FFB Forschungsinstitut

Freie Berufe

Fakultät II - Wirtschaft und Gesellschaft

Postanschrift:

Forschungsinstitut Freie

Berufe

ffb@uni-lueneburg.de

http://ffb.uni-

lueneburg.de

UniversitätL Ü N E B U R G

Two Scales, One Methodology -Expenditure Based Equivalence Scalesfor the United States and Germany

Joachim Merz, Thesia Garner, Timothy M. Smeeding,Jürgen Faik and David Johnson

FFB Discussion Paper No. 8

April 1994

Two Scales, One Methodology - Expenditure Based Equivalence Scales

for the United States and Germany

Joachim Men, Thesia Gamer, Timothy M. Smeeding, Jürgen Faik

and David Johnson*

Discussion Paper No. 8

April1994

ISSN 0942-2595

This paper evolves out of our current research with the hancial Support of the U.S. National Institute on Aging, Program Project #I-Pol-Ag09743-01, "The Well-Being of the Elderly in a Comparative Context. It is also available as the Cross-National Studies in Aging Program Project Paper No.8, All-University Gerontology Center, Maxwell School of Citizenship and Public AfFairs, Syracuse University, Syracuse, New York 13244-1090.

*Prof. Dr. Joachim Merz, University of Lüneburg, Department of Econornics and Social Sciences, Director of the Research Institute on Professions (!3orschungsinstitut Freie Berufe), Thesia Garner and David Johnson, U.S. Bureau of Labor Statistics, Prof. Timothy M. Smeediig, Project Investigator, Syracuse University, Maxwell School of Citizenship and Public AfFairs, Jürgen Faik, University of Franldürt.

Acknowledgements

The research for this paper is part of an National Institute on Aging Programm Project No. Pol-AG09743 at Syracuse University under the direction of Richard V. Burkhauser and Timothy M. Smeeding. The German CO-partners in this paper are Jürgen Faik, University of Frankfurt, and Joachim Merz, University of Lüneburg. The United States Partners in this paper are Thesia Garner and Pavid Johnson of the U.S. Bureau of Labor Statistics.

We would like to thank the former Sonderforschungsbereich 3 at the University of Frankfurt, financed by the German National Science Foundation, and the German Federal Statistical Office, Wiesbaden, for the opportunity to use an anonymized and reduced actual sample of the German Income and Consumption Suwey (EVS) as our German microdata base.

The fust dr& of this paper was presented to the Gerontological Society of America in November 1992. We would like to thank Patricia Ruggles and Stephen Jenkins for helpful comments on an earlier dr&. This paper was rewritten to reflect these comments. The manuscnpt was typed by Laura Griffin and Ester Gray of Syracuse University. The findings presented here do not represent the official position of the U.S. Department of Labor or the Bureau of Labor Statistics. The authors accept responsibility for all errors and omissions.

Forschungsinstitut Freie Berufe (FFB) Fachbereich Wirtschafts- und Sozialwissenschaften Universität Lüneburg Schamhorststraße 1, Gebäude 4 21335 Lüneburg Tel.: +0413 1/78 2051 Fax.: +04131/78 2059 Germany

Two Scales, One Methodology - Expenditure Based Equivalence Scales for the United States and Germany Joachim Merz, Thesia Garner, Timothy M. Smeeding, Jürgen Faik and David Johnson FFB-Discussion Paper No. 8, May 1994, ISSN 0942-2595

Abstract

Choosing an appropriate equivalence scale is a prerequisite for comparisons of economic well-being income distribution, inequality or poverty. This is true for country specific work or for cross-national comparisons. Researchers generally either use a country specific equivalence scale (social assistance, expert based, or poverty scales), or adopt a single scale for all comparison across countries. Here we follow a different approach. We use microdata to estimate equivalence scales based on a revealed preference consumption approach for West Germany and the United States. We review several approaches and rely on a complete demand system approach, which provides constant utility based equivalence scales using an extended linear expenditure system (ELES). The multiple equation expenditure system takes into account a full market basket with all its interdependencies and relative prices. Our consumption-based equivalence results are compared to alternative consumption based measures, expert based measures, and subjective based measures in use in both countries and to other scales used for cross-national comparisons.

JEL: I30, I32, D30, D31

Keywords: alternative equivalence scale, Germany, USA, distribution of income, inequality, poverty

Zusammenfassung

Die Wahl einer passenden Äquivalenzskala ist Voraussetzung für Vergleiche der Ein-kommensverteilung ökonomischer Wohlfahrt, Ungleichheit und Armut. Dies gilt vor allem für länderspezifische Analysen und/oder für länderübergreifende Vergleiche. Es werden von Forschern entweder eine jeweils landesspezifische Äquivalenzskala (Sozialhilfe, Experten basierte oder Armutsskalen) oder eine einzige Skala für einen mehrere Länder umfassenden Vergleich verwendet. Wir verfolgen hier einen unterschiedlichen Ansatz. Wir verwenden Mikrodaten um Äquivalenzskalen zu schätzen, die auf offenbarten Konsumpräferenzen für die alte Bundesrepublik und die Vereinigten Staaten basieren. Wir überprüfen verschiedene Ansätze und beziehen uns auf einen nachfragetheoretisch fundierten Systemansatz, der konstante nutzenbasierte Äquivalenzskalen über ein erweitertes lineares Ausgabensystem (ELES) liefert. Dieses multiple Ausgabegleichunmgssystem trägt einem vollen Warenkorb mit allen seinen Interdependenzen und relativen Preisen Rechnung. Unsere konsumbasierten Äquivalenzskalen werden mit alternativen Skalen, expertenbasierte und Skalen subjektiver individueller Einschätzung und anderen Skalen verglichen, die in beiden Ländern Verwendung finden und für länderübergreifende Vergleiche benutzt werden.

JEL: I30, I32, D30, D31

Schlagwörter: Alternative Äquivalenzskalen, Deutschland, USA, Einkommensverteilung, Ungleichheit, Armut

Cross-National Studies in Aging

Staff

Richard V. Burkhauser Project DUector Timothy M. Smeeding Project Investigator Sheng Zhu Project Programmer Loweii Lutz Project Programmer Detlef Jurkat Project Translator

Affiliated Researchers

Greg Duncan University of Michigan Richard Hauser University of FranldÜrt Joachim Merz University of Lüneburg Thesia Garner U.S. Bureau of Labor Statistics Barbara Boyle Torrey National Academy of Sciences

Research Staff

Deborah Bailey Research Assistant Barbara Bumca Research Assistant M a r ~ Daly Research Assistant Jack Lawrence Research Assistant Inge O'Comor Research Assistant J o b G. Poupore Research Assistant Stephen Rhody Research Assistant

Administrative and Support Staff

Margaret Austin Administrator and Budget Oficer Maruia W. Bomey Administrative Coordinator Esther Gray Administrative Secretary

Contact:

All-University Gerontology Center Maxwell School of Citizenship and Public Affairs

426 Eggers Hall, Syracuse University Syracuse, New York 13244-1090

United States of Amenca Telephone: (3 15) 443-2703

Internet: mbonney@maxwell.syr.edu Fax: (315) 443-1081

TWO SCALES, ONE METEIODOLOGY-I3XPENDlTUR.E BASED EQWALENCE SCALES FOR TBE UNITED STATES

Introduction

Equivaience scaies are used in measuring the economic resources (income, wealth)

available to persons in households of different &es and compositions. They are an integd

part of most economic weii-being comparisons involving income distribution, inequality and

poverty. Buhmann, Rainwater, Schmaus, and Smeediig (1988), has shown that different

national equivaience scales and approaches produce differences in the measurement of

household needs, and aiso in intergroup and international comparisons of poverty and income

position using different equivalence scaies. Hence, the equivalence scaie used can importantly

affect the outcome of such studies.

This paper offers an alternative to the current Set of equivalence scaies by using

microdata to estimate a Set of equivalence scales based on reveded preference for Wen

Germany and the United States, using the Same methodology for each nation. Our paper is

part of a joint United States and German research project to compare equivalence scales using

consistent methods and similar microdata from the household expenditure s w e y s of both

countries.

We review severai approaches to estimating these equivalence scales, but rely on a

complete demand system approach as specified by an extended linear expenditure system

(ELES) to provide constant utility based equivalence scaies. This multiple equation

expenditure system takes into account a fuil market basket with all its interdependencies and

relative prices.

For purposes of international comparison, most analysts rely on one equivaience scaie,

testing sensitivity of the basic results to the scaie chosen (e.g., Förster, 1993). When two

countries are used, for example the United States and Gemany, anaiysts try to use each

nation's own scale and test the sensitivity of the results by substituting one nation's scaie for

another's ,and vice versa a (e.g., Bukhauser, Duncan, and Hauser, 1991). We follow a

different approach, one which considers both the economic and ktftutional dierences of two

nations. Employing one nation's scale on another nations people would ignore differences in

the provision of "ment" goods, such as health care and education, across these two nations.

Our one methodology approach explicitly allows for national differences in consumption

weights and goods pnces to affect the resultant scaies. Moreover, comparisons of cross-

national and intra-national income dismbutions are supported by a consistent methodological

basis whereby adjustments for differences in consumption needs are determined by actual

consumption Patterns and not by expert judgements or by public opinion. We also compare

the equivalence scales enimated here to different scales irnplicit in German and United States

social policy toward the aged, in poverty measurement, and in other policy and program issues

where household size adjustments are called into play.

Our project began with Merz and Faik's (1992) estimates of several types of

consurnption based equivalence scales for Germany. These were the first such scales ever

estimated in Germany. A sirnilar research approach was used by Phipps and Garner (1992)

to cornpare the United States and Canada. The resulting equivalence scales were

"indistinguishable statistically or practically" (Phipps and Garne:, 1992, p. 18). However, the

results of our analysis do produce different equivalence scales for both nations. After

additional ad jmen t s , at the suggestion of the American Partners, we selected a Set of

methods, defuiitions and equations which were then re-estimated for both nations.

The paper is arranged as foiiows: the second section briefiy embeds our approach within

the general literature on equivalence scales. We then review revealed preference consumption

expenditure-based equivalence scaies and speciSr our Engel single equation expenditure

approach and the ELES complete demand System approach. Next we describe the microdata

bases, a sample of the most recently available West German Income and Consumption Survey

(EVS) for 1983 and eventually the U.S. Consumer Expenditure Survey (CEX) for 1983. The

empirical results are discussed and compared to other scaies in the literature in the last section.

