecturer (adjunct faculty) at the University of Macedonia, department of Marketing and Operations Management, at the city of Edessa, and also as a Lecturer at the department of Technology Management of the University of Macedonia, at the city of Naousa. He is also an adjunct Lecturer at the Open Greek University. Of all the common areas that derive from the functional interface of departments during the process of creating a technological innovation, the most sensitive area is the one that is called on to fit the product to the market. The interface, or in other words, the integration of Marketing and Research and Development departments, is considered to be of significant importance, since the successful implementation of the common activities becomes a particularly critical factor in assuring the offering of high tech products. In this paper, various models integrating Marketing and R&D are analysed, as well as their objectives, their methodology and their conclusions, each time considered with respect to the adopted viewpoint of High Technology. The findings highlight the common areas among the existing theories, and indicate that in order for a high tech company to survive, the functional departments of Marketing and R and D have to re-orient towards a more systemic approach, so that they can enjoy the benefits of their combined capabilities, and further expand their knowledge expertise to the direction of corporate goals accomplishment. Keywords: High Technology, Integration, Integration Models, Marketing and R and D,


I. Introduction Song and Parry (1997) consider that if high-tech enterprises care for being successful, they have to effectively bring together the R&D and Marketing departments. Numerous research works in this area demonstrate the critical role of an effective contact between these two departments (Griffin, Hauser, 1996), (Gupta, Raj and Wilemon, 1986), (Dutta, Νarasimhan and Rajiv, 1999).

Souder and Chahrabarti (1978) define integration as the symbiotic interrelation of two or more entities, which leads to the production of net benefits for them that exceeds the sum of the net benefits that the individual entities would have otherwise produced. On the other hand, Lawrence and Lorsch (1967) define integration as the process of obtaining unification of exertion between various subdivisions for the achievement of the organizational objective.

Millman observes that it is not the lack of innovative ideas or technical competences that are the main barriers responsible for the non-completion of an innovative product, or for its non-commercial exploitation, but the lack of Marketing orientation and the inefficient intradepartmental communication.

ACARD (1978) stressed that organisational problems, specialisation and lack of communication can only have a negative impact on innovations. Also the isolation and the seclusion of market concepts caused by the remoteness of Research and Development from Marketing, due to the expectation that new products will create new markets, act against innovative activities.

The objective of the paper is to critically combine and evaluate each suggested framework, so that the key points can be identified and encapsulate in a holistic manner the main factors that represent what constitutes the high tech field; an interrelated and dynamic environment, a network of unfavorable conditions that escalates constantly.

The paper is organized as follows: First, integration is studied as a critical variable and at the same time as the cohesive link between the perceived environmental uncertainties. In this part, emphasis is laid on the complexity that appears as a resultant of the many sources of uncertainty, and an initial summarized approach-as a figure-is provided. Successively, integration is seen through the angle of the gap between the required and the achieved levels, and the contribution of integration regarding the success of innovative products is evaluated. Finally, high tech markets and integration are approached through the scope of the demand and the supply driven attributes. Throughout the paper various aspects shaping the high tech arena are critically evaluated through the extensive literature review. II. Presentation – Analysis of Integration Models Integration as a variable of both the perceived environmental uncertainty and the adopted innovation strategy.

According to Souder and Moenaert (1990), every innovative process inherently includes 4 types of uncertainty. These are namely, the technological uncertainty, which refers to the optimal option of technology that will be adopted to use, the uncertainty regarding consumers’ needs, the uncertainty created by the competitor’s product offers for the same targeted public, and lastly the uncertainty proceeding those three, that is the uncertainty of distribution of resources, so as to reduce the three aforementioned uncertainties. In particular, according to Moenaert and Souder (1990) the uncertainty of resources is composed of human, financial, and technical resources, which shall be committed/ allocated in the most optimal way possible. The uncertainty of human resources, involves the uncertainty regarding the choice of the type of personnel of Research and Development and/or of marketing that shall staff the relating departments, their capabilities, the uncertainty concerning the suitable constructional and processing capabilities, or even the uncertainty regarding whether and to what extent the innovative process will enjoy the right and proper support from the top management.

Financial uncertainty is the uncertainty related to obtaining the funds that will be used for innovation development. On the other hand, technical uncertainty relates to the infrastructure needs required for innovation development. The uncertainties described above, may cause delays when


taking critical decisions and, consequently may lead to the ignorance and non-exploitation of opportunities. This is more intense particularly in the innovation arena, which is characterized by its turbulent and volatile environment, intense competition and most rapidly deteriorating technologies. It should be stressed that, while the three types of uncertainties (technological, competitive and of customers) are considered to stem from the external environment of the enterprise, the uncertainty of resources originates from the internal environment (1986). (see figure 1) Figure 1: Theoretical framework of integration as a function of perceived uncertainty

Source : Thomas Fotiadis ‘’Development of a decision-making procedure for the introduction of New Products of High Technology’’, Ph.D.dissertation, University of Macedonia, 2004


There are two ways of measuring the reduction of uncertainty for the above factors; qualitatively and quantitatively. The quantitative dimension measures the volume of information, while the qualitative measures the precision, the compatibility and consequently the usefulness of information relating its target. High quantity and low quality reduction of uncertainty means that the organisation possesses the basic knowledge for a wide spectrum of subjects, while the opposite means a deep knowledge of individual subjects. Of course, the objective of the organisation is its enrichment with deep and of wide range information. The smaller the gap between desirable and essential, and achieved information in each area is, the more successful the innovative effort will be (see figure 2). Figure 2: The crucial role of Information as a factor and a catalyst in the success of Innovative (high tech) products.


The central subject of most models that Saren (1984) calls “hybrid”, has to do with the transformation of ideas into marketable products (Crawford,1983), (Miaoulis and Laplaca, 1982). These models include levels of activities as well as decision gates (hence their characterization as hybrids). The activities involved in these models, can be considered as separate activities of information processing, while decision gates reflect on the official options of the enterprises concerning activities that can influence the flow of information between the activities.

According to Souder, the human resources allocated to a particular project are called on to have specific roles between these activities. (According to Souder, role is the expected behaviour attributed to a specific position in an organisation). Therefore, from the reduction of uncertainty viewpoint, these roles are connected to a list of information through which individuals shall work in order to decrease


uncertainty. Hence, the total of human resources that have been officially – in the framework of the organisation –called on to complete a project, form a project team.

Technological innovation is an amalgam and a result of information processing activities by almost all functional departments. Consequently, also the roles that the individuals are called on to have, are functionally determined, and the project team of the innovation is composed of functionally specialised subdivisions. According to Kotler (1984), the role of the Marketing department includes the analysis, the planning, the application and the control of programs which have been planned in order to create, build and maintain beneficial relations and exchanges with targeted markets, so as to achieve the organisational objectives. The responsibilities of the Research and Development departments include the development of the scientific-technological know-how and its applications, so as to achieve the organisational objectives. Therefore, the Marketing department’s staff is more responsible for the provision of information concerning the user’s needs, and for the provision of information regarding competitors and resources. The equivalent expected obligation of the Research and Development department is the provision of information with regard to technologies generally, but also with regard to technologies used by the competition. It is observed that the abovementioned is a reasonable consequence of the functional differentiation, which determines the type of information held and processed by the members of organisations and by the organisational subdivisions. Based on the roles of the Marketing and Research and Development departments, we can certainly come to the conclusion that they constitute the main factors of technological innovation (See figure 3).

Figure 3: Framework of the Communication Channel between Marketing and R&D.


According to Schneider (1987) and also Lawrence and Lorsch (1967), it is differentiation that

creates the need for integration. This becomes apparent, since market-related information is usually unknown to Research and Development, while at the same time Marketing does not have access to technological information. Therefore, Research & Development and Marketing become providers of known information to departments, which due to their functional roles do not have access to it, and respectively they become beneficiaries of information from departments that have access to it due to their assigned responsibilities. Both departments alternate successively, respectively and continuously in their roles of the provider and the beneficiary of the essential information. As a result, during the innovative process, there is the phenomenon of posterior interdependence, which happens precisely


because these types of information and their flow exist during the whole of the process of innovation (Thompson,1967). Based on those conditions integration is expected to have a decisive role in the innovative return (Cheng, 1983), Ames (1968) summarises that after all it is the high degree of customer orientation expected from the industries, which denotes integration and functional interdependence of the Research & Development and Marketing departments as critical variables for the enterprises. Through the communication of information, team members transfer information and knowledge to other members. Certainly, individuals do not only share information with the members of their functional subgroup, but also with members from other functional subgroups. The transmission of information becomes the vehicle that allows the individuals that participate in the organisation to have access in information that has already been obtained.

Generally, theories concerning the organisational adoption of innovation have been focused on a two-stage type of model: The first stage is launching and the second is implementation. In the first stage, all essential information on the evaluation of innovation is collected, and then the decision to adopt (develop) it or not is taken. The second stage, involves all activities necessary so that innovation is turned into a functional reality (Rogers, 1983). This two-stage process may also be extended to the process of technological product innovation in an enterprise. Souder and Moenaert (1990) consider that the two-stage process framework is both relative and also supportive to the study of integration of the Marketing and the Research and Development departments during the development process of innovative products. They define the stage of launching as the stage of planning an innovative product, and the stage of implementation as the stage of development. During the stage of planning, the organisation has set its objective to differentiate the product idea to be developed, and to determine whether the company shall invest resources in this idea. In the case of an affirmative answer to that issue, the organisation proceeds to the second stage, which is concerned with the transformation of the product idea into the actual product. The activities involved in this stage, aim both to the development of the product itself, as well as to the upcoming Marketing strategy that will be used.

A study by Victor and Blackburn (1987) shows that uncertainty is related to two variables: The variability, which refers to the number of special cases that is dealt with during the

accomplishment of an objective, and the analyzability, that refers to the extent the procedures, which determine the sequence of the steps necessary for the accomplishment of an objective, are known. In case technological innovation is considered from the reduction of uncertainty point of view, variability refers to the volume and to the rate at which uncertainty emerges. Analyzability reflects on the degree to which there exist procedures for uncertainty recognition and reduction. In addition, since technologically innovative products can be considered as a process of uncertainty reduction, then it logically follows that the degree of variability and analyzability changes during the project’s life cycle. And since change is the central characteristic of technological innovation, then both analyzability and variability are subject to changes. Therefore, from an uncertainty reduction viewpoint, the most crucial challenge is the best possible recognition of the relative potential uncertainty (i.e. reduction of variability), and the recognition of the tools for the reduction of uncertainty (i.e. increase of analyzability).

Perrow (1967) produced a matrix using analyzability and variability as dimensions, having two levels of intensity in each dimension that creates 4 possible combinations of high / low variability / analyzability.

The reduction of variability and the increase of analyzability during the project’s life are essential prerequisites for a successful technological innovation. The objective is the passing from a relatively low analyzability and a relatively high variability to the opposite direction of the corresponding variables. The wider the information and knowledge horizon of the enterprise is, the less unknown occurrences will emerge. At the same time, the organisation needs to discover or to select the right tools and processes for the reduction of uncertainty. The following graph illustrates Perrow’s matrix (see figure 4).


During the stage of planning, communication and transmission of information between the Marketing and Research and Development departments contribute to the reduction of variability and the recognition of suitable processes and tools.

According to Silver Cohen and Rainwater (1988), the contribution of knowledge is greater during the initial stages of life of an innovation. These stages are usually characterized by high perceived variability and low analysability. As the volume of information becomes bigger, variability tends to decrease, while analysability tends to increase. The challenge here is the decrease of variability and the increase of analysability at the stage of planning. This minimises the chances of experiencing unpleasant surprises at the stage of development. Moreover, it clarifies the picture of what should be done at the stage of development. Figure 4: Perrow’s matrix

Based on the above, Souder and Moenaert (1990) suggest that the success of the innovation of

technological products is characterized by the maximum decrease of variability and the maximum increase of analysability during the stage of planning.

According to the aforementioned proposal, the most efficient teams would accomplish both conditions, which can be graphically demonstrated as follows at figure 5:

Figure 5: Analyzability and Variability during the stages of the project life cycle.


Hence, if this behaviour can be described as an efficient risk reduction, then certainly the achievement of that behaviour would not be feasible without the communication between Marketing and the Research and Development departments.

The point of view adopted, considers that the reduction of the four types of uncertainty is positively related to the success of the innovation. The objectives - work of the Marketing and the Research and Development department - are subject to two variables; variability and analysability. The innovation is conceptually separated into two phases, the stage of planning and the stage of development. In the stage of planning, the transmission of innovative information decreases variability and increases analysability. The interdepartmental transmission of information creates convergence between the two groups, i.e. it prompts them to the same common objective, a fact that is positively connected to the reduction of uncertainty and hence to the successful result of the project’s progress. The result of integration is more positive than that of the sum of the individual efforts of departments since it produces synergies, and also includes the continuing decrease of variability and the continuing increase of analysability. Interdepartmental communication also improves the convergence of roles. Information requirements vary depending on the stage of development. Thus, the stage of planning is largely dependent on the exchange of innovative information, while the stage of development is largely dependent on the exchange of coordinative information.

The approach proposed by the previous discussion can be summarised in figure 6:

Figure 6: Model of the information-processing process of a technological innovation.

Model-framework of required and achieved integration and contribution of integration to the

success of innovative products.


Gupta, Raj and David Wilemon (1986), allege that the interdepartmental collaboration of the functional departments has an important role in the development process of new products. However, the integration of the Marketing department and the Research and Development department in a systemic way has the most decisive role, particularly when the development concerns innovative products.

Based on the argument of Wind and Robertson (1983) and Hutt and Speh (1984), the relation and the interdependence of Marketing to other functional departments has enjoyed little attention and the literature of Marketing has paid little attention to the network of interrelations that exists between the marketing and the functional departments of an enterprise, a surprising fact given the enormous importance of the interdependence and its significance that is considered an integral part of the Marketing Concept, attempt to create a framework.

Figure 7 presents a brief presentation of the theoretical platform and beneath the plan’s parameters are analysed to their components:

Figure 7: Study-model of the interface of R&D with Marketing.

Hence, referring to the perceived need for integration, we observe that it is a variable dependent on

the new product strategy the enterprise follows, but also to the perceived by the enterprise risk of the environmental uncertainty. The correlation is positive, i.e. the increase of the level of uncertainty and the adoption of more offensive strategies, involve an increase of the degree of awareness for the need for integration (Freeman, 1974).

The level of achieved integration directly depends on organisational variables, such as the structure of the wage system, the dislike or sympathy attitude and orientation of top management towards risk (whether the company can be characterized as a risk-lover or as a risk-avoider), the emphasis put on the need for integration of the Marketing and Research and Development departments, and the differences in culture and the “social blend” of the Marketing department and the Research and Development department.

The distance between the perceived or the adequate/desirable integration and the realised existing integration, influences the level of success of new –innovative products.


Literature (Gaibraith and Nathanson, 1978), (Lawrence and Lorsch, 1969) stresses that the uncertainty that stems from the environment combined with the organisational strategies are responsible for the determination of the need for collaboration and control. Hage (1980) considers that the real issue is the extent to which the environment itself determines the organisation and the extent to which the organisation itself can shape the environment.

Hage (1980), compiling Child’s (1972) position for strategic option, and that of Lawrence and Lorsch’s (1969) regarding the environmental composure, concludes that sometimes there is a high degree of strategic option and decision, while other times there are restrictions imposed by the environment.

Regarding the key factors that determine the required degree of integration, it is now obvious that it depends on the requirements of the organisational strategy and on the impositions created by the environmental uncertainty.

Environmental uncertainty is the second resultant that affects the degree of required integration. Here we refer to the ability of the enterprise to anticipate changes in the strategic choices of its competitors, the customers’ requirements for new products, and the technology, the emergence of new competitive forces in the market, and the new imposed restrictions regarding planning and product requirements. The environment-relating uncertainties, strengthen the need for processing information, and hence the need for co-ordination, collaboration and control between the organisational subgroups. The classic studies of Burns and Stalker (1961), Woodward (1965), Lawrence and Lorsch (1969) and Khandwalla (1972), point out the fact that the higher the uncertainty originating from the environment is, the higher is the degree of specialisation or differentiation inside the enterprise, respectively.

The second axis of Gupta, Raj and Willemon’s (1986) framework-model has to do with how much co-ordination – integration is achieved between the Research and Development department and the Marketing department. According to the scholars, the actual integration finally achieved in the framework of an enterprise is a function of the following three factors:

a) The organisational structure, b) The attitude of top Management towards the integration of the Marketing department and the

Research and Development department, and the action the highest management is determined to undertake in order to implement it, and

c) The social and the cultural differences between the Marketing department and the Research and Development department, which become noticeable during the development process of new products.

The organisational structure is explicable in terms of: a) Complexity: Complexity is a function of the number of specialised personnel and their

professionalism, inside the organisation. Consequently, the enterprises successful in innovation–given the increased probability of a large number of highly specialised staff– are expected to experience higher levels of complexity. A generally acceptable principle is that as the level of complexity increases, the level of difficulty for achieving the integration of different functional subgroups increases too in the framework of an organisation.

b) Formality: It is defined as the emphasis (inside the organisation) put on the conformity of the procedures and the rules during work. According to Kanugo (1979), a formal and sound framework is possible to cause alienation and tendencies of non-mixture between the specialised professionals. Moreover, Kahn et.al. (1964), found that although formal frameworks cause the reduction of uncertainties, probably they are also responsible for an increase of the conflict of the roles. On the other hand, Organ and Greene (1981), suggest that formalization does not necessarily have only clearly negative implications for the personnel of an enterprise. Deshpande (1982), in his study on the extent of use of information resulting from Marketing research, observes that less formally structured companies have greater chances of better utilizing research findings. On the other hand, John and Martin (1984) argue that both reliability and functionality of a plan’s result increase along with the increase of the formal frameworks of Marketing planning. In conclusion, formalization has double and conflicting roles and implications.


c) Concentration: It relates to the hierarchy of power and the degree of participation in decision-making. Concentration increases as the hierarchy level that makes decisions inside the organisation gets higher, and as the degree of participation of the remaining levels in the decision-making procedures, gets lower. Zaltman et.al. (1973) stress that strict emphasis on the hierarchy of power diminishes organisational innovation, since it encourages only making a positive feed-in of the return. The reduced participation in the decision taking process, can also act as a “wall” in the perceptions and ideas that emerge during the process of innovation development. On the contrary, wide participation enhances the commitment to the completion of the project, since it gives the feeling of “ownership” of some part of it. Negative relations between the concentration and the use of the results of Marketing research, and also between the reliability and functionality of plans, were found in the empirical research of Deshpande (1983) and John and Martin (1984), while Wind (1982) reports integration problems in centrally oriented structures. There are conflicting reports, regarding the dimension of concentration. Thus, while the studies of (Hage and Aiken 1970), (Palumbo, 1969), (Blau, 1973), (Daft and Becker, 1978), (Hage and Dewar, 1973), found a negative correlation between the degree of concentration and the results of innovations, Hage (1980) stresses that concentration can be positively correlated to innovation. According to him, the mechanistic organisations, which are characterized by small rates of change, also constitute the places where a radical innovation can take place, because therein it is very likely that a crisis will happen, precisely as it takes place in an organisational structure that is receptive to dictatorial practices (p. 243). Probably the fairest representation of truth is given by Galbraith and Nathanson (1978) who stated that a “good” structure does not exist, but the structure that can be characterized as optimal is the one that is adapted to the needs and the commands of the environment of objectives and aims. This theory explains why certain organisational structures are more efficient with higher levels of binding – adaptation – adoption of innovation, and also mostly productive regarding their research teams (1983). Questionable innovations require an organisational structure in which experts coming from different schools of thought and having different mentalities, can be integrated in team frameworks that function smoothly. The proposed structure is compatible - if not similar - to the findings for the electronics enterprises (Burns and Stalker, 1961), and to NASA, as described by Galbraith (1973).

The attitude of Top Management is the second variable that influences the level of achieved integration. Souder (1977), (1981) and Souder and Chakrabarti (1978) consider that the top management can either promote or prevent the development of a productive communication channel between the Research and Development department and the Marketing department. In the most effective cases of integration, Souder and Chakrabarti (1978), found that there was a common wage system for the departments of Marketing and Research and Development, and that both departments felt that the responsibility of the success or failure of a project, is divided equally and is common. Both (functional) groups had the feeling that their mutual collaboration enjoyed the management’s appreciation – a lot of evidence supported this fact.

Also, the attitude of top management regarding risk-taking was found that has a positive effect on the successful result of the innovation (Quinn, 1979), (Roberts, 1978), (Roberts,1979), (Roberts and Fusfeld, 1981), (Wind,1982). It is also further implied that encouragement for the development of good relations between Marketing and Research and Development regarding top management plays an important part.

The third, decisive factor for achieving integration, are the social and cultural differences between the leaders of the Marketing and R&D departments.

A systems approach of the Marketing and Research and Development departments. High Tech markets driven by demand and supply.

William Shanklin and John Ryans (1984) also explored the systemic approach of Marketing and Research and Development department for the highest possible beneficial result for the high technology enterprises. They do not consider the association of the two departments simply as an


alternative that leads high technology enterprises to the direction of the results’ optimisation, but as the unique direction and a one-way road.

In order to provide a deeper understanding of the high tech markets, they separate them based on whether they are driven by supply or demand (see figure 8).

Figure 8: Comparison of Supply-driven and Demand-driven high tech enterprises.


In the case where technological progress is the force that justifies the creation of markets and

demand, then the market is supply-driven. In this case, the role of Marketing is complementary and supporting, and it is mainly based on and shaped by business dexterity and instinct. In the case of a totally new product, a market should be created (Omni, 1981). The technology-driven high tech enterprises are closely connected to the kind of enterprises often referred to as innovation-driven enterprises. The central strategic objective of those enterprises is the achievement of lucrative commercial applications for their laboratorial findings. The department of Research and Development is the main instigator of Marketing and the Marketing’s direction of efforts. In other words, Marketing is called on to satisfy an optional need from the viewpoint that this need is created and comes in the surface following the discovery of the solution to the purchasing problem (i.e. after the satisfaction) of the need. Moreover, the innovations carried out in high tech enterprises of that type, may have a wider spectrum of applications, nevertheless this factor is either slow to become perceptible, or is completely ignored, leading to devastating consequences for the commercialisation of innovations.

On the other hand, there are the high tech enterprises, which are driven by demand, i.e. they shape their orientation based on existing demand as their main driver. This second case is the natural development of a normal route to maturity in the life cycle of those enterprises. In this case, there is an inversion of roles, since the department of Research and Development is requested to give technological solutions to purchasing problems already realised by consumers, which Marketing – as it owes to – has recognized and outlined. The undertaking of that conventional responsibility by Marketing includes more focus on the 4 p’s, less stress on the approach of business intelligence and instinct, and more focus on the creation and tightening of narrow cohesive channels of communication and smooth collaboration with the department of Research and Development. In the second case, the


results of the laboratories of the Research and Development come as an answer to the commands of Marketing.

Companies in high tech sector should effectively face the systematic challenge of organisational adaptation as high tech markets mature, i.e. to evolve from supply-driven to demand-driven. This transition, which is a result of an evolutionary course to maturity, is not a smooth and without-obstacles process. The reorientation of executives, who have been successful in the supply-driven high tech enterprises, is rather difficult and many times unfeasible. This transition requires the a priori awareness of the changes imposed by the factors that shape things inside the maturity framework. It also imposes reorganisation and reorientation according to the commands of a more customer-oriented approach. The success of this transition depends on whether a core capability transforms to become core rigidity, and whether the high tech company will become involved in a continuous and constantly growing state of losing contact with its customers. That is to say, the special core capability, which put the enterprise forward and gave it a comparative advantage in the first stages of its existence and development, can represent a critical factor of disorientation and literally a barrier at the period of maturity. III. Conclusions Literature in Marketing and Research & Development has examined the need for integration of the Marketing and R&D departments in High Tech markets, several times and from different points of view. Prevalent among the existing theories is the widely accepted common issue of suboptimisation which results as a chain reaction of the lack of integration among the operational departments-especially R&D and Marketing. The vast majority of the High Tech industry companies face obstacles in achieving intradepartmental cooperation – that being valid mostly for the MKT and R/D departments – due to innate factors that characterize the high technology environment. Due to the fact that these factors combine with other – innate as well – and particular to High Tech characteristics in an interrelated and dynamic way, a network of unfavorable conditions, escalates constantly.

That is the exact reason why the high-tech environment is described as volatile, turbulent, tumultuous and highly uncertain. The paper approached this status through the viewpoint of integration, that is through the provision of the systemic intradepartmental cooperation. What this paper attempted to achieve, was to encapsulate in a compact way the approaches provided by scientific research and focalize on the numerous findings in the high tech area. The underlining of the simultaneous coexistence of the multitudinous characteristics provided by the conducted literature review constitutes a solid assessment basis for further research attempts. The idea of the application of a strategic and concurrently systemic approach of operational intradepartmental coexistence and cooperation is prevalent almost throughout every research activity undertaken and described herein. This paper provided a multi-sided approach and in several occasions, synthesized relevant and substantial concepts from different knowledge fields in order to contribute a concluded and manifold view to the several high tech scholars. Defining terms of high tech have been analyzed; the realized degree of integration and intradepartmental cooperation/communication as innate and integral variables have been stressed out, while light has been shed on both qualitative and quantitative aspects. IV. References 1] Achrol, R.S., Review T. and Stern, L.W. (1983) The environment of marketing channel dyads:

a framework for comparative analysis. Journal of Marketing, 57, 4, 55-67. 2] Advisory Council for Applied Research and Development (ACARD), (1978) 3] Publication, Industrial Innovation, December , HMSO. 4] Ames, C.B. (1968) Marketing planning for industrial products. Harvard Business Review, 46,

5, 100-111.


5] Blau, P. M. (1961) The Organization of Academic Work. New York: Wiley. 6] Burns, T. and Stalker, G. M. (1961) The Management of Innovation. London: Tavistock. 7] Cheng, J. L. C.(1983) Interdependence and coordination in organizations: a role system

analysis. Academy of Management Journal, 26, 1, 156-162. 8] Child, J. (1972) Organizational structure, environment, and performance: the role of strategic

choice. Sociology 6, 2-22. 9] (1981) Creativity in Modern Industry, Omni, 6. 10] Crawford, M. C. (1983) New Products Management. Homewood : Irwin Press. 11] Daft, R. L. and Becker S. (1978) Innovation in organizations: adoption in school organization.

New York: Elsevier. 12] Deshpande, R. (1982) The Organizational context of market research use. Journal of

Marketing, 46, 91-101. 13] D.J. (1968) Power and role specifity in organizational theory. Public Administrative Review,

29, 237-248. 14] Dutta S., Νarasimhan O. and Rajiv S. (1999) Success in high technology markets: is marketing

capability critical. Marketing Science, 18, 4, 547. 15] Freeman, C. (1974) The Economics of Industrial Innovation. Baltimore, MD: Penguin. 16] Gaibraith, J.R. and Nathanson D.A. (1978) Strategy implementation: the role of structure and

process. St. Paul. MN: West. 17] Galbraith, J. R. (1977) Organizational Design, Reading. MA: Addison – Wesley. 18] Galbraith, J. (1973) Designing complex organizations reading mass: Addison – Wesley. 19] Galbraith, J. R. and Nathans, D. A. (1978) Strategy implementation: the role of structure and

process. St. Paul, MN: West. 20] Galbraith, J. R. and Nathanson D. A. (1978) Strategy implementation: the role of structure and

process. St. Paul. MN: West. 21] Griffin A. and Hauser J. (1996) Integrating R and D and marketing: a review and analysis of

the literature. Journal of Product Innovation Management, 13, 191-215. 22] Gupta A.K., Raj S.P. and Wilemon D.L. (1986) A model for studying R and D and marketing

interface in the product innovation process. Journal of Marketing , 50, 15-17 . 23] Hage, J. and Aiken M. (1970) Social Change in Complex Organizations, New York: Random

House. 24] Hage, J. and Dewar R. (1973) Elite Values versus Organizational Structure in Predicting

Innovation, Administrative Science Quarterly, 18, 279-290. 25] Hage, J. (1980) Theories of Organizations, New York: Wiley. 26] Hutt, M.D. and Speh T.W. (1984) The marketing strategy center: diagnosing the industrial

marketer’s interdisciplinary role. Journal of Marketing, 48, 53-61. 27] Jauch, L.R. and Kraft, K.L. (1986) Strategic management of uncertainty. Academy of

Management Review, 11, 4, 777-790. 28] Kahn, R.L., Wolfe, D.M., Quinn, R.P., Snoek J. D. and Rosenthal, R.A. (1964) Organizational

Stress. New York: Wiley. 29] Kanungo, Rabindra N. (1979) The concept of alienation and involvement revisited. Psychology

Bulletin, 86, 119-138. 30] Khandwalla, P. N. (1972) Environment and its impact on the organization. International

Studies of Management Organization, 2, 297-313. 31] Kotler, Ph. (1984) Marketing Management: analysis, planning control (5thed). London:

Prentice Hall. 32] Lawrence, P.R. and Lorsch J.W. (1969) Organization and environment: managing

differentiation and integration. Homewood, IL: Irwin. 33] Lawrence, P.R. and Lorsch, J.W. (1967) Differentiation and integration in complex

organizations. Administrative Science Quarterly, 12, 1, 1-47.


