Assessing Indirect Tax Reform in a Tourism-Dependent Developing Country

www.elsevier.com/locate/worlddev

World Development Vol. 33, No. 7, pp. 1183–1200, 2005
2005 Elsevier Ltd. All rights reserved Printed in Great Britain 0305-750X/$ - see front matter

doi:10.1016/j.worlddev.2005.04.005

Assessing Indirect Tax Reform in a Tourism-Dependent Developing Country NISHAAL GOOROOCHURN and CHRIS MILNER University of Nottingham, UK

*

Summary. — The paper investigates the effects of reform of the current structure of indirect taxes in Mauritius, a relatively tourism-dependent economy. It uses a computable general equilibrium model to explore the relative efficiency of changing sales tax rates on tourism and nontourism related sectors, and allowing for equity considerations. The efficiency of tax reforms are also distinguished for cases where tourist arrivals are exogenously set, and where they endogenously adjust to changes in relative prices. The simulation results show that the tourism sectors are currently undertaxed. Additionally, taxing tourism sectors is found to be the most socially efficient means of raising tax revenue.
2005 Elsevier Ltd. All rights reserved. Key words — Africa, Mauritius, tourism, CGE, tax reform

1. INTRODUCTION In recent years, there have been a number of sources of pressure for tax reform in developing countries, including domestically and from multilateral agencies. There are several studies which consider different aspects of tax reform including direct and indirect tax systems design, reducing burden on the poor, improvement in tax administration, and coordinating tax instruments. 1 This paper investigates indirect tax reform in a tourism-dependent developing economy, Mauritius. The efficiency and equity impact of changes in sales tax rates from the current tax structure for each sector are identified, with special attention of the modeling being given to tourism-related sectors. Recent figures show that tourism is now the second largest foreign exchange earner, after manufacturing, with sugar in third place. Tourism earnings as a percentage of total exports increased from 5.9% in 1975 to 14.9% in 1990 and to 20.2% in 2002. The tourism sector is now a major pillar of the Mauritian economy, with gross tourism receipts contributing to about 19% of GDP, 4.6% of employment, and generating more than 12% of tax revenue directly and indirectly (Durbarry, 2002). 2 International tourism plays a significant role in the economic development of many develop-

ing as well as developed countries (Dieke, 1993; Gray, 1987; Sinclair, 1998); offering a potential major source of foreign exchange receipts, employment, and increased economic activity in general. Consequently, tourism is a prime target for taxation in several economies, including Mauritius. As noted by Bird (1992), despite the importance of the tourism sector in developing countries, very few studies have been undertaken to analyze tourism taxation. Bird (1992) argues that developing countries tend to undertax their tourism sectors. In particular where there are distinctive and differentiated natural amenities and the demand for tourism products is relatively inelastic, there is scope to ‘‘extract’’ more tax revenue. Such increases in taxation may well also reduce environmental degradation, but they also risk reducing foreign earnings from tourism if demand for a specific country’s tourism product is more elastic than anticipated. This is in particular a risk in the long term if the distinctiveness of a specific tourism location decreases * The authors are grateful for help and comments on an earlier version of the paper received from participants of the ‘‘International Conference on Tourism Modelling and Competitiveness,’’ University of Cyprus, October 2003. Final revision accepted: February 11, 2005.

1183

1184

WORLD DEVELOPMENT

with the accumulation of tourist arrivals and the development of alternative destinations. Tax reform in tourism-dependent economies is of particular interest because of the distinctive characteristics of the tourism sector for tax modeling purposes, with tourism-related commodities being consumed by both domestic residents and international tourists. This characteristic offers policymakers the possibility of increasing tax revenue at a relatively lower (domestic) social cost than alternative means of tax generation. The risk facing such a tax policy is that tourists may substitute from higher to lower taxed destinations and that increased taxation of tourism-related commodities might adversely affect domestic consumers or specific groups of consumers. Capturing such linkages is difficult or impossible using partial equilibrium methods of analysis. Nevertheless, relatively little empirical analysis on this topic has used general equilibrium methods. There are a few studies that have used computable general equilibrium (CGE) models to analyze tourism (for example, Adams & Parmenter, 1995), but they do not investigate taxation of the tourism sector. An exception is Blake (2000) who uses a CGE model to evaluate the economic impacts of tourism taxation for the Spanish economy. This study explores the resource allocation and overall welfare effects of taxing international and domestic tourism activities, but wholly abstracts from distributional considerations. This study uses the CGE approach of Blake (2000), but applied to a developing country rather than an industrialized country and incorporates distributional considerations in the analysis. The above work reviewed does not explore equity effects. The model consists of 17 sectors and eight household types, and is applied to Mauritius to investigate indirect commodity tax reform in the presence of tourism in order to explore the relative benefits of taxing tourism-related and nontourism sectors. The tax system in Mauritius consists of mainly the income tax, corporate tax, sales tax, import tariffs, excise duty, gambling tax and hotel tax. Like most developing countries, the economy relies highly on indirect taxes, which consist of 78% of total tax revenue collected and represent around 60% of total government revenue. The structure of tax revenue in Mauritius is given in Table 1. In 2001, the sales tax was replaced by a Value Added Tax (VAT). The general VAT rate in Mauritius is currently 10%, but there are several specific commodities