Equivilence Scales for Welfare Comparisons: Aim, General Approaches, and Issues

Equivaience scales deflate household income according to the household type to

"calculate the relative amounts of money two different types of households require in order to

reach the same standard of living" (Muellbauer, 1977, p. 460). Given equal preference or

utility levels u for w o households and constant prices @), an equivalcnce scale (e) of a

household with composition (U) relative to that of some reference household with composition

(Uo) then is defmed as

e = c(u.p,a)Jc(u.p,ao) = Y I Y ~ , (1)

where C(.) is the cost function of reaching utility level u and y is the moncy income of the

respective household.

Economies of scale and differences in individual needs by age suggest that a per capita

measure of household income which gives equal weight to each person a crude equivalence

scale. Unadjusted household income implicitly contains yet another type of scaie-a zero

adjustment for differences in household size and composition. A behavioral based approach

to equivalence scales producc results which are more sensitive to such diierences.

Equivalence approaches can be divided into three general categories: expert, subjective,

and consumption based.' Expert b.ased equivalence scales are defined by physiological and

socio-cultural basic needs stated by some experts. Examples are "Zentimetergewichte" (height

* weight) (Engel, 1895), physiological and further basic needs (Rowntree, 1901), or basic food

expenditures (Orshansky, 1965). Subjective equivalence scales are based on individual surveys

asking either for the minimum income needed by a typical household or for the minimum

income for the respondent's own household (Kapteyn and van Praag, 1976; Kapteyn,

Kooreman and Willemse, 1988; van Praag et al., 1982; deVos and Gamer, 1991 are examples

of these).

Consurnption based equivalence scales rely on revealed preferences measuring actual

consumption expendimes of different household types. Single consumption equation methods

f k dealt with either absolute expenditures with specific adult and children goods (Rothbarth,

1943) or budget shares (Engel, 1557) where the incorne relation y/yo is given by identical

relative expenditures. Later, multiple consurnption equation rnethods encornpassing several

goods to capture different economies of scale in different goods were developed (Prais and

Houthakker, 1955, generalizing the Engel rnodel).'

More recently, the complete dernand systern approach has bein based on cost h c t i o n s

defined by rnicroeconornic theory (and its duality assumptions) and incorporaring the household

allocation problem for a full rnarket basket of expendimes (Barten, 1964; Gorman, 1976;

van der Gaag; and Smolensky, 1982). Though we ignore the issue here, recent research in this

area has also addressed the issue of inua-household allocation of resources via a household

production approach (Gronaq 1988).' . .

Revealed Preference Consumption Expenditure Based Equivalence Scales: Our Approach and

' Impiicit Choices

In this paper we concenmte on one revealed preference consumption based

method-the complete demand system ELES approach. As an expendime based model this

approach is behaviordy based and relies on actual expenditures of different household types

to estimate an equivaience scale, rather than using physiologicdy based needs (e.g., minimum

quantities of minerais or vitamins) or socidy and politicaily determined "needs." We

experimented with the Engle (1 975) single equation approach because of its use as a traditionai

reference in practice but decided to rely on the rnore general ELES approach as argued by

van der Gaag and Smolensky (1982) and Betson (1990).

Tne cornplete dernand systern approach is a rnore general approach than the Engel

approach, taking into account the consurnption of a full rnarket bashet satisfying individual

needs and preferences in a closed demand systern. Lluchos (1973) Ex~ended Linear

Expenditure System (ELES) the dernand system can be derived frorn rnavimization of a

lifetime utiliry function under a lifetime wealth constraint (Kakwani, 1980)~ The two period

intertemporal utility rnavimization problem which yields the Same results (van der Gaas and

Srnolensky, 1982) is

"th Ci ßi = 1, v,/mi > g, (goods: i = I ,..., n; periods: t = 1,2), where vti = expendime of

good i in period t, 6 = the subjective utility discount factor, T = interest rate, Öi = marginal

budget share, g, = subsistenct expenditures, and with m;:

mi = I + da 1 (4)

as commodity specific wei$Ihng factos yielding mi = 1 for the reference household with a=O.

Constrained ophka t ion yields the current penod linear expenditure vstem:

0 0 V . = a . + ß.- I

+ uia resp. I I I

I (3 v i = gi + u 1 .U + ßip(z - x j g j ) (i = 1, ..., n),

with (z - Ejg) as supernumerary income and

0 0 ai = gi - P i &. gj,

0 Pi = ßip, where p = [ ( I + 6)/(2 + 6)](2 + x)(1 + X), and

0 - 0.d. - ßiCjg,dik a i ~ - P I rk

as the elernents of the household cornposition coefficients s-vector ai, with goods i, j = I , ..., n

and household characterinics k = I, ..., s.

0 0 0 After estimation of a; , Pi and ai with Ci Pi = C. I ßip = p C . r~ P. = u the srnicnual

coefficients ßi and ,q. are given by

The dual of the utility rnavimization problern with its Stone-Geary utiliry function yields the

following cost function (Dearon and Muellbauer, 1980):

C , ) = C gi(i + d a ) + exp u - C ßiloeßy + C ß)og(i + d a ) ] . I

Finally, the aue, constant util* household equwalence scaie with respect to differentes

in household composition is given as in (1) by the hction of both households' cost functions

e = c(u,a)/c(u, a d .

The s t r u c d iduence of the household composition, gven by the s-vectos

di (i = I, ..., n), which is important to calculate the utility level, the cost function value, and

commodity specific weighting factors, can be derived via equation (6) by solving the linear

expenditure systemS

-1 A d = U:, with d k = A(,,,,)cuk, (4 k (k = 1, ...J),

0 I I with A(ij) = +(I - P i ) for i=j and -ßO OJ W . else; dk = (dlk ,..., dnk) and crk = (a ik ,..., anJ .

To caiculate ELES expendinire equivalenct scales, three important questions conctrning

the underlying approach need be answered:

Which basker of goods should we utilize? How should we incorporate household cornposition? Which resource or budget consaaint rneasure should be used?

Which Good, or \%ich Basket of Goods? Traditionally, food is the central

category fulfilling the rnost basic needs. Our food category compnses basic food, seni-luxury

food and rneals out of home. klany equivalence scales irnplicitly presented in Social

Assisrance stipends and other sirnilar minimurn consumption standards progams are based on

a basket of goods. We considered two baskets of goods: food, clothing, and shoes, and

housing and energy (goods basket I) and goods basket I plus body and health care (goods

basket 11), to describe basic standard of living for expenditures in indusbialized countries. We

selected goods basket I1 for our modeling. Food policy and goods basket I results are available

from the authors upon request

How Should We Incorporate Household Composition? In bringing demo~raphics

into the ELES model, we follow the Barten (1964) approach using a linear combination of

household composition dummies, a procedure which is comparable to the van der Gaag and

Smolensky (1982), United States approach.6 Here we can either specify a separate household

type and give each a dummy (i.e., two adults, one child household, etc.), or we can combine

a more or less homogeneous group (like the number of persons in age groups) to form a

polytomeous dummy variable. We follow the second approach in our model.'

Which Resource or Budget Constraint Bleasure Should be Used? As mentioned

above, the budget consuaint regressor might be either total expendimes or some measure of

household income capturing saving and dissaving processes. Our anaiysis will show the results

for both resource measures. T'he question of a permanent income measure to better capture the

durable expendime problem nill be discussed within the complete demand System.

Microdata: West German Income and Consumption Survey 1983 and the United States Consumer

Espenditure Survey 1986-87

Two databases were used to estimate our equivalence scales, the West German Income

and Consumption Survey (Einkommens- und Verbrauchsstichprobe, EVS) and the Unired

Stares Consumer Expendime Survey (CEX). The project staff identified a common Set of

variables and aggregation of variables for the purpose of estimating this Set of equivalence

scales.

West Germin Incume and Cousumptiou Survey 1983

The most recently available and extensive cross-section microdata base for household

economic research in Germany is the Income and Consumption S w e y (1983). Information

about this survey of more than 44,000 households (ail persons living together regardless of

mari-iage or birth Sam), with detailed expenditure and income microdatq is summ-ed in

Table 1A. To protect respondents' privacy an anonymized 96 percent random sample of the

original EVS (1983) was made avaiiable to us for our analysis, reducing our wnple to 42,752

units. This sampIe, was provided by the Sonderforschungsbereich 3 "Microanalytic Founda-

tions of Sociai Policy" at the Universities of Frankfurt and blamheim, financed by the Gernan

National Science Foundation. ' Our sample is restricted to German-headed households of four than seven members.

Household inforrnation consists of household characteristics, income, transfer and tau

information of a variety of sources. Consumption expenditures are aggegated into 20

categories. Additionally, socio-demogaphic information about each Person in every household

was also used. Variable defuiitions are presented in Table 2.4.

United Ststes Consumer Expenditure Survey

A basic description of rhe u'nired Sutes Consumer Expenditure Survey ( C E 3 dara

which underlies this report is contained in Table 1B. This survey is used to compute rhe

United Stares consumer pnce index and to collecr statistics on expenditures by various

household units. The sample used for this srudy was resmcted to consumer units participating

in four complete interviews (or interviews 2-5) in 1986-87. The sample included 5,073

consumer units. This sample was reduced to 4,972 consumer units when resuicted to units

with fewer than seven persons. For the analysis in which income was used as an explanatory

variable, the sample was W e r reduced to 4,373 by restricting it in order to calculate the

ELES System (see Garner and Blanciforti, 1993, for further information). Variable d e f ~ t i o n s

q e given in Table 2B.

TABLE 1A

WEST C E R i i W ' S iNCOME AND CONSUMFTION SURVEY (EINKOMMENS- UND VERBRAUCHSTICHPROBE, EVS) 1983

Legal basis: Federal sraristic: Bundessmistik-Gesen (BStatG) I 4 M k 1980: 196263, 1969, 1975, 1975, 1983 (1988).

Sample: Quota sample with voluntary parricipation (Euler, 1982). Observations: 0.2 percent of all private households in West Germany (ca 50,000 households

&oss)), 44,507 households fmally to analyze, reduced by 4 percent to 42,752 for our purposes.

Not included: households of foreignen, households in insrinitions, households with a monrhly net household income 2 250,000 DM. Remaining households represent c a 92 percrnt of all West German households.

Number of variables per household: 548.

P- --

Questionnaires/Methods: First interview (Gmdinrerview) January 1983:

Sociodemo,gaphics, durables available Over the year bookkerping (Haushalrungsb ücher):

Monthly (for 11 months) information (laufende Monaüanschreibungrn): one figure for an enrire respective rnonth (gathered in a four month booklet (Vieneljahresheft)):

all income figures irnporrant expendinires

One month of daily information (Feinanschreibung) by a smtified rotation procedure: daily information:

detailed smaller private consumption expendinires (open question) food and semi-luxury expendirures (open quenion)

Final interview (Schlußinterview) January 1984: Wealth (selected items) and savings.