34] Miaoulis, G. and Laplaca, P. (1982) A systems approach for developing high technology products. Industrial Marketing Management, II, 253-262.

35] Millman, A. F. Understanding barriers to product innovation at the R and D/ marketing interface., European Journal of Marketing, 16, 5.

36] Moenaert RK. and Souder WE. (1990) An information transfer model for integrating marketing and R+D personnel in new product development projects. Journal of Product Innovation Management , 7, 91-107.

37] National Science Foundation (1983) The process of technological innovation: reviewing the literature. Washington, DC: NSF.

38] Organ, D.W. and Greene C.N. (1981) The effects of formalization on professional involvement: a compensatory process approach. Administrative Science Quarterly, 26, 237-248.

39] Palumbo John, G. and Martin J. (1984) Effects of organizational structure of marketing planning on credibility and utilization of plan output. Journal of Marketing Research, 21, 170-183.

40] Perrow, C. (1967) A framework for the comparative analysis of organizations. American Sociological Review, 32, 2, 194-208.

41] Quinn, J.B. (1979) Technological innovation, entrepreneurship, and strategy, Sloan. Management Review, 20, 19-30.

42] Roberts, E. (1978) What do we know about managing R & D. Research Management, 21, 6 –11.

43] Roberts, E. (1979) Stimulating technological innovation – organizational approaches. Research Management, 22, 26-30. .

44] Roberts, E. and Fusfeld, A.R. (1981) Staffing the innovative technology based organization .Sloan Management Review, 22, 19-26.

45] Rogers, E. M. (1983) Diffusion of Innovation. New York: Free Press. 46] Saren, M. A. (1984) A classification and review of models of the intra firm innovation process,

R andD . Management, 14, 1, 11-24. 47] Schneider, S.C. (1987). Information overload–causes and consequences. Human Systems

Management, 7, 2, 143-153. 48] Shanklin, W. L. and Ryans, J. K. (1984) Organizing for high – tech marketing. Harvard

Business Review, 164 – 171. 49] Silver, S. D., Cohen, B. P. and Rainwater J. (1988) Group structure and information exchange

in innovative problem solving, In Lawler, E. J. and Markowsky, B (eds) Advances in Group Processes, Greenwich, Conn. : JAI Press, 5, 169-194.

50] Song X.M. and Parry, M.E. A (1997) Cross–national comparative study of new product development processes: Japan and the United States, Journal of Marketing, 61, 2, 1-18.

51] Souder, W.E. ( 1977) An Exploratory study of the coordinating mechanism between R & D and marketing as an influence on the innovation process. Final Report to the National Science Foundation. Pittsburgh: Univ. of Pittsburgh, School of Engineering.

52] Souder, W.E. (1981) Disharmony between R & D and marketing. Industrial Marketing Management, 10, 67-73.

53] Souder and A. K. Chakrabarti. (1978) The R&D / marketing interface: results from a empirical study of innovation projects. IEEE Transactions on Engineering Management, EM–25, 88-93.

54] Souder, W. E. Managing New Product Innovations. Lexington MA: Lexington Books 55] Thompson J. P. (1967) Organizations in Action. New York: McGraw – Hill. 56] Victor, B. and Blackburn. (1987) Determinants and consequences of tech uncertainty: a

laboratory and field investigation. Journal of Management Studies 24, 4, 387 – 403. 57] Wind, Y. (1982) Product Policy, Concepts, Methods and Strategy. New York: Addison –

Wesley.


58] Wind and Robertson, T. S. (1983) Marketing strategy: new directions for theory and research. Journal of Marketing, 47, 12-25.

59] Woodward, J. (1965) Industrial organization: theory and practice. London: Oxford Univ. Press.

60] Zaltman, G., Duncan, R. and Holbek, K. (1973) Innovations and Organizations. New York: John Wiley.


Is There A Link Between Strategic Human Resource Management Practices and Organizational Outcomes?

A Study of The Northern Cyprus Manufacturing Industry

Şerife Zihni Eyüpoğlu Dr., Faculty of Economics and Administrative Sciences, Near East University, Cyprus

Tel: 0392 2236464 ext.271/278 Fax: 0392 2236461 E-mail: [email protected]

Berna Kocaman

Prof.Dr., Faculty of Law, Ankara University, Turkey [email protected]

Abstract This study evaluates the links between non-managerial SHRM practices and organizational outcomes (organizational and market performance). Results based on a study of 24 manufacturing organizations in the Turkish Republic of Northern Cyprus indicate that these practices have a positive affect on organizational performance. However, only limited support is obtained for the hypothesized relationship between the practices and market performance.

Keywords: Strategic human resource management; non-managerial employees; organizational outcomes; bundle.

I. Introduction There is no question regarding the fact that employee’ qualities, attitudes, and behaviour in the workplace goes a long way in determining whether an organization will be successful or not. People (human resources) provide organizations with an important source of sustainable competitive advantage and the effective management of this human capital (employees), not physical capital, may be the ultimate determinant of organization performance and survival. Of course, this depends heavily on employee skills and commitment as key factors in the process of creating value. Accordingly, it is only logical for organizations to develop the productive potential of their human capital in order to achieve superior performance. Careful management of this human capital paves the way for the focus on human resource management in organizations. II. Strategic Human Resource Management The question of how HRM policies and practices are linked to organizational performance has been a subject of great interest to academicians, practitioners, and researchers for many years (Wright et al., 1999). Although much has been made of recent contributions there is still a great deal of uncertainty around the precise nature of these linkages (Mueller, 1996). In recent years references have been made to strategic human resource management (SHRM). The shift from HRM to SHRM has broadened the


focus of research from the micro analytic research to the more macro or strategic perspective (Delery and Doty, 1996). The strategic perspective has grown out of research with the desire to demonstrate the importance of human resource practices for organizational performance.

The underyling assumption of SHRM is that organizations adopting a particular strategy require HR practices that are different from those required from those required by organizations adopting alternative strategies (Jackson and Schuler, 1995). There is a growing body of empirical SHRM research, however, differences have arisen due to the alternative perspectives. Some researchers, such as the likes of Delaney, Lewin, and Ichniowski, 1989; Huselid, 1993,1995; Osterman, 1994; Pfeffer, 1994; and Terpstra and Rozell, 1993, have used a “universalistic perspective”, also known as the “best practices” approach to SHRM. These researchers posit that some HR practices are always better than others and that all organizations should adopt these best practices. Pfeffer (1994) argued that the greater use of 16 management practices, such as participation and empowerment, incentive pay, employment security, promotion from within, training and development, results in higher productivity and profits across organizations. Also, Osterman (1994) argued that a number of innovative work practices, such as teams, job rotation, quality circles, and total quality management, results in productivity gains for all organizations. The practices identified by these two researchers have been labelled “high performance work practices” (HPWP).

The second perspective of SHRM emphasized by other researchers is known as the “contingency perspective”. Contingency theorists include Butler, Ferris, and Napier (1991); Gomez-Mejia and Balkin (1992); Lengnick-Hall and Lengnick-Hall (1988). They argue that for an organization to be effective, its HR practices must be consistent with other aspects of the organization. More specifically, contingency theorists have attempted to show how a number of HR practices are consistent with different strategic positions and how these practices related to firm performance, the organization’s strategy being considered to be the primary contingency factor in SHRM literature (Balkin and Gomez-Mejia, 1987; Begin, 1993; Gomez-Mejia and Balkin, 1992).

The configurational approach is the third perspective to SHRM. In general, configurational theories are concerned with how the pattern of multiple independent variables is related to a dependent variable rather than with how individual independent variables are related to the dependent variable. In simpler terms, the configurational perspective of strategic human resource management argues that it is the pattern of human resource management practices that contribute to the attainment of organizational goals (Wright and McMahan, 1992). Thus the effectiveness of any human resource practice depends on its interrelationship with others; they do not stand on their own. Configurational theorists include MacDuffie (1995); Doty, Glick, and Huber (1993);Meyer, Tsui, and Hining (1993); and Arthur (1992). III. Prior Empirical Work Conducted One of the first major attempts to examine how effective management of human resources might contribute positively to organizational performance was made by Ferris, Russ, Albanese, and Martocchio (1990). Their study of the U.S. Construction Industry showed that firms that had human resource management departments were generally high performers (larger total sales volume), firms that had a higher percentage of their workforce unionized also performed better than firms with a lower percentage, and firms performed better when they engaged in more formalized strategic planning.

Also, based on questionnaire responses from human resource managers in the U.S. Steel minimills, Arthur (1994) concluded that the mills with “commitment” systems had higher productivity, lower scrap rates, and lower employee turnover than those with “control” systems.

Using the work of Miles and Snow (1984), who identified three pure types of strategies; defender, prospector, and analyzer, Bird and Beecher (1994) examined the linkages between business strategy and human resource management strategy in Japanese subsidiaries in the U. S. and discovered that the subsidiaries with matched combinations did outperform their “unmatched” counterparts, particularly with regard to human resource management related outcomes (employee morale, tenure, promotion, and turnover).


The first to carry out research on human resource “bundles”(systems) was MacDuffie (1995) who examined automotive assembly plants and concluded that innovative human resource practices affect performance not individually but as a group.

Huselid (1995) concluded that high performance work practices were associated with lower employee turnover, greater productivity, and higher corporate financial performance.

Delaney and Huselid (1995), using 590 profit and non-profit firms from the National Organizations Survey, found a positive association between HRM practices such as training and staffing selectivity, and perceptual firm performance measures. However, their results did not support that complementaries among HRM practices enhance firm performance. IV. Data, Hypothesis, and Methodology Data The main objective of this study is to ascertain the relationship between SHRM and organizational outcomes and to contribute to the literature in some respects. It attempts to ascertain whether the impact found for SHRM on organizational performance in other countries also holds for Northern Cyprus, and whether the SHRM practices lead to better outcomes.

This study has differentiated between HRM practices used for managerial and nonmanagerial employees. This is because most organizations use differents policies and practices for these two group of employees. Most of the studies previously conducted have lumped together HRM practices for different groups of employees or have only studied a certain employee category. Including both groups of employees in the study would have allowed to investigate whether or not the same or different HRM practices for managers and non-managers are associated with superior performance. However, this study chose to examine one group of employee, namely non-managerial employees, due to anticipated difficulties that would have been faced in data collection and the necessity of a much more detailed, time consuming, and lengthy study. Hypotheses Based on the review of the literature presented, this study proposes that a positive relationship exists between SHRM and organizational outcomes. The SHRM practices choosen for study are employee selection, training and development, participation, information sharing, performance appraisal, incentive compensation, and promotion. These seven practices are considered to be broad enough to capture the key dimensions of a SHRM system. Each variable will be tested against organizational outcomes and then a bundle of the variables will be tested against organizational outcomes.

Thus, the hypotheses tested are as follows; 1] There is a positive relationship between selection and organizational outcomes (organizational

and market performance). 2] There is a positive relationship between training and organizational outcomes (organizational

and market performance). 3] There is a positive relationship between participation and organizational outcomes

(organizational and market performance). 4] There is a positive relationship between information sharing and organizational outcomes

(organizational and market performance). 5] There is a positive relationship between performance appraisal and organizational outcomes

(organizational and market performance). 6] There is a positive relationship between incentive compensation and organizational outcomes

(organizational and market performance). 7] There is a positive relationship between promotion and organizational outcomes (organizational

and market performance).


8] A bundle of SHRM practices is positively related with organizational outcomes (organizational and market performance).

Sample and Data Collection The Northern Cyprus manufacturing industry constitutes the scope of the study. Based on the information given by the Northern Cyprus Chamber of Industry and the State Planning Organization (1998) there are 741 registered manufacturing organizations in Northern Cyprus which are divided into 9 sub-sectors. The organizations included in this study ideally would have then those that employed at least 40 employees, because this limit has been used in other studies related to this area (Osterman, 1994). However, these studies were conducted in, for example, the U.S.A., U.K., and Spain where the populations are great and where exist a great number of medium and large sized organizations. Due to the small population of Northern Cyprus in comparison to these countries and because a majority of the organizations are classed as small such a limit could not be imposed. Thus, the organizations included in this study had to employ at least 20 employees. Organizations employing less than 20 employees were excluded from the study because these organizations either may not have a well developed HRM function or have inadequate managerial resources to complete the questionnaire. Again, the manufacturing industry has been selected for study due to a majority of the empirical studies conducted on the link between HRM practices and organizational performance being in this industry. The number of organizations that met the criteria established (the total number of privately owned manufacturing organizations with at least 20 employees) for the study is 93.

A questionnaire was constructed to carry out the investigation. The questionnnaire contained items concerning HRM practices and was administered to the HR managers or the persons in charge of HRs in the organizations. The HR managers/persons responsible for the HRs were used to collect data so that it would be valid and reliable due to the fact that these persons have adequate knowledge about the organization’s HRM activities as well as having knowledge on most organizational aspects.

Ideally, multiple respondents/informants (HRM managers and non-managerial employees) should have been chosen because it is recognized that experiences are likely to vary between levels of staff; HRM practices may exist but only have relevance to a minority of those working for the organization, and data collected from employees at different levels of the organization makes it possible to access the rhetoric of what the HR group is trying to achieve as well as the reality experienced by employees (Legge, 1995; Truss et al., 1997). However, this would have complicated the study considerably.

The questionnaire items were constructed based on the items used in other similar studies, such as Delaney and Huselid (1996), Kalleberg and Moody (1994), and Huselid (1995). The items concerning HRM practices refer to non-managerial employees. The fact that a specific group of employees is referred to creates problems in as far as the generalization of the research results to other professions is concerned. However, limiting the occupation under study makes the comparison of units studied easier.

The original version of the questionnaire was in English. This questionnaire was translated into the local language (Turkish) and in order to pre-test it was administered to 3 managers from 3 different organizations. Issues assessed during the pilot test were instructions and statement clarity, questionnaire layout and appearance, and the length of the questionnaire. Improvements were made based on the comments of the managers whom participated in the pilot test. The revised questionnaire consisted of 56 items. The study was cross sectional. Measurement of Variables The HRM practices mentioned below have been chosen for study because by looking at previous empirical studies it appears that these practices are the ones most likely to elicit the appropriate employee behaviour in question in the study.

The SHRM Variables (Independent Variables): The first three HRM activities were assessed based on the works of Snell and Dean (1992). Selection was assessed using 6 items which captured aspect such as the importance given to selection, and to understand whether the organization makes


efforts into ensuring employee-job fit. A sample item is “Employees are administered an employment test prior to hiring”. Selection is measured by the variable SELECT. The effort made by the organization to train its employees also serves as an indication of its concern for the future of its workforce. Training was assessed using a 9 item scale which assessed the importance and effort given to training activities for non-managerial employees in the organization, whether training was just for job-related skills or did it also include the development of problem solving and decision making abilities, and whether the organization views training as an investment. A sample item is “The organization gives great priority to training its non-managerial employees”. Training and development is measured by the variable TRAIN. Employee participation was measured using a 4 item scale assessing the degree of which non-managerial employees contribute to the production and decision making processes, and whether supervisors delegate authority when necessary. A sample item is “Non-managerial employees have the opportunity to use personal initiative and judgements in their jobs”. Employee participation is measured by the variable PART. Based on the works of Huselid (1995), information sharing was assessed by a 4 item scale and measured the nature of communication between managerial and non-managerial employees in the organization by understanding whether periodical meetings between managerial and non-managerial employees were conducted about company issues, and whether survey were conducted in which non-managerial employees are asked about their degree of job satisfaction. A sample item is “Non-managerial employees are constantly kept informed about organization related matters through newsletters and periodic meetings”. Information-sharing is measured by the variable INFO. Assessments of the remaing three pracrtices were made according to the works of Snell and Dean (1992), and Delaney and Huselid (1996). Performance appraisal was assessed using a 7 item scale to measure the importance given to performance appraisal, the purpose of evaluation (administrative or developmental), and whether it was conducted on a regular basis. A sample item is “Formal performance appraisal programs exist in the organization”. Performance appraisal is measured with the variable PERF. Incentive compensation measured the degree of the linkage of performance and pay level, whether performance appraisal was linked to compensation, whether an employee’s KSAs has an influence on pay level, and whether compensation was relative or higher compared to other organizations doing the same kind of job. Assessment was made using a 7 item scale. “Compensation for non-managerial employees is closely tied to individual or group performance”. Incentive compensation is measured with the variable INCEN. Promotion measured the degree to which non-entry vacancies were filled from within the organization, the degree of managerial and administrative positions that were filled by people who entered the organization at an entry-level position, and whether promotion was based on merit. Assessment was made using a 5 item scale. A sample item is “Non-entry jobs are usually filled from within the organization”. Promotion is measured with the variable PROMO.

All of the mentioned HRM practices measured respondents views based on Likert scales ranging from 1, “strongly disagree” to 5, “strongly agree”. The Organizational Outcomes Variable (Dependent Variable): In order to study the effects of human resource management on organizational outcomes, different indicators have been used. According to Dyer and Reeves (1995), the indicators can be classified into three groups. First is outcomes directly related to human resource management, such as employee turnover, and employee morale (Huselid, 1995; MacDuffie, 1995). The second group is constituted by operational or bottom-line results with organizational productivity being used most frequently (Arthur, 1994; Huselid, 1995; Ichniowski et al., 1997; MacDuffie, 1995). Also, product quality has been used (Arthur, 1994; Huselid, 1995; Ichniowski et al., 1997; MacDuffie, 1995). The third group of indicators is constituted by financial performance measures. Some of these indicators are profit or total sales (Terpstra and Rozell, 1993), and profitability rates (Delery and Doty, 1996; Huselid, 1995). Also Huselid (1995) and Huselid et al. (1997) employ a market measure Tobin’s q. In other studies subjective indicators have been employeed for this type of measurement (Delaney and Huselid, 1996).


Based on the above literature and considering the scope of the study, this study measured the dependent variable organizational outcomes, and based on the works of Delaney and Huselid (1996) asked respondents to make subjective managerial judgements on how they thought their organization was performing compared to their competitors regarding;

organizational performance (quality of the product, development of new products, ability to attract essential employees, low employee turnover, management-employee relations, and employee-employee relations), and

market performance (market position/share, sales growth, and profitability compared to competitors).

This perceptual measure was used due to the reluctance of Northern Cyprus organizations to give

financial information about their organization’s performance. Also, perceptual measures of performance like this have been used in other similar studies conducted in the U.S. and U.K., such as Delaney and Huselid (1996), and Youndt et al (1996), and research has found measures of perceived organizational performance to correlate positively with objective measures of organizational performance (Powell, 1992). One of the limitations of existing studies is that they almost without exception used quantitative techniques.

Respondents made their judgments on a 5-point Likert scale ranging from 1, “poor” to 5, “excellent”.

Organizational performance is measured with the variable PERORG and market performance is measured with the variable MRKPER.

Control Variables: Looking at prior research, the control variables suitable for this study are organization size (Blau and Schoenherr, 1977; Snell, 1992), market competition (Delaney and Huselid, 1996), and organization age (Delaney and Huselid, 1996). However, upon analysis of the study environment it became evident that there were no major differences between organizations when it came to size, age, and degree of competition faced. Therefore it was decided that the study variables need not be controlled. Data Analysis In this study the data have been analyzed through the use of frequency distributions, correlations, multiple regression, means, and standard deviations obtained from the SPSS (Statistical Packages for Social Science). V. Results General Findings The process resulted in the completion of 24 questionnaires from 24 organizations. This resulted in a 25.5 % (24/93) response rate. The low response rate reflects the underlying difficulties in obtaining cooperation from organizations when it comes to organizational issues. However, it is felt that the data obtained is sufficient to discern general trends and relationships.

The respondent organizations in the study came from a variety of sub-sectors, namely food and beverages (25%); paper and paper products (17%); non-metallic mineral products (17%); chemical/petroleum and plastics (13%); textile (8%); metal products (8%); and forestry/furniture and fixtures (4%); and others (8%). Results for Organizational Performance Table 1 provides the most relevant descriptive statistics and the bivariate Pearson correlations for the main variables used in the study.


Table 1: Descriptive Statistics and Correlation Matrix for Study Variables (Organizational Performance)

Descriptive Statistics

3,8750 ,5872 244,0556 ,6249 243,7731 ,7249 243,9896 ,6776 243,6875 ,9303 24

PERORGSELECTTRAINPARTINFO

Mean Std. Deviation N Descriptive Statistics

3,8750 ,5872 243,7560 ,8271 243,7440 ,7722 243,7750 ,9066 24

PERORGPERFINCENPROMO

Mean Std. Deviation N

Correlations

1,000 ,443 ,501 ,309 ,349 ,530 ,511 ,624,443 1,000 ,548 ,048 ,524 ,320 ,370 ,448,501 ,548 1,000 ,253 ,757 ,602 ,502 ,538,309 ,048 ,253 1,000 ,240 ,475 -,068 ,293,349 ,524 ,757 ,240 1,000 ,470 ,545 ,418,530 ,320 ,602 ,475 ,470 1,000 ,248 ,474,511 ,370 ,502 -,068 ,545 ,248 1,000 ,665,624 ,448 ,538 ,293 ,418 ,474 ,665 1,000

, ,015 ,006 ,071 ,047 ,004 ,005 ,001,015 , ,003 ,411 ,004 ,064 ,038 ,014,006 ,003 , ,116 ,000 ,001 ,006 ,003,071 ,411 ,116 , ,129 ,010 ,377 ,082,047 ,004 ,000 ,129 , ,010 ,003 ,021,004 ,064 ,001 ,010 ,010 , ,121 ,010,005 ,038 ,006 ,377 ,003 ,121 , ,000,001 ,014 ,003 ,082 ,021 ,010 ,000 ,24 24 24 24 24 24 24 2424 24 24 24 24 24 24 2424 24 24 24 24 24 24 2424 24 24 24 24 24 24 2424 24 24 24 24 24 24 2424 24 24 24 24 24 24 2424 24 24 24 24 24 24 2424 24 24 24 24 24 24 24

PERORGSELECTTRAINPARTINFOPERFINCENPROMOPERORGSELECTTRAINPARTINFOPERFINCENPROMOPERORGSELECTTRAINPARTINFOPERFINCENPROMO

Pearson Correlation

Sig. (1-tailed)

N

PERORG SELECT TRAIN PART INFO PERF INCEN PROMO

As Table 1 indicates, and consistent with much prior research on the relationship between HRM practices and organizational performance the correlations between the HRM practices and organizational performance are all positive (at p<0.05). This providing preliminary partial support for the hypotheses 1, 2, 3, 4, 5, 6, and 7.

However, a more detailed analysis to test the hypotheses is required so regression analysis was

used. Table 2 indicates the regression results and Anova test for organizational performance. The model was highly significant with an R square of 0.547 and an adjusted R square of 0.349. It

can also be seen that six of the seven HRM practice coefficients are positive, and five of them (selection, participation, performance appraisal, incentive compensation, and promotion) are statistically significant providing further partial support for hypotheses 1, 3, 5, 6, and 7. The F-test indicates that the HRM practices jointly explain a significant amount of the variance in organizational performance.

It can also be seen that of HRM practices incentive compensation has the largest beta indicating that it has the largest effect on organizational performance followed by performance appraisal.

In general, the results indicated in Tables 1 and 2 suggest that the bundle of HRM practices is positively associated with organizational performance consistent with hypothesis 8.


Table 2: Regression Model and ANOVA Test for Organizational Performance

Model Summary

,740a ,547 ,349 ,4738Model1

R R SquareAdjustedR Square

Std. Error ofthe Estimate

Predictors: (Constant), PROMO, PART, INFO,SELECT, PERF, INCEN, TRAIN

a.

ANOVAb

4,339 7 ,620 2,761 ,044a

3,592 16 ,2247,931 23

RegressionResidualTotal

Model1

Sum ofSquares df Mean Square F Sig.

Predictors: (Constant), PROMO, PART, INFO, SELECT, PERF, INCEN,TRAIN

a.

Dependent Variable: PERORGb.

Coefficientsa

,471 1,029 ,458 ,653,222 ,203 ,237 1,094 ,290,106 ,245 ,130 ,432 ,672,172 ,187 ,198 ,917 ,373

-,199 ,184 -,316 -1,081 ,296,191 ,168 ,269 1,139 ,272,293 ,214 ,386 1,368 ,190

8,933E-02 ,184 ,138 ,487 ,633

(Constant)SELECTTRAINPARTINFOPERFINCENPROMO

Model1

B Std. Error

UnstandardizedCoefficients

Beta

Standardized

Coefficient

st Sig.

Dependent Variable: PERORGa.

Results for Market Performance Table 3 provides the most relevant descriptive statistics and the bivariate Pearson correlations for the variables used in the study.

Table 3 indicates that only three of the seven HRM practices (participation, incentive compensation, and promotion) correlate positively with market performance (at p<0.05), however the degree of association is weak.

Table 4 indicates the regression results and Anova test for market performance. The model indicates an R square of 0.427 and an adjusted R square of 0.176 suggesting that the

model was not significant. It can also be seen that four of the seven practice coefficients are positive, and three of them (selection, performance appraisal, and incentive compensation) are statistically significant providing support for hypotheses 1, 5, and 6. The F-test indicates that the HRM practices jointly do not explain a significant amount of the variance in market performance.


Table 3: Descriptive Statistics and Correlation Matrix for Study Variables (Market Performance)

Descriptive Statistics

3,6354 ,6915 244,0556 ,6249 243,7731 ,7249 243,9896 ,6776 243,6875 ,9303 243,7560 ,8271 243,7440 ,7722 243,7750 ,9066 24

MRKTPERSELECTTRAINPARTINFOPERFINCENPROMO

MeanStd.

Deviation N

Correlations

1,000 -,140 -,167 ,206 -,206 -,002 ,213 ,044-,140 1,000 ,548 ,048 ,524 ,320 ,370 ,448-,167 ,548 1,000 ,253 ,757 ,602 ,502 ,538,206 ,048 ,253 1,000 ,240 ,475 -,068 ,293-,206 ,524 ,757 ,240 1,000 ,470 ,545 ,418-,002 ,320 ,602 ,475 ,470 1,000 ,248 ,474,213 ,370 ,502 -,068 ,545 ,248 1,000 ,665,044 ,448 ,538 ,293 ,418 ,474 ,665 1,000

, ,257 ,217 ,167 ,167 ,496 ,159 ,420,257 , ,003 ,411 ,004 ,064 ,038 ,014,217 ,003 , ,116 ,000 ,001 ,006 ,003,167 ,411 ,116 , ,129 ,010 ,377 ,082,167 ,004 ,000 ,129 , ,010 ,003 ,021,496 ,064 ,001 ,010 ,010 , ,121 ,010,159 ,038 ,006 ,377 ,003 ,121 , ,000,420 ,014 ,003 ,082 ,021 ,010 ,000 ,24 24 24 24 24 24 24 2424 24 24 24 24 24 24 2424 24 24 24 24 24 24 2424 24 24 24 24 24 24 2424 24 24 24 24 24 24 2424 24 24 24 24 24 24 2424 24 24 24 24 24 24 2424 24 24 24 24 24 24 24

MRKTPERSELECTTRAINPARTINFOPERFINCENPROMOMRKTPERSELECTTRAINPARTINFOPERFINCENPROMOMRKTPERSELECTTRAINPARTINFOPERFINCENPROMO

Pearson Correlati

Sig. (1-tailed)

N

MRKTPER SELECT TRAIN PART INFO PERF INCEN PROMO

In general, the results obtained suggest that the bundle of HRM practices does not positively

associate with market performance, thus support is not found for the part of hypothesis 8 related to market performance.

In conclusion, full support was obtained for hypotheses 1,5, and 6, and partially for hypotheses 3, 7, and 8. VI. Conclusion and Discussion The study has examined the relationship between SHRM practices and organizational outcomes and adds to the growing empirical evidence suggesting that people are the preeminent organizational resource and the key to achieving outstanding performance (Arthur, 1994; Huselid, 1995; MacDuffie, 1995). As the analysis revealed, the non-managerial SHRM practices are found to have a positive affect on organizational performance indicating that the effective implementation of key non-managerial SHRM practices should be able to bring in higher levels of organizational performance. However, it would seem that the same SHRM practices indicate no such association with market performance.


