Tourism Analysis, Vol. 13, pp. 000–000 1083-5423/08 $60.00 + .00Printed in the USA. All rights reserved. Copyright 2008 Cognizant Comm. Corp.

www.cognizantcommunication.com

CONSTRUCTING A REGIONAL TOURISM SATELLITE ACCOUNT:

THE CASE OF QUEENSLAND

TIEN DUC PHAM, LARRY DWYER, and RAY SPURR

School of Marketing, University of New South Wales, Australia

This article presents the results of a scoping study to examine the feasibility of constructing Tour-ism Satellite Accounts (TSAs) for the regions of Queensland, Australia’s primary holiday state.The construction of a regional TSA should not be considered as the end of a process but rather asthe beginning of an ongoing process to unfold the importance of the tourism sector at a levelrelevant to policy makers. Regional TSAs may be expected to generate policy-relevant insights sothat tourism activity can be adequately nurtured and stimulated at the right time and right placefor sustainable regional economic development. This report presents findings regarding the mostsuitable methodology for the task, given the available data. The construction of a TSA at theregional level is a complex and potentially expensive task, and the aim of this study has been topropose a methodology that will have long-term application and that can be maintained and up-dated in an efficient and cost-effective manner. Although set in the context of regional Queensland,the discussion is of general interest for regional tourism stakeholders in destinations worldwide.

Key words: Tourism Satellite Account; Regional Queensland; Economic contribution

Introduction

The role of Tourism Satellite Accounts (TSAs)in providing an information base for assessing theeconomic contribution of tourism to an economyis now widely understood. Following the interna-tional agreement at the United Nations WorldTourism Organization (UNWTO) conference onTourism Statistics in Nice in 2000 on the conceptsand methodology for developing a TSA, this satel-lite account has been developed at the nationallevel in many countries (Libreros, Massieu, &

Address correspondence to Larry Dwyer, Ph.D., Qantas Professor of Travel and Tourism Economics, School of Marketing, AustralianSchool of Business, University of New South Wales, NSW 2052, Australia. Tel: +61 2 9385 2636; Fax: +61 2 9313 6337; E-mail;l.dwyer@unsw.edu.au

1

Meis, 2006). However, because tourism activitytends to be unevenly concentrated within a coun-try, coupled with the natural differences of the in-dustry structures of regions in the same country,tourism will certainly make different contributionsto the economy as a whole, or to the regionaleconomies, even if visitor expenditure volumesand patterns were similar among regions. Conse-quently, the national TSA cannot help us deter-mine the importance of tourism to different re-gions or provide any guidance to the regionaldestination management.

2 PHAM, DWYER, AND SPURR

The extensive involvement of governments intourism at a state or provincial and local level inareas such as planning, infrastructure provision,and marketing has led to a strong demand for in-formation of the kind provided in a TSA to bemade available at the state (province) as well asthe regional level (Jones, Munday, & Roberts,2003). In Australia, each state or territory has itsown governmental structure, and each encouragestourism development through its tourism depart-ment or office. The federal government noted thatif TSAs were available at the state, territory, andregional levels this would provide “a valuableinput to industry and government in terms of tour-ism’s impacts and help inform investment and pol-icy decisions by industry and government respec-tively” (Tourism Australia, 2003, p. 41).

In response to this stated need, the SustainableTourism Cooperative Research Centre (STCRC)in Australia has developed TSAs for all eight ofthe Australian states and territories (New SouthWales, Queensland, Victoria, Tasmania, SouthAustralia, Western Australia, Northern Territory,and Australian Capital Territory; see www.crctourism.com.au). For each state or territory, TSA pro-vides the following information:

• contribution of tourism to Gross State Product(GSP), total state employment, interstate trade,and international trade;

• a measure of tourism concentration in the con-ventional industries to reflect the interdepen-dency of tourism activity and the conventionalindustries;

• a detail breakdown all types of tourism contri-butions above by international, intrastate, inter-state, and outbound tourism;

• comparable measure of tourism contribution overtime, with other conventional industries, acrossstates and territories, and consistent with the na-tional aggregates.

Although TSAs have been developed for statesin several countries, the STCRCT State TSA proj-ect is believed to be the first set based on the state-level data and reconciled with the national TSA.

Notwithstanding the development of TSAs atthe Australian state and territory level, some statesare keen to further “drill down” into their econo-mies to develop TSA at the substate (regional)

level. One state that is particularly keen to developthe regional TSA is Queensland. As stated in theQueensland Tourism Strategy, “Understandingand quantifying the economic value of tourism areessential in order to attract investment. Govern-ment and industry resource allocation and invest-ment decisions must be based on detailed researchand accurate statistical data” (Tourism Queens-land, 2006, p. 17). To assist regional tourismstakeholders to plan, manage, and develop tourismat the regional level of Queensland, there is a needfor an explicit account of tourism activity at thislevel. However, this is a challenging task due tothe availability of data at the regional level. Butgiven the relative neglect of this topic in the re-search literature, the specific challenges involvedin such a task are not clear. Hence the need for ascoping study of the issues involved in producinga regional TSA for the state is inevitable. The re-sults of this scoping study have substantial impli-cations for tourism planning at the state and re-gional level both in Australia and internationally.