Further Information: Euler, M. 1982. Einkommens- und Verbrauchsstichprobe (EVS) 1985, in: W i c h a f t und Starisrik 611982, pp. 453-57.

Statistisches Bundesamt (1984 and various years), Fachserie 15, Wichaftsrechnungen, E iommens - und Verbrauchsstichproben, Heft 7. Aufgaben, Methode und Durchfiihtung, Sniugan und Mainz.

Y ' Wirtschaft und Statistik1 (WiSta), various years.

2

-

TABU 1B

UNITED STATES CONSUMER EXPENDiTTJRE SURVEY INTERVIEW, 1986-87

L q a I Basis and Justifiatioß for the Consumer Expenditure Suwey (Interview and Diary): To produce weighrs for the Consurner Rice index and to present statistics on the spendiig of connimer unirs.

Suwey Sample: National probabiliry sample, matified by primary sampling uni8 (PSü's) thar comin of counries (or pam thereof), goups of counties, or independent cities. The sample of households is desiged to represent the civilian noninstitutional population and a ponion of die instimtional population living 'in gouped quarters, includiig college and univenity housing, living in the four Census regions of the United States. The sample size is targeted at approximately 5,000 interviews per quarter or every three months. About 86 percent of the eligible sample unirs participated in an interview during the period for this study. The d e s i p is such ihat each consumer mit is to be interviewed once per quarter for five consecutive quaners, and zhen rotated out of die sample.

QuestionnaireMethod: During the initial personal interview, information is collected on demopphic and family chmcterisrics and on the inventory of major durable goods of each consurner unit. The second thmugh fifh interviews use uniiom questionnaires to collect household and memDer information and expendinire data for die previous three months in general. Detailed income data such as wage and salary earnings, unemplopent cornpensation, child suppon and alimony. and employent information on each household member, are also obtained in die second and fifth interviews. Asset and liabiliry dara are also collected in the fifth interview. Ninery to 95 percent of total consumer expenditures are collected using die Interview (USDL, 1990).

Further Information: Gamet, T. and Blancifoni, L. 1992. "Household Income Repon Completeness: An Analysis of U.S. Consumer Expenditure Survey Data," ASA Proceedings of the Secrion on Economic und Business Starutics 1991, Atlanra, G.4. Alexandria, VA: American Statistical Association.

U.S. Depamnent of Labor, Bureau of Labor Statistics, 1990. Comumer Expenditure Swvt"/, 1987, Builetin 2354. Washingon, DC: U.S. Government Pnnting Office, June.

Empirical Results .

We besjn with basic cornparative statistics describing each sample used in our anaiysis.

Regression and equivalence scaie results foIIow. Findy, our resuits are compared to other

commody used equivalence scaies in both nations.

German Microdata Descriptive Measures

Descnptive information based on weighted individual data representing a total

population of 23.5 miliion households in West Germany (1983) is given in Table 2A for seven

aggegated consumption expenditure categories: food, clothirig and shoes, housing and euer=,

transportation and cornrnunication, body and health care, education and entertainment, and

personal belongings and other goods and services (our goods basket) period. Table 3A gives

figures for these categories plus household net income and a computed remainder (household

net income minus private consurnption). As shown in Table 3A, housing and energy (19.6

percent) and food (15.6 percent) amount for the largest shares of income with body and health

care (3.6 percent) the smallest share of income. The variante: measured by the coeficient of

variation, is highest wirhin body and health care, the category with the lowest average

expenditures.

It should be noted that Table 3A o d y comprises households with positive values for

income. It includes 22 percent (100-33,146147,745) of households with a negative remainder,

indicating some dissaving or use of credit. Table 4.4 presents descnptive measures for the

household rypes we use in o w regression d y s e s . The breakdowns encompass single persons

and rnarried couples with and without children. A distribution of units by household size is

also presented.

1

TAIILE 38

DESCRII"i'1VE MEASUIIES O P EXPENDI'i'URE ANI) INCOME CATECORIES IN T l lE 1986 UNITED Sl'ATES C0NSl)RlER EXPENDITUltE SURVEY

(iii ilolliirs)

Ex~ieirililiies

Pood

Clotliinc aiid slioes

Iloiisine. and eiierav

Transporialion aiid coiiiniiinicatioii

ßodv and Iieallli care

Ediicatioii aiid eiiteriai~iiiieiit

Ollier coods and services

Reiiiaintler

Haskel I

ßasket II

Iloiiseliold net incoiiie

Tolal Expenditures

Reniarks: Only cases wliere coiisiiiiier iiiiit iiel iiicoiiie > 0, expeiiditiires 1-7 > 0. coiisiiiner units are complele incoine reporlers as defined by DLS, and consuiner ui\i(s participaie in ilie survey iiiicrview quarlers 2-5. Meiiii, slaiiilüril deviatioii aiid Sliiiean are based oii weiglited sample. All calegories are based oii loial s ~ n p l e of 4,323 rel,reseiitiiig a pol~olülioii of 69,545,216. I'ercc~il of Iioiiseliolds witli iiegalive or Zero expendiiures giveii in last column.

Legend: Mem = average expeiidiliires; pcrcenl = biidget sliarcs; reiiiiiiiider = Iiouseliold iiet iiicoine nliiiiis all expendiiures vl (i=1, ..., 7).

I'erceiit

15.64

3.69

26.20

19.89

5.61

6.59

11.73

10.65

45.53

51.13

100.00

89.35

Me;iii

3,939.95

930.56

6.601.55

5,012.23

1.412.54

1.66 1.63

2.955.46

2,684.01

11.472.06

12.884.60

15.197.94

22.513.92

Sliiiiil;ird I)cvl;~tioii

2.443.00

1.059.12

4,208.08

5.851.52

1.845.59

2.542.61

3.823.80

19.598.91

6.544.87

7.182.09

25.277.75

14,864.48

Sl;iiid;ir<l

I ) fv i i i l i i~ i i

1Mc;iii

0.620

1.138

0.637

1.167

1.307

1.530

1.294

7.102

0.571

0.557

1.003

0.660

Mciliiii~

3.564.00

638.65

5.954.45

2.922.00

1,028.40

98 1 .00

1,933.15

692.66

10,496.60

11,908.30

20.283.30

19,572.40

Miii

30.00

0.00

0.00

0.00

-4.313.00

0.00

0.00

-98,490.98

769.00

932.24

50.75

1,721 .OO

Max

28.294.00

11.055.26

73,890.94

44,349.09

45.268.00

72.637.00

96,874.43

794.837.19

88,910.43

92,132.43

897.3 13.00

159,94 1.68

Skewiiess

1.719

3.282

3.242

2.189

9.042

7.490

6.186

14.924

2.1 19

2.014

10.447

1.976

Kurlosis

7.297

17.891

28.709

5.659

159.849

129.725

93.075

505.420

11.154

9.749

280.222

7.781

Noii-

posillvc (pereenl)

0.00

3.28

0.07

0.30

1.02

1.71

0.62

46.43

0.00

0.00

0.00

0.00

TAIILE 4A

DESClllPTlVE MEASURES OF 1IOUSEIIOI.D COMPOSITION IN TIlE 1983 CERMAN CONSUMPTION SURVEY

Ilouseliold Type

Shigles I. all 2. siales, agcd 18 to 64 3. feiiiales, aged 65 or over

Sliigle Purc~it, agetl 18 to 64 4. and 1 cliild 5. aiid I cliild, aged 0 to 6 6. aiid I cliild, aged 7 10 17 7. und 2 cliildren

Murrled coulilcs 8. bolli aged 18 to 65 or ovcr 9. botli agcd 18 to 64 10. botli aged 65 or over

11. oiie aged I8 to 64 1 one aged 65 or ovcr

12. aiid I cliild 13. aiid 1 cliild, aged 0 to 6 14. aiid I cliild. agetl 7 lo 17 15. and 2 cliildreii 16. aiid 3 cliildreii 17. aiid 4 cliildreii

Niiiiihcr of Iloiisclioltls

7,574 4,702 2.872

497 112 385 183

10.419 6,421 2,382 1,613

5,53 1 2.653 2,878 . 5,876 1,329

202

Nuiiiber 01 Iloiiseliolds

7.4 1594e6 3.86477e6 3.55 114e6

309,630 80,583

229,047 99.65 1

5.97656~6 3.48383e6 1.60179e6 89 1,026

1.33976~6 1.12886e6 1.2 10We6 1.90751e6 448,031 73.619

Suiiilile

Perceiit of I'ersoiis

17.72 1 1 .OO 6.72

1.16 0.26 0.90 0.43

24.37 15.03 5.57 3.77

12.94 6.2 1 6.73

13.74 3.1 1

, 0.47

Pcrceiit of Iloiiseliolds

99.66 6 1.87 37.79

3.97 0.89 3.07 1.96

83.16 5 1.27 19.0 1 12.87

59.22 28.40 30.81 6138 43.59 26.79

Poliiilatioii

Percent of Persoiis

3 1.59 16.47 15.13

1.32 0.34 0.98 0.42

25.46 14.84 6.82 3.80

9.97 4.81 5.16 8.13 1.91 0.31

Percent 01 llouseholds

99.64 51.93 . 47.7 1

4.31 1.12 3.19 2.40

83.24 48.52 22.3 1 12.41

56.4 1 27.22 29.20 58.60 39.40 23.78

1'AnLE 411

I>ESCItII'TIVE MEASURES OF IIOUSEIIO1.I) COMI'OSITION IN TIIE 1986 UNITED STATES CONSUMER EXPEN1)ITURE SURVEY

Ilouseliold Tylie

Sliigles I. all 2. males, aged 18 to 64 3. feniales, aged 65 or over

S l n ~ l e Pareiii, ngcd 18 to 64

4. aiid I cliild 5. aiid I cliild, aged 0 io 6 6. and I cliilil, aged 7 io 17 7. aiid 2 cliildren

Married coiiples 8. boili aged I8 io 65 or ovcr 9. boili agcd 18 io 64 10. i~oili aged 65 or ovcr I I. oiie agcd 18 io 64 1

one aged 65 or over

12. md 1 cliild 13. and I cliild, aged 0 to 6 14. and 1 cliild. aged 7 io 17 15. and 2 cliildt'eii 16. aiid 3 cliildreii 17. arid 4 cliildreii