Table 1. Structure of tax revenue in Mauritius 1997/98

2002/03

$Mn

%

$Mn

%

Direct tax Income tax Corporate tax Taxes on property and other Indirect tax Sales tax Excise tax Tax on gambling Hotel tax Import duties

121.5 42.7 40.4 38.4

23 8 8 7

186.2 64.1 74.3 47.8

22 7 9 6

404.7 93.9 52.7 24.3 21.4 212.3

77 18 10 5 4 40

677.1 338.4 80.4 33.3 0.1 224.9

78 39 9 4 0 26

Total

526.1

100

863.3

100

that are exempted from sales tax or taxed at lower or higher rates. Following the introduction of the VAT, the contribution of sales tax/ VAT revenue to overall tax revenue has more than doubled, increasing from 18% in 1997/8 to 39% in 2002/03, while trade tax reliance has fallen from 40% to 26% over that period. 3 The paper explores marginal tax reforms using the concept of marginal excess burden. For the existing, nonoptimal structure of taxes and tax rates, the incremental welfare cost of raising extra revenue through marginal changes to sales tax rates on alternative sectors is measured. The results indicate that taxation of the tourism sector would be relatively efficient, and that taxation of the sector is currently below the constrained ‘‘optimal’’ level. This conclusion also holds when allowance is made for equity considerations as well as allocative efficiency issues. The remainder of the paper is organized as follows. Section 2 provides an overview of the structure of the CGE model used. In Section 3, the concept of the marginal excess burden (MEB) is explained, and its use in modeling marginal tax reform is explained. Estimates of MEBs based on efficiency grounds only are derived (Section 4) for each of the sectors of the economy. The sensitivity of these estimates based on different levels of inequality aversion is analyzed in Section 5. Finally, the summary conclusions of the paper are set out in Section 6. 2. THE COMPUTABLE GENERAL EQUILIBRIUM MODEL We use a single country CGE model, where Mauritius is taken as a small open economy,

ASSESSING INDIRECT TAX REFORM

and tourism is the only export sector for which the world price is not fixed. The model is static with the conventional neoclassical assumptions, and belongs to the same family of models as those described by Dervis, de Melo, and Robinson (1982) and Robinson, Naude, Ojeda, Lewis, and Devarajan (1999). The model explicitly captures 10 types of taxes, and the government balances the budget through adjustment to transfer income. The model reports marginal excess burden associated with different tax simulations in a tax-distorted economy. (a) Model overview This subsection provides an overview of the nested equation structure of the model (Table 9 provides a list of the equations of the model). 4 On the supply side, the model includes 17 sectors each consisting of profit maximizing firms with production functions in the form of a Leontief function of value added and intermediate inputs. The value-added component is a constant elasticity of substitution (CES) aggregate of labor and capital, each of which is homogeneous, mobile among sectors, internationally immobile, and fixed in supply. Labor is further decomposed into the unskilled, semi-skilled, and skilled categories, which are aggregated in a Cobb–Douglas (CD) fashion. Intermediate inputs, which are a CES composite of domestically produced and imported inputs following the Armington assumption, are aggregated in a Leontief function. Output of each sector is split between domestic sales and exports using a constant elasticity of transformation (CET) specification. The demand side consists of the household sector, the government, tourism demand, and investment demand. There are eight household groups in the model classified according to income level, and each household group maximizes utility in a multistage budget process. In the first CD nest, consumers decide on the mix between domestically produced and imported commodities, while in the second Armington CES nest, they decide on how to allocate expenditure across different sectors. Investment demand is formulated in a similar fashion. Tourism demand is formulated at two levels: microtourism demand and macrotourism demand. Microtourism demand is the demand for goods and services of tourists within the destination country, while macrotourism de-