The article is structured with four sections: (i)a context for the scoping study, (ii) the System ofNational Account (SNA) and the role of TSA, (iii)details of the technical tasks for constructing theTSAs for regions of Queensland, and (iv) a con-clusion with some policy implications of thestudy. Although the context for discussion is re-gions within Queensland, the results are of generalinterest to any state or province worldwide thatwishes to develop TSA at the regional level.

The Context: Tourism in Queensland

As indicated in Table 1, tourism consumptionin Queensland was $19.20 billion in 2003–2004,equivalent to 25% of the total tourism consump-tion in Australia ($75.8 billion). Table 1 presentstwo broad measures of tourism activity, one fromthe demand side (consumption) and one from thesupply side (gross product), in both Queenslandand Australia. It can be seen that tourism plays amore important role in the economy of Queens-land than in the Australian as a whole. The tableillustrates clearly the important role of tourism increating relatively more employment in Queens-land than the sum of all industries, 23% versus

CONSTRUCTING A REGIONAL TSA 3

Table 1Tourism Comparison: Queensland and Australia 2003–2004

Queensland ShareQueensland Australia in Australia

Total tourism consumptiona,c ($ million) 19,200 $75,798 25%Share in total consumption 16.8% 11.8%

Total consumptionb ($ million) $114,155 $643,059 18%

Tourism GSPa/GDPc ($ million) $8,112 $35,262 23%Share in GSP/GDP 5.6% 4.2%

Total GSP and GDPb ($ million) $145,156 $841,351 17%

Persons employed in tourisma,c (’000 persons) 103.9 448.6 23%Share in total employed persons 5.5% 4.6%

Total employed personsd (’000 persons) 1,880.8 9,651.3 19%

aSpurr et al. (2007).bABS (2007a).cABS (2006).dDS data.

19%, when we compare Queensland to Australiaas a whole.

In Queensland, tourism activities are concen-trated in 14 different regions as follows.

• Gold Coast• Brisbane• Sunshine Coast• Hervey Bay/Maryborough (Fraser Coast South

Burnett)• Southern Downs• Toowoomba Golden West• Bundaberg• Fitzroy• Mackay• Whitsundays• Northern (Townsville)• Tropical North Queensland• Outback• Other

Figure 1 superimposes and presents two maps.The first is a map of Statistical Divisions (definedby the Australian Bureau of Statistics), which aredivided by red boundaries. The second map oftourism regions is color coded. As seen, the twomaps do not precisely align to each other, as tour-ism regions tend to pick out only regions withhigh tourism activity. As a result, the Other regionis used to make up for differences between thetourism region map and the actual size of Queens-land.

Queensland faces increasing competition forboth domestic and international visitation fromemerging destinations such as China, India, Af-rica, and Southeast Asia. In response, the recentlyformulated Queensland Tourism Strategy (Queens-land Tourism, 2006) lays the foundation for thecoordinated and sustainable development of tour-ism in the state, giving the industry and govern-ment the vision, goals, targets, and actions to meetthe challenges and opportunities facing the indus-try over the next 10 years. In particular, theQueensland Tourism Strategy emphasizes that thetourism industry needs to adopt a strategic ap-proach to achieving long-term sustainability throughpreserving existing markets, increasing visitor ex-penditure, enhancing product and industry devel-opment, and acknowledging the important roles ofthe different regions of the state in tourism devel-opment. An essential element in this strategy isenhancing regional tourism delivery with provi-sion of good information about the economic con-tribution of tourism to the various geographic re-gions that comprise the state.

Given the increasing pressure being placed onlocal and regional tourism structures to deliver abroad range of tourism development initiatives, aknowledge of the economic contribution to thestate from each of its tourism regions has becomea policy imperative. In particular, such knowledgeis basic to the formulation of the Destination Man-

4 PHAM, DWYER, AND SPURR

Figure 1. Tourism regions, Queensland.

agement Plans identifying the key tourism issuesand opportunities within each region.

It was in this context that a scoping study wasattempted to explore the issues involved in devel-oping some form of Tourism Satellite Accountingfor each of the regions of Queensland.

The System of National Accountsand the Role of TSA

Input–Output Tables

The System of National Accounts (SNA) is aframework that compiles and reconciles supplyand demand data for goods and services in aneconomy. At the industry level, the Input–Output

(I-O) tables reflect the most comprehensive rela-tionship between many areas of an economy.While a full explanation of the I–O framework isbeyond the scope of this article, a simple represen-tation of an I–O table is presented in Figure 2.

The I–O table contains a number of measuresto compare both the contribution of industries toan economy and the performance of an economyas a whole. For example, the most common mea-sures as previously mentioned are GVA and GDP,which can be calculated as:

GVA = P1 + P2 + P4

GDP = GVA + P3

CONSTRUCTING A REGIONAL TSA 5

What Is a TSA

A common problem with measuring the eco-nomic significance of tourism spending is that thetourism sector or tourism industry does not existexplicitly among the j industries or the c commod-ities in the I–O tables. The “tourism sector” isused to describe the consumption of tourists, whorequire a wide range of goods and services thatconstitute only a part of the consumption reportedin the I–O tables. Thus, both the terms tourism“sector” and tourism “industry” are defined veryloosely. In this report, the term industry or com-modity has been used for those that are alreadywell defined in the current SNA and the term sec-tor for tourism where it includes more than oneconventional industry or commodity.