Niiiiilier of Iloiisrliulils

1,07 1 615 456

97 26 7 1 82

96 1 573 276 112

334 1 69 165 399 195 44

Niiiiiber o l Ho~iseliolds

17,367,574 9,613,607 7,753,968

1,626,684 475,543 1,151,141 1,425,646

16,34 1,054 9,610,367 4,785,127 2,625,739

5,426,280 2,808,559 2,617,720 6,240,610 2,877,207 612,879

Saniple

I'ercciit 01 l'ersoiis

24.77 14.23 10.55

2.24 0.60 1.64 1.90

22.23 13.25 6.38 2.59

7.73 3.91 3.82 9.23 4.5 1 1.02

Percciit oT Iloiiseliolds

100.00 57.42 12.58

7.51 2.02 5.52

10.59

74.67 44.52 21.45

8.70

43.15 21.83 21.32 56.60 54.30 33.59

Fopulalloii

Percenl of Fersons

24.97 13.82 11.15

2.34 0.68 1.66 2.05

23.50 13.86 6.88 3.78

7.80 4.04 3.76 8.97 4.14 0.88

Fercenl o l Houseliolds

100.00 55.35 44.65

7.46 2.04 5.28

11.58

74.99 44.24 2 1.96 12.05

44.09 22.82 21.27 56.69

54.13 34.92

United States Mcrodata Descnptive Measures

Information presented here is based on approximately 69.55 miIlion United Stares

connimer units in 1986-1987 with four i n t e ~ e w s in Tables 3B and 4B closely paraiiering the

German descriptions in Tables 3A and 4A. The Same seven aggregate data categones available

for Germany have been replicated in the United States by aggregation of the more detailed

United States expenditure categories (See Table 2B). Weighted mean totd consumption

expenditures in the United States in 1986-87 e q d %23,301. Consumer unit weighted mean

net income (icome before taxes minus income and all property taxes and payrnents for Social

Security and Railroad Retirement) is $26,180 (Table 3B). Housing and energy account for the

largest percentage of income share (23.78 percent). Transportation and communication follow

with 20 percent. Food cornes in third at 15.5 percent. The mailest income share is ailocated

to clothing and shoes (3.75 percent) (Table 3B).

The weighted sample is cornposed rnostiy of rnarried couple consurner units (60.7

percent) while singles represent 23.3 percent of the total sarnple. Elderly single persons

represent 9.34 percent, while married couples, with both persons aged 65 or older, represent

7.8 percent. Married couples, with both persons aged 18 to 64 years with one to four children

represent approximately 31 percent of the sample (Table 4B).

ELES Complete Demand System Approach

Tne fuil market basket in our ELES estimates encompasses seven expendihire

categories: food, clothing and shoes, housing and energy, transportation and communication,

body and Iiraith care, education arid entertainment, and personal belongings and other goods

and services. By the Statisticai Office's definition in Germany, these expenditures describe

private consumption; cimilar c o m c t i o n was rnade for the United States.

In the theoreticai approach with equation (3, the income measure is intertemporal

wealth, z, hcorporating saving and dissaving processes. Our proxy, household net income

(rather than total expenditures = personal consumption) is incorporated in the estimates which

follow.

The ELES complete demand vstem reduced form coefficients, as in equations (3)

through (7, were estimated equation by equation ushg OLS following the Zeiher (1962)

seemingiy unrelated regression approacbg These results are shown in Tables 5A and 5B and

6A and 6B. The goodness-of-fit measured by the adjusted R~ shows a range from 8 percent

(body and health care) to housing and enera (46 percent), which is quite good for a Cross

section analysis. The seven categories encompass private consumption expenditures with a

total marginal propensity to consumer Ci bi = 0.46 indicating a high remainder marginal

propensity to consume. Since the remainder captures-besides saving and dissaving-a variety

of other e~~endin i res '~ and the so-cdled "natistical diierence" (survey errors concernhg

total expenditure minus income), a relatively high remainder value is plausible.

The ELES equivalence scales depend on a selected income level of the reference

household type [z -> vi -> u -> C()]." It is an empirical question whether the scale is

positively or negatively correlated with the income level because the ELES full market basket

approach includes both basic goods (with an income elasucity normdly < 1) which have an

opposite effect. Thus, the empirical results in Tables 7A and 7B contain differential effects

accordhg to different income Ievels.

Despite wide income ranges used, from subsistence to 1.5 times the median, the

corresponding equivalence scales do not vary greatly by income level. This result corresponds

with the findings of van der Gaag and Smolensiq (1982) based on the U.S. Consurner

Expenditure Survey 1972173 and with the United States results given in Table 7B. Dierences

TAnLE 5A

CEIlRlANY E1,ES: REGRESSION 1IESUI;I'S ACCORI>INC T 0 SOCIODEMOCRAPIIIC VARIAULES

Expeiidilure Ciitcgory

Food

Cloiliing aiid slioes

Iloiising nnd energy

Traiisporlatioii aiid coiiiiiiiiiiicnlio~i

Dody aiid healili care

Ediicaiion and eiiieriainiiie~il

Ollier goods nnd services

Reiiiarks: ( I ) t nol sig~iilicn~il al IIic 0.1 percciil Icvel; * iioi sigiiificniil iil ilie I perceiit level; iill ollier paraiiielers are sigiiificanl al llie 0.1 perceiil Ievel. (2) süniple size: ii = 42,745 (iiel incoiiie > 0; cxl>eiiililiircs 1101 rcslricled).

Legend:. a = coiisiaiii; I> = iiei iiicoiiie; iiiiiiiber of persoiis iii nge classes: C, (aged 0-6). c2 (aged 7-17), C, (aged 18-64), c4 (aged 65t); cS = funlly staius of lioiisehold Iieiid (> 18 yenrs; iiinrried = I, i~oiininrried = 0).

11'

0.386

0.285

0.455

0.160

0.076

0.196

0.200

OLS l'iiriiiiielcrs

11

1.874.126

528.963

3.163.374

797.649

281.848

1,000.9 14

1,210.206

11

0.052

0.042

0.08 1

0.093

0.034

0.059

0.IOI

C,

36.406'

-163.700

73 1.327

-78.658'

262.974

153.282

-52 1.295

c2

95 1.087

214.425

771.158

-221.692

-43.807.

274.5 12

-339.497

C,

1,489.878

202.843

577.949

572.515

- 174.425

-80.440'

-461.801

c4

1,160.554

-77.277'

724.564

-785.277

93.548'

-490.151

-47 1.898

c5

1,020.596

167.526 .

37.043'

413.006

2 18.454

-63.089'

643.002

1'Alll.E 511

UNITEI) STATES ELES: ItliCltESSION ItESULTS ACCC~RDINC T 0 S0CIOL)EMOCItAI'IIIC VARIAULES

Expeiiditure Calegory

Food

Cloiliing and slioes

liousing and energy

Traiisportaiion and coniinunicaiion

ßody and Iiealth care

Educaiion and enlertairiineiii

Oiher goods and services

Reniarks: (I) + nol sigiiiiicaiil ai ilie 0.1 perceili level; ++sigiiificaiii at .05 ievei; . t t + sigiiilicanl ai .I0 level; all oilier parainelers sigiiiticant a1 .001 level. (2) saiiiple size: ii = 4,323 (iiet iiicoiiie > 0; exlieiitliiiires iioi resiricied).

Legend: a = coiistaiii; b = nei iiicoine; iiiiiiiber of Iiersoiis iii age cl;isses: cl (agetl 0-6). cz (aged 7-17). C, (aged 18-64), c4 (aged 65.1.); = fanlily siaius of hoiiseliold Iiead (Z 18 years; iiiarried = I, iioiiiiiarried = 0).

Adj. 1t2

.449

.275

,288

,203

.097

.I99

.252

0I.S Pnrnaielers

n

1,154.49

184.84

3.509.39

852.77

409.76

308.75

833.53

Ii

0.035

0.017

0.074

0.055

0.010

0.036

0.070

ci

-53.36'

9.09'

- 133.87'

-377.66'

- 4 4 . 3 ~ ' ~

49.76'

-15.46'

e2

520.5 1

157.62

155.49'"

1 12.90'

29.34'

190.13

-77.45'

C,

766.76

152.27

405.32

1,362.33

164.65

148.33''

41.12'

c4

385.98

-36.65'

539.71

45.01'

747.20

-250.89''

77.8 1'

C,

680.14

64.95ttC

827.84

1,179.20

345.76

454.02

508.74

'VAIl1.E 6A

CEIlhIANY ELES: ItECllESSION I1ESUI.TS ACCOllDlNC T 0 IlOUSEllOLD SlZE

Expeiiillliire Ciilceory

Food

Clotliing antl slioes

Iloiisiiig aiiil energy

Trniisporluiion piid co~iiiiiiiiiicülioii

nody aiid Iiealili eure

Ediicnlion aiid eniertaiiiiiieiit

Otlitr gootls and services

~ ~ ~ ~ ~ ~ k s : ( I ) i. 1101 sig~iificiiiit iit tlie 0.1 Iicrcciii level; * iioi sig~iiii~iiiii iil IIie I perceiit level; iill ollier paraineters are significanl al Ilie 0.1 ~~erceiit Ievel. (2) siiiiiplc sk~c: 11 - 42,745 (iict iiicoiiic > 0; exlieiiililiires iioi rcsiricled).

I.egeiiJ: a = coiistaiil; b - iiel iiicoiiie; niiiiilier of persons iii age classes: C,, ..., c5: 2, ..., 6 persons (as 011 diiinmies). .

1t2

0.365

0.275

0.454

0.149

0.069

OLS I';iraiiieters

P

2,1195,300

549.4115

3.759.605

536.352

286.838

723.237

616.185

I)

0.050

0.044

0.079 ~

0.099

0.032

0.058

0.101

C I

2,370..102

268.335

732.628

826.912

-114.635' 182.398 409.590 453.955

306.345 -213.830' -555.749 -961.142

c2

3,473.498

505.348

1,534.393

1,525.575

270.713' 0.189'

-1,545.565 0.200

58.259' -16.408' P-

C)

4,221.367

7 17.076

2,095.531

1,478.343

-322.157' --

c4

5,162.813

716.480

2,710.411

1,398.042

c5

5,937.595

573.637

3,285.845

1,109.1 14

TAIILE 61)

UNlTED STATES ELES: ItECRESSION RESULTS ACCORDINC T 0 IIOUSEHOLD SIZE

Expeiitliture Cniegory

Food

Clotliing aiid slioes

Housing md energy

Transporiatioo and communicaiioii

Dody and Iiealih care

Education and entertainnient

Oilier goods and sewices

Remarks: (I) + 1101 sigiiificaiii ni tlie 0.1 I~erCeiit Ievel;++Sigiiificiili( n1 .05 Icvei;.kl-i. sigiiilicalii al .I0 level; all otlier paraineters sigiiificaiil at ,001 level. (2) saiiiple size: ii = 4,323 (iiel incoiiie > 0; expeii~liliires iiol resiricietl).