There are some limitations that represent opportunities for future research. Firstly, a rigorous test of the model would require longitudinal data about HRM adoption. However, the data and the empirically tested hypotheses pertain to the cross-section.

The fact that the study was limited to the manufacturing industry means that the study’s conclusions cannot be made applicable to all professions and sectors. However, several researchers (Huselid, 1995; Pfeffer, 1994) have argued that it is unlikely that the best HR practices vary across industries.

Also, another important limitation is the relatively small sample size, and the use of a single channel of data collection may be a source of bias (Salancik and Pfeffer, 1977; Podsakoff and Organ, 1986).

Finally, the study is also limited because it included only a small set of the population of SHRM practices currently being used by organizations. Although the SHRM practices included were derived from existing theory, other SHRM practices may also have important impacts on organizational outcomes. Table 4: Regression Model and ANOVA Test for Market Performance

Model Summary

,653a ,427 ,176 ,6278 ,427 1,700 7 16 ,179Model1

R R SquareAdjusted RSquare

Std. Errorof theEstimate

R SquareChange F Change df1 df2

Sig. FChange

Change Statistics

Predictors: (Constant), PROMO, PART, INFO, SELECT, PERF, INCEN, TRAINa.

ANOVAb

4,691 7 ,670 1,700 ,179a

6,307 16 ,39410,997 23

RegressionResidualTotal

Model1

Sum ofSquares df Mean Square F Sig.

Predictors: (Constant), PROMO, PART, INFO, SELECT, PERF, INCEN,TRAIN

a.

Dependent Variable: MRKTPERb.

Coefficientsa

1,276 1,364 ,936 ,3638,4E-02 ,270 ,076 ,311 ,760-,135 ,325 -,142 -,417 ,682,557 ,248 ,546 2,245 ,039-,455 ,244 -,612 -1,863 ,081

6,8E-02 ,222 ,081 ,304 ,765,821 ,284 ,916 2,889 ,011-,355 ,243 -,465 -1,458 ,164

(Constant)SELECTTRAINPARTINFOPERFINCENPROMO

Model1

B Std. Error

UnstandardizedCoefficients

Beta

StandardizedCoefficients

t Sig.

Dependent Variable: MRKTPERa.


Bibliography 1] Arthur, J.B. (1994) “Effects of human resource systems on manufacturing performance and

turnover”, Academy of Management Journal, Vol. 37, pp.670-687. 2] Arthur, J.B.(1992) “The link between business strategy and industrial relations systems in

American steel minimills”, Industrial and Labour Relations Review,Vol. 45, pp.488-506. 3] Begin, J.P. (1992) “Comparative HRM:A systems perspective”, The International Journal of

Human Resource Management, Vol. 3, No.3, pp. 379-408. 4] Bird, A. and Beecher, S. (1994) “Links between business strategy and human resource

management strategy in U.S-based Japanese subsidiaries:An empirical investigation”, Journal of International Business Studies, Vol. 1, pp. 24-46.

5] Blau, P.M. and Schoenherr, R. (1977) The Structure of Organizations, New York: Basic Books. 6] Delaney, J.T. and Huselid, M.A. (1996) “The impact of HRM practices on perceptions of

organizational performance”, Academy of Management Journal, Vol. 39, pp. 949-68. 7] Delaney, J.T., Lewin, D. and Ichniowski, C. (1989) Human Resource Policies and Practices in

American Firms, Washington, DC:US Government Printing Office. 8] Delery, J.E. and Doty, H. (1996) “Models of theorizing in strategic human resource

management:Tests of universalistic, contingency, and configurational performance predictions”, Academy of Management Journal, Vol. 39, pp. 802-835.

9] Dyer, L. and Reeves, T. (1995) “Human resource strategies and firm performance:What do we know and where do we need to go?”, The International Journal of Human Resource Management, Vol. 6, pp. 656-670.

10] Ferris, G.R., Russ, G.S., Albanese, R. and Martocchio, J.J. (1990), “Personnel/human resources management, unionization, and strategy determinants of organizational performance”, Human Resource Planning, Vol. 13, pp. 215-27.

11] Huselid, M.A. (1995) “The impact of human resource management practices on turnover, productivity, and corporate financial performance”, Academy of Management Journal, Vol. 38, pp. 635-670.

12] Huselid, M.A., Jackson, S.E. and Schuler, R.S. (1997) “Technical and strategic human resource management effectiveness as determinants of firm performance”, Academy of Management Journal, Vol. 40, pp. 171-88.

13] Ichniowski, C., Shaw, K. and Prennushi, G. (1997) “The effects of human resource management practices on productivity”, American Economic Review, Vol. 87, pp. 291-313.

14] Jackson, S.E. and Schuler, R.S. (1995) “Understanding human resource management in the context of organizations and their environments”, In M.R. Rosenzweig and L.W. Porter (eds.), Annual Review of Psychology, Vol. 46, pp. 237-264. Palo Alto, CA:Annual Reviews.

15] Kalleberg, A.L. and Moody, J.W. (1994) “Human resource management and organizational performance”, American Behavioural Scientist, Vol. 37, pp. 948-962.

16] MacDuffie, J.P. (1995) “Human resource bundles and manufacturing performance: Flexible production systems in the world auto industry”, Industrial and Labour Relations Review, Vol. 48, pp. 197-221.

17] Meyer, A.D., Tsui, A.S. and Hining, C.R. (1993) “Configurational approaches to organizational analysis”, Academy of Management Journal, Vol. 36, No. 6, pp. 1175-1195.

18] Miles, R.E. and Snow, C.C. (1984) “Designing strategic human resource ystems”, Organizational Dynamics, Vol. 13, pp. 36-52.

19] Mueller, F. (1996) “Human resources and strategic assets: An evolutionary resource-based theory”, Journal of Management Studies, Vol. 33, No. 6, pp. 757-785.

20] Osterman, P. (1994) “How common is workplace transformation and who adopts it?”, Industrial and Labour Relations Review, Vol. 47, pp. 173-188.

21] Pfeffer, J. (1994) Competitive Advantage Through People, Boston:Harvard Business School Press.


22] Podsakoff, P.M. and Organ, D.W. (1986) “Self-reports in organizational research: Problems and prospects”, Journal of Management, Vol. 12, pp. 531-44.

23] Powell, T.C. (1992) “Organizational alignment as competitive advantage”, Strategic Management Journal, Vol. 13, pp. 119-134.

24] Salancik, G.R. and Pfeffer, J. (1977) “An examination of need-satisfaction models of job attitudes”, Admininistrative Science Quarterly, Vol. 22, pp. 427-56.

25] Snell, S.A. (1992) “Control theory in strategic human resource management: The mediating effects of administrative information”, Academy of Management Journal, Vol. 35, pp. 292-327.

26] Snell, S.A. and Dean, J.W. (1992) “Integrated manufacturing and human resource management”, Sloan Management Review, Vol. 23, pp. 47-61.

27] Wright, P.M. and McMahan, G.C. (1992) “Theoretical perspectives for strategic human resource management”, Journal of Management, Vol. 18, pp. 295-320.

28] Wright, P.M., McCormick, B., Sherman, S. and McMahan G.C. (1999) “The role of human resource practices in petro-chemical refinery performance”, International Journal of Human Resource Management, Vol. 10, pp. 551-571.

29] Youndt, M.A., Snell, S.A., Dean, J.W. and Lepak, D.P. (1996) “Human resource management, manufacturing strategy, and firm performance”, Academy of Management Journal, Vol. 39, pp. 836-866.


The Effect of Relative Earnings Performance on Firms’ Accrual Decisions: Evidence from France

Ramzi Benkraiem Management Research Centre

Department of accounting The University of Toulouse Graduate School of Management – France

Tel: 00 33 (0)6 78 48 83 09 E-mail: [email protected]

Abstract This study empirically examines the relationship between relative earnings performance (REP) defined against industry and discretionary accruals (DA). We assume that the degree of DA depends systematically on REP. Our sample consists of 768 firm–year observations over the period 2000–2003. Three test methods are used: mean accrual difference tests, mean accrual distribution tests across 10 REP portfolios and regression analysis. After controlling a potential “backing out problem”, our results support the hypothesis that (1) poor REP firms tend to choose income increasing DA while good REP firms tend to make the opposite; (2) the variations of these DA depend directly and negatively on the variations of REP and (3) the negative relation between DA and REP is particularly more striking for firms with poor REP than for firms characterized otherwise. These results suggest that REP is an important factor affecting managers’ accounting practices.

Keywords: Relative earnings performance, Pre-managed earnings, Operating cash flows, Discretionary accruals, French accounting practices.

I. Introduction Because of certain latitude offered by accounting standards, managers can fit their intentions by selecting a number of accrual choices for evidence of “earnings management”. These choices, while being individually in conformity with the GAAP1, are entirely directed towards the achievement of precise objectives. Considering the recent financial scandals2, earnings management is always a topical concern not only in France but also in other countries.

Several theoretical and empirical studies tried to analyse managers’ incentives to manage reported earnings3. In this framework, Martson and Craven (1998) specify that investors as well as financial analysts’ pressures push managers to take short-term oriented accrual choices. Particularly, when financial performances do not meet specific targets, managers can be temped to adjust their earnings in order to reduce the visibility of their weakness (Burgstahler and Dichev, 1997; Degorge, Patel and Zeckhausser, 1999; Moherle, 2002; Mard, 2004). For instance in the French context, Chalayer and Dumontier (1996) provide evidence on the adjustment of accruals upwards during hard periods and

1 The French accounting standards tend to approach the IASB norms. Starting in 2005, French quoted companies have to apply the international standards for the consolidated accounts. 2 Some financial scandals fall under the simple violation of GAAP. We are not interested by such cases. We only consider the discretionary practices which respect the accounting standards. 3 For a review of this literature, see Stolowy and Breton (2004).


downwards during good periods. Moreover, Saboly (2001) and Djama (2003) investigate the relation between financial distress and earnings management. They report that firms modulate their accounting numbers prior to failure.

Our study extends this body of research by investigating whether the REP affects managers accounting practices. As mentioned by Bagnoli and Watts, (2000; p. 378) “previous theoretical work has generally viewed earnings management as a single firm’s one shot disclosure, rather than as a choice that may depend, in part, of the disclosure of its rivals”. There exists little evidence in France on how firms in an industry manage accruals based on their REP. Therefore, the purpose of this paper is to empirically examine the relationship between the REP and DA. For that, we employ three test methods: mean accrual difference tests, mean accrual distribution tests across 10 REP portfolios and regression analyses.

Our results support the hypothesis that (1) poor REP firms tend to choose income increasing DA while good REP firms tend to make the opposite; (2) the variations of these DA depend directly and negatively on the variations of REP and (3) the negative relation between DA and REP is particularly more striking for firms with poor REP than for firms characterized otherwise.

The remainder of the paper proceeds as follows: section 2 reviews the background and develops the hypothesis. Section 3 specifies the research design. Section 4 reports and discusses the empirical results and section 5 concludes with the principle contributions of the paper. II. Background and Hypothesis Development The REP is supposed to inform us about the firms’ performances using industry as a benchmark. The examination of such a variable can be of particular interest. Indeed, Porter (1980) notes that the structural characteristics of industries are a determining factor of a firm’s performance. Holmstrom (1982) adds that companies draw an advantage when comparing their performances with those of rival companies operating in the same industry. Theoretically, the REP displays aspects of performance that managers cannot control. Hence, it can provide a better basis for managers’ work evaluation.

In this direction, investors and creditors often use the REP when making decisions to allocate their funds. This REP constitutes to some extent a comparison criterion. It enables them to better evaluate and judge the firm’s performance. Moreover, it is also likely to be used in determining executive compensation (Bannister and Newman, 2003). This could constitute strong incentives for managers to adjust their own earnings in view of their REP and thus to draw a comparative advantage.

Pyo and Lustgarten (1990) specify that one firm’s earning decisions are affected by other firms’ performance information. They evoke a co-movement of intra-industry information transfer. Trueman (1990) shows that the manager gains from earnings adjustment when these earnings are highly correlated with those of industry and when the industry displays good performances. Park and Ro (2004) empirically confirm these results. They find a higher level of accruals adjustment for firms which have earnings that are highly correlated with industry than for those characterised otherwise.

Defond and Park (1997) empirically test Fundenberg and Tirole’s (1995) theory concerning the explanation of the managers’ accounting choices. They take as a performance indicator the pre-managed earnings compared to the industry median pre-managed earnings. They consider that the performance is poor (good) when the pre-managed earnings are below (above) this median. Then, they study the relation between REP and DA for a two-year period. Finally, they come to the conclusion that managers choose income-increasing (decreasing) accounting policies when their REP is poor (good). The degree (and not the direction) of DA also depends on future expected REP.

In the same way, De Albornoz and Alcarria (2003) are directly inspired by Defond and Park’s (1997) analysis. They check the same hypothesis in the Spanish context. After controlling a potential “backing out problem”1, they provide similar results. However, they find that DA depend partially on 1 The association between DA and pre-managed earnings may increase the risk of statistical inferences. The authors control this possibility and find that their results do not seem to be affected by this type of problem.


future REP. They state that “current performance considered on its own is a more important explanatory variable of discretionary accounting choices than expected future performance”.1

Consistently, Peasnell et al. (2000) study among other things the impact of REP on working capital accruals. Their results, for a set of English firms, show that mangers facing poor performances choose income-increasing DA.

Chung and al. (2005) are also interested in the impact of REP on DA, but they rely on another measurement. Indeed, they follow the same principal of REP calculation (the concerned company performances minus the industry median), but they replace the pre-managed earnings by the operating cash flows. They confirm a negative association between the REP and DA.

Consistent with these previous researches, it would be logical to expect managers to be attentive to the REP defined against industry. Thus, they would choose either increasing or decreasing accounting policies depending on the evolution of such a variable. In fact, when REP is poor, the managers seem more justified to boost reported earnings in order to reduce such performances’ visibility. On the other hand, when REP is good, they appear more incited to store up income-increasing options for use in future poor times.

In addition, a fall of the REP should be expressed by a higher pressure on firms’ managers. Consequently, greater incentives to manage earnings upwards may occur and vice versa. So, the DA mean variations (i.e. the extent of earnings adjustment) should be directly conditioned by the reciprocal variations of the REP.

Finally, considering the interest carried by investors and financial analysts for good performances, the mangers should be more prompted to increase reported earnings when the REP is poor than to reduce them in the opposite case. III. Research Design Sample Table 1 summarises the sample selection procedure of this study. The accounting figures are available on AMADEUS. This database provides in particular consolidated financial statements of 629 French firms listed on the Paris stock market. The starting sample is composed of all the French companies available over the period 2000-2003. Financial, insurance and holding companies (65.00-70.99 and 74.15 NACE2 codes) are excluded. Their specific accounting characteristics make the estimation of their discretionary accruals problematic (Peasnell et al., 2000). Firms with insufficient data for accruals determination are deleted. Furthermore, to ensure efficient estimation of discretionary accruals, industries (defined according to the two-digit NACE code) with less than ten observations in a year are removed (Koh, 2003). This selection process yields a final sample of 768 firm-year observations distributed on 10 non-financial industries. Table 1: Sample selection procedure

1 De Albornoz and Alcarria (2003), p. 460. 2 The NACE-code is an industry classification chart which is comparable to the US or UK SIC

Criteria NB of firm-years Non-finance-related firm-year observations on AMADEUS (2000 – 2003) Less Firms with missing data for accruals determination Selected firms Less Industries with fewer than 10 observations per year Final sample

1428

(261)

1167 (399) 768


Measurement of Variables Discretionary Accruals We focus on earnings management through DA which are assumed to reflect subjective accounting choices made by managers. For that, we initially have to compute total accruals (TA). Along the lines of prior research (Healy, 1985; Jones, 1991), we use the balance sheet approach as follows:

TAit = ( CAit – Cashit) – ( CLit – STDit) – DEPit (1)

Where for firm i in year t: CAit is the change in current assets; Cashit is the change in cash and cash equivalents; CLit is the change in current liabilities; STDit is the change in long-term debt included in current liabilities; and DEPit is the depreciation and amortization expense.

Total accruals can be decomposed into discretionary and nondiscretionary accruals. However, these two components are not directly observable. We therefore need a model to separate them. The modified Jones (1991) model proposed by Dechow et al. (1995) is one of the most frequently used in current research. In particular, the cross-sectional version using industry and fiscal year combination is considered to perform better than the time-series counterpart (Bartov et al., 2000). Consequently, we use the following cross-sectional version of the modified Jones model:

TAit /Ait-1 = β1 (1/ Ait-1) + β2 [( REVit – ARit)/Ait-1] + β3 (PPEit /Ait-1) + εit (2)

Where for firm i in year t: Ait-1 is the lagged total assets; REVit is the change in sales revenues; ARit is the change in account receivables; PPEit is the gross property plant and equipment; and εit is the regression residual.

Nondiscretionary accruals (NDA) that arise from the normal activity of the firm are defined as the fitted values from Eq. (2). DA are considered as the difference between TA and the fitted value, NDA. Thus, the firms’ DA are computed as:

DAit /Ait-1 = (TAit /Ait-1) – (NDAit /Ait-1) (3)

DAit /Ait-1= (TAit /Ait-1) – [b1 (1/ Ait-1) + b2 (( REVit – ARit)/Ait-1) +

b3 (PPEit /Ait-1)]

Where b1, b2 and b3 are the OLS regression estimates of β1, β2 and β3 obtained from the original Jones model (Peasnell et al., 2000; Bartov et al., 2000). Relative Earnings Performance (REP) To determine the REP, we call upon two measurements derived from the pre-managed earnings and the cash flow from operations.

The pre-managed earnings (PME) can be calculated as follows (Subramanyam, 1996):

PMEit /Ait-1 = (IOit /Ait-1) – (DAit /Ait-1) (4)

Where for firm i in year t: IOit is the income from operations (income before discontinued operations and extraordinary items).

The pre-managed earnings estimate the firms’ earnings which have not been subjected to any managerial discretion. On the basis of that, it appears possible to exploit such a variable to determine the REP. Following Defond and Park (1997), and De Albornoz and Alcarria (2003), we consider whereas that the pre-managed earnings compared to the industry median pre-managed earnings are able to inform us about the firms’ REP. Our first indicator of REP is thus written as follows1:

REPit /Ait-1 = (PMEit /Ait-1) – Median (PME/Ait-1) (5)

To support this first approach, we have also recourse to a second measurement of the REP. We keep the same principle of determination. However, we replace the pre-managed earnings by another

1 All variables are scaled by lagged total assets to reduce heteroscedasticity.


performance’s indicator namely the cash flow from operations (Chung and al., 2005). Our second indicator of the REP is written then as follows:

REPit /Ait-1 = (OCFit /Ait-1) – Median (OCF/Ait-1) (6)

Where OCFit is the operating cash flow and Median (OCF/Ait-1) is the industry median operating cash flows. Test Methods Mean Accrual Difference Tests The first method consists in using the two REP measurements already developed to divide the sample into two groups according to whether the REP is poor or good. Taking as a basis the first measurement of the REP, we consider a dummy variable which takes the value 1 if [(PMEit /Ait-1) – Median (PME/Ait-1) < 0] and 0 if not. We proceed in the same way for the second measurement. We consider an additional dummy variable which always takes the value 1 if [(OCFit /Ait-1) – Median (OCF/Ait-1) < 0] and 0 if not.

Obviously, value 1 indicates a poor REP, whereas, value 0 indicates a good REP. Thereafter, we examine whether mean accruals are different between poor REP firms and good REP firms. Also, we carry out mean accrual difference tests between the different groups. Under the null hypothesis of no earnings management, the mean accruals would be similar for the two portfolios of firms. Nevertheless, under the hypothesis of earnings management, they would be different. More exactly, we expect to see positive mean accruals for poor REP firms, negative mean accruals for good REP ones and significant mean accrual differences between them. Mean Accrual Distribution Tests The second method consists in subdividing the sample into several portfolios of firms and analysing the mean accrual distributions through them. We choose to subdivide our sample into 10 groups according to the magnitude of the REP.

More concretely, that amounts to gathering the companies in deciles of [(PMEit /Ait-1) – Median (PME/Ait-1)] and [(OCFit /Ait-1) – Median (OCF/Ait-1)]. The managers of firms forming part of the highest (lowest) deciles seem more (less) incited to adjust DA upwards (downwards). In other words, if the REP is an influential factor, then mean accruals will be positive (negative) for the top (bottom) groups. Simultaneously, they will gradually decrease as we progress from the highest to the lowest REP deciles (group number 1 -the highest- points out the smallest REP, while group number 10 -the lowest- represents the largest REP). Regression Analyses The regression analyses allow an examination of the relationship between the REP and DA in a more direct way. Indeed, they make it possible to project DA on REP as well as some control variables and to test if there are significant differences across different portfolios. Thus, we develop the following linear regression models (Eq. (7-10)):

DAit = β0 + β1 Poor (PME)it + β2 [Poor (PME) x ((PME) – Median (PME))]it + β3 SIZEit + β4

DEBTit + β5 OCFit-1 + εit (7)

DAit = β0 + β1 Poor (OCF)it + β2 [Poor (OCF) x ((OCF) – Median (OCF))]it + β3 SIZEit + β4 DEBTit + β5 OCFit-1 + εit (8)

DAit = α0 + α1 Good (PME)it + α2 [Good (PME) x ((PME) – Median (PME))]it + α3 SIZEit + α4 DEBTit + β5 OCFit-1 + εit (9)

DAit = α0 + α1 Good (OCF)it + α2 [Good (OCF) x ((OCF) – Median (OCF))]it + α3 SIZEit + α4 DEBTit + α5 OCFit-1 + εit (10)


Where for firm i in year t: Poor (PME) and Poor (OCF) are dummy variables equal to 1 when (PMEit) and (OCFit) are below their respective industry medians, otherwise they are coded 0. [Poor (PME) x ((PME) – Median (PME))]it and [Poor (OCF) x ((OCF) – Median (OCF))]it are respectively the interactions of Poor (PME), Poor (OCF) and the REP (which express the only poor REP). Good (PME) and Good (CFO) are dummy variables coded 1 if (PMEit) and (OCFit) are above their respective industry medians, and 0 if not. [Good (PME) x ((PME) – Median (PME))]it and [Good (CFO) x ((OCF) – Median (OCF))]it are respectively the interactions of Good (PME), Good (CFO) and the REP (which express the only good REP).

SIZEit is the log of market capitalisation at the end of the fiscal year; DEBTit is the total debt divided by total assets; CFOit-1is the lagged operating cash flow divided by total assets.

When firms’ performances are below (above) the industry median, the situation should stimulate the income-increasing (decreasing) accounting choices. We expect positive coefficients β1 (Eq. (7-8)) and negative coefficients α1 (Eq. (9-10)).

In parallel, being either poor or good, the REP should be negatively related with DA. In other words, an increase of REP should go with a decrease of DA. Thus, we anticipate negative coefficients β2 and α2 (Eq. (7-10)).

Besides, we try to assess whether mangers are more prompted to increase reported earnings when the REP is poor than to reduce them in the opposite case. For that, we compare the coefficients (β2 and α2) as well as the explanatory power (adjusted R square) of our four models. If it is the case, then we expect to observe β2 > α2 and higher adjusted R squares of Eq. (7-8) than those of Eq. (9-10).

We also control for size, debts, and lagged operating cash flows in Eq. (7-10) (Watts and Zimmerman, 1986; Chung and al., 2005; Cheng and Warfield, 2005; Becker et al., 1998; Dechow, 1994; among others). However, since previous research has not reported a strict consensus on signs, we do not predict them. The “Baking-Out” Problem The use of DA in computing one of the partitioning variables and at the same time as a dependent variable may increase a potential statistical bias called “the backing-out” problem (Lim and Lustgarten, 2002; Elgers et al., 2003). Estimating pre-managed earnings by subtracting estimated DA may cause a mechanical association between the dependant and explanatory variables in favour of earnings management hypothesis. In other words, the association between DA and REP may be negative even if these DA are measured with error. As mentioned above, De Albornoz and Alcarria (2003) use the pre-managed earnings to compute the REP. To check the validity of their findings, they replicate a method advocated by Lim and Lustgarten (2002). In fact, they recalculate the pre-managed earnings using the nondiscretionary accruals (NDA) instead of DA. Indeed, the NDA do not reflect subjective accounting choices made by managers to modulate reported earnings. So, under the hypothesis of earnings management, they expect to see different results when using these two components of accruals. The results confirm their prediction. As such, they conclude that their findings do not seem to be affected by the “the backing-out” problem. To examine whether our results are driven by this type of problem, we apply the same analysis. IV. Empirical Results Descriptive Statistics Summary statistics for variables used in the analysis are reported in Table 2. The income from operations (IO) and operating cash flow (OCF) are on average positive. Their means are respectively 7.45 % and 9.56 % of lagged total assets. Thus, firms in the sample tend to be profitable. The mean value of total accruals (TA) is negative (-2.11 % of lagged total assets) because of the weight of depreciation and amortization. Nondiscretionary accruals (NDA) are on average negative (-2.7 % of lagged total accruals) with a standard deviation of 9.10 %. Discretionary accruals are rather


manipulated upwards. Their mean is positive and their standard deviation is 17.38 %. It appears that the “variability” of DA is higher than its counterpart of NDA. The variable SIZE records the largest standard deviation. Mean leverage (DEBT) in the sample is 5.36 % of lagged total assets. Lagged cash flows from operations (CFOit-1) have a mean and median receptively of 8% and 10 %. They are rather close to those of the current year. Table 2: Descriptive statistics for variables

Percentiles Mean Std. Deviation 25 50 75 IO 0.074555 0.1231654 0.006750 0.062208 0.139977OCF 0.095694 0.1917020 0.045315 0.103667 0.1768901/ Ait-1 0.000070 0.0009032 0.000006 0.000020 0.000045( REV – AR) 0.174371 0.6666361 -0.032056 0.075712 0.253075PPE 0.359485 0.2927371 0.167003 0.319962 0.475568TA -0.021138 0.1979602 -0.098149 -0.034549 0.030808NDA -0.027408 0.0910116 -0.068255 -0.033702 0.002632DA 0.006269 0.1738696 -0.065026 0.004356 0.067759PME – Median (PME) -0.000795 0.1850734 -0.067756 0.000000 0.077051OCF – Median (OCF) -0.011224 0.1908057 -0.061703 0.000000 0.064116SIZE 3.849422 1.7898879 2.551204 3.532208 4.999389DEBT 0.053605 0.0752319 0.011318 0.026377 0.060919CFOit-1 0.083310 0.2919546 0.037586 0.107758 0.183557 Mean Accrual Difference Tests The sample is broken down into two groups according to whether the REP is poor or good. Table 3 exposes the mean DA for each portfolio of firms. Table 3: Mean discretionary accruals

PME<Median(PME) N Mean Std.

Deviation t P

DA Poor REP: 1 381 0.064859 0.1953998 6.479

0.000

Good REP: 0 387 -0.051412 0.1255313 -8.057 0.000

OCF<MED(OCF) N Mean Std. Deviation t P DA Poor REP: 1 389 0.054603 0.1979492 5.440 0.000

Good REP: 0 379 -0.043339 0.1274711 -6.619 0.000

Taking as a basis the pre-managed earnings, we notice a positive and significant DA mean (6.48 %) for the worst performing firms. On the other hand, we observe a negative and also significant DA mean (-5.14 %) for the best performing firms. The results resting on operating cash flows go in the same direction. In fact, the recorded DA means display practically similar values.

As expected, managers of firms where pre-managed earnings and operating cash flows undershoot (overshoot) the respective industry medians make income increasing (decreasing) DA. At this level, Table 4 reports a complementary examination by presenting the mean accrual differences between the two different groups.


Table 4: Mean DA difference between poor and good REP firms

Levene's Test for Equality of Variances t-test for Equality of Means

95% Confidence Interval of the Difference

PME<Median(PME) F Sig. t df Sig. (2-tailed)

Mean Difference

Std. Error Difference Lower Upper

Equal variances assumed

1.259 0.262 9.826 766 0.000 0.116271 0.0118329 0.0930419 0.1394994 DA

Equal variances not assumed 9.794 646.470 0.000 0.116271 0.0118715 0.0929594 0.1395819

Levene's Test for

Equality of Variances t-test for Equality of Means

95% Confidence Interval of the Difference

OCF<Median(OCF) F Sig. t df

Sig. (2-tailed)

Mean Difference

Std. Error Difference Lower Upper

Equal variances assumed

1.758 0.185 8.129 766 0.000 0.097942 0.0120483 0.0742905 0.1215935 DA

Equal variancesnot assumed 8.173 664.930 0.000 0.097942 0.0119834 0.0744121 0.1214719

Levene’s test1 for equality of variances as well as the P value (Sig. 2-tailed) allows us to analyse

the reported results. The mean DA difference between poor and good performers is 11.62 % for the first REP indicator and 9.79 % for second. The results found starting from these two indicators are rather close and lead to the same conclusions of Table 3. Mean Accrual Distribution Tests We subdivide the 768 firm–year observations into 10 portfolios based on the REP ranks. In other words, we classify all sample firms, initially in deciles of PME – Annual industry median (PME), then in the second time in deciles of OCF – Annual industry median (OCF).