While tourism activity generates revenue justlike other conventional industries, the SNA pro-vides no direct means to measure the contributionof tourism to an economy, let alone a comparisonwith other industries or across countries. From the

Figure 2. The structure of I–O tables. HH: household final consumption, INV: investment, GOV: gov-ernment final consumption, EXP: exports. Notation in the table is adopted from the Australian Bureauof Statistics.

I–O representation (Fig. 2), tourism consumptionis included among the final demand vectors (HHand EXP), and goods and services supplied totourism activities are included among associatedindustries such as Food and Drink, and Accommo-dation, Hotel, and Cafe. The need for an addi-tional accounting system to address these issues isinevitable.

During the last decade there has been a rapidexpansion in the development of TSA around theworld (Libreros et al., 2006). The term satelliteaccount refers to an additional account system,which is used in conjunction with the SNA tomeasure the size and significance of a sector thatis not adequately defined in the SNA. The Organi-zation for Economic Cooperation and Develop-ment (OECD), the Commission of the EuropeanCommunities EUROSTAT, the United NationsWorld Tourism Organization (UNWTO), and theUnited Nations (UN) jointly endorsed and stan-dardized the TSA framework in 2001 (Eurostat,OECD, UNWTO, & UN, 2001). They provided a

6 PHAM, DWYER, AND SPURR

Recommended Methodological Framework (RMF)and encouraged all countries to follow the RMFframework as closely as possible in the construc-tion of TSA so that the tourism sector can be com-pared across countries and over time.

The RMF is quite comprehensive, covering anextensive range of definitions and terminologies tobe adopted in the construction of a TSA. Essen-tially, the RMF focuses on bringing both demandand supply of the tourism sector together in a setof tables. The procedure to construct a TSA is infact a process of extracting the hidden componentsof tourism expenditure data from the supply anddemand sides in the I–O system and presentingthem in a format compatible with other conven-tional industries in the I–O tables so that directcomparisons can be made. Ideally, if the informa-tion is available, the TSA will contain as muchinformation on the tourism sector as those conven-tional industries and commodities from I–O ta-bles. To what extent this can be achieved dependson the availability of data in each country or econ-omy. In the last few years, the RMF has been go-ing through an updating and revising process torefine its methodologies. Noticeably, the revisingprocess pays more attention to the treatment of thepackage tours and the analysis of the analysis oftransactions developed by travel agency and touroperators (WTO, 2004).

It should be noted at this point that there hasbeen some discussion in the tourism research liter-ature of whether a TSA for small regions shouldstill be classified as a “satellite account”—indeed,a “satellite to what?” when the I–O data at thesubstate (regional) level are not readily availablefrom the official government statistical organiza-tions. The World Tourism Organization (UNWTO)would prefer that the term TSA not be used in aregional context for fear that it may degrade thecredibility of the terminology. As this study setsout only to explore the technical aspect of the con-struction of an account to measure the size andcontribution of the tourism sector at the regionallevel, we have continued to use the term TSA inthis article for simplicity. Because the concept ofTSA at the national level has been well estab-lished, using it conveys the tasks of this scopingstudy much easier.

The Role of TSA

Apart from enabling direct comparison of thetourism sector with other industries in an econ-omy, the existence of a TSA has a much moresignificant role in practice. Government agenciesinternationally have long recognized that a TSAcan serve as a tool for improving strategic man-agement, marketing, and planning for the tourismsector and enhance the effectiveness of industrypolicies. The policy relevance of a TSA can bedescribed as:

• TSA can identify the type of tourists and theirimpacts on tourism expenditure for a destina-tion. On this basis, appropriate marketing cam-paigns can be established.

• Tourist expenditure patterns can reveal whichcommodity or industry a destination needs to befocused on when there is an increase in demandby tourists, or more precisely a type of tourist.In this way, a TSA provides critical informationfor informing policy on investment and assess-ing public funding policies.

• TSA is also relevant to labor force policies in-cluding in relation to education and training.

• TSA provides a continuous flow back and forthbetween the supply and demand aspects of thetourism sector. It provides information about theinteraction of tourism with other industries inthe economy. The TSA thus becomes an impor-tant tourism forecasting and modeling tool forimpact analysis.

Who should use TSA? Obviously, governmentat all levels should be among the major users toprovide the right ambient conditions for the tour-ism sector. In response, elements of the privatesector for which tourism activity is important andtourism industry associations should be able torely on the TSA data to inform their decisions. Tothe extent that it continues to be updated using aconsistent and reliable methodology and inputs ofdata, a TSA can have considerable long term valueand applications, particularly in supporting publicpolicy making and by providing up-to-date and re-liable information to all users of tourism economicdata.