Legeiid: a = coiisiaiit; L> = nel iiicoiiie; ii~iiiiber of licrsoiis i i i nge classes: c l (aged 0-6). c2 (aged 7-17), C, (aged 18-64), cq (aged 65-1.); cS = faiiiily siatus

of liouseliold Iiead (> 18 years; iiiarried = I, ooiininrrietl = 0).

Adj. li2

0.410

0.263

0.280

0.168

0.038

0.191

0249

OLS I'nrnnieiers

n

1,7 16.890

262.3 16

3,881.633

1,550.050

848.838

3 15.460

891.964

Ii

0.0'IO

0.0 18

0.077

0.066

0.010

0.038

0.072

C!

1,033.552

96.075' ' 1,063.808

1,500.583

475.737

172.379++*

290.279"

c2

1,702.495

307.270

1,148.046

2,76 1.367

3 13.509

705.790

490.853''

C,

2,206.106

506.446

1,463.440

3,254.873

345.532

933.714

396.2 17"

c4

2,594.1 10

610.111

1,370.837

3,014.674

423.165

1,107.180

254.020t

c5

3,014.980

55 1.756

1.3 13.250

3,472.505

335.442"

596.053"

123.552'

L

TABLE 7A

ELES: EQUIVALEYCE SCliLES FOR GERtvLiri

Household Type

Sigle 1. all 2. q e d 18 to 64 3. aged 65 or over

Single Parent, Aged I8 to 64 4. and 1 child 5. and 1 child. aged 0 to 6 6. and 1 child. aged 7 to 17 7. and 2 childrcn

Married Couples 8. all 9. both sged 18 to 64 10. both ased 65 or over 11. one aged 18 to oJione ased 65 or over

Mnrried Couple, Both Aged 18 to 64 17. and 1 child 13. and 1 child aged 0 to 6 14. and 1 child. sged 7 to 17 15. and 2 children 16. and 3 children 17. and 4 children

18. 1 penon 19.2 penons 20. 3 penons 21.4 persons 22.5 pcnons 23. 6 penons

'Subsistenc: level for household types 1-17: 17.n2.17 DM.

' ~ o w c r incomc is households with incorne lcvek of 16.W m 18.W DM about 80 pm:nc of die median. 'Upper income is houscholds wich incomes 1.5 times die median or 34,OCO DM.

Subsistence'

100.00 1m.n 80.49

1 16.03 10751 120.90 129.29

148.75 154.28 109.74 131.01

167.54 159.03 172.42 180.80 194.06 207.37

100.00 147.8 1 173.92 189.98 20 1 .04 199.38

16.081.08 DM;

R e f m c e

LoweP 20J0135

100.00 10285 79.03

115.49 107.10 120.22 127.92

148.19 153.83 106.22 130.33

166.5 1 158.22 171.23 179.04 191.42 203.68

100.00 147.83 173.64 189.43 199.81 196.89

subsistence

hcome Level

Median 22,757.11

100.00 102.89 78.25

115.21 106.88 119.86 127.21

147.90 153.59 104.35 130.20

165.97 157.79 170.58 178.1 1 190.03 201.75

100.00 147.85 173.40 188.94 198.72 194.67

level for household

(in Dhl)

hIean 2494l.21

100.00 10292 77.74

115.02 106.83 119.62 126.73

147.70 153.43 103.12 179.i9

162.61 157.50 170.16 177.19 159.10 200.47

100.00 147.86 173.23 188.61 197.99 193.19

types

Upperc 3435.67

100.00 103.00 76.29

114.49 106.32 118.96 125.39

147.15 157.98 99.66

128.62

164.59 156.70 168.98 175.75 186.51 196.59

100.00 147.90 172.77 187.70 195.96 189.06

18-73:

-

7

TABLE 7B

ELES: EQUIVALENCE SCALES FOR FOR iMTED STATES

Eousehold Type

Single 1. all 2. q e d 18 to64 3. q e d 65 or over

Single Parent, Aged 18 to 64 4. and 1 child 5. and 1 child, aged 0 to 6 6. and 1 child, s e d 7 to 17 7. and 2 children

Married Couple 8. all 9. hoch aged 18 to 64 10. both q e d 63 or over 11. one ased 18 CO 64lone -ed 65 or over

Married Couple, Both Aged I8 ro 63 12. and 1 child 13. and 1 child, aged 0 CO 6 14. and 1 child, q e d 7 to 17 15. and 2 children 16. and 3 children 17. and 4 children

18. 1 penon 19.2 penons 70. ä penons 21. 4 penons 22. 5 persons 22. 6 PC~SOW

aSubsinence level for household wes b~ower is 80 percent of che median. 1 '1 'Upper is 150 percrnt of die median. 11

,!

Subsistencea

93.60 100.0 85.11

1062- 94.44

110.57 111.47

158.71 168.98 13921 154.10

171.33 163.43 179.56 176.42 182.96 185.51

100.00 148.93 178.47 196.19 199.02 19937 I

Reference Income

~ o w e r ~

93.40 100.0 84.41

10622 94.49

1 10.53 11 1.44

158.67 169.10 13839 153.93

171.47 163.62 179.63 176.54 185.03 18534

100.00 149.03 179.27 196.97 199.64 199.41

Level (ii

Median

93.03 100.0 83.12

10613 94.59

11051 11 1.37

158.59 169.3 1 136.88 153.63

171.72 163.96 179.77 176.76 183.15 185.40

100.00 149.15 1 ~ 0 1 ; 197.9 1 200.40 199.45

1-17: S7,000; subsistence level for household rypes 18-23: 9 2 0 0 . ji

US. dollan)

M a n .

92.71 100.0 81.99

10673 94.67

110.47 11132

158.52 169.49 135.57 153.36

171.94 16426 179.89 176.94 1831.5 185.45

100.00 14925 181.07 198.73 201.05 199.49

Uppere

9254 100.0 8139

10673 94.71

110.45 11119

158.48 169.59 134.86 15322

172.05 164.42 179.95 177.05 183.30 185.48

100.00 149.30 1~1.52 i 199.17 j;

201.40 199.51 ii ,

in equivalence scales by household type are discussed in the next section, where we compare

our d t s with the results pr~ented in the literature.

Comparinp Equivalence Scales

Biihmann et ai. (1988) present equivalence scaie sensitivity estimates across ten

countries using the Luxembourg Income Study GIS) data base. W~th difFerent methods they

focus on international comparisons using various rlpes of equivalence scales for each of four

types of general methods: consumption, expert program, expert statisticai, and subjective.

They present a wide range of resuits. We will demonstrate a comparison of our results to

discuss in particular differences which are due to selected methods of measurement in G m a n y -

and in the United States.

Consumption Based Results

i h e market basket ELES approach produced the Set of equivdence scdes found in

Tablz 8. Since die ELES equivalence scales do not really vary according to the income level,

a natural level to be taken is the arithmetic mean of the sample's household net income.

Comparing the German and the United States ELES approaches, we find that the household

size values (Table 8, lines 18-23) are in fairly close concordance. Overall, the mauimum

Pattern of diierences by household size done is 5 percent or less. Among various age and

child groups the differences show a consistent partern by goup; united Stares values for single

parents ( l i es 3-7) are lower, while for couples without children (lines 8-1 1) they are higher.

United States aged couples (lme 10) have the highest relative value, where the United States

results are 36 percent above the German results. i h e next step is to compare these

methodologically consistent equivaience scdes to other types employed in the literature. . .

TABLE 8

COMPARING EQUnrA.LENCE SCALE RESULTS FOR GERMANY AND THE UNITED STATES, 1986-1987

5. and 1 Child, Aged 0 to 6

9. boch aged 18 CO 64 10. borh aged 63 or over 11. one aged 18 to 641

iMarried Couples, Both Aged 18 Co 64

13. and 1 child, aged 0 to 6 14. and 1 child, aged 7 to 17 15. and 2 childrenb 16. and 3 childrenb

Aiternate Equivalence S d e s

Different types of G e r m azid United States equivaience scales are widely used in each

nation. In each case different groups have appIied each of the different methodoloJies at

different times to reach a Set of resuits. We present four speciüc sets of comparisons here:

one each for the subjective and expert stathical scaies methods, and two expert program

scales-one used primariiy for famiiies with children, the other for the elderly (Table 9).

We begin by compariug the subjective scales derived from answers to the "minimum

income question" WQ). Here f d e s were asked the minimum amount after taxes which

the government by means of a sociai security System, should provide for their household if

they had no other income. The German resuits are taken fIom a European survey as reported

in van Praag et ai. (1982). The United States results were obtained by deVos and Garner

(1991). The results are presented for each scaie and the differente between them expressed

as the ratio of the United States to the German amount. Results were produced by household

size aione for Germany, and for household size and age in the United States. At the boaom

of Table 8 (lines 1-6), we see that the United States results are consistently Iarger than the

German resuits by a factor of 19 to 56 percmt.