1185

mand captures tourist demand as represented by tourist arrivals. The microtourism demand is modeled like household demand, with each tourist maximizing his/her utility given a budget constraint. This generates demand of a representative tourist already at the destination. Total tourism demand is then the microdemand multiplied by the number of tourists in the country. The latter is captured by the macrotourism demand, which incorporates the effect of changes in the relative international price of tourism on tourist arrivals on the assumption that other destination prices are fixed. The price of tourism is a weighted average of the price of tourism commodities in Mauritius divided by the exchange rate. Changes in the price of tourism change tourist arrivals and consequently aggregate tourism demand. Therefore, changes in the price of tourism goods change tourism export demand, resulting in a less than perfectly elastic export demand. The model treats government consumption as a standard demand rather than as public goods. 5 Normally, an increase in taxation causing an increase in government revenue will increase government consumption which in turn will crowd out effectively private consumption, resulting in a reduction in welfare of the household sectors in addition to the one caused by higher taxes. To avoid this effect, government consumption is fixed in the model, and the budget is balanced using adjustment to transfer income, following Feltenstein (1984a, 1984b). The dead weight loss of taxation is measured as the difference between the reduction in consumer surplus and the increase in government tax revenue. If the increase in government revenue is transferred to the household sector (via transfer income), change in household welfare will then measure the dead weight loss of a given tax simulation. However, one limitation of using transfer income is that the distributional effect of taxation is sensitive to the proportions in which transfer income is allocated to different household groups. This formulation of transfer income is one of the few different government closures used in CGE. Alternative ways of balancing the budget include domestic borrowing, foreign borrowing, and government consumption. Taylor (1983, 1989), Taylor and Lysy (1979), and Bourguignon, Branson, and de Melo (1989) demonstrate how the choice of closure rules affects the distributional effects of policy changes. In our case, the formulation of transfer income

1186

WORLD DEVELOPMENT

tends to favor the poorer household groups. In this study, we are concerned about the relative efficiency (equity adjusted) effects of taxing different sectors, with the same closure rule used as each sector is subject to the same tax experiment in turn. Therefore, the absolute equity effect associated with the chosen government closure rule will only differ across alternative experiments to the extent that the amount of tax revenue generated differs across the experiments. The robustness of the results is, however, checked using sensitivity analysis in Section 5. With the small country assumption, the world prices of imports and exports are assumed to be fixed, but import prices will change when tariffs and the exchange rate change. With exogenously fixed foreign savings (balance of payment deficit fixed at zero), the equilibrating mechanism at work is the exchange rate. Changes in the exchange rate will alter the relative price of imports and exports, and consequently imports and exports will adjust to restore equilibrium. The model is also ‘‘savings-driven,’’ with total investment adjusted to be equal to savings. The exchange rate is used as the nume´raire, and is measured such that a rise in the exchange rate represents a depreciation of the domestic currency. (b) Data and policy instruments The data for the Mauritian CGE model consist of a social accounting matrix (SAM) for the year 1997. An aggregate SAM for Mauritius is given in Table 8. The production activities are further decomposed into 17 sectors, and no distinction is made between commodities and sectors. There are five sectors that are related to tourism, and they are Hotel and Restaurant (95.4%), Transport and Communication (61.4%), Retail and Wholesale Trade (2.7%), Other Manufacturing (12.1%), and Other Services (12.3%). The tourism ratio, measured as the proportion of tourism demand out of total demand, is given in brackets. A detailed description of the SAM for Mauritius can be found in Gooroochurn (2002). Our model includes all the main categories of tax including income, corporate, sales, property, and trade taxes. All taxes are treated in an ad valorem fashion. In each case, the average rate of taxes are calculated by dividing the observed tax paid by the observed tax base. Average income tax rates are calculated for each household group. 6