The construction of a TSA should not be con-sidered as the end of a process with the compila-

CONSTRUCTING A REGIONAL TSA 7

tion of tourism expenditure data. In fact, the con-struction of such an account should be consideredas the beginning of an ongoing process to unfoldthe importance of the tourism sector at the mostrelevant level to policy makers and provide themwith policy relevant insights so that the tourismsector can be adequately nurtured and stimulatedat the right time and right place for sustainableregional economic development.

Currently, two sets of TSA at the state level forQueensland are available for 2003–2004. OneTSA was developed by the Office of Economicand Statistical Research (OESR, 2006). Anotherhas been developed by STCRC (2007). The differ-ences between the TSA are quite minor and eitherwould form a suitable benchmark for the construc-tion of the regional TSA. STCRC (2007) has beenused as the benchmark in this scoping study, sinceit is currently being updated to take into accountof recent revisions in the national TSA (AustralianBureau of Statistics [ABS], 2006).

Constructing a Regional TSA

The construction of TSA for regions within anystate could adopt either a bottom-up or top-downapproach (Jones, 2005). The bottom-up approachbuilds a TSA of a region directly—the TSA of theState is a simple aggregation of all regions in thestate. Although it can be quite difficult, this ap-proach has the advantage of requiring one stage ofconstruction only and TSA at both levels can beachieved simultaneously. Once the TSAs of re-gions are derived, no further reconciliation or ad-justment to the regional TSA is needed to matchtargets at the state level.

But if the state TSA is readily and officiallyavailable, it is possible to take advantage of thestate TSA and apply the corresponding regionalshares to disaggregate the state TSA into TSA ofregions. This is the top-down approach. It is likelyto be considerably less resource intensive and it ismuch simpler in terms of maintaining consistencyand adding-up conditions between the state andthe regional TSA. Any further estimates indepen-dently generated by regions are also an additionalimportant source to refine the TSA of regions.

Given the availability of Queensland TSA(OESR, 2006; STCRC, 2007), a natural procedure

to construct a TSA for Queensland regions wouldbe a top-down approach. Figure 3 illustrates thetop-down approach to derive the regional TSA.The figure presents a simplified procedure althoughin practice, there is more work involved betweensteps and these steps will be discussed in detail inthe following sections.

As seen in Figure 3, there are three sources ofdata to construct TSA for regions of Queensland.These are (1) a state TSA, (2) regional share data,and (3) regional I–O tables.

State TSA

As indicated, two sets of TSA at the state levelfor Queensland are available: one from OESR(2006) and one from STCRC (2007). The stateTSA provides data on the following items:

• Tourism GSP by type of visitor• Tourism GSP for Queensland• Tourism GSP by tourist and business/other vis-

itor• Shares of factor income, tourism and non-

tourism• Share of factor income by industry• Tourism factor income by type of visitor• Tourism factor income by broad purpose of visit• Tourism employment by status• Tourism employment by industry• Tourism employment by industry by type of

visitor• Tourism full-time equivalent employment by in-

dustry and type of visitor• Tourism employment by industry and broad

purpose of visit• Visitor expenditure by visitor origin• Visitor expenditure in Queensland by expendi-

ture item

TRA Data (Regional Shares)

Data on regional shares of tourism expenditureare available from Tourism Research Australia(TRA). TRA is the main government research unitthat conducts a comprehensive survey of the tour-ism sector in Australia. These surveys are de-signed to obtain not only characteristics and travelbehaviors of domestic travelers within Australiaand overseas visitors to Australia but also expendi-

8 PHAM, DWYER, AND SPURR

Figure 3. Flow chart of the construction process.

ture data related to their travel. Data from TRA areclassified into three main groups with estimates ofexpenditure for each:

• Domestic day travelers• Domestic overnight travelers• Overseas visitors

These data can be used to derive regionalshares and, depending on the variables or the typeof expenditure in the State TSA, the correspondingshares are calculated. These shares are then ap-plied to the State TSA to obtain the tourism con-sumption at the regional level.

Regional I–O Tables

A set of I–O tables for regions of Queenslandis available from two sources. One is the Centreof Policy Studies (COPS) at Monash Universitywhile the other is the University of Queensland,which has undertaken a major economic model de-velopment in recent years including an Economet-ric Input–Output Model at the national level anda regional CGE model for all statistical divisionsof Queensland.

Because the I–O database for regions ofQueensland contain single output industries, theexpenditure of tourists by commodities can beconsidered as the outputs for the tourism purposeof the same industries. For each industry, this out-put level for the tourism purpose is used to calcu-late the proportion of tourism purpose in the totaloutput supplied by the industry. This proportion isoften defined as a tourism ratio. Once the tourismratio is available, it is assumed that the industryuses a uniform technology to produce goods (or

services) for both tourism and nontourism pur-poses. This assumption allows us to derive furtherinformation on gross value added, gross regionalproduct, and employment levels for the tourismsector.