The second and third Set of scaies for Gemany are those implicit in their Sociai

Assistance Regulations. These scaies have also been used as German poverty line equivaience

scaies by the OECD and by German researchers (Hauser and Nouvertne, 1980; and Hauser and

Fischer, 1986). Thus, one German scaie is used for both progammatic (social assistance) and

statirticd (puverty) uses. i h e comparison scales for the United States came kom the oilicial

poverty lie for the mtisticd s d e (U.S. Bureau of the Census, 1989) and from the m t i o d . .

median benefit levels for the AFDC program in the United States for the progi-ammatic scale

(Green Book, 1992). In contxast with the subjective scaies, the expert program and statistical

C

TABLE 9

COMPARING EQUIYALENCE S C M X RESULTS: SUBJEClTVE A i EWERT Sc.4U.S FOR G m AND iJNi'ED STATES, 1983

Holls&oId Type

S i e 1. al l 2. aged 18 to & 3. aged 65 or over

Sigle Parent, Aged 18 to 64 4. and 1 child 5. and 1 child aged 0 to 6 6. and 1 child aged 1 CO 17 7. and 2 children

Married Couple 8. all 9. both aged 18 ro 64 10. boh aged 65 or over 11. one aped 18 to &/one aged 65 or over

Married couples, both aged 18 to 61 12. and 1 child

13. arid 1 chiid P:& o CO 16 14. and 1 child, aged 7 to 17 15. and 2 children 16. and 3 chiidren 17. and 4 children

Household N e

18. 1 pmon 19. 2 people 20. 3 peoplc 21.4 pcopk 2 7 5 P m p ' i ~ ~ 23. 6 pcople

Gennany

100 100 100

120

120

135

145 154 162

I00 120 135 145 154 162

Expert Program .%der, Social Subjective

United StaLes

100 I06 84

143

146

I72

195 211 193

100 143 172 195 29 1 193

Germany

100

161 145 171 222

180 181

242 1 6 252 303 564 475

100 18 1 242 303 364 425

W e s

Dine~nee

I00 106 84

119

120

127

134 I37 119

I00 119 136 I27 134 119

Assisiana:

United States

100

144 144 144 171

144 144

174 174 174 203 236 270

100 144 174 203 2ä6 270

Diiereae

100

89 99 84 78

80 80

>

72 n 69 67 65 64

100 80 72 67 65 64

e

1 TABLE 9 (CONT.)

Household Type

Sig!e 1. all 2 aged I8 U, 64 3. aged 65 or o v a

S i e Parat, &ed 18 to 64 4.andlchüd 5. aad 1 child. aged 0 to 6 6.andlchüd,agedlt017 7. and 2 chüdren

filarried Couple 8. alI 9. both aged 18 to 64 10. both aged 65 or over 11. one aged 18 to Wone ased 65 or ovez

UIarried coupls, both aged IS to 63 12. and 1 child 13. and 1 chiid. aged 0 ro 16 14. and 1 child, q e d 7 to 17 15. and 2 childru 16. and 3 chiidrcn 17. and 4 chiidru

Hourehold N e 18. 1 penon 19. 2 pmpie 20. 3 pmple 21. 4 people 22. 5 peopie 73. 6 peopls

Sam:: Gern from Van Praag. er d. (1982); (De B~11dcsmunster. 1990); median mte AFDC bcnefirs (Gretn Book. 1992); German S o d Asismc. Same as C: li UiS. Povaq linc macrix b m U.S. B u m af rhe C i ~ n i s (1989): German Sacial Retiremcnt ( d a Bu-. i!

1990); U.S. S o d Seaxiry Adminisdon (Green Book. 1992): Differcnc U (USrUGcrmany) = 100. 'I I!

I I

Expert

Germany

100 100 100

161 145 177 222

180 181 180

242 226 5 2 303 3.54 475

I00 181 242 303 5.54 425

U.S. from DeVos and G- (1991); German Social Asismw 1

ExperL m w a m -es, ~oeia l I Statirrui Scaies,

United States

100 102 94

135 135 135 158

128 132 119

158 158 158 199 235 263

100 128 157 20 1 238 268

Germany

100

133

167

167

2GO

233 267 300

100 167 200 233 267 300

Poveny:

Difierence

100 102 94

84 93 79 71

71 73 66

65 70 63 66 65 62

100 71 65 69 65 63

Retirement.

United Stater

100

M

M

150

nc

IW.

IW.

na

1GO 150 na na na na

/I

merenee 1 ,!

100

1

I

I i

1 I ! i

I I

! i

il j/

100

i 1 11

d e s for the United States are less than those for Germany for each household size. In other

words, the additional costs of extra' household members (children or aduits) beyond the Grst

member are implicitly much higher in Germany than are the additional costs in the United

States. United States vaiues are fiom 64 to 80 percent of German "progam" vaiues and 63

to 71 percent of German "statistical" vaiues. Moreover, the difference increases systematically

with household size.

The importance of these differences cannot be minimized. Essentialiy they say that if

a single Person "needs" $100 a month to be nonpoor or at a sociai assistanceIAFDC ,ouarantee,

four penons need $303 in Germany and oniy $201-S203 in the United States-a fd one-third

difference. The impact of these differences on poverty measurement or income adjusted for

differences in household size is enormous. If each counw wes its "own" scale in these cases,

the income requirements of larger size households will be consistently larger in Germany as

compared to the United Stares. This is of particular importance in studies which compare the

economic well-being of children relative to older people in the two countries. Because children

live on average in larger households than older people, the smaller the returns to scale, the

relatively worse off children will appear. The German scale wili make children appear much

worse off relative to the Arnerican scale. Because their method of calculations is not held

coustant, it is difficult to decide which-if either-of these scales is most appropriate for Cross-

national comparisons.

The final Set of scales uses the German and United States implicit scales for social

retirement. The oniy major category for retirement benefits in the United States is for aged

couples versus individuals (single, widows, or widowen, survivors or retirees per se). Here . .

the difference between the scales is oniy 10 percent, much closer than the other United

StatedGerman couplings, but still significant For instance, Smolensky et al. (1988) found that

the differente between the United States social retirernent scaie and United States poverty iine

scaie produced a 72 to 33 percent differente in poverq rates among singe elderly women in

the United States.

International Eqnivalence Scaie Approaches

The German-U.S. comparisons can be further expanded to Cover types of equivalence

scaies used in the cross-national comparative literature on poverty and income inequality.

W e the range of scaies in use is wide, four recent studies have used a Set of equivaience

scaies which are almost identicai. Represent equivalence scaies as some power Parameter for

which household size is raised, i.e., as in Buhmann et d. (1988), recent studies on poverty for

the United States (Ruggles, 1990), for the OECD (Forster, 1993), and for the European

Cornmission (Hagenaars, Zaidi and de Vos, 1992), and on income inequality for OECD

(Atkinson, et al., 1993) all used formulae which resulted in a household size coefficient of E

= .5. Table 10 compares this common internationaI equivalence scale to the ELES scales

produced in this report. The differences are large, for both United States and Germany at

larger household sizes, and for United States alone at smaller sizes.

Concludig Remarks

Our equivalence scale study based on actuai connimption expenditure rnicrodata using

the constant utility based ELES approach provides a vaiiety of interesting results with regard

to different household composition effects in both coun~es . Additional discussion and

exarnination of specific groups, e.g., older people manied couples and single mothcrs, and

children are clearly in order. The r d t s diffcr from those scaies which are impiicit in the

German social potiticai discussion and Sociai Assistance Regulations and within the United

States sociai weIfare System. Because our equivalence scales are behavioraily based on broad

COMPARING EQUIVAJXNCE SCALE RESULTS FOR GERM4iW AND THE UNiTED STTES WiTH A COMMON

INERNATIONAL S C D

Household Type

Household size

18. 1 person

19.2 people

20. 3 people

21. 4 people

22. 5 people - 23. 6 people

-Y'

100

aELES Gemiany: ar mean of reference income level based on Tables A3 and A4. b~~~~ United States: at mean of reference income level based on Tables A3 and A4. CCommon Inrernacional Scde (see text).

148

173

189

198

193

United statesb

100

Common international ScaieC

100

149

181

199

20 1

200

141

173

200

224

245

and representative samples in both nations, these results should be considered in the respective

sociai politicai discussion, in both counti-ies.

Our results lead us to believe:

1. Comparing the ELES German with United States results, differences in cross- national d t s from our ELES scales appear to be much smailer than do the results from other pairwise similar approaches.

2. Because there are differentes in each couplet of approaches, particularly for the nonregression based r e d e in Table 9, one might expect that using the "same genre" of scale, e.g., German Social Assistance for G m , and United States Social Assistance or poverty for the United States, would produce different results. Recent research by Bmkhauser et al. (1991) indicates that this is the case. In fack a large amount of the motivation for this project denves fiom the differences which such choices make in policy relevant results.

3. Substituting one "international" equivalence scale across a couplet of other approaches (Table 10) is liable not to solve many problems. Whiie some "average" scaie will be between outlier estimates, the average will still be far fiom the two poles.

4. The ELES results which apply the Same methods and market basket data are based on identical methodologies but produce slightly different results when only household size is considered, and larger differences by age and other charactenstics.

5. When compared with a common international comparisons scale, the differences are much larger than between the United States and German ELES scales produced here.

In the future we intend to compare our scaies to other international and national approaches.

But we expea that, given the alternatives, an approach which holds method constant and which

makes the underlying data as comparable as possible is the best approach to follow.

Endnotes

1. Two recent surveys on equivalence scales and their uses in inequality and poverty measurernenL Coulter, Coweil, and Jenkins (1992), and Buhmann et al. (1988), divide the topic into five categories: econometric, subjective, budget standard, social assistance, and programmatic equivalence scales. Poilak and Wales (1979) in generai discuss welfare comparisons and equivaience scales. For further recent equivalence scales o v e ~ e w s , for example, see Klein (1986, 1990), and Bradbiny (1992b).

2. The resulting identincation problem of caicularing (n) good specinc scaies and one general scale out of information from (n) available goods can be approximated by exogenously setting one scaie or by iterative solutions (Singh and Nagar, 1973; McClements, 1977).

3. See1 and Hartmeier (1990) estimate household production based equivalence scales to develop standard times for household activities.

4. The identification problem here is solved by the following Barten's (1964) approach to incorporate household characteristics in a demand system (Xakwani, 1977).

5. Since A is independent of the household characteriitics, the inverse of A, A - I , only needs to be computed once to calculate ail s vectors dk giving the household composition intluence for the entire expenditure system by D(,) = (d [,..., dJ.

6. The FELES approach by Merz (1993a), is functionalizing important ELES Parameters by socio-demographic factors. Computations with a single variable "household size" defme proportional effects, which however, should be reveaied by the analyzed behavior and not by a given functional form. Van der Gaag and Smolensky (1982), for example, take log of family size in their overall (ELES) regression specification.

7. Another generai possibility to incorporate the household composition is to run separate regressions for separate subgroups given by each household rype (see Merz, 1980, pp. 60-62.

8. The opportunity to use this unique microdata base as provided by Professor Dr. R Hauser, University of Frankfurt and by Gerrnan Federal Statisticai Office, Wiesbaden.

9. The results for food therefore have to be similar to the above Engel approach. However, because the system approach requues a subsarnple with all categories' expenditues und household income > 0, the sample size and thus the estimated coefficients will differ.

10. Consisting of voluntary social security coniributions, other income eansfers (gifts, - . automobile tax, other -es, garden renf etc.), wealth accumulation expenditures

(expenditures for sociev building deposits, shares, savings), and mortgage payments, interestc, etc.; for details, see Statistisches Bundesamt (1983).

11. The common utiiity level for the reference household as weii as to the household of specinc interest is chosen .to be u = uo with the characteristics of the reference household. Pollak and Wales (1979) and Blundeii and Lewbel (1991) stressed the point that any utiiity based equivalence scale is not imique because of the utiiity function transformation properties. Blackorby and Donaldson (1991) show how mique scales can be determined.