Table 5 reports the variations on average as well as the standard deviations (second line) of the income from operations, operating cash flow, total and discretionary accruals. For the PME - Median (PME) portfolios, the average of OCF are -20.20 % for the first group and 12.81 % for the last. The OI means for the same groups are 1.78 % and 16.12 %, respectively. The mean variations of IO, except the second decile’s case, seem positively related to those of REP. Simultaneously, the magnitude of TA means decreases as we progress from the first to the last REP group. However, these TA display on average values of 21.99 % for the first group, - 5.85 % for the sixth, - 4.47 % for the eighth, and 8.90 % for the tenth. The reciprocal variations of TA and REP do not seem completely stable. DA means are 21.99 % for the first and – 14.22 % for the last. As predicted, the extent of earnings management (the DA mean variations) seems to be conditioned by the reciprocal variations of the REP. Furthermore, the relationship between the variations of these two variables appears entirely stable.

On the other hand, the OCF - Median (OCF) portfolios report practically similar results. In particular, TA means record values of 21.99 % for the first REP group, -5.68 % for the sixth, - 5.04 % for the seventh and -5.52 % for the eighth. Here also, the reciprocal variations of TA and REP do not seem completely steady. The DA means decrease from the highest (23.59 %) to the lowest (-13.38 %) group. Also, the DA mean’s variations seem negatively related to those of REP. On the whole, the reciprocal variations of DA and REP appear the most regular. Our interpretation of this result is that an

1 The interpretation of DA mean difference can be done only in the case of equality of variances. In other words, that amounts to rejecting the null hypothesis of no equality of variances. This null hypothesis would be rejected if the significance of Levene’s test is higher than the significance level α (1%, 5% or 10%).


increase of the REP is expressed by a lower pressure on managers to overvalue their reported earnings by adjusting the DA and vice versa. Table 5: Mean accrual distribution tests across 10 REP portfolios

Deciles of OCF-MED(OCF) Income from Operations Operating Cash flow Total Accruals Discretionary Accruals 1 0.037452 -0.253842 0.291294 0.235927 (0.1390250) (0.3821660) (0.4184337) (0.3634852) 2 0.013594 0.005244 0.008350 0.045570 (0.0863393) (0.0422731) (0.0915341) (0.1041736) 3 0.020779 0.044066 -0.023287 0.019094 (0.0923005) (0.0358557) (0.0929921) (0.0943468) 4 0.039676 0.076334 -0.036658 0.013025 (0.0817303) (0.0329470) (0.0783237) (0.0744828) 5 0.036389 0.086872 -0.050483 -0.024111 (0.0906499) (0.0273725) (0.0928534) (0.0842158) 6 0.058347 0.115157 -0.056810 -0.018416 (0.0887251) (0.0340372) (0.0951048) (0.0886573) 7 0.094765 0.145252 -0.050487 -0.007718 (0.0885076) (0.0331243) (0.0882504) (0.0789387) 8 0.119527 0.174764 -0.055236 -0.017964 (0.1105093) (0.0347783) (0.1145881) (0.1070743) 9 0.137608 0.217277 -0.079669 -0.045007 (0.1296966) (0.0401924) (0.1252648) (0.1278864) 10 0.195183 0.343541 -0.148358 -0.133811 (0.1696689) (0.1086424) (0.1888168) (0.1752978) Total 0.074555 0.095694 -0.021138 0.006269 (0.1231654) (0.1917020) (0.1979602) (0.1738696)

Regression Analyses

Deciles of PME-Median(PME) Income from Operations Operating Cash flow Total Accruals Discretionary Accruals 1 0.017864 -0.202091 0.219954 0.250935 (0.1262250) (0.4053171) (0.4385556) (0.3532794) 2 0.005488 0.012986 -0.007497 0.056224 (0.0850389) (0.0800761) (0.1171185) (0.0860748) 3 0.013701 0.038013 -0.024312 0.017565 (0.0931984) (0.0822969) (0.1363942) (0.0893912) 4 0.043771 0.076500 -0.032729 0.017547 (0.0890616) (0.0524206) (0.0982589) (0.0833499) 5 0.045352 0.087921 -0.042570 -0.008743 (0.0850937) (0.0607489) (0.1017167) (0.0838267) 6 0.057385 0.115966 -0.058580 -0.012182 (0.0817080) (0.0565082) (0.0946249) (0.0789167) 7 0.085926 0.138019 -0.052093 -0.018744 (0.0924649) (0.0537942) (0.1052021) (0.0853097) 8 0.116709 0.161448 -0.044739 -0.028881 (0.0950226) (0.0542995) (0.1076764) (0.0858485) 9 0.145292 0.213789 -0.068497 -0.061809 (0.1287717) (0.1041990) (0.1607068) (0.1305161) 10 0.218629 0.307721 -0.089092 -0.142284 (0.1612797) (0.1281596) (0.1957465) (0.1790416) Total 0.074555 0.095694 -0.021138 0.006269 (0.1231654) (0.1917020) (0.1979602) (0.1738696)


Table 6 reports the results of the OLS regressions for Eq. (7-10). The coefficients β1 and α1 are significant with the expected positive (for β1) and negative (for α1) signs. This evidence suggests that the variables Poor and Good (PME; OCF) capture mangers’ incentives for income increasing and income decreasing DA, respectively.

The coefficients β2 and α2 are on the whole significant and negative. Whether they are poor or good, the REP is negatively related to the DA. These results confirm those summarised in table 5. However, the coefficients β2 of Eq. 7 and 8 (-0.949/- 0.925 and -0.815/ -0.811) are higher than those of Eq. 9 and 10 (-0.708/-0.530 and -0.727/-0.587); moreover, the adjusted R squares of the first two Eq. (0.585/0.636 and 0.551/0.601) are also higher than those of the second (0.145/0.223 and 0.108/0.177). Table 6: Regression analyses

DAit = β0 + β1 Poor (PME)it + β2 [Poor (PME) x ((PME) – Median (PME))]it + β3 SIZEit + β4 DEBTit + β5 OCFit-1 + εit (7)

DAit = β0 + β1 Poor (OCF)it + β2 [Poor (OCF) x ((OCF) – Median (OCF))]it + β3 SZEit + β4 DEBTit + β5 OCFit-1 + εit (8)

Adj. R2 0.585 0.636 Adj. R2 0.551 0.601 F 1083.013 268.936 F 941.977 231.873 P 0.000 0.000 P 0.000 0.000 Durbin-W 1.921 1.960 Durbin-W 1.977 2.020

DAit = α0 + α1 Good (PME)it + α2 [Good (PME) x ((PME) – Median (PME))]it + α3 SIZEit + α4 DEBTit + β5 OCFit-1 + εit (9)

DAit = α0 + α1 Good (CFO)it + α2 [Good (CFO) x ((OCF) – Median (OCF))]it+ α3 SIZEit + α4 DEBTit + α5 OCFit-1 + εit (10)

Adj. R2 0.145 0.223 Adj. R2 0.108 0.177 F 131.296 45.072 F 94.036 34.099 P 0.000 0.000 P 0.000 0.000 Durbin-W 1.909 1.961 Durbin-W 2.029 2.043 Where for firm i, in year t: Poor (PME) 1 if PMEit < Annual industry median (PME), 0 otherwise; Poor (OCF) 1 if OCFit < Annual industry median (OCF), 0 otherwise; Good (PME) 1 if RNMit > Annual industry median (PME), 0 otherwise; Good (OCF) 1 si OCFit > Annual industry median (PME), 0 otherwise;

B P Value

B P value

Intercept 0.039 0.000 -0.022 0.130

Good (OCF) -0.060 0.000

Good(OCF)x(OCF-MED) -0.727 0.000 -0.587 0.000

SIZE 0.023 0.000

DEBT -0.180 0.020

OCFt-1 0.047 0.024

B P value

B P value

Intercept 0.044 0.000 -0.006 0.689

Good (PME) -0.077 0.000

Good (PME)x(PME-MED) -0.708 0.000 -0.530 0.000

SIZE 0.023 0.000

DEBT -0.252 0.001

OCFt-1 0.036 0.073


OCF – MED (OCF) OCFit - Annual industry median (OCF); PME – MED (PME) PMEit - Annual industry median (PME); SIZE Ln of market capitalisation; DEBT Ratio total debts to total assets; OCFt-1 Lagged operating cash flow divided by total assets.

This supports the prediction that the mangers are more prompted to increase reported earnings when the REP is poor than to reduce them in the opposite case. This could be explained in particular by a substantial pressure exerted by investors and financial analysts on firms’ mangers reporting poor REP.

The coefficients of the control variables are significant at conventional levels. The coefficient of lagged operating cash flow is positive. The size and debt variables, consistent with prior research (Chung and al., 2005; Becker et al., 1998; Defond and Park, 1997), have positive and negative signs, respectively. The “Baking-Out” Problem We perform additional analyses to assess the robustness of our results. We reproduce the analysis of Lim and Lustgarten (2002). Thus, we redefine the pre-managed earnings as income from operations less nondiscretionary accruals (IO – NDA). Consistent with this reasoning, we restate the REP indicator as (IO – NDA) – annual industry median (IO – NDA).

Subsequently, we reiterate the mean accrual distribution tests and the regression analyses according to the redefined REP. If the initial findings are driven by the “backing-out” approach, then we would expect to observe similar results using either (IO – DA) or (IO – NDA). Table 7 presents mean accrual distributions in total discordance with those displayed in table 5. Especially, the DA means are positive for the first four deciles and negative for the last six. Consistently, these results are confirmed with those reported in Table 8. In fact, the coefficients of the redefined variables using (IO – NDA) display opposed signs to those of the respective variables reported in Table 6. These findings suggest that the initial results are not the simple consequence of the “backing out” problem. Thus, they provide evidence supportive of earnings management. V. Conclusion This paper empirically investigated French non-financial firms to determine if the relative earnings performance defined against industry affects the degree of earnings management. The examination of the REP is of a particular interest. Relatively few researches, particularly in France, have considered earnings management as a choice that may depend, in part, of the disclosure choices of rival firms.

To determine the REP, we called upon two measurements. We compared the pre-managed earnings and the cash flow from operations to their respective annual industry medians. We focused on earnings management through DA which are assumed to reflect subjective accounting choices made by managers. For that, we relied on the cross-sectional version of the modified Jones model using industry and fiscal year combination. Our sample consisted of 768 firm–year observations over the period 2000–2003. We broke down this sample into several groups according to the REP ranks. Then, we used three test methods: mean accrual difference tests, mean accrual distribution tests across 10 REP portfolios and regression analyses. After controlling a potential “backing out problem”, the results support the hypothesis that (1) poor REP firms tend to choose income increasing DA while good REP firms tend to make the opposite; (2) the variations of these DA depend directly and negatively on the variations of REP and (3) the negative relation between DA and REP is particularly more striking for


firms with poor REP than for firms characterized otherwise. On the whole, these findings suggest that the REP is an influential factor affecting firms’ accounting decisions. Table 7: Mean accrual distribution tests across 10 redefined REP portfolios

Table 8: Regression analyses using redefined variables

DAit = β0 + β1 Poor (Redefined)it + β2 [Poor (Redefined) x ((IO – NDA) – Median)]it + β3 SIZEit + β4 DEBTit + β5 OCFit-1 + εit

DAit = α0 + α1 Good (Redefined)it + α2 [Good (Redefined) x ((IO –NDA) –Median)]it + α3 SIZEit + α4 DEBTit + α5 OCFit-1 + εit

Adj. R2 0.025 0.054 Adj. R2 0.053 0.068 F 20.294 9.683 F 43.991 12.276 P 0.000 0.000 P 0.000 0.000 Durbin-W 2.034 2.075 Durbin-W 2.049 2.074

Deciles of [(IO-NDA) - Median (IO-NDA)] Income from Operations Operating Cash flow Total Accruals Discretionary Accruals 1 -0.080594 -0.097735 0.017141 -0.041540 (0.1197366) (0.4244222) (0.4651983) (0.3970833) 2 -0.014113 0.051622 -0.065735 -0.074784 (0.0902446) (0.1098557) (0.1622133) (0.1128105) 3 0.022725 0.067762 -0.045038 -0.032937 (0.0544000) (0.1234381) (0.1283780) (0.1207020) 4 0.033745 0.079278 -0.045533 -0.015049 (0.0550182) (0.1271031) (0.1387249) (0.1239745) 5 0.040814 0.088417 -0.047602 0.003022 (0.0545386) (0.0931963) (0.0994848) (0.0965727) 6 0.067748 0.107268 -0.039520 0.008407 (0.0539565) (0.0779124) (0.0908674) (0.0775919) 7 0.100979 0.092181 0.008798 0.051403 (0.0598907) (0.1451224) (0.1721730) (0.1392288) 8 0.124934 0.155726 -0.030792 0.018245 (0.0672952) (0.1106278) (0.1168768) (0.1046671) 9 0.166095 0.163488 0.002607 0.051866 (0.0761855) (0.1723389) (0.1940522) (0.1765172) 10 0.278036 0.234880 0.043156 0.094480 (0.1394965) (0.1532325) (0.1688550) (0.1449092) Total 0.074555 0.095694 -0.021138 0.006269 (0.1231654) (0.1917020) (0.1979602) (0.1738696)

B P value

B P value

Intercept 0.023 0.019 -0.004 0.814

Poor (Redefined) -0.053 0.001

Poor (Redefined) x ((IO – NDA) – Median)

0.351 0.000 0.084 0.393

SIZE 0.011 0.004

DEBT -0.051 0.538

OCFt-1 0.016 0.493

B P value

B P value

Intercept -0.018 0.013 -0.065 0.000

Good (Redefined) 0.030 0.051 Good (Redefined) x ((IO – NDA) – Median)

0.530 0.000 0.353 0.000

SIZE 0.004 0.011 0.004 DEBT 0.082 -0.038 0.648 OCFt-1

0.022 0.010 0.644


Where for firm i, in year t: Poor (Redefined) 1 if ( IOit – NDAit) < Annual industry median ( IOit – NDAit), 0 otherwise; Good (Redefined) 1 if ( IOit – NDAit) > Annual industry median ( IOit – NDAit), 0 otherwise; The other variables are as defined previously. VI. References 1] Bagnoli M., Watts S.G., 2000. The effect of relative performance on earnings management: a

game theoretic approach. Journal of Accounting and Public Policy 19 (4-5), 377-397. 2] Bannister J.W., Newman H.A., 2003. Analysis of corporate disclosures on relative performance

evaluation. Accounting Horizons 17 (3), 235-246. 3] Bartov E., Ferdinand A.G., Tsui J.S.L., 2000. Discretionary accruals models and audit

qualifications. Journal of Accounting and Economics 30 (3), 421-452. 4] Burgstahler D., Dichev I., 1997. Earnings management to avoid decreases and losses. Journal

of Accounting and Economics 24 (1), 397-413. 5] Chalayer S., Dumontier P., 1996. Performance économique et manipulations comptables : une

approche empirique. Actes du Congrès de l’AFC, mai, Valenciennes. 6] Cheng Q., Warfield T.D., 2005. Equity incentives and earnings management. The Accounting

Review 80 (2), 441-476. 7] Chung R., Firth M., Kim J.B., 2005. Earnings management, surplus free cash flow and external

monitoring. Journal of Business Research 58 (6), 766-776. 8] De Albornoz G.B., Alcarria J.J., 2003. Analysis and diagnostics of income smoothing in Spain.

European Journal of Accounting 12 (3), 443-463. 9] Dechow P., Sloan R., Sweeney A., 1995. Detecting Earnings Management. The Accounting

Review 70 (2), 193-225. 10] Defond L., Park C.W., 1997. Smoothing income in anticipation of future earnings. Journal of

Accounting and Economics 23 (2), 115-139. 11] Degeorge F., Patel J., Zeckhauzer R., 1999. Earnings management to exceed thresholds. Journal

of Business 72 (1), 1-35. 12] Djama C., 2002. Le risque de faillite modifie-t-il la politique comptable ?. Actes du Congrès de

l’AFC, mai, Louvain. 13] Elgers P.T., Pfeiffer R.J., Porter S.L., 2003. Anticipatory income smoothing: a re-examination.

Journal of Accounting and Economics 35 (3), 405-422. 14] Fundenberg D., Tirole J., 1995. A theory of income and dividend smoothing based on

incumbency rents. Journal of Political Economy 103 (1), 75-93. 15] Healy P., 1985. Evidence on the effect of bonus schemes on accounting procedure and accrual

decisions. Journal of Accounting and Economics 7, 85-107. 16] Holmstrom B., 1982, Moral hazard in teams. Bell Journal of Economics 13 (2), 324-340. 17] Jones J., 1991. Earnings management during import relief investigation. Journal of Accounting

Research 29 (2), 193-228. 18] Koh P.S., 2003. On the association between institutional ownership and aggressive earnings

management in Australia. The British Accounting Review 35 (2), 105-128. 19] Lim S., Lustgarten S., 2002. Testing for income smoothing using the backing out method: a

review of specification issues. Review of Quantitative Finance and Accounting 19 (3), 273-291. 20] Mard Y., 2004. Les sociétés françaises cotées gèrent-elles leurs chiffres comptables afin

d’éviter les pertes et les baisses de résultats. Comptabilité Contrôle Audit 10 (2), 73-98. 21] Martson C., Craven B., 1998. A survey of corporate perceptions of short-termism among

analysts and fund managers. The European Journal of Finance 4, 233-256. 22] Mc Nichols M., Wilson G., 1988. Evidence of earnings management from the provision for bad

debts. Journal of Accounting Research 26 (Supplement), 1-36.


23] Moehrle S., 2002. Do firms use restructuring charge reversals to meet earnings targets?. The Accounting Review 77 (2), 397-413.

24] Park M.S., Ro B.T., 2004. The effect of firm-industry correlation and announcement timing on firms’ accrual decisions. The British Accounting Review 36, 269-289.

25] Peasnell K.V., Pope P.F., Young S., 2000. Accrual management to meet earnings targets: UK evidence pre- and post-Cadbury. The British Accounting Review 32, 415-445.

26] Porter M.E., 1980. Competitive strategy. Free Press, New York. 27] Pyo Y., Lustgarten S., 1990. Differential intra-industry information transfer associated with

management earnings forecast. Journal of Accounting and Economics 13 (4), 365-379. 28] Saboly M., 2001. Information comptable et défaillance des entreprises: le cas français.

Comptabilité Contrôle Audit 7 (2), 67-86. 29] Stolowy H., Breton G., 2004. Accounts manipulation: A literature revue and proposed

conceptual framework. The review of Accounting and Finance 3 (1), 5-66. 30] Subramanyam K., 1996. The pricing of discretionary accruals. Journal of Accounting and

Economics 22, 249-281. 31] Trueman B., 1990. Theories on earnings announcement timing. Journal of Accounting and

Economics 13 (3), 285-301. 32] Watts R.L., Zimmerman J.L., 1986. Positive Accounting Theory. Prentice-Hall, Englewood

Cliffs, New Jersey.


An Application of Casetti’s Expansion Method to a Variable Coefficient Regression Model of Electricity Demand: Simulation

Results of Alternative Estimation Methods

Roger L. Burford Emeritus Professor

Louisiana State University 1282 East Stanwick Place, Baton Rouge, LA 70810

E-mail: [email protected] Tel: (225) 766-4882

Susan M. L. Zee

Department of Business Administration and Economics Southern University at New Orleans, 6801 Press Drive, New Orleans, LA 70126

E-mail: [email protected] Tel: (504) 813-9929

Abstract Almost 40 years ago Emilio Casetti (1969, 1972, 1986, 1991, 1995, 1997, and 1999) introduced what he called “the expansion method” as an approach to the study of geographic diffusion of economic growth from a growth pole into the hinterlands. He used the method as a means of combining spatial and temporal analysis of economic growth. The present paper presents the results of a Monte Carlo study on the statistical problems of estimating systematically varying coefficients for price and income in a simplified model of demand for electricity use by rural residential electricity consumers. As the model is defined, it utilizes Casetti’s expansion method to define regression coefficients for price and income which are themselves functions of electricity price, income, and lagged demand. While our model is framed in the setting of residential electricity demand, our emphasis is on selection of an appropriate parameter estimation method, given the statistical problems that are raised by this formulation of the model. This study uses a Monte Carlo simulation approach to examine the statistical behavior of coefficients estimated by several different types of estimators in an effort to identify estimation procedures which can yield coefficient estimators having the most statistically desirable distributional characteristics, especially for relatively small samples. What is most desirable is specified in terms of relative bias, mean square errors, and normality of the coefficient distributions. Key words: Variable coefficient regression model, Monte Carlo simulation, Hildreth-Houck estimator, Robust regression, Factor regression, Ridge regression.


I. Introduction This paper develops a realistic, variable coefficient regression model for electricity demand. Through a Monte Carlo simulation, the study determines the most desirable estimation procedures. It should be noted that it is not the purpose of the present exercise to develop and test an ideal model of electricity demand. Residential electricity demand has been studied extensively over many years by many different people [see, for example, Taylor (1975), Houthakker (1980), Bohi (1981), Hartman (1983), Howry and Varian (1984), Engle, Mustafa, and Rice (1992), McKean and Winger (1992), and McQueen, Hyland, and Waston (2004, 2005)]. Rather, our purpose is simply to use electricity demand as a vehicle to test the usefulness of several alternative estimation methodologies to reliably estimate the systematically varying coefficients in our original model by reliably estimating the fixed coefficients of the expanded model.

It is generally agreed that the most important determinants of average residential electricity demand per customer include some "appropriately defined" price variable, income, weather variables, and probably availability and prices of competing energy sources. Other appropriate variables might include numbers of people per household, sizes and characteristics of housing, and numbers, types, and ages of cooling and heating equipment and appliance stocks.

For the present study, the price variable used is the average price paid per KWH. The issue of whether average price or marginal price should be used has been discussed extensively in the literature, and it has generally been concluded that the marginal price is the appropriate one to use, since consumers are charged according to a rate schedule and the price paid for an additional KWH is determined at the margin by the rate schedule. There are two reasons for using average price rather than marginal price in the present study. The first is that the consumers are not generally aware of the rate schedule and do not know what price they are paying at the margin. They generally do know (or can easily determine) the average price. Furthermore, McKean and Winger have shown that, if all variables are stated in logarithmic form (as is done in most studies of electricity demand, and is done in the present study), the estimated price elasticity is the same whether average or marginal price is used; only the intercept of the demand equation changes. The income variable used is average per capita household income in the area served by the electric company. The data used are annual.

All of the studies mentioned to this point assume that the coefficients in the model (the elasticities) of price and income, as well as any other variables included, remain constant over the period of study. This is clearly an untenable assumption. Moreover, it makes no real sense to assume that the coefficients, if variable, are random, or follow a random walk. Intuitively, it makes much more sense to assume that the coefficients vary through time (and across space) in some systematic way, as functions of prices, incomes, and possibly other economic, attitudinal, or technological variables relevant to the time and place. This paper presents a specific hypothetical functional form of the relationships between price and income elasticities and other variables, hence, its relation to Casetti’s expansion method. II. The Model Let yij represent the natural logarithm of the average annual kilowatt hours of electricity consumed by residential households in year i and in region j. Define x1ij to be the natural logarithm of the average price paid by the consumer per KWH and x2ij to be the natural logarithm of the average per capita annual income of people in region j in year i. For simplicity, we assume that these are the only relevant explanatory variables.

yij = β0 + β1ijx1ij + β2ijx2ij, (1) where

β1ij = b10 + b11x1ij + b12yi-1,j + e1ij (2) and


β2ij = b20 + b21x2ij + b22yi-1,j + e2ij. Substituting Equations (2) into Equation (1), the expanded model becomes

yij = β0 + b10x1ij + b11x1ij2 + b12x1ijyi-1,j + b20x2ij + b21x2ij

2 + b22x2ijyi-1,j + uij. (3) where

uij = e1ijx1ij + e2ijx2ij. (4)

This, clearly, is an example of the classic expansion method as defined by Casetti. β1ij represents the price elasticity of demand for electricity in region j in year i and β2ij represents the income elasticity of electricity demand in region j in year i. Both price and income elasticities are thus defined to be systematically variable across time and across region. Price elasticity is defined to be a function of current price in region j and KWH in region j lagged one year. Income elasticity is defined similarly to be a function of current income in region j and KWH lagged one year.

The justification for Equations (2) is essentially as follows: It seems clear on the surface that price elasticity itself is affected by the level of price. Indeed, by definition, if the demand curve is downward sloping (or upward sloping for that matter) elasticity changes with price. The same is true of the relationship between quantity demanded and income. The justification for stating the elasticities as functions of the lagged dependent variable is essentially that the existing level of electricity consumption is determined, not only by the variables included explicitly in the model, but also by a host of other economic, attitudinal, and technological factors which change through time and vary through space. The lagged dependent variable is, to a degree, a proxy variable for all the factors leading to its current value. In addition, use of the lagged dependent variable is intended to reflect habitual behavior (inertia); a reluctance to change usage habits quickly as prices and incomes change.

As the model is defined by Equations (1) – (3), it is expected that E(u) = 0, but Var(u) =/ σ2I. That is to say, the variances of the residuals are heteroscedastic since u is defined as a

function of x1 and x2. In addition, as the model is defined, there is a high degree of multicollinearity among the "independent" variables of the model, and the fact that two of the independent variables are stated as functions of the lagged dependent variable introduces a stochastic element in the independent variables as well. Thus, several of the basic assumptions of OLS are violated, necessitating the search for another estimation method. III. Related Statistical Models There is a substantial body of literature in econometric and statistical journals dealing with variable coefficient regression. This model is unlike most of the other variable coefficient regression models which have been discussed extensively in the literature. Most of the models discussed in the literature can be divided into random coefficient models with stationary parameters, random coefficient models with non-stationary parameters, and random coefficient models with some systematic components.

Considering models with parameters generated by stationary stochastic processes, Dielman (1983) presented a comprehensive survey to that time of statistical methods for studying the individual entities in a pooled cross-sectional and time series database. The methods included were seemingly unrelated regressions, dummy variable models, error component models, and random coefficient regression (RCR) models. There are two widely known methods for dealing with RCR models. The first deals only with shifts in the intercept term associated with individuals and/or time periods. The second method introduces variations in both slope and intercept coefficients.

Early models considering stationary random variations of coefficients include the models of Hildreth and Houck (1968) and Swamy (1970, 1973). The Hildreth and Houck model considered problems where coefficients represent a stationary stochastic process varying through time, while the Swamy model assumes the coefficients to be a stationary process across individuals in a cross-sectional study. Several of these papers have dealt with asymptotic properties of the two estimators [e.g.


Crockett (1985) and Carter and Yang (1981, 1986)] while others have used Monte Carlo techniques to study small sample properties [e.g., Froehlich (1973), Raj (1975), Dent and Hildreth (1977), Mehta, Lee, and Swamy (1986), Swamy and Thurman (1994), Swamy, Tavlas, and Chang (2005)].

One of the more important contributions to the literature related to nonstationary RCR models is that of Cooley and Prescott (1973a, b, c, 1976). Applications of the Cooley-Prescott model include those of Roll (1972), Cooley (1975), Lamotte and McWhorter (1976, 1977), Machak, Spivey, and Wrobleski (1985) and Cooley and Prescott (1995).

The third class of models includes systematically time varying, but non-stochastic, coefficients. The earliest work in this group is switching regressions by Quandt (1958 and 1960). Other models of this type include those of Farley and Hinich (1970), Quandt (1972) and Goldfeld and Quandt (1973).

There are two models in the literature which are particularly relevant to the present research. One is by Belsley (1973a, b) and the other is by Singh, Nagar, Choudhry, and Raj (1976).

The Belsley model has the following form: K yt = Σ βktxkt + εt, t = 1, 2, ..., T k=1 and R βkt = Σ ΓkrZrt + ukt, t = 1, 2, ..., T r=1 k = 1, 2, ..., K R ≤ K where yt is the t'th observation on the dependent variable, and the vector Zt is the t'th observation on a subset of the independent variables xkt. Conceptually, Belsley’s model is essentially the same as Casetti’s expansion method. The primary difference between the approaches of Belsley and Casetti is that, while the “expansion method” was viewed by Casetti as a modeling approach to realistically explain changes through time and space, without considering the special statistical problems that might result from its use, Belsley was more concerned with the statistical properties of the estimators of the coefficients for this model. Belsley then assumed that E(εε') = σ2I and that ukt is independent over time with σ2 = Ω. He then simplified the problem by assuming that σ2 = 0 and then developed a recursive procedure similar to the Kalman Filter to estimate βkt. Further study on the Belsley Model can be found in Belsley (1984) and Foschi, Belsley, and Kontoghiorghes (2003).