Expenditure Allocation: Demand Side

For each region the TRA has expenditure tour-ism data covering Domestic Day Travelers, Do-mestic Overnight Travelers, and Overseas Visi-tors. Although domestic and overseas visitors mayvisit many individual locations (stops) during acomplete trip (journey), the expenditure data col-lected from the surveys are for the entire journey,not for individual stops. TRA adopts a very com-prehensive procedure to allocate expenditure toeach stop of the journey (Carter & Collins, 2005).

While in the actual TRA survey the list of ex-penditure items is relatively detailed, at the re-gional level, the sets of published expenditureitems are much more aggregated compared withwhat are available at the national as well as thestate level. This is due to TRA’s criteria for releas-ing meaningful data, particularly that with lowRelative Standard Error (RSE) (Carter & Collins,2005). Table 2 displays expenditure items avail-able for all three TRA expenditure data groups.

Unfortunately, the process of using TRA datato disaggregate the state TSA into regional expen-diture share data is not as straightforward as Fig-ure 3 might imply. The expenditure items for eachgroup do not match directly with the expenditureproducts of the state TSA (see Table 4); thus, sep-arate mapping and data treatment for each groupare inevitable.

CONSTRUCTING A REGIONAL TSA 9

While the TRA data (Table 2) refers to expen-diture items, the TSA estimates expenditure ontourism characteristic and connected products. Inthe TSA, the definition of tourism follows the dis-tinction between tourism characteristic and con-nected products, and tourism characteristic andconnected industries as defined in the AustralianTSA (ABS, 2007a, pp. 26–27). Table 3 providesa mapping between TSA products and TRA ex-penditure items.

Because of the mismatch of data classifications,the task of deriving regional expenditure is pre-sented separately for each group of travelers sothat an appropriate mapping procedure can be per-formed. To differentiate data from two sources,hereafter Items are referred to as data from TRAand Products are referred to as data from stateTSA.

Domestic Overnight Expenditure. A first prob-lem with this group is that the expenditure item“Package Tour” needs to be disaggregated into itscomponent expenditure items. The Package Touritem includes a range of goods and services pur-chased by domestic overnight visitors. These in-clude airfares, bus/coach fares, car hire, train fares,ship/boat fares, any other transport, accommoda-tion, convention fees, restaurant meals, entertain-ment charges, other hire fees, or any other expen-diture (Carter & Collins, 2005). Unpublished TRAwork (Carter & Collins, 2005) indicates that 60%of the Package Tour is comprised of airfares, withthe remaining 40% for all other items. These ratiosare applied by TRA in order to decompose the

Table 2TRA Expenditure Items by TravelingGroup Classification

Domestic InternationalDomestic Day Overnight Tourism

Transport fare Package tour Accommodation, food& package and beverages

Food and drink Accommodation OtherEntertainment Food and drinkFuel AirfaresShopping Other transportOther Entertainment

FuelShoppingOther

Table 3Mapping TSA Products and TRA Expenditure Items

TRAExpenditure

Queensland TSA Products Items

Tourism characteristic productsTravel agency and tour operator Travel agency

servicesTaxi fares Other transportLong distance passenger Airfares

transportationMotor vehicle hire and lease Other transportAccommodation services AccommodationTakeaway and restaurant meals Food and drinkShopping (including gifts and Shopping

souvenir)Tourism connected products

Local area passenger transportation Other transportRepair and maintenance of motor Other

vehiclesFuel FuelFood products Food and drinkAlcoholic beverages and other Food and drink

beveragesMotor vehicles, caravans, boats, etc. Other transportRecreational, cultural, and sports Entertainment

servicesGambling and betting services EntertainmentEducation OtherActual and imputed rent on holiday Other

housesOther tourism and services Other

Package Tour item. STCRC (2007) breaks thePackage Tour up using different shares. Neverthe-less, both sources, TRA and STCRC, display asimilarity in the way that their shares are applieduniformly to all states.

In the regional context, the cost per night ofother goods and services in the package will bemore likely to vary from region to region. Thus,using uniform shares for all regions of Queenslandwill remove the differences of regional cost of vis-iting.

Here, the Package Tour item is assumed to con-tain Travel Agency, Accommodation, Food andDrink, Airfares, Other Transport, Entertainment,and Fuel. Following the STCRC (Spurr, Dwyer,Forsyth, Ho, Pambudi, & Hoque, 2007) at the statelevel, it is assumed that the share of Travel Agencyexpenditure in domestic overnight package is ap-proximately 5%, and this is the only share applieduniformly across all regions of Queensland. Forthe remaining 95% of Package Tour, it is sug-

10 PHAM, DWYER, AND SPURR

Figure 4. Decomposition of package tour item.

gested to distribute it among other expenditureitems (TRA only) using region-specific shares ofAirfares, Accommodation, Food and Drink, OtherTransport, Entertainment, and Fuel. This way ofdistribution certainly takes into account the differ-ences between regional costs of visiting. Figure 4illustrates how the Package Tour item in the TRAdata is decomposed and added to other items in aregion.

Once the Modified TRA data are derived, thesedata are used to calculate regional shares that willbe subsequently used to disaggregate the stateTSA into regions of Queensland. Table 4 illus-trates an example of the derived regional shares.