References

Barte.& A. 1964, "Family Composition, Prices, and Expenditure Patterns," in: P. Hart et al. (eds.), Economehic Analysis for Natratronai Economic P l m i n g , 16th Symposium of the Colston Society, London, p. 277-97.

Betson, D. 1990, Alternative Estimates of the Cost of Chilhen fiom 1980-1986 C o r n e r e e n d i t u r e Survey. Washington, DC: Office of the Assistant Secretary for Pl&g and Evaluation, U.S. Department of Health and Human Services.

Betson, D., et al. 1992, "Trade-Offs Implicit in Child Support Guidehes," Jownai of Polis. AnaZysis und ~Mamgement, W , No. 1:l-20.

Blackorby C. and D. Donaldson, 1991, "Adult-Equivalence Scaies and the Econornic Implementation of interpersonal Comparisons of Weli-Being," Working Paper, University of British Columbia, Vancouver.

Blundeli, R and A. Lewbel, 1991, "The Information Content of Equivalence Scaies," J o d of Econometrics, 49-68.

Bojer, H., 1977, "The Effect on Consumption and Household Size and Composition," European Economic Review, 9, 169-93.

Bradbuy, B., 1992% "Meaniring the Cost of Children," SPRC Discussion Papers, No. 32, The University of New South Wales, Kensington, Austraiia

Bradbury, B., 1992b, "The Theory of Family Size Equivalence Scaies: An Overview," Paper prepared for the Luxembourg Income Study CIS) Summer School, Waiferdange, Luxembourg, July 20-3 1.

Buhmann, B., Rainwater, L., Schmaus, G. and T.M. Smeeding, 1988, "Equivalence Scaies, WeU-Being. inequality, and Poverty: Sensitivity Estimates Across Ten Countries Using the Luxemboug Income Study (LIS) Database," The Review of Income and Wealth, 115-142.

Bukhauser, R, G. Duncan, R Hauser, and R Berntsen. 1991. "Wife or Frau, Women do Worse: A Comparison of Men and Women in the United States and Germany Foliowing Marirai Disruption," Demogaphy, 28, No. 3 (August): 353-360.

Der Bundesminister fik Arbeit und Soziaiordnung (ed.), 1990, Übersicht über die Soziale Sichterhei~ Born (mirneo).

Coulter, F.A.E., CoweU, F.A. and S.P. Jenkins, 1992, "Dierences in Needs and Assessment of Income Distributions," Bulletin of Economic Research, 44, No. 277-124.

Deaton, A. and J. Muelibauer, 1980, The Economics of Connuner Behavior, Cambridge.

Deato9 A. and J. MueIlbauer, 1986, "On MeanrRog Child Costs: Wib Appiications to Poor Counüies," J o d of Politicd Economy, 9 4 : n W .

DeVos, IC and T. Garner, 1991, 'An Evaluation of Subjetive Poverty Denaitions: Compariug F k d t s &om the U.S. and The Netherhds," Review of Income and Wedth, 37, No. 3:267-286.

Engel, E., 1857, "Die Productions- und Consumptionsverhältnisse des Königeichs Sachsen," Zeitschnz des Statrsrischen Bureaus des Königlich Sächsischen Ministeriums des Imern, Nr. 8 und 9.

Engel, E. 1895, "Die Lebenshaitugskosten belgischer Arbeimehmerfden n-uher und jet*" Bulletin de Z'Imiihrt InternahrnahomZ de Siufkitique, Tone E, Premiere Livraison, 3-15.

Euler, M., 1982, "Einkommens- und Verbrauchsstichprobe W S ) 1983," Wirtschq? m d Sta&ik, 611982, S. 433-37. -

Förster, M. 1993. "Comparing Poverty in 13 OECD Countries: Traditional and Synthetic Approaches," Luxembour,o Income Study Working Paper No. 100. October.

Gorman, W.M. 1976, Trick with Utility Functions, in: Artis, M. and R Nobay (eds.), Essays in Economic Analysis, Cambndge.

Green BooS 1992, U.S. Committei on Ways and bleans.

Gronau, R, 1988, "Consumption Technology and the InMarnily Distribution of Resources: Aduit Equivaience Scaies Reexamined," Journal of PolizicaZ Economy, 96:IlSi-205.

Hauser, R and I. Fischer, 1986), "The Relative Economic Status of One Parent Families in Siu Major i ndha l i s ed Countries," SB3 Discussion Paper No. 187, University of FrankfurS West Germany.

Hauser, R and U. Nouvertne, 1980), "Poverty in Rich Coun~es," SB3 Discwion Paper NO. 39, Universiry of FrankfurS West Germany.

Kakwani, N.C., 1977, "On the Estimation of Consumer Unit Scaies," The Review of Economics and Starisiics, 59507- 10.

Kakwani, N.C., 1980, "lncome Inequality and Poverty. hiethods of Eshmauon and Policy AppIications," New York

Kapteyn, A. and B.M.S. van P-, 1976, "A New Approach to the C o m c t i o n of Family . . . Equivalence Scalti," European Economic Rsview, 7313-35.

Kapteyn, A., Kooreman, P. and R Willemse, 1988, "Some bIethodological Issues in the hplementation of Subjective Poverty Lines," Jownal of Human Resources, 23222-42.

Klein, Th., 1986, ÄquivaIenzskaIen - Ein Literatumrvey, Sfb 3 - Arbeitspapier Nr. 195, F-eirn.

Klein, Th., 1990, "Zur wohifahrtsgerechten Bemessung von Sozialeinkommen," H-rtschqF und Wissenschq?, 5:274-36.

Leser, C.E.V., 1963, "Forms of Engel Functions," Econometrica, 31:694-703.

Lluch, C., 1973, "The Extended Linear Expenditure System," European Economic Review, 42 1-32.

McClements, L.D., 1977, "Equivdence Scales for Children," Jownd of Public Economics, 8:191-210.

Merz, J., 1980, Die Ausgaben privater Haushalte - Ein mikroökonome~sches Modell für die Bundesrepublik Deutschland, Frankfürt am Main/New York

Merz, J., 1983a, "FELES: The Functionaiized Extended Linear Expenditure Systern-Theory, Estimation Procedures and Application to individual Household C o m p t i o n Expenditures involving Socioeconornic and Sociodemopphic Characteristics," European Economic Review, 2339-94.

Merz, J., 1983b, "Der E iduß sozioökonomischer Größen auf die individuelle private Nach- frage nach dauerhaften Konsumgtern - Eine Anwendung der diskreten Entschei- dungsmodelle LOGIT und TOBIT," Zeitschrift &r Wirrschafr- und Sozialwissen- schaften, 103:23j-755.

Merz, J., arid Jiirgen Faik, 1997, "Eqivalencr Scales Based on Revealed Preferenct Consumption Expendinire Mi~licrodata-The Case of West Germany," Discussion Paper No. 4, Forschungsinstitut Freie Berufe, University of Liineburg, Depmen t of Economic and Social Sciences, Liineburg (Germany).

Muellbauer, J., 1977, "Testing the Barten Model of Household Composition Effects and the Cost of Children," The Economic Journal, 57:460-87.

OECD, 1982, The OECD L i s t of Social Indication, OECD, Paris.

Oahansky, M., 1965, "Counting the Poor: Another Look at the Poverty Profile," Social Senulnulq Bulletin, 285-29.

Poiiak, R and T. Waies, 1979, "Welfare Comparisons and Equivalence Scales," dmerican Economic Review, (Papers and Procerdings), 69:216-71.

P&, S.J. and ES . H o u W e r , 1955, The Analysis of Fmily Budgets, Cambridge.

Rothbarth, E., 1943, "Note on a Method of Determining Equivaient Income for FamjJia of Different Composition" In C. Madge (eci.), Wir-Time Pattern of Sang & Eiqenditure, Cambndge,

Rowntree, B.S., 1901, Poverryr A Study of Town Life, London

Scheffter, M., 1991, Haushaitsgöße und Privater Verbrauch - zum Einfiuß einer steigenden K i n d 4 auf den Privaten Verbrauch, FranlGua/Paris.

Seel, B. und E. Hartmeier, 1990, "Wohlfahrtsgerechte Einkommen und MXmurnstandards - der Beitrag des 'Standardmodells'," Hm'rtschafS und Wissenschaft, 5:U7-47.

Singh, B. aud AL., Nagar 1973, "Determination of C o r n e r Unit Scaies," Econometnca, 41.:347-55.

Statistisches Bundesamt (1984 and various other years), Fachserie 15, WirtschaftsIechnungen, Einkommens- und Verbrauchsstichproben, Heft 7, Aufgaben, Methode und Durchführung, Stuttgart und Mainz

Statistisches Bundesamt 1983, Code - Verzeichnis der Ausgaben für die Einkommens- und Verbrauchsstichprobe 1983, Wiesbaden.

~sak lo~lou , P. 1991, Estimation and Comparison of Two Simple Models of Equivaience Scales for the Cost of Children," i%e Economic Journal, 101:343-57.

U.S. Bureau of the C e m , 1989, "Poverty in the United States," Current Population Reporrs, Senes P-60, No. 163. February.

van der Gaag, J. and E. Smolensky, 1982, "True Household Equivalence Scaies and Charactenstics of the Poor in the United Siates," The Review of Income and Wealrh, 28:17-28.

van Praag, B.M.S., Hagenaars, A.J.hL and H. van Weeren, 1982, "Poverty in Europe," The Review of Income and Wealrh, 28, No. 3545-59.

Wichaf t und Statistik Cri;iSta), various years.

Zeliner, A., 1962, "An Efncient Method of Estimating Seemingly Unrelated. Regressions and Tests for Aggregation Bias," Journal of rhe American Statistical Association, 573384- 368.