The systematic random coefficient regression model proposed by Singh, et al. (1976) is of the following form: K yt = Σ βktxkt, t = 1, 2, ..., T, k=1 and βkt = bk + akfkt + ekt, where yt is an observation on the dependent variable at time t, xkt is a nonstochastic observation on the k'th independent variable at time t, βkt is a vector of systematically time-varying coefficients to be estimated, bk and ak are parameters to be estimated, fkt is a function of time, and ekt is the random error term. Singh, et al. simplified the model by assuming that fkt = t and developed a four-step modified Hildreth-Houck generalized least squares estimator for βkt. Again, Singh’s model can be considered to be an example of the expansion method, where the only expansion variable used is time. IV. Estimation of the Model For purposes of the statistical tests, the “true” form of Equation (3), to be estimated was assumed to be


yij = 7.1449 -1.5215x1ij -.024744x1ij2 +.16404x1ijyi-1,j +.13935x2ij +.016231x2ij

2 -.00068635x2ijyi-1,j + .0372zuij, (5)

where zuij is a randomly generated standard normal variable, .0372 is the standard deviation of uij , and where uij is as defined in Equation (4). Note, the coefficients used in Equation (5) were determined by fitting a least-squares regression line to a set of observed data including average annual KWH, average price per KWH for a set of 12 rural electric cooperatives over a period of 25 years each. These coefficients were assumed, for purposes of the simulation, to be the true values of the coefficients. For purposes of the simulation, “estimated” values of yij were computed by plugging values of the “independent” variables into the equation, along with randomly generated standard normal values of zuij. The simulated values of the βkij were then estimated using the several different estimation methods tested.

As noted previously, several of the required OLS assumptions are known to be violated in the model specified in Equations (1) to (3). The two chief violations are (1) that the residuals (uij) are defined to be heteroscedastic (functions of price and income) and (2) an extremely high degree of multicollinearity exists in the model. For example, several of the "independent" variables in the model have R-squares with other independent variables as high as .9997 and variance inflation factors are as high as 3,751 when applied to an actual data set involving 12 separate regional electric companies over a period of some 25 years. In addition, two of the independent variables are functions of lagged values of the dependent variable, introducing a possible problem of stochastic regressors and probably also autocorrelation of residuals. Thus some alternative to OLS is necessary.

Several different estimation techniques have been explored in this study, including the estimated generalized least squares methods of Hildreth and Houck, the modification of the Hildreth-Houck estimator proposed by Singh, et al., an estimated generalized least squares estimator proposed by Goldfield and Quandt, a modification of the Goldfield-Quandt estimator proposed by Amemiya, several robust weighted generalized least squares estimators, factor regression as proposed by Scott (1966, 1969), and a ridge regression estimator developed by Lawless and Wang [Lawless and Wang (1976), and Lawless (1978)]. This particular ridge estimator was chosen over others based on the results presented by Edlund (1990), and its ease of implementation.

The primary objective of the estimated generalized least squares estimators is to overcome the heteroscedasticity problem. The robust weighted regression estimator is also designed to overcome the effects of heteroscedasticity as well as possible non-normality of residuals (particularly, heavy tails), and the effects of outliers. However, the results of our experiments imply that heteroscedasticity problems, while they do exist, are much less serious in their effects on coefficient estimators than is multicollinearity. The ridge regression estimator used is designed as a means of overcoming the multicollinearity problem, as is the factor regression estimator.

In an effort to simultaneously overcome heteroscedasticity and multicollinearity problems, the data matrix of independent variables was transformed into an orthogonal matrix by using a singular value decomposition algorithm (the Q-R algorithm). Each of the estimation methods which had been previously tried were then applied to the transformed matrix and the resulting estimators were retransformed to their original scales.

To test the relative bias and efficiency of each of these methods, a Monte Carlo study was designed in which 1000 samples each of size 25, 50, and 75 observations were randomly generated according to Equation (5), and with error terms as defined in Equation (4). Then each of the estimation techniques was applied to estimate the coefficients of the model. Relative biases, mean square errors, and normality tests for the estimated coefficient distributions were computed for each of the experimental runs.

From the simulation study, it was concluded that the Lawless-Wang ridge regression estimator offers by far the best promise for the present model. [See Tables 1-6.] This conclusion is based on a comparison of the relative biases, the mean square errors of the distributions of coefficient estimates, and the normality tests for the distributions of coefficients. In terms of biases and mean square errors, the factor regression was the next best performer, but the coefficient distributions do not tend toward


normality for this estimator. Performance of the Hildreth-Houck estimator using orthogonal data matrices is also quite acceptable. It was expected that use of a weighted robust regression estimator with orthogonally transformed data would perform well. While its performance is better than OLS with orthogonal data, however, it does not compare well with either the Hildreth-Houck estimator with orthogonal data or the Lawless-Wang ridge estimator.

The simulation experiments reported on here were conducted using the GAUSS language developed by Aptech Systems, Inc. The particular advantage of using GAUSS for the simulations involving complex matrix manipulations is that (1) none of the existing standard statistical packages contain the routines needed to compute some of the estimators evaluated, or are not flexible enough to permit one to write his own code, and (2) GAUSS is designed for the user to write code for complex matrix operations as compact and simple matrix expressions. V. Results of the Present Study The data in tables 1 through 6 present the basic results of the study in terms of the relative performance of the estimators in the simulation studies. It is clear from these tables that all of the estimators are biased for small samples. However, as sample sizes increase, the biases for most of the estimators tend to disappear. While, for the factor regression model, the bias tends to be relatively small for small samples, the bias does not decrease systematically for larger sample sizes. The Lawless-Wang ridge estimator has the smallest bias, both for small samples and for larger samples. The OLS estimates tend to be more biased than the ridge estimator for all sample sizes. Similarly, all of the other estimators have substantially larger biases than the ridge estimator. While the Hildreth and Houck estimator is more biased, this bias is reduced substantially by fitting the Hildreth-Houck model to orthogonal data.

Similarly to bias, the mean square errors of all other estimators are several fold larger than those for the ridge estimator. Using the orthogonal transform did reduce the mean square errors substantially for the Hildreth-Houck estimator and the Singh, et al. estimator, but even with the orthogonality transformation their mean square errors are substantially larger than those of the ridge estimator. The mean square errors for the factor regression model are also quite small, but they do not decrease systematically for larger sample sizes.

Also, the Kolmogorov-Smirnov normality test statistic seems to decrease consistently for the ridge estimator, more so than for any other estimator. A K-S statistic larger than .030 indicates significant non-normality at the .05 level. Table 1: Summary Statistics from Simulation Runs Lawless-Wang Ridge Regression Estimator

n=25 n=50 n=75

Coeff

Bias MSE Norm Bias MSE Norm Bias MSE Norm

b10 .224 1.217 .023 .080 .757 .020 .0001 .072 .020 b20 -.162 .554 .023 -.066 .384 .013 -.004 .046 .013 b11 -.167 .740 .018 -.094 .319 .023 -.001 .118 .018 b21 -.227 1.082 .020 -.083 .726 .015 -.013 .090 .022 b12 -.126 .263 .024 -.083 .127 .016 -.011 .044 .022 b22 .021 .021 .024 .017 .011 .015 -.001 .005 .019

Table 2: Summary Statistics from Simulation Runs OLS and Orthogonal OLS


n=25 n=50 n=75 Coeff Bias MSE Norm Bias MSE Norm Bias MSE Norm

b10 OLS Orth

.921 .901

81.722 78.365

.018 .019

.214 .337

17.72 19.30

.020 .022

-.179 -.296

6.187 6.335

.028 .022

b20 OLS Orth

2.408 2.408

575.75 546.99

.018 .018

-.700 1.251

217.95 197.34

.025 .018

-.620 -.115

30.171 29.700

.026 .013

b11 OLS Orth

.128 .126

1.095 1.043

.032 .034

-.001 .063

.194 .222

.022 .019

-.001 .017

.068 .071

.036 .018

b21 OLS Orth

-.146 -.145

2.333 2.217

.018 .017

.064 -.077

.866 .791

.025 .019

.052 .024

.121 .119

.037 .017

b12 OLS Orth

-.215 -.211

.640 .637

.014 .013

.034 -.091

.164 .197

.013 .021

.014 .012

.075 .080

.019 .022

b22 OLS Orth

.040 .032

.114 .114

.014 .012

-.017 .019

.032 .004

.013 .022

.018 -.020

.015 .016

.017 .021

Table 3: Summary Statistics from Simulation Runs Factor Regression Estimator

n=25 n=50 n=75

Coeff Bias MSE Norm Bias MSE Norm Bias MSE Norm b10 1.098 1.352 .203 .966 .991 .232 1.007 1.039 .151 b20 .191 .075 .090 .218 .065 .090 .236 .063 .069 b11 -.030 .003 .204 -.047 .003 .232 -.042 .022 .115 b21 .002 .000 .095 .003 .000 .073 .004 .000 .067 b12 -.158 .025 .174 -.056 .002 .092 -.161 .026 .105 b22 .046 .003 .112 .005 .003 .054 .051 .003 .056

Table 4: Summary Statistics from Simulation Runs Hildreth-Houck and Orthogonal HH

n=25 n=50 n=75 Coeff Bias MSE Norm Bias MSE Norm Bias MSE Norm

b10 HH Orth

-1.92 1.16

388.81 2.342

.044 .017

-1.17 1.14

93.98 1.656

.033 .019

-1.64 1.11

33.689 1.371

.035 .018

b20 HH Orth

1.258 .767

3290.7 1.272

.014 .013

-1.44 .730

1300.8 1.680

.014 .030

-.911 .759

149.41

.692

.023 .013

b11 HH Orth

.033 .129

4.097 1.061

.028 .027

-.120 .047

.906 .270

.021 .016

-.071 .039

.348 .039

.025 .017

b21 HH Orth

.044 .013

13.806 .277

.013 .023

.156 .101

5.180 .119

.020 .021

.118 .134

.591 .068

.033 .029

b12 HH Orth

.129 .190

3.879 2.982

.027 .023

.190 -.134

.931 1.187

.026 .034

.209 -.192

.406 .118

.021 .032

b22 HH Orth

-.109 .016

.710 .133

.027 .025

.065 .005

.184 .052

.024 .020

-.066 -.009

.075 .023

.027 .026

Table 5: Summary Statistics from Simulation Runs Robust Regression Estimator from Orthogonal Data


n=25 n=50 n=75

Coeff Bias MSE Norm Bias MSE Norm Bias MSE Norm b10 1.047 85.49 .016 .226 22.53 .019 -.296 7.270 .022 b20 2.375 612.63 .020 1.051 225.00 .017 -.266 31.96 .022 b11 .143 1.102 .029 .055 .248 .022 .014 .081 .020 b21 -.151 1.504 .023 -.061 .906 .018 .034 .130 .021 b12 -.242 .729 .018 -.074 .221 .022 .014 .092 .012 b22 .005 .131 .017 .011 .042 .023 -.021 .018 .016

Table 6: Summary Statistics from Simulation Runs Singh, et al and Orthogonal SNCR

n=25 n=50 n=75 Coeff Bias MSE Norm Bias MSE Norm Bias MSE Norm b10 SNCR Orth

-.715 1.20

187.413.

370

.032 .021

-1.57 1.13

112.79 2.439

.035 .022

-.95 1.04

51.571 1.602

.038 .027

b20 SNCR Orth

-2.70 .945

1538.6 6.807

.017 .028

-3.04 .915

1457.8 5.273

.043 .044

-.542 .739

176.85 1.347

0.31 0.32

b11 SNCR Orth

.082 .075

2.115 1.431

.018 .020

-.507 .044

.888 .757

.043 .032

.006 .055

.488 .349

.037 .023

b21 SNCR Orth

-.132 -.151

6.216 6.713

.022 .026

.257 -.131

5.926 5.067

.042 .042

.069 .015

.707 .830

.040 .040

b12 SNCR Orth

-.158 -.138

2.071 .746

.035 .033

.306 .059

1.096 .463

.043 .052

.091 -.014

.731 .307

.036 .063

b22 SNCR Orth

-.003 -.006

.385 .316

.038 .033

-.108 -.005

.226 .224

.040 .045

-.054 -.014

.146 .138

.038 .055

Note: Bias in Tables 1 through 6 is relative bias. It is defined as the mean of the difference between the 1000 estimated values of the respective coefficients generated in the simulation and the actual pre-specified values, divided by the actual. MSE is the mean of the squared values of these same differences. The Normality test used is the Kolmogorov-Smirnov test of goodness of fit of the 1000 estimated coefficients to a normal distribution. VI. Summary A model such as the one applied here, like those of Casetti, Belsley, and Singh, assume that in many real modeling problems it is appropriate to model coefficients as systematically varying parameters, either over time or space, or both, rather than as constant parameters. However, use of Casetti’s expansion method, or the equivalent formulations of Belsley or Singh, will usually result in serious violations of the basic assumptions which are necessary for OLS estimates to be the best linear unbiased estimates available.

The results of the simulation experiments reported here suggest that the Lawless-Wang ridge regression estimator may be the most reliable estimator to use in models such as this.

It was anticipated that application of a robust weighted regression estimator or an estimated generalized least squares estimator to account for heteroscedasticity along with an orthogonalizing transform (such as the Q-R algorithm), to remove multicollinearity, might prove to be the most reliable estimator. While these do perform much better with orthogonal data than with the original data, and much better than OLS, the Lawless-Wang ridge estimator is still the most reliable (smallest biases, smallest mean square errors, and most consistent approach to normality) in the simulation studies.


VII. References 1] Belsley, D. A. (1973a). On the determination of systematic parameter variation in the linear

regression model. Annals of Economic and Social Measurement. 2: 487-494. 2] Belsley, D. A. (1973b). A test for the systematic variation in regression coefficients. Annals of

Economic and Social Measurement. 2: 495-499. 3] Belsley, D. A. (1984). Demeaning conditionaing diagnostics through centering. The American

Statistician. 38: 73-77. 4] Bohi, D. R. (1981). Analyzing demand behavior: a study of energy elasticities. The Johns

Hopkins University Press. Baltimore. 55-91. 5] Burford, R. L., and Zee, S. (1990). Variability of price and income elasticities for electricity

over time and space. Presented to the Western Regional Science Association, Molokai, Hawaii. 6] Burford, R. L., and Zee, S. (1994). Explorations in estimation methodologies for a simplified

model of residential demand for electricity with systematically varying coefficients. Proceeding of the annual meeting of the Decision Sciences Institute in Honolulu, Hawaii.

7] Carter, R. L., and Yang, C. K. (1981). Large sample inference in random coefficient regression models. American Statistical Association: Proceedings of the Social Statistics Section. 305-308.

8] Carter, R. L., and Yang, C. K. (1986). Large sample inference in random coefficient regression models. Communications in Statistics. 15A: 2507-2525.

9] Casetti, E., and Semple, R. K. (1969). Concerning the testing of spatial diffusion hypotheses. Geographical Analysis. 1: 254-259.

10] Casetti, E. (1972). Generating models by the expansion method: applications to geographic research. Geographic Analysis. 4: 81-91.

11] Casetti, E. (1986). The dual expansions method: an application to evaluating the effects of population growth on development. IEEE Transactions on Systems, Man, and Cybernetics. SMC-16: 29-39.

12] Casetti, E. (1991). The investigation of parameter drift by expanded regressions: generalities and a ‘family planning’ example. Environment and Planning A. 23: 1045-1061.

13] Casetti, E. (1995). Spatial mathematical modeling and regional science. Papers in Regional Science. 4: 3-11.

14] Casetti, E. (1997). The expansion method, mathematical modeling, and spatial econometrics. International Regional Science Review. 20: 9-33.

15] Casetti, E., and Can, A. (1999). The econometric estimation and testing of DARP models. Journal of Geographical Systems. 1: 91-106.

16] Cooley, T. F. (1975). A comparison of robust and varying parameter estimates of macro-economic models. Annals of Economic and Social Measurement, 4.

17] Cooley, T. F., and Prescott, E. C. (1973a). Tests of an adaptive regression model. Review of Economics and Statistics. 55: 248-250.

18] Cooley, T. F., and Prescott, E. C. (1973b). An adaptive regression model. International Economic Review. 14: 364-371.

19] Cooley, T. F., and Prescott, E. C. (1973c). Systematic (non-random) variation models, varying parameter regression: a theory and some applications. Annals of Economic and Social Measurement. 2: 463-473.

20] Cooley, T. F., and Prescott, E. C. (1976). Estimation in the presence of stochastic parameter variation. Econometrica. 44: 167-184.

21] Cooley, T. F., and Prescott, E. C. (1995). Economic growth and business cycles. Frontiers of Business Cycle Research. Princeton University Press. 1-38.

22] Crockett, P. W. (1985). Asymptotic distribution of the Hildreth-Houck estimator. Journal of the American Statistical Association. 80: 202-204.


23] Dent, W. T., and Hildreth, C. (1977). Maximum likelihood estimation in random coefficient models. Journal of the American Statistical Association. 72: 69-72.

24] Dielman, T. E. (1983). Pooled cross section and time series data: a survey of current statistical methodology. The American Statistician. 37: 111-122.

25] Edlund, P. (1990). Ridge estimation of transfer function weights. Communications in Statistics. 19 (2): 451-468.

26] Engle, R. F., Mustafa, C., and Rice, J. (1992). Modeling peak electricity demand. Journal of Forecasting. 11: 241-251.

27] Farley, J. U., and Hinich, M. J. (1970). A test for a shifting slope coefficient in a linear model. Journal of the American Statistical Association. 65: 1320-1329.

28] Foschi, P., Belsley, D. A., and Kontoghiorghes, E. (2003). A comparative study of algorithms for solving seemingly unrelated regression models. Computational Statistics & Data Analysis. 44: 3-35.

29] Froehlich, B. R. (1973). Some estimators for a random coefficient regression model. Journal of American Statistical Association. 68: 329-335.

30] Goldfeld, S. M., and Quandt, R. E. (1973). The estimation of structural shifts by switching regressions. Annals of Economic and Social Measurement. 2/4: 475-485.

31] Hartman, R. S. (1983). The estimation of short-run household electricity demand using pooled aggregate data. Journal of Business and Economic Statistics. 1: 127-135.

32] Hildreth, C. and Houck, J. P. (1968). Some estimators for a linear model with random coefficients. Journal of the American Statistical Association. 63: 584-595.

33] Houthakker, K. S. (1980). Residential electricity revisited. The Energy Journal. 1: 29-41. 34] Howrey, E. P., and Varian, H. R. (1984). Estimating the distributional impact of time-of-day

pricing of electricity. Journal of Econometrics. 26: 65-82. 35] LaMotte, L. R., and McWhorter, A., Jr. (1976). A test for the presence of random coefficients

in a linear regression model. American Statistical Association: Proceedings of the Business and Economics Section. 400-405.

36] LaMotte, L. R., and McWhorter, A., Jr. (1977). Estimation, testing, and forecasting with random coefficient regression models. American Statistical Association: Proceedings of the Business and Economics Section. 814-817.

37] Lawless, J. F. (1978). Ridge related estimation procedures: theory and practice. Communications in Statistics. A7 (2): 139-164.

38] Lawless, J. F., and Wang, P. (1976). A simulation study of ridge and other regression estimators. Communications in Statistics. A5 (4): 307-323.

39] McKean, J. R., and Winger, W. D. (1992). Simultaneous equation estimates of electricity demand for the rural south: revenue projections when prices are administered. Journal of Forecasting. 11: 225-240.

40] McQueen, D., Hyland, P., and Watson, S. (2004). Monte Carlo simulation of residential electricity demand for forecasting maximum demand on distribution networks. IEEE Transactions on Power Systems. 19: 1685-1689.

41] McQueen, D., Hyland, P., and Watson, S. (2005). Application of a Monte Carlo simulation method for predicting voltage regulation on low-voltage networks. IEEE Transactions on Power Systems. 20: 279-285.

42] Machak, J. A., Spivey, W. A., and Wrobleski, W. J. (1985). A framework for time varying parameter regression modeling. Journal of Business and Economic Statistics. 3: 104-111.

43] Mehta, J. S., Lee, R. Y., and Swamy, P. A. V. B. (1986). An experimental study of estimators associated with a random coefficient regression model. American Statistical Association: Proceedings of the Business and Economics Section. 561-566.

44] Quandt, R. (1958). The estimation of the parameters of a linear regression system obeying two separate regimes. Journal of the American Statistical Association. 873-880.


45] Quandt, R. (1960). Tests for the hypothesis that a linear regression system obeys two separate regimes. Journal of the American Statistical Association. 324-330.

46] Quandt, R. (1972). A new approach to estimating switching regressions. Journal of the American Statistical Association. 67: 306-310.

47] Raj, B. (1975). Linear regression with random coefficients: the finite sample and convergence properties. Journal of the American Statistical Association. 70: 127-137.

48] Roll, R. (1972). Interest rates on monetary assets and commodities price index changes. Journal of Finance. 27: 251-277.

49] Scott, J. T., Jr. (1966) Factor analysis regression. Econometrica. 34: 552-562. 50] Scott, J. T., Jr. (1969). Factor analysis regression revisited. Econometrica. 37: 719. 51] Singh, B., Nagar, A. L., Choudhry, N. K., and Raj, B. (1976). On the estimation of structural

change: a generalization of the random coefficient regression model. International Economic Review. 17: 340-361.

52] Swamy, P. A. V. B. (1970). Efficient inference in a random coefficient regression model. Econometrica. 38: 311-323.

53] Swamy, P. A. V. B. (1973). Criteria, constraints, and multicollinearity in random coefficient regression models. Annals of Economic and Social Measurement. 2: 429-450.

54] Swamy, P. A. V. B., and Thurman, S. (1994). Exchange rate episodes and the pass-through of exchange rates to import prices. Journal of Policy Modeling. 16: 609-623.

55] Swamy, P. A. V. B., Tavlas, G. S., and Chang, I. (2005). How stable are monetary policy rules: estimating the time-varying coefficients in monetary policy reaction function for the US. Computational Statistics & Data Analysis. 49: 575-590.

56] Taylor, L. D. (1975). The demand for electricity: a survey. Bell Journal of Economics. 6: 74-110.


Comparative Economics of Rice Production in Pakistan: A Price Risk Analysis

Mohammad F. Hussain, Sofia Anwar and Zakir Hussain1

Abstract This intent of this paper was to estimate the comparative economics of Rice crop in Pakistan to analyze different price risk scenarios in the WTO regime. The extent of policy distortion and agricultural protection was also explored in the study. The data pertained to Basmati rice from Punjab and for Irri rice from two provinces Punjab and Sindh over the three years period (2001-2003). The Policy Analysis Matrix (PAM) was used for the analysis. The risk prices of Basmati and Irri rice, fertilizers (DAP and Potash) were further utilized to estimate the comparative advantage and competitiveness of rice crop in free trade regime. The static analysis results showed that Pakistan has comparative advantage in the production of Basmati rice at export parity price. The price Risk analysis also supported these results. In the free trade scenario, Pakistan is likely to maintain competitiveness in Basmati rice market. On the other hand Pakistan has no comparative advantage in production of Irri rice at the risk price, in future free trade regime. However, Sindh is likely to have comparative advantage in Irri rice production at risk price.

I. Introduction The world economic scenario is set for change under free trade regime, increasing competition and relative competitiveness of different countries. The structural changes and policy reforms in Pakistan has profound implications for national economy. Pakistan agricultural economy is subjected free trade regimes. Therefore, the information of comparative economics and competitiveness of agricultural commodities is useful tool to to maintain viable terms of trade in agriculture. Pakistan rice is one of the major cash and export crop and until recently enjoyed comparative advantage in “Basmati rice”. The rice area in Pakistan is more than 2 million ha, occupying 18 per cent of the acreage under cereals and 10 per cent of the total cropped area. The rice area showed increasing trend with a growth rate of 0.48 percent in year 2002 over the year 1991. However, production trend witnessed leapfrog jump from 4.9 Mt (metric tons) in the year 1991to 6.3 Mt in the year 2002 showing a 30.38 percent increase over the period. The rice contributes 17 per cent to overall production of food grains with its contribution in the value added by major crops is around 16 per cent production (GOP, 2003).

Of the total rice area, more than 60 per cent is grown in Punjab, showing 0.82 percent growth rate from 1991 to 2002. The production in Punjab followed the national trend with a growth rate of 2.87 percent for the same period (IRRI, 2004). Nearly 50 per cent of the total rice area is under fine varieties while in Punjab 78 percent of total rice hectare is under fine varieties (APCom, 2002 Khan, 2001). Pakistan rice yield is below 3 ton per ha with the exception of two years i.e.1999 and 2000. The world average rice yield is nearly 3.9 ton/ha. Among major rice exporting countries, leading rice exporters has double yield than Pakistan

1 The authors are respectively graduate student, Texas A&M, University, Agricultural Economist, Government of Punjab and Professor and Chairman, Department of Economics , University of Sargodha.


Rice being cash crop, is one of the major contributor to country’s GDP. It accounts for 5.4 percent of value added in agriculture and 1.3 percent in GDP (GOP 2003-04). The rice export from Pakistan increased from 1.204 million ton (Mt) to 3.016 Mt from 1992 to 2002 and recorded an annual growth rate of 13.4 percent over the period. Rice exports showed large improvement during 2002 (23.9 per cent growth over 2001) and surpassed the annual export target (US $460 million) by 20.8 per cent (SBP, 2004).

he present rice pricing policy in Pakistan is market oriented and is exposed to vagaries of supply and demand forces. Pakistan rice especially basmati rice is facing stiff competition from India, Thailand and USA and lost the traditional glory of aromatic “Basmati” rice.. The changing economic situation warrants not only the assessment of current status but also determine the future potential of country competitiveness and comparative advantage in rice production.

In the review of previous studies, Appleyard (1987) concluded that in case of rice Punjab has smaller DRC value than Sindh and both have comparative advantage in rice production. Ali (1992) showed that producers of basmati rice were heavily taxed. The extent of negative transfers to producers was very small when world prices were translated into export parity rather than import prices. Longmire and Debord (1993) indicated that strong comparative advantage prevailed in the production of basmati rice. Akhtar (2004) showed that basmati rice possessed comparative advantage while farmers of IRRI rice were in a situation of comparative disadvantage and the price structure discriminated against basmati rice. II. Material and Methods The study was based on secondary data and the time series data of the cost production (COP) was obtained from Agriculture Price Commission (APCom); the policy organ of the Ministry of Food, Agriculture and Livestock, Pakistan. The COP data were collected for basmati rice from Punjab and for Irri rice from two provinces Punjab and Sindh, for three harvesting years i.e. 2000-01 to 2002-03.

The Policy Analysis Matrix (PAM) was selected as analytical framework. The basic format of the PAM is a matrix of two way accounting identities; one set defining profitability and the other defining the difference between private and social values of a commodity system (Monk and Pearson 1989). The measurement of comparative advantage and policy distortions in rice crop was made through approaches i.e. Domestic Resource cost (DRC), Nominal Protection Co-efficient (NPC) and Effective Protection Co-efficient (EPC). In the static analysis, PAM was developed based on crop budgets at both financial and economic prices. The economic prices were estimated based on three years average world prices of rice and fertilizers. All other tradable inputs were weighted by premium (1.138) i.e. ratio of shadow exchange rate to official exchange rate.

The Risk is a situation in which actuarial odds could be calculated from observing the frequencies of outcomes (Newberry and Stiglitz, 1981). In this study risk prices of rice and fertilizer were utilized in estimation of economic prices .The effect of changing the output price is equivalent to changing yields and thereby quantity by the same proportion as revenue is the product of price and quantity (Kannapiran and Flemming, 1999).