Table 3 suggests two ways to form the regionalTSA from the state TSA values. One way is to

Table 4A Hypothetical Example of Regional ExpenditureItems (%)

StateExpenditure TRAItems Reg 1 Reg 2 Reg 3 → Total

Travel agency % % % 100%Accommodation % % % 100%Food and drink % % % 100%Airfares % % % 100%Other transport % % % 100%Fuel % % % 100%Shopping % % % 100%Entertainment % % % 100%Other % % % 100%

aggregate products in the state TSA by groups tomatch with the TRA expenditure items. Subse-quently, the regional TSA is derived by applyingthe regional shares from Table 4 to the aggregatedState TSA. The approach we favor is to carry allstate TSA products from the state level down tothe regional level. In this approach, the values ofthe state TSA products are the control totals forthe regional level. All extra information derivedfrom the regional shares using TRA data are justa means to transform the state TSA products intoregional tourism expenditure while maintainingthe control totals at the state level.

Column 2 in Table 5 contains values of all stateTSA products; columns 4 to 7 show the corre-sponding regional values that are derived byapplying the implied TRA shares to state TSAdata. These shares are taken from Table 4. As il-lustrated, after allocating all of the state TSA intoregions, the sum of regional TSA (column 8)should add up to the original TSA values at thestate level (column 2). Table 5 reflects the top-down approach that is suggested for the wholescoping study, which is in a way similar to themethod for constructing the state TSA: valuesavailable at the higher level (state) remain un-changed, with the lower level (region) shares ap-plied to disaggregate the values at the higher level(state) into values at the lower level (regions).Once more data become available over time thisapproach provides the opportunity to go back andrefine the composition of products in a group.

CONSTRUCTING A REGIONAL TSA 11

Table 5Disaggregating State TSA Down to Regional TSA Using TRA Expenditure Shares

AllQueensland TSA Products Values TRA Expenditure Items Reg1 Reg2 Reg3 → Regions

Tourism characteristic productsTravel agency and tour operator services X1 Travel agency shares x11 x12 x13 . . . X1Taxi fares X2 Other transport shares x21 x22 x23 . . . X2Long distance passenger transportation X3 Airfares shares x31 x32 x33 . . . X3Motor vehicle hire and lease X4 Other transport shares x41 x42 x43 . . . X4Accommodation services X5 Accomodation shaes x51 x52 x53 . . . X5Takeaway and restaurant meals X6 Food and drink shares x61 x62 x63 . . . X6Shopping (including gifts and souvenirs) X7 Shopping shares x71 x72 x73 . . . X7

Tourism connected productsLocal area passenger transportation X8 Other transport shares x81 x82 x83 . . . X8Repair and maintenance of motor vehicles X9 Other shares x91 x92 x93 . . . X9Fuel X10 Fuel shares x101 x102 x103 . . . X10Food products X11 Fod and drink shares x111 x112 x113 . . . X11Alcoholic beverages and other beverages X12 Food and drink shares x121 x122 x123 . . . X12Motor vehicles, caravans, boats, etc. X13 Other transport shares x131 x132 x133 . . . X13Recreational, cultural, and sports services X14 Entertainment shares x141 x142 x143 . . . X14Gambling and betting services X15 Entertainment shares x151 x152 x153 . . . X15Education X16 Other shaes x161 x162 x163 . . . X16Actual imputed rent on holiday houses X17 Other shares x171 x172 x173 . . . X17

Other tourism services X18 Other shares x181 x182 x183 . . . X18

Domestic Day Visitor Expenditure. In this sec-tion, we follow the top-down procedure that hasbeen applied in the above Domestic Overnightsection to derive the regional tourism expenditurefor the Domestic Day group. But first of all, theTransport Fares and Packages item has to be allo-cated to appropriate expenditure items. Data fromABS (2006) reveals that the nature of DomesticDay expenditure has an insignificant amount ofTravel agency, in general only 5% of the total ex-penditure of Travel agency, Taxi fares, Long dis-tance passenger transportation, Motor Vehiclehire and lease, and Local area passenger trans-portation. It is recommended that only 5% of theTransport Fare and Package item is extracted forTravel Agency. The remaining 95% is then dis-tributed as Transport in the region (Fig. 5).

These (modified TRA) regional data will beused subsequently to derive regional shares to dis-aggregate the expenditure in the Queensland TSAfor the Domestic Day group into regions of Queens-land.

International Visitor Expenditure. As seen inTable 2, TRA data only offer two broad expendi-ture items for each region of Queensland, namelyAccommodation, Food and Beverages, and Other.

Fortunately, region-specific expenditure patternscan be obtained from TRA. These patterns are pre-sented in terms of item shares in the total regionalexpenditure. These expenditure items include:

• Accommodation• Food and drink• Fuel• Transport• Shopping• Entertainment• Education• Other

Given the sum that TRA provides for interna-tional visitor expenditure of each region, it is rela-tively easy to derive all eight expenditure itemsfrom the expenditure pattern. This step is illus-trated in Figure 6.