Veröffentlichungen des Forschungsinstituts Freie Berufe (FFB) der Universität Lüneburg:

1 FFB-Jahresberichte FFB-Jahresbericht 1990, FFB-Jahresbericht 1991, FFB-Jahresbericht 1992.

2 FFB-Buchreihe, ISSN 0937-3373 Sahner, H.; Henmann, H., Rönnau, A. und H.-M. Trautwein, 1989, Zur Lage der Freien Bemfe 1989, Teil 1, 426 Seiten,

ISBN 3-927816-01-9, LUneburg. Sahner, H., Hemnann, H., ~ ö n n a u , Ä. und H.-M. Trautwein, 1989, Zur Lage der Freien Bemfe 1989, Teil 11, 955 Seiten,

Teil I und Teil I1 zusammen (Restauflage), ISBN 3-927816-02-7, Liineburg. Preis: DM 65, - Sahner, H., Herrmann, H., Rönnau, A. und H.-M. Trautwein, 1989, Zur Lage der Freien Bemfe 1989, Teil 111, 167

Seiten, (vergriffen), ISBN 3-927&16-04-3, LUneburg. Preis: DM 45,- Sahner, H. und F. Thiemann, 1990, Zukunft der Naturheilmittel in Europa in Gefahr? 81 Seiten,

ISBN 3-927816-06-X, Liineburg. Preis: DM 10.- Trautwein, H.-M., Donner, H., Semler, V. und J. Richter, 1991, Zur tariflichen Beriicksichtigung der Ausbildung, der

Bereitstellung von Spitzenlastreserven und der Absichemng von Beschiiftigungsrisiken bei Seelotsen mit dem Anhang Steuerliche Aspekte des tariflichen Normaleinkommens der Seelotsen, 183 Seiten, ISBN 3-927816-07-8, LUneburg. Preis: DM 35.-

Sahner, H., 1991, Freie Bemfe in der DDR und in den neuen Bundesländern, 177 Seiten, lSBN 3-927816-09-4, LUneburg. Preis: DM 48,-

Burmester, B., 1991, Ausbildungsvergleich von Sprachtherapeuten, 54 Seiten, ISBN 3-927816-10-8, Liineburg. Preis: DM 15.-

Sahner, H. und A. Rönnau, 1991, Freie Heilbemfe,und Gesundheiübe~fe in Deutschland, 653 Seiten, ISBN 3-927816- 1 1-6, Liineburg. Preis: DM 110,-

Forschungsinstitut Freie Bemfe (FFB) (Hg.), erstellt vom Zentmm zur Dokumentation für Naturheilverfahren e.V. (ZDN), 1992, 1993, Dokumentation der besonderen Therapierichtungen und natürlichen Heilweisen in Europa, Bd. I, 1. Halbband, 842 Seiten, Bd. I, 2. Halbband, 399 Seiten, Bd. 11, 590 Seiten, Bd. 111, 272 Seiten, Bd. IV, 419 Seiten, Bd. V., 1. Halbband, 706 Seiten, Bd. V., 2. Halbband, 620 Seiten, ISBN 3-88699-025-7, Liineburg (nur zu beziehen über das Zentrum zur Dokumentation für Naturheilverfahren e.V. (ZDN), HufelandstraDe 56, 45147 Essen, Tel.: 0201-74551) Preis: DM 750,-

3 FFB-Reprintreihe Merz, J., 1991, Microsimulation - A Suwey of Principles, Developments and Applications, International Journal of

Forecasting 7, S. 77-104, ISBN 0169-2070-91, North-Holland, Amsterdam, FFB-Reprint No. 1, Department of Economics and Social Science, University of Lüneburg, Liineburg.

Henmann, H., 1993, Regulation of Attorneys in Germany: Legal Framework and Actual Tendencies of Deregulation, Faure, M., Finsinger, J., Siegers, J. und R. Van den Bergh (ed.), Regulation of Profession, S. 225-245, ISBN 90- 6215-334-8, MAKLU, Antwerpen, FFB-Reprint No. 2, Department of Economics and Social Science, University of LUneburg, Liineburg.

Trautwein, H.-M. und A. Rönnau, 1993, Self-Regulation of the Medical Profession in Germany: A Survey, Faure, M., Finsinger, J., Siegers, J. und R. Van den Bergh (ed.), Regulation of Profession, S. 249-305, ISBN 90-6215-334-8, MAKLU, Antwerpen, FFB-Reprint No. 3, Department of Economics and Social Science, University of Liineburg, Liineburg.

Merz, J. and K. G. Wolff, 1993, The Shadow Economy: lllicit Work and Household Production: A Microanalysis of West Germany, Reprint FFB-No. 4, Department of Economics and Social Science, University of Lüneburg, LUneburg.

Merz, J., 1003, Wnildcl iii dcti Friicii Uiruiiin - Zum Forschungsbcitrag des Porscliuiigsikisituts Freie Berufe (FFB) Universität Liineburg, Bundesverband der Freien Berufe (Hg.), Jahrbuch 1993 - Der Freie Beruf, S. 164-173, Bonn, FFB-Reprint Nr. 5, Fachbereich Wirtschafts- und Sozialwissenschatten, Universität Liineburg, Liineburg.

. Bes~~hera~chri f t : S~hsmhorststr. 1 Vorstand: Uni".-Prof. Dr. Joachim Mon IDiiekIoil

Gebäuds 4 Univ.-Prof. Dr. Harald Hemnann 21335 Lüneburg Uni".-Prof. Dr. Heinr Sahnsi Tel.: 04131 178-2051 Fax: 04131 178-2059

4 FFB-Diskussionspapiere, ISSN 0942-2595 Forschungsinstitut Freie Bemfe, 1992, Festliche Einweihung des Forschungsinstituts Freie Bemfe am 16. Dezember

1991 im Rathaus zu Lüneburg, FFB-Diskussionspapier Nr. 1, Fachbereich Wirtschafts- und Sozialwissenschaften, Universität Liineburg, Lüneburg.

Merz, J., 1992, Time Use Dynamics in Paid Work and Household Activities of Married Women - A Panel Analysis with Household Information and Regional Labour Demand, FFB-Discussion Paper No. 2, Department of Economics and Social Science, University of Lüneburg, Lüneburg.

Merz, J. and J. Faik, 1992, Equivalence Scales Based on Revealed Preference Consumption Expenditure Microdata - The Case of West Gemany, FFB-Discussionpaper No. 3, Department of Economics and Social Science, University of Lüneburg, Lüneburg.

Merz, J., 1993, Statistik und Freie Bemfe im Rahmen einer empirischen Wirtschafts- und Sozialforschung, Antritts- vorlesung im Fachbereich Wirtschafts- und Sozialwissenschaften der Universität Lüneburg, FFB-Diskussionspapier NI. 4, Fachbereich Wirtschafts- und Sozialwissenschaften, Universität Liineburg, Lüneburg, (erschienen in: Bundesverband der Freien Bemfe (Hg.), Jahrbuch 1993 -Der Freie Bemf, S. 3 1-78, Bonn).

Merz, J., 1993, Microsimulation as an Instmment to Evaluate Economic and Social Programmes, FFB-Discussionpaper No. 5, Department of Economics and Social Science, University of Lüneburg, Liineburg.

M e q J., 1993, Market and Non-Market Labor Supply and Recent German Tax Reform Impacts - Behavioral Response in a Combined Dynamic and Static Microsimulation Model, FFB-Discussionpaper No. 6, Department of Economics and Social Science, University of Lüneburg, Liineburg.

Krickhahn, Th., 1993, Lobbyismus und Mittelstand: Zur Identifikation der Interessenverbände des Mittelstands in der Bundesrepublik Deutschland, FFB-Diskussionspapier NI. 7, Fachbereich Wirtschaffs- und Sozialwissenschaften, Universität Lüneburg, Liineburg.

Merz, J., Gamer, Th., Smeeding, T. M., Faik, J. and D. Johnson, 1994, Two Scales, One Methodology - Expenditure Based Equivalence Scales for the United States and Germany, FFB-Discussionpaper No. 8, Department of Econo- mics and Social Science, University of Liineburg, Lüneburg.

Merz, J., 1994, Microsirnulation - A Survey of Methods and Applications for Analyzing Economic and Social Policy, FFB-Discussionpaper No. 9, Department of Economics and Social Science, University of Liineburg, Lüneburg.

Merz, J., 1994, Microdata - Adjustment by the Minimum Information Loss Principle, FFB-Discussionpaper No. 10, Department of Economics and Social Science, University of Liinebuq, Lüneburg.

Merz, J., Widmaier, U. and H. Niggemann, 1994, What makes the Difference between Unsuccessful and Successful Firns in the German Mechanical Engineering Industry? A Microsirnulation Approach Using Data from the NIFA- Panel, FFB-Discussionpaper No. 11, Department of Economics and Social Science, University of Lüneburg, LUneburg.

5 FFB-Dokumentationen Merz, J., 1992, ADJUST - Ein Programmpaket zur Hochrechnung von Mikrodaten nach dem Prinzip des minimalen

Informationsverlustes, Programm-Handbuch, Dokumentation Nr. 1, Fachbereich Wittschafts- und Sozialwissen- schaften, Universität Lüneburg, Lüneburg. Available also in English as:

M e q J,, 1993, ADJUST - A Program Package for the Adjustment of Microdata by the Minimum Information Loss Principle, Program-Manual, Documentation No. le, Department of Economics and Social Science, University of Lüneburg, Lüneburg.

6 Sonstige Arbeitsberichte, ISSN 0175-7275 Sahner, H., 1988, Die Interessenverbände in der Bundesrepublik Deutschland - Ein Klassifikationssystem zu ihrer Erfas-

sung, Arbeitsbericht Nr. 41, Universität Lüneburg, ISSN 0175-7275, Lüneburg, Sahner, H., 1989, Freie Bemfe im Wandel, Arbeitsbericht NI. 59, Universität Lüneburg, ISSN 0175-7275, Lilneburg, Rönnau, A., 1989, Freie Berufe in Niedersachsen - Numerische und wirtschaftliche Entwicklung; Bedeutung als Arbeit-

geber, Arbeitsbericht Nr. 60, Universität Lüneburg, ISSN 0175-7275, Lüneburg, Matusall, V., Kremers, H. und G. Behling, 1992, Umweltdatenbanken - vom Konzept zum Schema Arbeitsbericht NI.

1 12, Universittt Lüneburg, ISSN 0176-7275, Lüneburg.

7 Sonstige Bücher Spahn, P.B., Galler, H.P., Kaiser, H., Kassella, T. und J. Merz, 1992, Mikrosimulation in der Steuerpolitik, 279 Seiten,

ISBN 3-7908-061 1-0, Berlin. Prcis: DM 85,-

Auf Anfrage werden die FFB-Jahresberichte, FFR-Reprints, FFR-niskiissinnspapiere. FFB-Dokumentritinnen lind sonstige Arbeitsbcrichte kostenlos zugesandt. Die aufgcführtcn Bücher können im Forscliuiigsiiistitut besielli werden.

AascSrij(: Forschvngsiniil~~l Freie Bemfe. Universitdt Lüneburg. Compos, Schornhorstslr I. Gebaude 4 21335 Lüneburg. Telefon: 01131/78-2051, Fax 04/31/78-2059