The application of risk analysis allowed some ex-ante assessment of potential policy change. In the study risk analysis was carried out, utilizing the two software’s i.e. the “Best Fit” and “@ Risk 4.5.3.” III. Analytical Results In the static analysis, the Policy Analysis Matrix (PAM), in Table 1 showed higher profit at economic prices as compared to financial prices indicating the competitiveness of Basmati rice in world market. The NPI ratio was 0.93 implying that tradable inputs were not taxed in the Punjab. The farmers were paying the prices closely to the world market. The value of NPC 0.78 indicated that basmati paddy was under priced rather than receiving border prices. The value of DRC 0.60 showed Punjab has


comparative advantage in production of basmati paddy. The value added to basmati per acre-inch of water estimated was Rs. 150.68. Rice is highly water consumptive crop and cost of Irrigation especially tube well is becoming prohibitive, thus lowering the comparative advantage of this commodity. Table 1: Policy Analysis Matrix for Basmati Paddy in Punjab/Pakistan

Revenue Production costs Profit

Tradable Non-tradable NPI= 0.93

Private prices 10174.71 3969.35 5010.05 1195.30 NPC= 0.78

Social prices 13014.26 4274.65 5268.68 3470.93 EPC= 0.71

Divergence -2839.56 -305.30 -258.63 -2275.62 DRC= 0.60

Value added 8739.61

Value added /acre inch of water 150.68

Unit values in Pakistani Rupees

In the case of Irri rice, the results of Policy Analysis Matrix (PAM) portrayed that farmers were earning less revenue at social prices than at financial prices (Table 2). Thus Pakistan has lost out comparatively in IRRI production. The nominal protection coefficient (NPI) 0.90 indicated farmers paid less than open market for tradable inputs, pinpointing small protection provided by government. There was implicit subsidy only on Irrigation water. The NPC 1.41 represented higher price of output at domestic level as compared to world market. This suggested market forces were favoring Irri rice. Table 2: Policy Analysis Matrix for IRRI Paddy in Pakistan Revenue Production costs Profit

Tradable Non-tradable NPI= 0.90

Private prices 8870.95 3255.28 4497.89 1117.79 NPC= 1.41

Social prices 6304.21 3620.41 4571.86 -1888.05 EPC= 2.09

Divergence 2566.74 -365.13 -73.97 3005.84 DRC= 1.70

Value added 2683.81

Value added /acre inch of water 45.49

The EPC has a value of 2.09, and reinforced the conclusions drawn through NPI and NPC. The IRRI producer was earning higher return due to interventions provided by the GOP, as an incentive to encourage the farmers to produce more. The DRC was 1.7, an indicator of comparative disadvantage in IRRI production for the export purposes. The value added is 45.49 per acre-inch of water that calls for a judicious use of this scarce resource.

In order to develop risk policy analysis matrix, risk prices of inputs (DAP and Potash fertilizer) and Basmati rice were estimated. The program estimated the Risk Tnormal distribution as best-selected probability distribution, shown in Fig 1. The Risk function estimate gave the mean risk price as US $ 174.84 per ton with standard deviation of 24.38. This risk function estimate was further truncated to define the boundaries for resultant value during simulation. The mean price obtained through Tnormal


function was US $ 176.89. This risk price of DAP provided the Import parity price (IPP) Rs. 20499.99 per ton and Rs. 20.50 per Kg at the farm level.

In the same way “Normal distribution” of time series world price of potash was estimated. The mean value of Risk price estimated from this probability distribution was US $198.5 per ton (Fig 2). The Risk price added with freight charges, insurance price, financial and handling charges, marketing costs at the port and from port to market, transportation cost and taxes, gave import parity price (IPP) of Rs.22524 per ton and Rs.22.50 per 40 Kg. The final results were generated through hyper cube Monte Carlo simulations through @ Risk program (Table 3 and Table 4).

The PAM was modified to include economic tradable costs based on IPP of DAP and Potash separately. Economic cost of tradable inputs was further projected for the next four years. Figure 1: Normal Distribution of Time Series World Price of DAP Fertilizer

Normal(175.681, 24.487)

Val

ues

x 10

-2

dollar per ton

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

140

150

160

170

180

190

200

210

220

230

< >90.0%135.40 215.96

BestFit Trial VersionFor Evaluation Purposes Only

Figure 2: Normal Distribution of Time Series World Price of Potash Fertilizer

Normal(198.560, 10.961)

Value

s x 10

-2

dollar per ton

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

175

180

185

190

195

200

205

210

215

220

< >90.0%180.53 216.59


Mean = 198.560

Mean = 198.560

Inpu

tFit


Figure 3: Normal Distribution of Time Series FOB Price of Basmati Rice Normal(452.494, 35.323)

Value

s x 10

-2

dollar per ton

0.0

0.5

1.0

1.5

2.0

2.5

400

420

440

460

480

500

520

< >90.0%394.4 510.6


Mean = 452.49

Mean = 452.49

Inpu

tFit

The time series data for the years 1990 to 2002 on FOB price of Basmati rice were utilized in the “Best Fit” a companion product of @Risk program to determine the probability distribution of Basmati rice (Fig 3). The best-selected probability distribution of basmati rice was Risk Tnormal. The Risk Tnormal probability distribution has four arguments that were mean, standard deviation, minimum and maximum.

The Risk Tnormal probability distribution of FOB price of basmati rice was entered in the @ Risk in the form of formula. The @ Risk performed the analysis by employing hypercube Monte Carlo Simulation analysis technique. The risk was measured as the dispersion from the mean value of US $452.49 per ton with a standard deviation of 35.32. The value of standard deviation showed oscillating price trend of Basmati price in the world market (Fig 3). The product of risk price of US $452.49 per ton and shadow exchange rate provided export parity price (EPP) of paddy in market as Rs.527.15 per 40 Kg.

The final results of analysis are presented in the Policy Analysis Matrix (PAM) given in Table 3. The economic revenue at future price was estimated Rs.13214.35 per acre which was considered at the first year of free trade. In the next 4 years, projections were made on the basis of this economic revenue by adding the risk function in it as error term. The original economic revenue was taken as base for the revenue in subsequent year and risk function estimate was included here to add the error. The values of economic return were Rs. 12837, Rs.13293, Rs.13746.35 and Rs. 114198.5 and Rs.14651 in the year 2004 through 2008 respectively. The NPC for inputs in the first year was estimated as 0.87 showing that farmer was paying prices closer to world market. The economic tradable cost was calculated on the basis of shadow exchange rate (SER). All other tradable inputs were weighted by 1.138 (Ratio of SER to OER). The economic tradable cost based on fertilizer prices indicated higher market prices than world prices. This difference was caused by high freight charges of in-land transportation and government taxes. In the next 4 years NPI showed a declining trend because of fixed financial prices. The increase in world price of DAP, will cause an increase in domestic market prices and NPI value for next 4 year may show an increase and become closer to unity.

The NPC for output (NPC) was 0.79 in the first year and showed a decreasing trend up to 4th year. It explained variation between domestic and foreign price of output would increase, showing under pricing of domestic output price, due to implicit taxation and resource transfer from Basmati Rice.

The EPC was 0.75, and reaffirmed the results of NPI and NPC. During the study period, growers were paying nearly world prices for inputs but output (basmati) was under priced in domestic market. Basmati crop showed an over all implicit taxation, non-protection and possible resource transfer to


other sectors. The decreasing trend in EPC suggested implicit taxation will increase and showing disincentives to producers, provided domestic prices are not improved. In the ensuing WTO regimes, farmers are likely to achieve world prices, resulting in an increase of EPC. That will provide an incentive to growers to produce more Basmati Rice. The DRC, an indicator of comparative advantage was 0.62 and decreased up to 0.51 in the fifth year. The declining value of DRC indicated an increasing trend in the comparative advantage level of Basmati cultivation.

This suggested that in the ensuing free trade regime Pakistan would be able to maintain its competitiveness in rice market at the given level of all other factors. Table 3: Risk Price Policy Analysis Matrix with DAP for Basmati Paddy in Punjab


Tradable Non-tradable

Private prices 10174.71 3969.35 5010.05 1195.30 NPI NPC EPC DRC

Social prices 12841.36 4537.98 5154.28 3149.10 0.87 0.79 0.75 0.62

Social prices 2 13293.85 4712.82 5154.28 3426.75 0.84 0.77 0.72 0.60

Social prices 3 13746.35 4887.66 5154.28 3704.40 0.81 0.74 0.70 0.58

Social prices 4 14198.84 5062.51 5154.28 3982.05 0.78 0.72 0.68 0.56

Social prices 5 14651.33 5237.35 5154.28 4259.70 0.76 0.69 0.66 0.55

Divergence 1 -2666.65 -568.63 -144.23 -1953.80

Divergence 2 -3119.15 -743.47 -144.23 -2231.45

Divergence 3 -3571.64 -918.31 -144.23 -2509.10

Divergence 4 -4024.13 -1093.15 -144.23 -2786.75

Divergence 5 -4476.63 -1268.00 -144.23 -3064.40

Table 4: Risk Price Policy Analysis Matrix with Potash for Basmati Paddy in Punjab




Social prices 12841.36 4553.42 5154.28 3133.66 0.87 0.79 0.75 0.62

Social prices 2 13293.85 4743.70 5154.28 3395.87 0.84 0.77 0.73 0.60

Social prices 3 13746.35 4933.98 5154.28 3658.08 0.80 0.74 0.70 0.58

Social prices 4 14198.84 5124.26 5154.28 3920.30 0.77 0.72 0.68 0.57

Social prices 5 14651.33 5314.55 5154.28 4182.51 0.75 0.69 0.66 0.55

Divergence 1 -2666.65 -584.06 -144.23 -1938.36

Divergence 2 -3119.15 -774.35 -144.23 -2200.57

Divergence 3 -3571.64 -964.63 -144.23 -2462.78

Divergence 4 -4024.13 -1154.91 -144.23 -2724.99

Divergence 5 -4476.63 -1345.19 -144.23 -2987.21

In the Table 4 economic tradable cost was Rs. 3852.15 per acre based on Risk price of Potash. The

second row of PAM indicated social profitability, indicative of competitiveness in open market. The


NPI was 0.89, showing farmers were paying input prices closer to world market and there was no subsidy. The NPI discounted by 1.138 (premium) gives NPI 1.05 showing higher market prices than world prices. The higher market price was due to high transportation, freight charges and government taxes. The NPI showed declining trend in the subsequent years. The NPC for output (basmati) was 0.77; showing domestic market price was lesser than border price. On the basis of projected economic values, NPC slightly increased in 2nd year and then went on decreasing. It indicated increasing variance between domestic and border price in the coming years. Thus, distortion in market and border prices must be corrected through marketing efficiency and intelligence. The EPC was 0.72, indicative of implicit taxation on Basmati Rice, non-provision of support and resource transfer from this commodity to others. The DRC was 0.6 and decreased onward showing increasing comparative advantage and continued to do so in coming scenario due to specialization in this crop.

The Risk Price for IRRI Rice was estimated in the same way, as discussed in previous section. The normal distribution of time series FOB price of IRRI, plotted by “Best fit” is given in Fig 4.This Risk function indicating a mean Risk price of US $190.47 per ton generated the risk price of IRRI rice randomly. The EPP of paddy was indirectly calculated form rice price, by considering all marketing, processing cost and rice to paddy ratio as Rs. 200 per 40 Kg. In the Table 5 and table 6, Policy Analysis Matrix (PAM) for IRRI Rice provided results based on the cost of production estimates averaged over two provinces i.e. Punjab and Sindh. The results in both PAMs showed almost similar values therefore are discussed jointly.

The economic revenue indicated in second row was less than economic production costs and represented a net economic loss, thus indicating that Pakistan will have no competitiveness in export market. The NPI was less than 1 i.e. 0.85 indicative of lower input prices in domestic market than international market. This showed an implicit protection provided to these inputs. The price dispersion was going to increase in the next four years given the current level of market price. The NPC was 1.17; showing growers were earning higher profits at domestic price than international price. This indicated protection was provided to IRRI rice and incentive were being provided to farmers to enhance production of IRRI Rice. The EPC also confirmed this finding for first year and maintained the same magnitude in the projected values of output and input. The DRC was 1.17 and indicated a strong comparative disadvantage of Pakistan in IRRI production. The PAM results were similar to static analysis indicating comparative dis-advantage of Pakistan in IRRI production. Figure 4: The Normal Distribution of Time Series FOB Price of IRRI Paddy

Normal(190.475, 20.712)

Value

s x 10

-2

dollar per ton

0.0

0.5

1.0

1.5

2.0

2.5

3.0

160

170

180

190

200

210

220

230

240

< >5.0%90.0%156.41 224.54


Mean = 190.475

Mean = 190.475

Inpu

tFit


Table 5: Risk Price Policy Analysis Matrix with DAP for IRRI Paddy in Pakistan




Social prices 1 7328.82 3911.56 4444.44 -1027.18 0.83 1.21 1.64 1.30

Social prices 2 7519.29 4086.40 4444.44 -1011.54 0.80 1.18 1.64 1.29

Social prices 3 7709.77 4261.24 4444.44 -995.91 0.76 1.15 1.63 1.29

Social prices 4 7900.24 4436.09 4444.44 -980.28 0.73 1.12 1.62 1.28

Social prices 5 8090.72 4610.93 4444.44 -964.65 0.71 1.10 1.61 1.28

Divergence 1 1542.14 -656.41 52.43 2146.11

Divergence 2 1351.66 -831.25 52.43 2130.48

Divergence 3 1161.19 -1006.09 52.43 2114.85

Divergence 4 970.71 -1180.94 52.43 2099.22

Divergence 5 780.24 -1355.78 52.43 2083.58

Table 6: Risk Price Policy Analysis Matrix with Potash for IRRI Paddy in Pakistan

Revenue Production costs Profit Tradable Non-tradable


Social prices 1 7138.34 3884.27 4444.44 -1190.37 0.84 1.24 1.73 1.37

Social prices 2 7328.79 4031.84 4444.44 -1147.48 0.81 1.21 1.70 1.35

Social prices 3 7519.24 4206.68 4444.44 -1131.87 0.77 1.18 1.70 1.34

Social prices 4 7709.69 4381.52 4444.44 -1116.27 0.74 1.15 1.69 1.34

Social prices 5 7900.14 4556.36 4444.44 -1100.66 0.71 1.12 1.68 1.33

Divergence 1 1732.61 -629.12 52.43 2309.30

Divergence 2 1542.16 -776.69 52.43 2266.42

Divergence 3 1351.71 -951.53 52.43 2250.81

Divergence 4 1161.26 -1126.37 52.43 2235.20

Divergence 5 970.81 -1301.21 52.43 2219.59

IV. Conclusions and Recommendations Rice is an important economic crop of Pakistan contributes more than 5 percent to agriculture GDP. The country is major exporter of fine rice (Basmati) and facing stiff competition from other exporting countries. Therefore, assessment of its competitiveness is imperative to reap the benefits of free trade in the future. Thus the study was aimed at analyzing the changing comparative advantage of rice over time and its implications for trade development .The specific objectives of the study were to: reveal and analyze the agricultural comparative advantage of rice and assess its implications on its trade; determine the extent of agricultural protection and policy distortion in rice market in Pakistan; investing in technological, institutional, infrastructure development and human resources to increase competitiveness, through risk analysis; and provide relevant information to researchers and policy makers for better formulation of food, agriculture and trade policies.


The results showed that tradable inputs were not taxed for rice cultivation. Farmers almost paid close to world prices for the inputs and only a nominal protection was provided in the form of indirect subsidy on irrigation water. Despite of this factor, basmati rice was under priced rather than receiving border prices. The value of DRC was 0.60 for basmati rice indicating Pakistan has comparative advantage in production of basmati paddy. In the case of Irri, the DRC was 1.27, an indicator of comparative disadvantage in IRRI production at export parity price. The Value added for basmati and Irri per acre-inch of water was estimated as Rs. 150.68 and for Irri Rs. 45.88, respectively.

In the Risk analysis, projected economic tradable costs showed an increasing trend up to 5th year. The Nominal protection coefficient for inputs in the first year was estimated 0.87. It showed that farmer was paying prices closer to world market. The nominal protection co-efficient for output (NPC) was 0.79 in the first year and goes on decreasing up to 5th year. This explains under pricing of domestic output. It is indicative of implicit taxation and resource transfer from Basmati Rice. The EPC was 0.75, as compared to 0.68 in static model. It reaffirms the results of NPI and NPC.

In the coming WTO regime, farmers are likely to achieve world prices, resulting in an increase of EPC. The DRC, an indicator of comparative advantage was 0.62 and decreased up to 0.51 in the fifth year. The declining value of DRC indicates an increasing trend in the comparative advantage level of Basmati cultivation.

The risk analysis showed that Pakistan would not enjoy competitiveness in Irri trade in world market. The NPI was estimated 0.83 indicative of lower input prices in domestic market than international market. The price dispersion is going to increase in the next four years given the current level of market price. The NPC was 1.21; showing that growers will be earning higher profits at domestic price than international price at given market price. The EPC also confirmed this finding for first year and maintained the same magnitude in the projected values of output and input. The DRC was estimated 1.30 and indicated comparative disadvantage of Pakistan in Irri production at risk price in free trade scenario. V. Recommendations 1] The basmati rice has comparative advantage both in static and dynamic context. The

stakeholders should improve the quality of both commodities in order to remain competitive in the international market.

2] The private sector is not congruent to the market mechanism; therefore the powerful syndicate (REAP) may be regulated through anti-monopoly authority and or commercial courts. The market imperfections must be removed through marketing efficiency and institutionalization of market intelligence.

3] The electronic commerce must be encouraged at production, processing and marketing of rice trade. The grading and standardization of products in line with international standards must be ensured for rice.

4] The value added per acre-inch of water shows the criticality of this vital input. The present flat rate system is allocatively neutral leading to misallocation of this scarce resource. Therefore, the water pricing of this input is imperative for its rational allocation.

5] The introduction of high technology in handling and processing rice will help reduced the cost of marketing.

VI. References 1] Akhtar, W., S. Muhammad., M. Waqar and M.A. Kahlid. 2004.“Economic Incentives and

Comparative Advantage in Rice-Wheat System in Punjab” socio-economic research studies 2002-03 (Federal) Pakistan Agricultural Research Council. Pp: 21-38


2] Ali, S. 1992 “Producer And Consumer Subsidy Equivalents Of Agricultural Policies In Pakistan: Concept Measurement and Implication.’’ Pakistan Jour. of Agri. Economics vol. 1, No. 1(1992) Pp: 1-23.

3] APCom (Agricultural Prices Commission).2002. “Support Price Policy”. Reports on Wheat, Rice, Seed Cotton and Sugar cane. Agricultural Prices Commission, Islamabad.

4] Appleyard, D. R. 1987. “ Report On Comparative Advantage”. Agricultural Prices Commission (APCom series No.61) Islamabad

5] GOP(Government of Pakistan). 2001, 2002, 2003. “Economic Survey, 2001-02 to 2003-04. Economic Advisor’s Wing Finance Division. Ministry of Finance. Islamabad.

6] IRRI (International Rice Research Institute).2002, 2004. http://www.IRRI.org 7] Kannapiran, C. A. and E. M. Flemming.1999. “Competitiveness and Comparative Advantage

of Tree Crop Smallholdings in Papua New Guinea”. Working Paper Series in Agricultural and Resource Economics.

8] Khan,S.R.A.2001.“Rice__ Wheat Cropping System”. Dawn Business, http://dawn.com/2001/05 9] Longmire. J. and P. Debord. 1993. “Agricultural Pricing and Comparative Advantage in

Pakistan: An Update to 1991-92”. Report prepared for the South Asian Division of the World Bank Washington, D.C.

10] Masters, W. A and A. W. Nelson.1995. “Measuring the Comparative Advantage of Agriculture Activities: Domestic Resource Cost and Social Cost Benefit Ratio”. Amer. Jour. Agri. Econ. 77 ( May 1995): 243-250

11] Monke, E. and S. R. Pearson. 1989. The Policy Analysis Matrix for Agricultural Development. Ithaca, N.Y., U.S.A.: Cornell University Press.

12] Newberry, D.M.G. and J.E. Stiglitz.1981. “The Theory of Commodity Price Stabilization: A Case Study in The Economics Of Risk”. Oxford University Press, New York.

13] State Bank of Pakistan(SBP) 2004.”Statistical Bulletin”. Statistics Department. State Bank of Pakistan.

14] USDA (United States Department of Agriculture).1996. “Managing Risk in Farming: Concepts, Research.


The Predictability of Amman Stock Exchange Performance: A Univariate Autoregressive Integrated Moving

Average (ARIMA) Model

Mohammad Al-Shiab* Finance & Banking Department, Mutah University, Jordan

P.O.Box 7, Finance & Banking Dept., Business School, Mutah University, Mutah, Jordan E-mail address: [email protected]

Abstract The study examines the univariate ARIMA forecasting model using Amman Stock Exchange (ASE) daily general index over the period 4/1/2004 – 10/8/2004, with out-of-sample testing undertaken on the following 150 days of the data. Different diagnostic tests were performed to reach the best model fits the data, moreover. It is predicted that ASE will continue to grow by 0.04% in the coming 150 days starting from 11/8/2004. The findings revealed, moreover, that ASE future performance could be classified into three major coming periods; firstly, when the growth will be significantly positive. Secondly, when there will be a high fluctuation performance. Thirdly, when the growth will be steady by being around 0.04%.

I. Introduction It is argued that economic development means a continuing rise in the tempo of economic activity, which inevitably brings in its wake an increased demand for, and supply of capital, waiting to be channeled through a stock exchange. The ASE was established in March 1999 as a non-profit, private institution with administrative and financial autonomy. It is authorized to function as an exchange for the trading of securities. The ASE is committed to the principles of fairness, transparency, efficiency, and liquidity.

In Jordan, the stock exchange organization and all the development processes in action have an unfavorable public image and they have failed to win the complete confidence. It is well to recognize that in Jordan stock exchange had evolved in the past more like private clubs catering to the needs of its members only, and not as a public institutions functioning in the wider public interests. As a result, the stock exchange still remains an institution in which only a microscopic section of the country is interested. Inadequate disclosure practices by Jordanian companies listed in the stock market and, consequently, more speculative activities in the securities market might explain the lack of public interest to some extent.

As a consequence, Jordanian government adopted a comprehensive capital market reforming policy, which aimed at building on the previous experience, boosting the private sector, expanding and diversifying the national economy, and improving regulation of the securities market to reach international standards. Among the most important features of the new orientation were institutional changes in the capital market, use of international electronic trading, settlement and clearance systems,


elimination of obstacles to investment, and strengthening capital market supervision to reach optimum transparency and safe trading in securities, in line with globalization and openness to the external world. In 1997, therefore, a new securities law was enacted to reflect the development of systems and the sophistication of new products and participants. The enactment of the Temporary Securities Law, No. 23 of the year 1997, was a landmark; indeed, it was a qualitative leap and a turning point for the Jordanian capital market. Its aim was to restructure and regulate the Jordanian capital market, and to complete its infrastructure in consistency with international standards, in order to secure transparency and safe trading in securities (The financial market observes international standards of fair practice in the orderly transaction conduct of the market).

To comply with international standards and best practices since 1998, the ASE works closely with the Jordan Securities Commission JSC on surveillance matters and maintains strong relationships with other exchanges, associations, and international organizations. The exchange is an active member of the Union of Arab Stock Exchanges, Federation of Euro-Asian Stock Exchanges (FEAS), an affiliate member of the World Federation of Exchanges (WFE), and an affiliate member of the International Organization for Securities Commissions (IOSCO).

The above significant changes in the way of running the capital market compared to past history where the market suffered from sever operational and informational problems has raised up the issue whether such improvements will be reflected on the capital market future performance.

Forecasting is an important part of econometric analysis, more specifically when talking about the capital market and its performance. In this respect the Autoregressive Integrated Moving Average (ARIMA), popularly known as the Box-Jenkins methodology, one of the most popular forecasting methods. However, research in the forecasting field has largely centered on three alternative approaches, namely; the first has largely been based around vacancy rates (Wheaton 1987; Wheaton and Torto 1988), while the second has tended to use demand and supply variables in reduced form models. In addition, the third studies have examined the use of forecasting models in the context of rental forecasting (Brooks and Tsolacos 2000; Brooks and Tsolacos 2001; McGough and Tsolacos 1994; McGough and Tsolacos 1995a; Stevenson and MacGarth 2003; Tsolacos 2002; Tsolacos et al. 1998). In the last decade there has been an increasing amount of empirical research on property markets (Chandrashekaran and Young 2000; Geltner and Mei 1995; Karakozova 2004; Siviatanides 1998; Siviatanides et al. 2001; Siviatanidou and Siviatanides 1999; Viezer 1999). Consequently, it could be strongly argued that the empirical evidence based on this issue has largely been limited proving the significance of the study especially in an emerging capital market such as the Jordanian one since, to the best knowledge of the researcher, this is the first study to forecast in general, and to forecast the ASE future performance in particular. The technique of forecasting used is not conventional, moreover.

The rest of the paper organized as follows: Section II presents the literature, section III describes the study methodology adopted, Section IV presents the empirical results and its discussion, and the final section provides concluding comments. II. Literature Review Searching the literature, it can be said that empirical studies that used univariate ARIMA models were very rare forcing the researcher to focus on the theory of this technique. One of the earliest papers found developed by (Wilson et al. 2000) which focuses on predicting turning points in securitized real estate return series in the US, the UK and Australia. These authors employ ARIMA and exponential smoothing models to identify turning points and compare the performance of these models with a spectral technique. The former two models study the behavior of returns and forecast turning points from the time domain whereas the latter provides a view from the frequency domain. In this way, the authors attempt through spectral analysis to identify hidden cycles of differing lengths and amplitudes in the return series. The findings of this study showed that the spectral analysis, which assumes a four-


year cycle in the data, performs well in capturing turning points in all quarterly series of real estate returns. Moreover, the study found that spectral analysis identified more clearly turning points than ARIMA and outperformed the exponential smoothing models, which as expected did not predict the turning points. Further to the evidence on the ability of different models to predict turning points, this study alerts researchers about the need to study the cyclical behavior of real estate returns from the frequency domain point of view.

Leong and Maki (2000) considered the long-run neutrality hypothesis using Australian data. A reduced form autoregressive integrated moving average (ARIMA) model is used with both quarterly seasonally unadjusted and adjusted Australian real GDP and nominal money supply to test the neutrality hypothesis. Using two measures of money stock, namely; M1 and M3, it is shown that the hypothesis is supported using M1 as a measure of money supply, that is, changes in M1 have no effects on changes in real output. However, the long-run neutrality hypothesis is rejected using M3, in that changes in M3 significantly affect changes in real output. Theses results for Australian indicate the sensitivity of the outcome to the type of money supply used.

In the UK context, (Brooks and Tsolacos 2001b) focus on a single real estate return series and they explicitly examine the forecasting performance of three different methodologies. The first forecasting methodology is the unconditional mean of a trailing sample of 200 observations of real estate returns. Such a model is identical to a random walk with drift in the log of the prices. The other forecasting methodologies are n ARMA model and vector autoregressive model (VAR). Along with real estate returns, two other variables were included in the VAR system, namely; the term structure of interest rates and the gilt-equity yield ratio. Brooks and Tsolacos used these methodologies to produce 160 out-of-sample forecast for horizons of up to 6 months. The forecast evaluation criteria showed that the long-term mean and the ARMA specification are more appropriate for forecasting in the short-term (i.e. one month ahead) than the more complex VAR model. However, the latter model produces the highest proportion of correct return sign predictions. For the six-month ahead forecasts, the long-term means return produces the most accurate forecasts under all evaluation measures.

Another study developed by (Brooks and Tsolacos 2003) explored the issue of forecasting securitized real estate returns for five European countries, namely; the UK, Belgium, the Netherlands, France, and Italy. Within a VAR framework, it is demonstrated that the gilt-equity yield ratio is in most cases a better predictor of securitized returns that the term structure or the dividend yield. In particular, investors should consider in their real estate return models the predictability of the gilt-equity ratio in Belgium, the Netherlands, and France, and the term structure of interest rates in France. Predictions obtained from the VAR and univariate time-series models are compared with the predictions of an artificial neural network models. It is found that, whilst no single model is universally superior across all series, accuracy measures and horizons considered, the neural network model is generally able to offer the most accurate predictions for 1-month horizons. For quarterly and half-yearly forecasts, the random walk with a drift is the most successful for the UK, Belgium and Dutch returns and the neural network for French and Italian returns. Although this study underscores market context and forecast horizon as parameters relevant to the choice of the forecast model, it strongly indicates that analysis should exploit the potential of neural networks and assess more fully their forecast performance against more traditional models.

(Stevenson and MacGarth 2003) examine four alternative rental-forecasting models in the context of the London office market. The forecasting ability of an ARIMA model, a Bayesian Vector Autoregression approach, an OLS based single equation model, and a simultaneous equation model. The models are estimated using the CB Hiller Pasrker London Office index over the period 1977-1996, with out-of-sample testing undertaken on the following three years of data. Diagnostic testing is also conducted on the alternative models. The findings reveal that the Bayesian VAR model produces the best forecasts, while the ARIMA model fails to pick up on the large uptake in rental values during the testing period.