Once all eight new TRA expenditure items arederived, the same approach to map TRA expendi-ture items to state TSA products is used to calcu-late regional expenditure data for the overseas vis-itor group. That means we apply the regionalshares to the state TSA to disaggregate the statevalues into the regional values.

12 PHAM, DWYER, AND SPURR

Figure 5. Decomposition of transport fares and packages.

Intraregional, Interregional, and InterstateTourism Expenditure Disaggregation. The twomatrices of expenditure for Domestic Overnightand Domestic Day include expenditure of touristsfrom within each region, from other regions ofQueensland, and also from other states. The struc-ture of the TRA database enables identification ofhome–destination pairs so that the number of Visi-tor Nights (VN) for Domestic Overnight and tripsfor Domestic Day can be obtained.

Using the number of VN alone to split expendi-ture of Domestic Overnight into intraregional, in-terregional and interstate tourism expenditure willnot provide an accurate system to locate expendi-ture into three sources: own region, from the restof Queensland, and interstate; it only reflects the

Figure 6. Decomposition of TRA expenditure items.

quantity side of the whole picture. The differencesin the cost of visiting are not taken into account.This implies that three nights in the SunshineCoast will cost the same as three nights in the Out-back of Queensland. In order to capture the priceeffect, the regional cost indicator of the IterativeProcedure should also be used in conjunction withthe number of VN. The regional cost indicator isavailable from the TRA data.

Let

VNij: the number of visitor nights for a touristfrom home region i to destination j,

R j: regional cost indicator of the destination region j,EXPij: expenditure in destination j from home re-

gion i.

CONSTRUCTING A REGIONAL TSA 13

EXPij = VNij * R j

when i = j, EXPij is intraregional component of re-gion j; i = other regions of Queensland, EXPij isinterregional component of region j; and i = re-gions of other states, EXPij is the interstate compo-nent of region j.

It is important to note that R j is constant in re-gion j regardless of where the visitors come from.Once EXPij is calculated, the shares can be applieduniformly to all TSA commodities in the region jto derive three sources of tourism expenditure. Asimilar approach should also be used for the Do-mestic Day. For the purpose of compiling regionalTSA, the three sources of expenditure (own re-gion, the rest of Queensland, and interstate) aresufficient.

Thus far, the three sets of tourism expenditurefor all regions, which is part of the TSA from thedemand side (consumption), have been presented.Figure 7 is a reproduction of Figure 3 indicatingwhere we are in the whole scheme. In the nextstep, these expenditures will be applied to anotherprocess to derive the supply side of the regionalTSA, for example GVA and Gross Regional Prod-uct (GRP).

Production Allocation: Supply Side

The tourism expenditure data collected by theTRA survey represent the amounts paid by the vis-itors. This means that the values of the goods and

Figure 7. Reproduction of Figure 3 indicating status in the process.

services are measured at the purchasers’ prices,which include all margins (e.g., transportationcosts) on top of the production cost (basic valueof the output) that initially are incurred by the cor-responding industries.

Before proceeding to derive the supply sidedata for the regional TSA, all margins and importshave to be removed from the consumption data.The purpose of removing these margins is to ob-tain the right magnitude of the output level so thatthrough the use of input–output ratios it is possi-ble to correctly derive the contribution of valueadded components of the industry producing tour-ism products. The margin rates are available fromthe national I–O tables. The step to calculate thetotal output level measured at basic prices is sim-ple: multiply the rates by the consumption data toget values of margins; these margin values arethen subtracted from the consumption data, the re-mainder is the basic price component.

The basic price component is the output of thecorresponding industries in a region. At this point,there are two alternatives to calculate the rest ofthe supply side data for the regional TSA. Oneapproach is to apply these regional output levelsas regional shares to the state TSA supply sidedata to calculate the corresponding regional supplydata for TSA.

However, it is not always the case that similarindustries in different regions would have thesame cost structure. Similar industries in different

14 PHAM, DWYER, AND SPURR

Figure 8. Reproduction of Figures indicating end of the process.

regions interact with the rest of the region, the restof Queensland as well as the rest of Australia indifferent ways, even though they produce verysimilar products.

The other approach is to use the regional I–Odatabase. The database contains specific cost struc-ture (input/output ratios) for each industry. Be-cause there is no empirical evidence to indicatethat the same industry would employ different

Figure 9. A cost-effective mechanism to produce regional TSA.

technologies to produce outputs for tourism andnontourism activities, it is reasonable to assumethat the input/output ratios are broadly similar forboth. Thus, for a given level of output, GVA andGRP can be derived directly.

Given the output level (the basic price compo-nent), the input/output ratios can be used to derivecompensation of employees (COE), gross operat-ing surplus (GOS), employment, commodity taxes

CONSTRUCTING A REGIONAL TSA 15

on intermediate materials used by the industry,and the net taxes on production. Employment canbe estimated by applying tourism ratios to thenumber of employed (full time equivalent) personsof the corresponding industries to calculate thenumber of people employed in the industries fortourism purpose (ABS, 2007b).