(Karakozova 2004) presents an econometric study of office returns determination in the Helsinki area, a small European market, over the 30-years period from 1971 to 2001. Particularly, the study investigates the variation in office capital growth, which is the most volatile component of office total return in the Helsinki market using three alternative models, namely; a regression model, an error correction model (ECM), and ARIMA model with exogenous explanatory variable (i.e. ARIMAX). The study also evaluates the forecasting performance of the alternative specifications. The results indicate that the ARIMAX models, incorporating past values of capital growth and growth in service sector employment and in the gross domestic product, are able to pick up shocks present in the data, and thus provide the best forecasting tool for office returns in Helsinki. The ECM models, which incorporate long-run information, can not satisfactorily model the irregulaties in the Helsinki office market, such as the boom-bust cycle of the 1980s and 1990s and thus are suitable for modeling and forecasting only part of the Helsinki market. it is predicted that office capital returns will continue to grow in real terms by 0.1% on average in the 2002-2005 period. This implies that real office total returns will grown average at the rate of 5.7% over the same time period. III. Methodology and Data Selection Generally speaking, there are five approaches to economic forecasting based on time series data, namely; exponential smoothing methods, single-equation regression models, simultaneous-equation regression models, autoregressive integrated moving average models (ARIMA), and vector auto-regression (VAR).

The uniqueness of the ARIMA models compared to the others is that the emphasis is not on constructing single-equation or simultaneous-equation models but on analyzing the probabilistic, or stochastic, properties of economic time series on their own under the philosophy let the data speak of themselves. Unlike the regression models, in which Υt is explained by K explanatory variable, the Box-Jenkins methodology allow Υt to be explained by past or lagged values of Y itself and stochastic error terms. For this reason, ARIMA models are sometimes called atheoretic models because they are not derived from any economic theory where it is the basis of simultaneous-equation models.

It said that Υt follows a first-order autoregressive or AR(1), stochastic process, if the following is equation true:

(Υt – δ) = α1 (Υt -1 - δ ) + ut (1)

where δ in the mean of Υ and where ut is uncorrelated random error term with zero mean and constant variance σ² (i.e. it is white noise). The above model indicates that the value of Y at time t depends on its value in the previous time period and a random term, where the Y values are expressed as deviations from their mean value. In other words, the model indicates that the forecast value of Y at time t is simply some proportion of its value at time (t-1) equal α1 plus a random shock or disturbance at time t. If the value of Y at time t depends on its value in the previous two time periods, however, it said that Υt follows a second-order autoregressive or AR(2). In general, the model can be written as:

(Υt – δ) = α1 (Υt -1 - δ ) + α2 (Υt-2 - δ ) + ….. + αp (Υ t-p - δ ) + ut (2)

in equation (2) it is said to be a pth-order autoregressive or AR(p) process. It has to be clarified that the AR(p) process is not the only mechanism that may have generated Υ.

If Y at time t is equal to a constant plus a moving average of the current and past white noise error terms, it is said that Y follows a first-order moving average or MA(1) process expressed as:

Υt = µ + β0ut + β1ut-1 (3)

In general, the model can be written as:

Υt = µ + β0ut + β0ut -1 + β0ut -2 + …. + β0ut -q (4)

In equation (4) it is said to be a pth-order moving average or MA(q) process.


It is quite likely that Y has a characteristics of both AR and MA which is called ARMA and Y, therefore, follows an ARMA(q, p) process written as:

Υt = θ + α1 Υt -1 + α2 Υt-2 + ….. + β0ut + β0ut -1 + β0ut -2 + …. + β0ut -q (5)

Where θ represents a constant terms. The time series models discussed earlier are based on the critical assumption that that the time

series explored are stationary. In other words, the mean and variance for the series are constant and its covariance is time invariant. The logic behind the stationary assumption is that if the estimated model to be used for forecasting is stationary, then it must be assumed that the features of this model are constant through time, and particularly over future time periods. Thus, the simple reason for requiring stationary data is that any model which is inferred from these data can itself be interpreted as stationary or stable, therefore, providing valid basis for forecasting. It could be strongly argued that many financial and economic data series are non-stationary, that is, they are integrated, however. For example, if a time series is integrated of first order (i.e. it is I(1)), its first differences are I(0), that is stationary. Generally speaking, if a time series is I(d) after differencing it d times, I(0) can be obtained. If we have to difference a time series d times to make it stationary and then apply the ARMA(q, p) model, we say that the original time series is ARIMA(q, d, p), that is, it is an autoregressive integrated moving average time series. Stationary issue will be checked by adopting the Dickey-Fuller, Augmented Dickey-Fuller, and Phillips-Perron unit root tests.

In order to be able to adopt the proper model, the q, d, and p values should be determined. To do so, correlogram and partial correlogram,1 hopefully, provide the needed aid in this task which will be used in this paper. Consequently, the model to be tested can be rewritten as:

Υ*t = θ + α1Υ*t-1 + α2Υ*t-2 +…+ αpΥ*t-p + β0ut + β1ut-1 +… + β1ut-q (6)

Where Υ* denote the d difference of a time series. Having identified the appropriate q and p values, estimating the parameters of the autoregressive

moving average terms included in the model will be in hand using the E-views. But, still the chosen model has to be checked to make sure it fits the data reasonably well, for it is possible that another ARIMA model might do the same job as well. That is why adopting Box-Jenkins ARIMA modeling needs considerable skills to choose the right ARIMA model. Some of those tests required are R-squared, Adjusted R-squared, Akaike info criterion (AIC), Shwarz criterion (SIC), Durbin-Watson (D-W), and Jarque-Bera for the residuals. The model should be chosen that one which has the lowest AIC and SIC on one side, but it has the highest R-squared and Adjusted R-squared on the another and its D-W more close to (2).

It is argued by econometricians, moreover, that volatility clustering is one of the special problems involved in forecasting prices of financial assets (Gujarati 2003), an issue will be checked by adopting the Autoregressive Conditional Heteroscedasticity (ARCH LM) model for the residuals.

By passing the earlier stages MINITAB will be employed for forecasting the Amman Stock Exchange General Index over the coming period after choosing the most model fits the data reasonably well. IV. The Empirical Results and Discussion Following the methodology presented in the above section, Table 1 shows the correlogram and partial correlogram of the ASE daily general index and first difference daily general index. From this table it has to be argued that the ACF of the ASE daily general index declines gradually and it is up to 23 lags are individually statistically significantly different from zero, for they all are outside the 95%

1 The partial autocorrelation ρkk measures correlation between time series observations that are k time periods apart after controlling for correlations at intermediate lags (i.e. lags less than k). In other words, partial autocorrelation is the correlation between Yt and Yt-k after removing the effect of the intermediate Y's since in time series data a large proportion of correlation between Yt and Yt-k might due to the correlations they have with the intervening lags Yt-1, Yt-2, …, Yt-k+1. Therefore, the partial correlation ρkk removes the influence of these intervening variables.


confidence pounds.1 But at all other lags, they are not statistically different from zero. The same rule could be used for the PACF which drops dramatically after the first lag from 0.92 to -0.192 and all its values after lag one are statistically insignificant. Table 1: Correlogram and Partial Correlogram 4/1/2004 – 10/8/2004

ASE Daily General Index ASE First Difference Daily General Index Lag ACF PACF ACF PACF

1 0.920 0.920 0.204 0.204 2 0.818 -0.192 -0.156 -0.206 3 0.737 0.114 -0.100 -0.022 4 0.674 0.024 -0.052 -0.059 5 0.628 0.074 -0.038 -0.039 6 0.597 0.058 0.046 0.047 7 0.554 -0.095 0.042 0.000 8 0.502 -0.019 -0.215 -0.237 9 0.474 0.141 -0.108 0.007

10 0.460 0.017 -0.009 -0.067 11 0.449 0.34 -0.019 -0.053 12 0.441 0.018 -0.037 -0.062 13 0.428 -0.001 -0.104 -0.152 14 0.423 0.109 0.010 0.046 15 0.423 0.004 0.092 0.052 16 0.414 -0.053 0.132 0.041 17 0.388 -0.073 0.214 0.205 18 0.339 -0.141 0.032 -0.045 19 0.285 -0.008 -0.126 -0.057 20 0.248 0.037 -0.112 -0.042 21 0.235 0.059 0.125 0.102 22 0.217 -0.082 -0.005 -0.102 23 0.203 0.069 -0.012 0.062 24 0.184 -0.045 -0.039 -0.078 25 0.166 0.046 -0.134 -0.025 26 0.158 -0.008 -0.017 0.065 27 0.156 -0.026 0.093 0.050 28 0.140 -0.081 0.016 -0.042 29 0.120 0.015 -0.070 0.031 30 0.104 -0.005 0.060 0.081 31 0.075 -0.119 0.090 0.053 32 0.043 -0.021 0.053 0.007 33 0.006 -0.085 0.133 0.091 34 -0.054 -0.136 0.011 -0.028 35 -0.117 -0.033 -0.104 -0.021 36 -0.157 0.019 -0.129 -0.090

The above clarification highlights the fact that the ASE daily general index time series is not

stationary and it has to be transferred to a stationary situation where we can apply the Box-Jenkins methodology. Such a move can be recognized by looking at the ACF and PACF of the ASE first difference daily general index time series presented in Table 1 since no observed trend in this series can be seen suggesting that the first difference of the time series is stationary. Following the robustness approach, Dickey-Fuller, Augmented Dickey-Fuller, and Phillips-Perron unit root test were employed as its results presented in Table 2. 1 The approximate 95% confidence limits for ρk are -0.2089 and +0.2089 (for more details see Gujarati, 2003, PP.485-488).


Table 2: ASE Daily General Index Stationary Test1

Unit Root Test Computed Value Critical Value at 1% level Dickey-Fuller -2.428876 -3.4749

Augmented Dickey-Fuller -2.919549 -3.4752 Phillips-Perron -2.642382 -3.4749

Since the computed values for all unit root tests, in absolute values, are less than the critical values at 1% level of significance, it could be concluded that ASE daily general index time series is not stationary confirming the results presented in Table 1. Running the same tests for the ASE first difference daily general index, it seems that the series become a stationary one as reported Table 3. Table 3: ASE First Difference Daily General Index Stationary Test

Unit Root Test Computed Value Critical Value at 1% level Dickey-Fuller -9.823964 -3.4752

Augmented Dickey-Fuller -9.419886 -3.4755 Phillips-Perron -9.867652 -3.4752

Moving further, the next step is to find the p and q that fits the series very well. One way to accomplish this is to consider the ACF and PACF and the associated correlograms of a selected number of ARIMA processes. Since each of these stochastic processes exhibit typical patterns of ACF and PACF, we can identify the time series with that process if the time series fits one of these patterns. Since ACF and PACF of ASE first difference daily general index do not show any typical pattern to conclude p and q values, the researcher decided to choose a collection of ARIMA(p,d,q) models for different values for p and q. Following the procedures explained in the methodology, therefore, four standards used for determining the most accurate ARIMA model from the collection exercised, namely; R-square, Adjusted R-square, Akaike info criterion (AIC), and Shwarz criterion (SIC). Appendix 1 present the mentioned tests ranked according to the p and q values used.

Clearly, Appendix 1 indicates that model number 30 is the most accurate one when p = 4, d = 1, and q = 5 according to the four standards used. The results of the univariate ARIMA(4,1,5) model are shown in Table 4. Table 4: Univariate ARIMA(4, 1, 5) model Estimators: ASE First Difference Daily General Index is the Dependent Variable

Variable Coefficient Std. Error t-Statistic Prob. C 1.118764 1.610363 0.694728 0.4884

AR(1) 0.520680 0.100598 5.175849 0.0000 AR(2) -0.728015 0.078818 -9.236705 0.0000 AR(3) 0.732310 0.088661 8.259701 0.0000 AR(4) -0.367145 0.124839 -2.940961 0.0038 MA(1) -0.234329 0.121806 -1.923785 0.0565 MA(2) 0.578561 0.110389 5.241128 0.0000 MA(3) -0.858537 0.125385 -6.847218 0.0000 MA(4) 0.110965 0.155356 0.714263 0.4763 MA(5) 0.182852 0.111753 1.636220 0.1041

1 It has to be mentioned that one truncation lag is found to be enough for having insignificant residuals autocorrelation.


R- square 34.28% AIC 8.962 Adjusted R- square 30.44% SIC 9.146 D-W 2.003 F-Statistic 8.125

Prob.( F-Statistic) 0.000

As it can be seen from Table 4 that the coefficient of AR(1), AR(2), AR(3), AR(4), MA(2), and MA(3) are highly significant. R- square and Adjusted R- square are not high enough could be explained as a result of loss of information about the long-run relationship among variable (Pindyck and Rubinfeld 1991). The value of Durbin-Watson, moreover, strongly support the view that there is no positive or negative first order serial correlation. In addition, the F-Statistic value is high enough to be strongly significant indicating the fact that ARIMA(4, 1,5) is well fits the data and suitable for accurate forecasting.

A very strict assumption in choosing the most accurate model is that the error terms are assumed to have white noise (i.e. zero mean and constant variance), uncorrelated, and normally distributed. Jarque-Berra test employed (see Appendix 1) along with its probability showing that the residuals are normally distributed (Jarque-Berra value 0.161616 and its probability 0.922371). In addition, the ARCH LM test for autocorrelation in the error variance of the chosen model was conducted showing that the error terms are not serially correlated since χ² = 2.463423 and its probability higher than even 10% level of confidence (Prob. 0.116525), consequently, residuals do not contain ARCH effects.

It has to be mentioned that D-W statistic tests only for the first order autocorrelation and it requires both an intercept in the regression and no lagged dependent variables among the regressors. While an important and widely used test for the autocorrelation, it is still only a partial test. A more general approach to testing for serial autocorrelation is to compute the autocorrelations and partial autocorrelations of the residuals up to any specified number of lags, however. The ljung-Box Q statistic tests for serial autocorrelation by summarizing the autocorrelations to see whether they are zero; that is, that the series is white noise. If the series is the residuals from ARIMA estimation, the number of degrees of freedom is the number of autocorrelations less the number of autoregressive and moving average terms previously estimated. Using SPSS software, Table 7 shows that the residuals autocorrelations equal to zero at different chosen number of lags since p-values are greater than 10% confirming the fact that our model the most appropriate one among the collection models and fits the data quite well. Table 5: ASE Forecasted Daily General Index 11/8/2004 – 27/3/20051

Lags Ljung-Box (χ²) Degrees of Freedom P-Value 6 2.023 1 0.959 7 10.786 2 0.217 8 10.904 3 0.282

10 14.154 5 0.225 12 14.336 3 0.280 14 21.079 9 0.134 16 22.214 7 0.136 18 26.436 13 0.118 30 39.784 25 0.134 32 41.026 23 0.132

1 See Appendix 2 for more details concerning the Autocorrelations and Partial Autocorrelations.


Thus our model would take the following form:

Y*t = 1.118764 + 0.520680Y*t-1 - 0.728015Y*t-2 + 0.732310Y*t-3 - 0.367145Y*t-4 + εt - 0.234329εt-1 + 0.578561εt-2 - 0.858537εt-3 + 0.110965 εt-4 + 0.182852 εt-5

Once the model chosen passed the early steps, a forecast can be obtained for the original time series using p and q founded. Our k-period forecast of ƒ YT would be given by:

ƒYT (K) = ƒYT + ƒY*T(1) + ƒY*T(2) + … +ƒY*T(K) (7)

where ƒYT is the forecast of Y and ƒY* is the forecast ∆Y. The results of forecasting of Y over the next 150 days covering the period 11/8/2004 – 27/3/2005 (e.g. the period allowed by the MINTAB software is 150 since the total number of observations entered 151) are given in Appendix 3.

As it can be seen from Appendix 3 that ASE performance could be classified into three major coming periods. Firstly, covering the period 11/8/2004 – 16/8/2004 when the growth will be significantly positive. Secondly, covering the period 17/8/2004 – 18/10/2004 when there will be a fluctuation performance since the growth will go up and down. Thirdly, covering the period 19/10/2004 – 27/3/2005 when the growth rate will be fixed (i.e. 0.04%) highlighting the potential of quite slow growth of ASE capital market. V. Conclusion This study has predicted the ASE performance over the coming 150 coming days. The ASE daily general index were used by a time series techniques, namely; Univariate Autoregressive Integrated Moving Average (ARIMA) model. Different diagnostic tests also performed to reach the best model fits the data. The findings revealed, moreover, that ASE future performance could be classified into three major coming periods; Firstly, covering the period 11/8/2004 – 16/8/2004 when the growth will be significantly positive. Secondly, covering the period 17/8/2004 – 18/10/2004 when there will be a high fluctuation performance. Thirdly, covering the period 19/10/2004 – 27/3/2005 when the growth will be steady by being around 0.04%. Overall, it is predicted that ASE will continue to grow by 0.04% in the coming 150 days starting from 11/8/2004. VI. References 1] Brooks, C.; and S. Tsolacos (2001b), "Forecasting Real Estate Returns using Financial

Spreads," Journal of Property Research, PP. 235-48. 2] ---- (2003), "International Evidence on the Predictability of Returns to Securitized Real Estate

Assets: Econometric Models Versus Neural Networks," Journal of Property Research, PP. 133-55.

3] Brooks, C.; and S.; Tsolacos (2000), "Forecasting Models of Retail Rents, Environment and Planning," 1825-39.

4] ---- (2001), "The Trading Profitability of Forecasts of the Gilt-Equity Ratio," International Journal of Forecasting, 11-29.

5] Chandrashekaran, V.; and M.S. Young (2000), "The Predictability of real Estate Capitilization Rates," in ERES Conference. Santa Barbara.

6] Geltner, D.; and G. Mei (1995), "The Present Value Model with Time-Varying Discount Rates: Implications for Commercial Property Valuation and Investment Decisions," Journal of Real Estate Finance and Economics, PP. 119-35.

7] Gujarati, Damodar N. (2003), Basic Econometrics (4th ed. ed.): McGraw-Hill. 8] Karakozova, Olga (2004), "Modelling and Forecasting Office Returns in the Helsinki Area,"

Journal of Property Research, PP. 51-73.


9] McGough, T.; and S. Tsolacos (1994), "Forecasting Office Rental Values using Vector Autoregressive Models," in The Proceedings of the Cut Edge Property Research Conference. Royal Institution of the Chartered Surveyors, London.

10] ---- (1995a), "Property Cycles in the UK: An Empirical Investigation of the Stylised Facts," Journal of Property Finance, PP. 45-62.

11] Pindyck, Robert S.; and Daneil L. Rubinfeld (1991), Econometric Models and Economic Forecasts (3rd ed. ed.): McGraw-Hill, Inc.

12] Siviatanides, P. (1998), "Predecting Office Returns: 1997-2001," Journal of Real Estate Finance, PP. 33-42.

13] Siviatanides, P.;, J.; Southard, W.; Torto, and W. Wheaton (2001), "The Determinants of Appraisal Based Capitalization Rates," Journal of Real Estate Finance, PP. 27-38.

14] Siviatanidou, R.; and P. Siviatanides (1999), "Office Capitalization Rates: Real Estate and Capital Market Influences," Journal of Real Estate Finance and Economics, PP. 297-322.

15] Stevenson, Simon; and Oliver MacGarth (2003), "A Comparison of Alternative Rental Forecasting Models: Empirical Tests on the London Office Market," Journal of Property Research, PP. 235-60.

16] Tsolacos, S. (2002), "A Symmetric Responses of UK Real Estates Returns to the Business Cycle," in ERES Conference. Glasgow, Scotland.

17] Tsolacos, S.;, G.; Keogh, and T. McGough (1998), "Modeling, Use, Investment, and Development in the British Office Market," Environment and Planning, 1409-27.

18] Viezer, T.W. (1999), "Econometric Integration of Real Estate's Space and Capital Markets," Journal of Real Estate Research, PP. 503-19.

19] Wheaton, W. (1987), "The Cyclical Behavior of the National Office Market," Journal of American Real Estate and Urban Economics Association, 281-99.

20] Wheaton, W.; and R. Torto (1988), "Vacancy Rates and teh Future of Office Rents," Journal of American Real Estate and Urban Economics Association, 430-6.

21] Wilson, P.;, J.; Okunew, G.; Ellis, and D. Higgins (2000), "Comparing Univariate Forecasting Techniques in Property Markets," Journal of Real Estate Portfolio Management, PP. 283-306.




Appendix 3: ASE Forecasted Daily General Index 11/8/2004 – 27/3/2005

UAFORCST LAFORCST AFORECST Forecasted Growth % Date Forecasted Day No.

2931.18 2841.64 2886.41 - 11-8-2004 1 2961.13 2821.27 2891.2 0.17 12-8-2004 2 2981.45 2817.58 2899.52 0.29 15-8-2004 3 2995.46 2821.33 2908.4 0.31 16-8-2004 4 3001.47 2818.48 2909.98 0.05 17-8-2004 5 3009.91 2822.52 2916.22 0.21 18-8-2004 6 3015.03 2825.35 2920.19 0.14 19-8-2004 7 3016.42 2823.94 2920.18 0.00 22-8-2004 8 3021.31 2827.49 2924.4 0.14 23-8-2004 9 3022.68 2827.52 2925.1 0.02 24-8-2004 10 3023.25 2826.11 2924.68 -0.01 25-8-2004 11 3026.47 2827.97 2927.22 0.09 26-8-2004 12 3026.54 2825.91 2926.22 -0.03 29-8-2004 13 3027.78 2824.34 2926.06 -0.01 30-8-2004 14 3030.42 2824.34 2927.38 0.05 31-8-2004 15 3030.97 2820.88 2925.93 -0.05 1-9-2004 16 3033.59 2818.96 2926.27 0.01 2-9-2004 17 3036.54 2817.14 2926.84 0.02 5-9-2004 18 3038.55 2812.87 2925.71 -0.04 6-9-2004 19 3042.61 2810.46 2926.54 0.03 7-9-2004 20 3046.3 2807.22 2926.76 0.01 8-9-2004 21

3049.89 2802.74 2926.31 -0.02 9-9-2004 22 3055 2799.95 2927.47 0.04 13-9-2004 23

3059.39 2795.94 2927.66 0.01 14-9-2004 24 3064.15 2791.77 2927.96 0.01 15-9-2004 25 3069.76 2788.8 2929.28 0.05 16-9-2004 26 3074.59 2784.66 2929.63 0.01 19-9-2004 27

3080 2781.06 2930.53 0.03 20-9-2004 28 3085.66 2778.1 2931.88 0.05 21-9-2004 29 3090.68 2774.28 2932.48 0.02 22-9-2004 30 3096.28 2771.29 2933.79 0.04 23-9-2004 31 3101.73 2768.45 2935.09 0.04 26-9-2004 32 3106.76 2765.15 2935.95 0.03 27-9-2004 33 3112.24 2762.68 2937.46 0.05 28-9-2004 34 3117.35 2760.04 2938.7 0.04 29-9-2004 35 3122.28 2757.29 2939.78 0.04 30-9-2004 36 3127.49 2755.18 2941.33 0.05 3-10-2004 37 3132.26 2752.77 2942.51 0.04 4-10-2004 38 3137.02 2750.49 2943.76 0.04 5-10-2004 39 3141.89 2748.62 2945.25 0.05 6-10-2004 40



3146.36 2746.44 2946.4 0.04 7-10-2004 41 3150.94 2744.53 2947.73 0.05 10-10-2004 42 3155.47 2742.79 2949.13 0.05 11-10-2004 43 3159.72 2740.84 2950.28 0.04 12-10-2004 44 3164.1 2739.17 2951.64 0.05 13-10-2004 45

3168.35 2737.52 2952.94 0.04 14-10-2004 46 3172.43 2735.77 2954.1 0.04 17-10-2004 47 3176.63 2734.26 2955.45 0.05 18-10-2004 48 3180.65 2732.68 2956.67 0.04 19-10-2004 49 3184.62 2731.09 2957.85 0.04 20-10-2004 50 3188.64 2729.68 2959.16 0.04 21-10-2004 51 3192.51 2728.17 2960.34 0.04 24-10-2004 52 3196.37 2726.72 2961.55 0.04 25-10-2004 53 3200.25 2725.37 2962.81 0.04 26-10-2004 54 3204.01 2723.93 2963.97 0.04 27-10-2004 55 3207.8 2722.58 2965.19 0.04 28-10-2004 56

3211.56 2721.28 2966.42 0.04 31-10-2004 57 3215.23 2719.92 2967.58 0.04 1-11-2004 58 3218.95 2718.66 2968.8 0.04 2-11-2004 59 3222.61 2717.4 2970 0.04 4-11-2004 60 3226.23 2716.12 2971.17 0.04 7-11-2004 61 3229.88 2714.93 2972.4 0.04 8-11-2004 62 3233.46 2713.71 2973.59 0.04 9-11-2004 63 3237.03 2712.51 2974.77 0.04 10-11-2004 64 3240.61 2711.38 2976 0.04 17-11-2004 65 3244.14 2710.21 2977.18 0.04 18-11-2004 66 3247.66 2709.09 2978.38 0.04 21-11-2004 67 3251.18 2708 2979.59 0.04 22-11-2004 68 3254.65 2706.9 2980.77 0.04 23-11-2004 69 3258.13 2705.84 2981.98 0.04 24-11-2004 70 3261.58 2704.8 2983.19 0.04 25-11-2004 71 3265.01 2703.75 2984.38 0.04 28-11-2004 72 3268.44 2702.75 2985.59 0.04 29-11-2004 73 3271.84 2701.75 2986.8 0.04 30-11-2004 74 3275.22 2700.77 2987.99 0.04 1-12-2004 75 3278.6 2699.82 2989.21 0.04 2-12-2004 76

3281.95 2698.87 2990.41 0.04 5-12-2004 77 3285.29 2697.93 2991.61 0.04 6-12-2004 78 3288.62 2697.03 2992.82 0.04 7-12-2004 79 3291.92 2696.12 2994.02 0.04 8-12-2004 80 3295.22 2695.24 2995.23 0.04 9-12-2004 81 3298.5 2694.38 2996.44 0.04 12-12-2004 82



3301.77 2693.52 2997.64 0.04 13-12-2004 83 3305.02 2692.68 2998.85 0.04 14-12-2004 84 3308.26 2691.85 3000.06 0.04 15-12-2004 85 3311.49 2691.04 3001.26 0.04 16-12-2004 86 3314.7 2690.24 3002.47 0.04 19-12-2004 87 3317.9 2689.45 3003.68 0.04 20-12-2004 88

3321.09 2688.67 3004.88 0.04 21-12-2004 89 3324.26 2687.92 3006.09 0.04 22-12-2004 90 3327.42 2687.16 3007.29 0.04 23-12-2004 91 3330.57 2686.43 3008.5 0.04 26-12-2004 92 3333.71 2685.7 3009.71 0.04 27-12-2004 93 3336.84 2684.98 3010.91 0.04 28-12-2004 94 3339.96 2684.28 3012.12 0.04 29-12-2004 95 3343.06 2683.59 3013.33 0.04 2-1-2005 96 3346.15 2682.91 3014.53 0.04 3-1-2005 97 3349.24 2682.24 3015.74 0.04 4-1-2005 98 3352.31 2681.58 3016.94 0.04 5-1-2005 99 3355.37 2680.92 3018.15 0.04 6-1-2005 100 3358.42 2680.28 3019.35 0.04 9-1-2005 101 3361.46 2679.65 3020.56 0.04 10-1-2005 102 3364.5 2679.03 3021.76 0.04 11-1-2005 103

3367.52 2678.42 3022.97 0.04 12-1-2005 104 3370.53 2677.82 3024.18 0.04 13-1-2005 105 3373.53 2677.23 3025.38 0.04 16-1-2005 106 3376.53 2676.65 3026.59 0.04 17-1-2005 107 3379.51 2676.07 3027.79 0.04 18-1-2005 108 3382.49 2675.51 3029 0.04 24-1-2005 109 3385.46 2674.95 3030.2 0.04 25-1-2005 110 3388.41 2674.41 3031.41 0.04 26-1-2005 111 3391.36 2673.87 3032.62 0.04 27-1-2005 112 3394.31 2673.34 3033.82 0.04 31-1-2005 113 3397.24 2672.82 3035.03 0.04 1-2-2005 114 3400.16 2672.3 3036.23 0.04 2-2-2005 115 3403.08 2671.8 3037.44 0.04 3-2-2005 116 3405.99 2671.3 3038.64 0.04 6-2-2005 117 3408.89 2670.81 3039.85 0.04 7-2-2005 118 3411.78 2670.33 3041.06 0.04 8-2-2005 119 3414.67 2669.85 3042.26 0.04 9-2-2005 120 3417.55 2669.39 3043.47 0.04 13-2-2005 121 3420.42 2668.93 3044.67 0.04 14-2-2005 122 3423.28 2668.48 3045.88 0.04 15-2-2005 123 3426.14 2668.03 3047.08 0.04 16-2-2005 124



3428.99 2667.59 3048.29 0.04 17-2-2005 125 3431.83 2667.16 3049.5 0.04 20-2-2005 126 3434.66 2666.74 3050.7 0.04 21-2-2005 127 3437.49 2666.32 3051.91 0.04 22-2-2005 128 3440.31 2665.91 3053.11 0.04 23-2-2005 129 3443.13 2665.51 3054.32 0.04 24-2-2005 130 3445.94 2665.11 3055.52 0.04 27-2-2005 131 3448.74 2664.72 3056.73 0.04 28-2-2005 132 3451.53 2664.34 3057.94 0.04 1-3-2005 133 3454.32 2663.96 3059.14 0.04 2-3-2005 134 3457.1 2663.59 3060.35 0.04 3-3-2005 135

3459.88 2663.23 3061.55 0.04 6-3-2005 136 3462.65 2662.87 3062.76 0.04 7-3-2005 137 3465.41 2662.52 3063.96 0.04 8-3-2005 138 3468.17 2662.17 3065.17 0.04 9-3-2005 139 3470.92 2661.83 3066.38 0.04 13-3-2005 140 3473.67 2661.5 3067.58 0.04 14-3-2005 141 3476.41 2661.17 3068.79 0.04 15-3-2005 142 3479.14 2660.84 3069.99 0.04 16-3-2005 143 3481.87 2660.53 3071.2 0.04 17-3-2005 144 3484.59 2660.22 3072.4 0.04 20-3-2005 145 3487.31 2659.91 3073.61 0.04 21-3-2005 146 3490.02 2659.61 3074.82 0.04 22-3-2005 147 3492.73 2659.31 3076.02 0.04 23-3-2005 148 3495.43 2659.02 3077.23 0.04 24-3-2005 149 3498.13 2658.74 3078.43 0.04 27-3-2005 150


Managerial Leadership Values across Cultures

Osarumwense Iguisi 1 Osarumwense Iguisi, Senior Research Fellow

Euro-African Management Research Centre, Maastricht-the Netherlands. E-mail: [email protected]

Abstract There has been general upsurge in cultures and managerial research in the last decades or so. Despite this fact, empirical studies on culture dimensions to management practices in Sub-Saharan Africa have been limited in numbers and scope. In Africa, we have very limited knowledge about its cultural values and the consequences it poses for leadership and motivation. In order to bring this very crucial and important issue as area for research priority to African scholars, business and management researchers, consultants and development experts, work-related values were studied in four European countries (France, Italy, Netherlands and Scotland) and one African (Nigeria) country through questionnaire. A major research question was whether the results could help to explain the disappointing economic development of African countries, including Nigeria. The findings do confirm profound differences in cultural values among the five countries involved in the research project. About motivation-the collective interest plays a more important role in Nigeria. In terms of preferred type of manager, the four European countries preferred the consultative type while the Nigerian respondents show a preference for the persuasive-paternalistic manager type. The participative manager type was more often rejected than preferred by the respondents.