Depending on the purpose of use, care must betaken when employing TSA data as the derivationof the regional TSA is based on survey data,which are subject to sample errors. Analysis ofsample errors in TRA data is beyond the scope ofthis article. The objective here is merely to dem-onstrate how all available data can be put togetherto derive the TSA data for regions of Queensland.The authors are of similar view to Carton and Col-lins (2005), who state that they are

reluctant to lay down general rules governing useof the regional expenditure estimates. This ispartly because one size does not fit all, given theinherent variability of the estimates, but mainlybecause use of given estimates is a risk manage-ment issue, which is dependent on the circum-stances. For example, if a substantial amount ofinvestment or funding is to be allocated, then amore cautious or conservative approach would becalled for. However, for other purposes, less reli-able estimates may be acceptable. This is matterfor individual judgment, depending on the con-text. (Section 4.1)

Given the data limitations that exist even inconstruction of national TSA, clearly data at theregional level must be treated with extra caution.

Where to Now?

Because the existence of regional I–O tables iscritical to the construction of the regional TSA(Fig. 8), it is most important that the regional I–Otables be kept up to date (supply side). The tour-ism expenditure data for the demand side obtainedfrom TRA should have a consistent format tomake use of the automatic routine to produce theregional TSA. This will make sure that the processcan produce the regional TSA for 2003–2004 andthe subsequent years.

If the regional I–O tables are updated on a reg-ular basis, the majority of tasks required to con-struct the regional TSA can be undertaken with

minimal effort, and as a result, the cost of produc-ing regional TSA in subsequent years can be rela-tively cheaper. It is the lack of the regularly pub-lished regional I–O tables that presents difficultiesin the production of TSA at the regional level(Fig. 9).

Conclusion

In many destinations around the world, smallerregions below state and province level have be-come increasingly interested in developing toolsto assess the economic contribution of tourism totheir jurisdictions. One such tool is to develop aregional Tourism Satellite Account. While the la-bel may be criticized, the intent is clear—the tour-ism regional account is expected to provide thesame type of information to a region as the TSAdoes at the national level. This article has indi-cated how TSA may be constructed for regions ofQueensland, Australia’s major holiday destinationstate.

The role of the regional TSA is very importantfor its policy relevance, which is directly relatedto specific regions for marketing campaigns, in-forming policy on investment and assessing publicfunding, education, and training relevant to laborforce policies, and modeling and forecasting im-pacts of tourism on an economy. Thus, the con-struction of the regional TSA is the beginning ofan ongoing process to understand how the tourismsector can have impacts on a region and how poli-cies can help develop tourism.

References

Australian Bureau of Statistics. (2006). Australian nationalaccounts: Tourism satellite account, 2004–05. Cat No.5249.0. Retrieved from http://www.abs.gov.au/AUSSTATS/abs@.nsf/ProductsbyReleaseDate/14D3E16F68F9ABEFCA2572D70018B969?OpenDocument

Australian Bureau of Statistics. (2007a). Australian na-tional accounts: State accounts, 2006–07. Cat No.5220.0. Retrieved from http://www.abs.gov.au/AUSSTATS/abs@.nsf/DetailsPage/5220.02006-07?OpenDocument

Australian Bureau of Statistics. (2007b). Estimating state/territory tourism employment, parts 1 and 2. Uun-published report to the Department of Industry, Tourismand Resources, February, Canberra.

Carter, P., & Collins, D. (2005). Travel expenditure by do-mestic and international visitors in Australia’s regions.Canberra: Tourism Research Australia.

16 PHAM, DWYER, AND SPURR

DX Data, ABS Labour Force Statistics, extracted in April2008

Eurostat, OECD, UNWTO, & UN (2001). Tourism satelliteaccount: Recommended methodological framework. Ma-drid: UN World Tourism Organization.

Jones, C. (2005). Tourism satellite accounts: The regionalperspective. In World Tourism Organization Conference,The Tourism Satellite Account (TSA): UnderstandingTourism and Designing Strategies, Iguacu Falls, Argen-tine/Brazil/Paraguay, October 3–6.

Jones, C., Munday, A., & Roberts, A. (2003). Regional tour-ism satellite accounts: A useful policy tool? Urban Stud-ies, 40(13), 2777–2794.

Libreros, M., Massieu, A., & Meis, S. (2006). Progress intourism satellite account implementation and develop-ment. Journal of Travel Research, 45, 83–91.

Office of Economic and Statistical Research. (2002). Thecontribution of international and domestic visitor expen-diture to the Queensland regional economies: 1998–1999. Prepared for Tourism Queensland and the Depart-ment of tourism, Racing and Fair Trading QueenslandGovernment, Treasury Department, Brisbane.

Spurr, R., Dwyer, L., Forsyth, P., Ho, T.V., Pambudi, D., &Hoque, S. (2007). Tourism satellite account for Queens-land 2003–2004. Gold Coast: Sustainable Tourism Co-operative Research Centre.

Tourism Queensland. (2006, November). Queensland tour-ism strategy: A ten year vision for sustainable tourism.Brisbane: Queensland Government.

World Tourism Organization. (2004). Enzo Paci papers onmeasuring the economic significance of tourism (Vol. 4,pp. 151–175). Madrid: Author.