In view of these differences, Western individualist and participative management models may not be very appropriate for Africa in general and Nigeria in particular. The suggestion is made to look for suitable African management models by studying the relatively more successful local companies and institutions.

I. Introduction

The importance of culture for effective management in Africa has become increasingly obvious in recent years as many of the expectations of African organizations and institutions created and managed along lines of Western textbooks and models have failed to achieve expected results of economic growth and sustainable development. In most or all projections of economic development, Sub-Saharan African Countries score poorly. Africa figures as the poor relative in the world family of nations and seems to be condemned to remain so for the foreseeable future. In official statistical data, Sub-Saharan African countries nearly always show up at the negative end. This situation has therefore brought management in most countries of Sub-Saharan Africa under severe criticisms for poor performance. In Africa, there are a few examples if any, of very successful (Western type) indigenous organizations


yet. Among several reasons for this dramatic situation, a lack of adequate and appropriate indigenous management takes a prominent position. The noticeable lack of success of many African formal organizations created and managed along lines of Western theories and models can be attributable to the fact of the African elite and managers ignorant of African historical and cultural conditions. Thus, while industrial organizations have no doubt, contributed somehow to the development and progress of Sub-Saharan African countries, I am of the opinion that the economic performances of these countries since decolonization have been disappointing compared to the achievements of other developing countries. Yet, the one factor, which consistently emerges from experience is that, African businesses not created and managed along lines of Western theories and models continue to flourish and experiencing positive growth situations. And often, only marginally successful westernized economic sector coexists with this revitalized traditional sector that goes largely unrecognized by statistical indicators.

It follows that where committed resources, both human and material, are not achieving set objects, there is the need therefore, to re-examine management theories and models, how they have been applied, and why they are not achieving set objectives or have failed. This is with the view to modifying and adapting them taking into consideration the impact of cultural values and the local environment on management processes and practices.

The objective of this research is to draw attention to the relevance of cultures to management with the sole purpose of contributing to a culturally feasible managerial leadership and motivation. Leadership and motivation practice in Africa requires identification of the 'growth-positive' and 'growth-negative' culture based-factors. Culture and Management Discourse The last decade has brought a renaissance of interest in cultural phenomena in societies and organizations. Researchers from a variety of disciplines (anthropology, psychology, sociology, and organization behavior) have provided a range of theoretical and analytical studies. Perhaps because of the different epistemological, methodological and political orientations that distinguish these disciplines, the literature remains theoretically unintegrated – in a state of conceptual chaos.

It is essential in any domain of research to define clearly the topic being studied, so that it is clear to readers what is and what is not being discussed. Before reviewing the relevant literatures about culture and cultures in organization, and the impact of culture on management, we must first agree on a workable definition of culture. Culture is a common word and like all common words it comes with much conceptual baggage, much of it vague, some of it contradictory. II. Definitions and Review of Classical Concepts of Culture According to Adler (in Blunt 1992: 189), ‘the cultural orientation of a society reflects the complex interaction of the values, attitudes, and behaviors displayed by its members’. Schein (1985:9) defined culture as ‘the pattern of basic assumptions-invented, discovered, or developed by a given group as it learns to cope with its problems of external adaptation and internal integration-that has worked well enough to be considered valid and, therefore, to be taught to new members as the correct way to perceive, think, and feel in relation to those problems’. Allaire and Firsirotu (1984) described culture as a set of ideas shared by members of a group.

Tayeb (1988:2) defines culture as ‘a set of historically evolved learned values, attitudes and meanings shared by the members of a community that influences their material and non-material ways of life. Members of the community learn these shared characteristics through different stages of the socialization processes of their life, such as a family, religion, formal education and society as a whole.

Culture consists of patterns, explicit and implicit of and for behavior acquired and transmitted by symbols, constituting the distinctive achievements of human groups, including their embodiments in artifacts; the essential core of culture consists of traditional (i.e. historically derived and selected) ideas


and especially their attached values; another systems may on the one hand, be considered as product of action; on the other, as conditioning elements of future action.

Cultures, from the anthropological viewpoint, exist to alleviate anxiety, to control the uncontrollable, to bring predictability to the uncertainty, and to clarify the ambiguous. Barley (1991) defines culture as the product of sharing socially constructed systems of meaning that allow members to make sense of their immediate, and perhaps not so immediate, environment that bring about clarity into employees’ working lives.

From a more dynamic perspective, culture is conceived as being made up of relations, rather than as stable suggests that every individual embody unique combination of personal, cultural systems of form and substance (Haastrup, 1996). This implies that national cultures, corporate cultures or professional cultures, for example, are seen as symbolic practices that only come into existence in relation to, and in contrast with, other cultural communities. People’s cultural identity constructions and their social organizations of meaning are, in other words, contextual (Fog Olwig and Haastrup, 1997). This relational approach to culture and the idea of culture and social experiences and thus that ultimately any individual behavior in the workplace becomes cultural. Management and Leadership Theories There are as many definitions of leadership as there are theorists who have attempted to define the terms. Leadership can be defined as an interpersonal relationship in which the actions or activities of groups are stimulated, controlled and directed by someone. Zaleznik (1990) suggests that leaders, but not managers, are charismatic and can create a sense of excitement and purpose in their followers. A manager's role is to bring order and consistency through planning, budgeting, and controlling. Leadership, on the other hand, is aimed at producing movement and change (Kotter 1990, 1996) in people and organizations.

This research study does not dwell on the distinction between the two concepts but uses the term manager or leader interchangeably to indicate a managerial leader. Leadership Theory Several writers have considered why some leadership styles frequently produce better results than others, yet almost all of them agree that the styles used should depend not only on the leader but also the led. Among eminent authors of seminal research in the area of leadership theory are Rensis Likert (1967); Fred Fiedler (1967); Robert Blake and Jane Mouton(1964).

In this study, all leadership theories will not be reviewed; rather, there will be specific focus on Likert’s (1967) four styles of leadership: (1) exploitative-authoritative, (2) benevolent-authoritative, (3) consultative, and (4) participative, which frequently found in the practice of leadership in Africa. Investigation of these leadership styles make it possible to show whether this classical Western management theory can be applied in African societies. This is not to say that other leadership theories cannot be put through the same review as the selected but chosen one is the most frequently applied by managers in Africa. III. Likert’s Leadership Theory A pioneer in theory formulation, Likert developed a four-style approach to leadership ranging from a very flexible and democratic style to one that is highly autocratic.

Likert’s theory focuses on styles of management in which leadership is conceived to be the main component. His questionnaire (Likert 1967, pp. 3-10) has ‘leadership processes’ as one of the seven variables. The other variables are:

1. Motivational forces 2. Communications, 3. Interaction-influence


4. Decision-making 5. Goal setting 6. Control processes

“Leadership processes” is used as a classification variable yielding the four styles of management:

(1) exploitative-authoritative, (2) paternalistic-authoritative, (3) consultative, and (4) participative. These styles are based on varying degrees of trust and confidence that each exhibits towards the subordinates. The following is a brief description of the four styles.

Style 1: Exploitative-Authoritative: This style has no confidence and trust in subordinates. It relies on centralized decision making from the top of the organization. Subordinates are not involved in any important decision-making. Downward communication is the main means of transmitting information within the organization in this style.

Style 2: Paternalistic-Authoritative: The relationship between superior and subordinate in this style resembles that of master-servant. Leaders express a condescending confidence and trust towards subordinates. An informal organization may develop within this style that does not always oppose the formal organizational goals.

Style 3: Consultative: While leaders have substantial but not absolute amount of confidence in subordinates, they still prefer to maintain control over most decisions. The top-level leaders make strategic decisions. The informal organization that usually develops within the formal organization may have an ambivalent attitude towards the formal organizational goal.

Style 4: Participative: The participative leadership style is characterized by complete confidence and trust in subordinates. Decentralized decision-making differentiates this style from the other three styles. Control is decentralized throughout the organizational hierarchy. There is a great overlap between formal and informal organization often they are one and the same.

Likert (1967) deals with the functional nature of the participative style and the dysfunctional nature of the autocratic style. Likert argued that the participative style should be effective for all kinds of organizations, tasks and situations.

According to Likert, research findings, based primarily in the United States, support his argument that leaders who use the participative style of leadership were more effective, not only in terms of achievement of production goals but in maintaining high worker’s morale within and between departments in an organization, than were leaders who use a style leaning more towards autocratic. IV. Critical Reflections on Managerial Leadership Discourse The theoretical developments that emanate from management discourse affect the scientific research on leadership and motivation in organizations. In chronological order, it is possible to distinguish four different approaches. Each has its own focus and preference for a method of research (cf. Bryman 1996).

According to Koot (2002) Firstly, there is the approach that focuses on determining characteristics of leadership, which dominated research up till the end of the 1940s. Researchers working from this perspective primarily looked at what distinguishes those who are leaders from those who are not. They usually depart from the assumption that good leaders are born and that the right ones simply have to be selected. Research within this approach focuses on three categories of characteristics of physical, capacity (intelligence, lingual skills, etc), and personality traits (such as how conservative someone is, whether he is introvert or extrovert and the amount of self-confidence he has). At the end of the 1940s, evaluations of what the research on characteristics of leadership had implied (for example Stodgill, 1948) showed that hardly any durable conclusions could be drawn from them. This led to a strong decline in the interest for studies on the characteristics of leadership. However, there were a few exceptions, such as Locke’s study on successful leaders (1991). He established that leaders are characterized by a sense of passion and high amounts of self-confidence and integrity.


The second type of leadership research is the style-of-leadership approach, which had its peak in the 1950s to the 1970s. The implicit idea supported by most of researchers in this approach was that leadership can be taught and that, therefore, attention must be paid to training and education. The suggestion was to determine how certain styles affect motivation, cooperation, achievements, satisfaction or commitment among employees. This was mostly established using the survey method. In practice, this meant that questionnaires’ regarding a leader’s behavior was handed out to his subordinates. A famous group of researchers working in this tradition was the group around Stodgill of the Ohio State University in the United States of America. The great amount of research done by this group led, among other things, to the (tentative) conclusion that a combination of ‘initiative-structuring style’ and a ‘companionable style’ can best guarantee that an organization will reach its goals. In the latter style, the leader must primarily be concerned about and win the trust of the employees, in the former the leader indicates precisely what his employees should do. At a later stage, serious doubts arose regarding suggested generalization (in time, place and space) of this perspective.

Consecutively, the contingency approach emerged and became particularly popular in the 1970s. The researchers studying leadership from this point of view, aimed to further specify the relationships that had been found earlier between leadership styles and independence variables, such as satisfaction, achievement, and cooperation with employees. They did this by elaborating on situational factors. Relationships were elaborated according to the type of organization vis-à-vis the extent to which a leader exercises control in a certain situation. One author that became especially famous for his extensive and systematic research on the relationship between variables task structure, the relationship between leaders and their employees and the extent of control held by a leader in a certain situation was Fiedler (1967). He collected more and more ‘evidence’ supporting the assumption that task-oriented leaders are most effective in situations with both a very large and a very small amount of control, whereas relationship—oriented leaders are most effective in moderately controlled situations. His data collection method was a refined version of the one used by the Ohio-researchers. He also used the survey method, but the questions he asked were quite indirect and the amount of possible answers were relatively large. This gave other authors the chance to criticize his conclusion and method. Researchers after him would always be able to ‘prove’ that relationships are slightly different in such and such situation. This effect is rather obvious, considering the fact that the contingency approach assumes that the effectiveness of leadership and the type of organization depend on the situation and that universal statements are almost impossible to make. Remarkable is that hardly any of the contingency researchers seems to be aware of the contradiction that lies within this: they depart from a situational approach and end up acting from a universalistic perspective seemingly without realizing it.

The increasing amount of criticism on the work of authors like Fiedler according to Koot (ibid) also caused a decline in the interest in elaborating on the relationships between already existing types of leadership styles and effectiveness. This decline in interest coincided with growing doubts regarding the usefulness of the rational-instrumental approach of management and organization. Simultaneously, new ideas on leadership developed into a movement that is often called the new leadership approach. At the heart of this approach lies the idea that creating an organizational identity and common patterns of meaning, which is referred to as corporate culture in the jargon, is an important task for leaders. As managers of meaning, leading managers are responsible for formulating mission statements and the corporate values that goes along with it. This cultural approach to management and organization arose due to the differences between the success of the Japanese and western economies, which was apparent at that time and established by many authors. Peter and Waterman’s (1982) book: In Search of Excellence, in which an analysis of the fifty successful American organizations showed that organizational culture is the most important factor determining success or failure in, no time led to a hype of attention to culture and the new leadership approach among organizational experts and managers.

This hype caused authors and researchers attention to be drawn away from the middle management levels. Furthermore, the trend of success stories created a situation, in which there was hardly any


evaluation of organizational change projects that failed and of the behavior of leaders with poor achievements. Related to this was the tendency to produce normative texts on management leadership and organization, which often lacked any type of systematic empirical support (cf. Watson & Harris, 1999; Knights & Wilmott, 1999). Soon the majority of publication consisted of prescriptive and quite abstract texts, which hardly paid any attention to the daily reality of the average manager.

Furthermore the emphasis was placed on models that were supposedly applicable to every situation, regardless of time and space and thus without any distinction concerning regional or national culture, branches and the profit or non-profit sector. At first glance, this seems surprising, as the supporters of the contingency approach had already criticized authors like Fayol for their ‘one best way’ strategies. One can wonder why people continue to hold on to the idea that culture can be ‘made’ and that reality can be controlled. The few evaluations that have taken place show that much of the planned cultural change has ended up in fiasco (see among others Koot and Hogema, 1990; Van der Leo and Giljam, 1995; Van Aken and Srikwerda, 1997; and Hope and Hendry, 1996). The theoretical question here is: are there certain basic assumptions upon which the dominant western theoretical discourse on management and organization is based, of which people are hardly aware, but that are nonetheless quite resilient? For answers to this and other questions, it is pivotal to do more in-depth analysis of the characteristics of the prevailing ways of thinking, speaking and writing on managerial leadership in non-western cultures. V. The Research Settings This survey study was carried out in five different European and African countries. The European countries are France, Italy, Netherlands and Scotland, while the African country involved in the research is Nigeria. VI. Research Methodology The research aimed at comparing management and employee’s perception, attitudes and values across the five countries of France, Italy Netherlands, Scotland, and Nigeria. For this purpose, written survey questionnaire was used. The Research and Development Unit of Euro-African Management Research Centre based in Maastricht developed the survey questionnaire. The questionnaire used as a base the 'Value Survey Model' developed by Hofstede (1982) for cross-national comparison of work-related values, and the Cross-Cultural Management Survey Model developed by Iguisi (1993) for cross cultural comparison on cultures and appropriate management. The questionnaire tried to obtain a fair representation of the opinions of two categories of respondents: Managers (everybody leading the work of others), non-managers (higher educated employees). The questionnaire contains items about the manager's work-related values and perceptions. Only the questions found significantly relevant for the understanding of the effect of culture on managerial practices are reported and examined in this study. The rationale for this research approach was that they reflect the managers and employees point of view, and thus should be able to reveal the common and uncommon managerial characteristics and the local culture. A focus on culture in the anthropological sense of the term is characteristics of E-AMARC research approach. In analyzing my data, I have treated the ordinal data as "quasi-interval" answers. This permits me to use mean scores and percentages of a particular question for a variety of statistical treatments.


VII. The Research Population Table 1 shows a number of relevant demographic data from the groups of respondents across the five countries.

In order to correct for the imbalance in the composition of the samples among managers (man.) and non-managers (nman.), I will always base the job-groups and country-wide comparisons on [man + nman] / 2, i.e. on a straight mean of the scores for the two groups, given equal weight to each. Table 1: Demographic Data

Table 1 allows a number of conclusions that should be taken into consideration in interpreting my data: VIII. Managerial Leadership From the review of managerial leadership theories, a description was taken of Likert four types of managers/leaders (Likert 4 management/leadership styles). The descriptions are 1. autocratic manager, 2. paternalistic manager, 3. consultative manager, and 4. democratic manager. The VSM developed by Hofstede (1982) asked for the preferred type and the actual type to which one's manager most closely corresponds. In the Cross-Culture Management Survey questionnaire (Iguisi, 1993), a new question was added about the most rejected type of manager "which one of the four types of managers would you strongly prefer not to work under?"


Table 2: Types of managers preferred, actual and rejected across cultures

Styles France Italy Scotland Netherlands Nigeria pre act rej pre act rej pre act rej pre act rej pre act rej % % % % % % % % % % % % % % % autocratic 0 35 76 5 28 66 10 35 52 0 12 71 7 31 60 paternalistic 24 20 6 30 34 1 27 31 4 24 36 0 40 39 3 consultative 49 20 6 48 32 11 52 25 6 56 39 0 35 15 18 democratic 27 16 12 17 6 22 11 9 38 18 9 19 18 14 19

The results in Table 2 have been compared among the five countries. The autocratic manager is more than 60 percent rejected by the five countries' respondents. The consultative manager is strongly preferred by the Dutch and Scottish respondents, while moderately preferred by the French and Italian respondents. The Nigerians on the other hand moderately preferred to work under the paternalistic type of managers. When I analyzed for the manager that the respondents would strongly preferred not to work under (rejected), I found that half of the Scottish respondents strongly rejected the democratic type of manager. For about one in three Italians, Dutch, Nigerians and the French, the democratic manager is the most often rejected type. Generally, the democratic manager is the more often rejected among the five countries respondents than preferred. IX. Discussions and Conclusions This research project describes the results of a survey study of five countries; France, Italy, Scotland, Netherlands and Nigeria.

This research was based on leadership styles developed by Likert. It is necessary to point out that the four leadership styles described do not necessarily provide exhaustive leadership styles but as I already pointed out, the four styles are useful insofar as the vast majority of people are able to express preference for one of the four styles. To quote Sandler and Hofstede (1980), “In general, managers, who are seen as exhibiting a distinctive style of leadership, are also considered more effective in promoting confidence and satisfaction among employees than managers who are not seen as having a distinctive style”.

The results presented in Tables 2 confirmed that the cultures of France, Italy, Scotland and Netherlands as measured by the work values and desires of the respondents are somehow different.

With managerial leadership, the empirical evidence presented in 2 reveals that the European respondents strongly attributes the most positive characteristics to a manager who is seen as consultative, while the Nigerian respondents strongly attributed the most positive characteristics to a manager who is paternalistic in attitude. Likert’s advocate for a participative management style being the most effective in all situations has no validation in this study and others (Iguisi, 1994). In general, the participative leadership style is the more often rejected than preferred among the five country respondents. The boss-subordinates relationship is a socio-cultural phenomena which bears resemblance to more fundamental relationships earlier in life; that of parents and child, and teacher and pupil. Both as bosses and subordinates, people can be expected to carry over values and norms from their early life experiences as children and school pupils. As family and school environments differ strongly among cultures, we can expect to find the traces of these differences in management styles as evidenced in my research results in Table 2.

This study thus casts serious doubts on the validity of the dominant Western universal perspectives in managerial leadership across cultures. It has been shown that cultural values pose serious challenges to African-Nigerian managers’ ability to adopt practices that can improve the effectiveness of leadership in their organizations and societies. While some of the modern management values fit very well with the African traditional values in respect to general management, new empirical research are


emerging that tends to expose African-Nigerian managers to new ways of looking at managerial leadership. This study introduces a fresh perspective and methodology into the study of managerial leadership styles in Africa and invites academics, management and organizational scholars, anthropologists and researchers to rethink the premise of their culture, management discourse and research concepts.

The debate today, therefore, is whether cultural values can become the foundational myth of modern and effective management in Africa. On the other hand, whether modern management is only possible once the African cultural values are no longer as important to Westernized African managers and elites as it is construed in most management and leadership settings in Africa.

The disparity between idealism and realism and the demands of effective managerial leadership realities in this study warns against untrammelled and unthinking transference of Western-based universal models and the practice of leadership in African cultures. The study calls for caution in the practice of Westernly developed leadership models that advocate universality in the practice of leadership and for the importation and imposition of training and education practices that draw uncritically on Western management theories and models without due sensitivity to the cultural differences and specificities of cultural values on how leadership is conceived of and practiced across cultures. "Western" models, such as American style "participative management", may not be very appropriate for managing indigenous organizations in Africa, but the problem is that there are hardly any alternative role models available yet of the African manager of the future, who would proudly retain the inheritance of his/her cultural values but at the same time is able to function in an effective organization on a competitive market. If one follows the news, this problem seems to exist in politics as much as in industry.

In conclusion, the empirical evidences that result from this research have shown that the different management theories of leadership in the form they have been developed and applied in the West may not or partially fit culturally in Africa. In theorizing managerial leadership that would be appropriate and effective in African organizations, there is the need to reflect on the cultures and assumptions of Western management theories, compare Western assumptions about social and cultural values with African cultural values and rebuild the theories or models through experimentation. In doing this, there is a need for the application of history and anthropological concepts to the field of management theorizing in Africa in order to help understand how African organizations and institutions worked in the pre-colonial era. Before the advert of colonial administration, the old African villages and towns had effective public administrative mechanism, which the village and town heads, chiefs and kings administered. The use of anthropological concepts in this context will help in the development of appropriate and effective ways of managerial leadership.

The similarities and differences among the five country respondents suggest that it make sense to study and compare cultural values, beliefs, perception and attitudes among countries, regions and sub-cultures within the same country.

X. References 1] Adler, N.J. (1986). International Dimensions of Organizational Behavior. Boston, MA: Kent

Publishing Co. 2] Argyris, C. and Schon D (1974) Theory in Practice. San Francisco: Jossey-Bass Bales R F and

F L Strodbeck (1952). Phases in Group Problem Solving. Journal of Abnormal Social Psychology, 46, 485-95.

3] Blake, R R. and Jane S (1964) Mouton. The Managerial Grid. Houston: Gulf.; 1964

4] Blunt, P.& M. Jones (1992). Managing organizations in Africa. Berlin & New York: Walter de Gruijter.

5] Decenzo D.A (2001) Fundamentals of Management: Essentials Concepts and Appication,. Prentice Hall, New Jersey.

6] DuBrin A.J (2000) Applying Psychology: Individual&Organizational Effectiveness.


7] Prentice Hall, Upper Saddle River, NJ 07458.

8] Fiedler, F.E.(1967) A theory of Leadership Effectiveness. New York: McGraw Hill.

9] Hall R.H., and W. Xu (1990) Research Notes: Run Silent, Run Deep-Cultural Influences on Organizations in the Far East. Organization Studies 11/4: 569-576.

10] Hall, D.T., and Nongaim, K.E. (1968) “An Examination of Maslow’s Need Hierarchy in an Organizational Setting,” Organizational Behavior and Human Performance, pp.12-35. February

11] Heller F A (1973) Leadership, Decision-Making and Contingency Theory. Industrial Relations, 12. 183-99.

12] Henricks, M. (1995) “Motivating Force,” Entrepreneur, pp. 70-72. December.

13] Heszberg, F. (1968) ‘One more time: how do you motivate employees?’ Havard Business Review, January-February, 53-64.

14] Hofstede G (1980) Culture's Consequences: International Differences in Work-Relate Values. Newbury Park CA: Sage.

15] Hofstede G. (1991) Cultures and Organizations: Software of the mind. London: McGraw Hill.

16] Hofstede G. and M.H. Bond (1988). `The Confucius connection: ‘From cultural roots to economic growth', In Organization Dynamics, 16, pp.4-21).

17] Iguisi O (1994). Appropriate Management in an African Culture. In: Journal of Management in Nigeria. Vol., 30 No.1, 16-24.

18] Iguisi O (1997). The Role of Culture in Appropriate Management and Indigenous Development in Africa. Paper delivered at the African seminar on Culture Dimensions to Appropriate management and Sustainable Development in Africa. UNESCO Publications, Paris

19] Iguisi O. (1998). Supporting Entrepreneurship in New Strategic Environment (SENSE): Cultural Perspectives to Entrepreneurial Effectiveness Across Cultures. EU-ADAPT/E-AMARC Research Publications.

20] Iguisi O. and Hofstede G (1993). Industrial Management in an African Culture. IRIC University of Limburg Press, Maastricht.

21] Iguisi, O (1993). Cross-Culture Management Survey. Institute for Research on Intercultural Cooperation: University of Limburg, Maastricht-NL.

22] Keaveney S.M. (1995). “Working Smarter: The Effects of Motivation Orientations on Purchasing Task Selection and Retail Buyer Performance.” In Journal of Business and Psychology, Spring 1995, p. 253.

23] Kunkel J H (1970). Society and economic growth: A behavioural perspective of social change. New York: Oxford University Press.

24] Lawler III, E.E., and Suttle, J.L. (1972) “A Causal Correlational Test of the Need Hierarchy Concept,” Organizational Behavior and Human Performance, pp. 265-87.

25] Likert R (1961). New Patterns of Management. New York: McGraw Hill,

26] Likert R (1967). The Human Organization. New York: McGraw Hill.

27] Maslow A. (1960) Motivation and Personality (New York: Harper, 1954). -(ed.) New Knowledge in Human Values. New York: Harper.

28] Marcum, J.W. (1999) “Maslow on Management,” National Productivity Review, p. 82. Summer

29] McClelland D. (1961). The Achieving Society, Princeton, NJ: van Nostrand

30] Nicholls, J (1987) ‘Leadership in organization: Meta, macro and micro’. European Management Journal 6: 16-25.

31] Stephen P. Robbins (2001). Managing Today. Prentice Hall, New Jessey

32] Tricer, H.M. and J.M. Beyer (1987), ‘Cultural leadership in organization’. Fontainebleu.

33] Vroom V. (1964) Work and Motivation, NY: John Wiley & Sons, Inc.

34] Vroom V.H. and P Yetton (1973) Leadership and Decision-making. Pittsburg: university of Pittsburg Press.

35] Zemke, R. (1998) “Maslow for a New Millennium,” Training, pp. 54-58 December.

Date post:	19-Nov-2023
Category:	Documents
Upload:	mcny
View:	0 times
Download:	0 times

Comparative Economics of Rice Production in Pakistan: A Price risk Analysis

Documents