Tourism is not identified as a distinct sector in the input–output table of Mauritius, and hence, several adjustments are required to capture tourism in the model. It is worth noting that tourism is treated as a demand phenomenon and is identified only on the demand side. It is assumed that there is an additional group of final demand, the tourists, who consume goods and services produced by the traditional sectors, and that there is no specific sector producing only for the tourists. In the I–O table of Mauritius, tourism demand lies partly in household consumption and partly in exports. After extracting tourism demand from the latter two components, we also need to differentiate between tourism demand that is met from domestic production and that by imported goods. Specific details on the above can be found in Gooroochurn (2002). The calibration process is based on Harberger’s convention, which assumes that all prices in the model equal to one. All share and shift parameters are calibrated from the SAM. The Armington elasticity of substitution ðrM i Þ and the value added elasticity of substitution ðrVA i Þ, given in Table 2, are taken from the GTAP database (Hertel, 1997). Other elasticity values are taken from previous studies or guesstimated judiciously. The elasticity of substitution between different goods in consumption is set to 1, and the elasticity of transformation between exports and domestic consumption is fixed at 2. Tourism price elasticity of demand is set to 1.8. Sensitivity analyses (reported in

Table 2. Elasticity values from GTAP Sectors

rVA i

rM i

Sugarcane Foodcrops and fruits Livestock, poultry, and fishing Other agriculture Mining and quarrying Sugar milling EPZ textiles EPZ nontextiles Other manufacturing Electricity, gas, and water Construction Wholesale and retail trade Restaurants and hotels Transport and communication Financial services Government services Other services

0.24 0.24 0.24 1.24 0.2 1.12 1.26 1.26 1.26 1.26 1.4 1.68 1.68 1.68 1.26 1.26 1.26

2.2 2.2 2.8 2.2 2.8 2.2 2.2 2.6 3.6 2.8 1.9 1.9 1.9 1.9 1.9 1.9 1.9

ASSESSING INDIRECT TAX REFORM

1187

Gooroochurn, 2004) are carried out on the elasticity values and the results are found to be robust. Table 10 lists the parameters, endogenous and exogenous variables used in the model.

(oV/oti) to the change in government revenue (oR/oti) for a given tax change, and is expressed as oV =oti ; ð1Þ ki ¼ oR=oti

3. MODELING THE EFFECTS OF MARGINAL TAX REFORM

where V is the household welfare, R represents government revenue, and ti, the tax rate for sector i. ki can be regarded as the marginal cost in terms of social welfare of an extra unit of government revenue, and the sign depends on those of the numerator and the denominator. oR/oti shows the slope of the Laffer curve, and hence, its sign depends on the position of the initial tax rate for commodity i on the curve. Normally, the curve is upward sloping initially, and then downward sloping for a higher tax rate. The sign of oV/oti is also ambiguous. Within a basic partial equilibrium framework without externalities, that is, where social costs (benefits) equal private costs (benefits), an increase in ti will increase welfare if ti is initially negative and will decrease welfare if ti is initially positive. However, under a general equilibrium framework, the above may not always hold. A change in the tax rate of a commodity will also affect the demand for other commodities via changes in final and intermediate demand (with complementarity and substitutional relationships possible). This will in turn affect prices of other goods, and hence, the net effect on welfare is difficult to discern. Changes in the factor market also complicate matters. However, the following holds: Assuming that there is an optimal tax rate, ~ti for sector i, then an increase in the tax rate for sector i (ti) will lead to an increase (decrease) in welfare for ti < ð>Þ~ti , that is, increasing the tax rate for an undertaxed (overtaxed) sector will increase (decrease) welfare. Given that we are changing the tax rate of one sector at a time in our simulations, the optimal tax rate for a given sector is assumed to be set under the assumption that everything else in the model is fixed, and this includes the initial tax rates of other sectors. In our simulations, oR/oti always turns out to be positive when increasing the tax rate. This is because existing tax rates are below the maximum revenue rate. Therefore, the sign of MEB depends principally on the sign oV/oti. For interpretation purposes, a higher MEB is more desirable since it reflects either a higher welfare gain (if MEB is positive) or a lower dead weight loss (if MEB is negative). 8 Therefore, if ki > kj taxing sector i to raise an additional dollar is more efficient than taxing

The relative efficiency of tourism taxes is explored using the concept of MEB per additional dollar of tax revenue. MEB measures the incremental welfare cost of raising extra revenues from an already existing distortionary tax and holding other taxes constant. This is in contrast to estimating welfare costs of taxes that are calculated by completely replacing a tax with an equal yield lump sum tax. Estimating the MEB is along the lines of the marginal tax reform literature which emerged after the influential paper by Ahmad and Stern (1991). Unlike the optimal tax literature, which attempts to derive those tax rates that minimize welfare loss for a given tax revenue, marginal tax reform takes the existing tax system as given and identifies the effect of tax reform at the margin. To investigate the relative efficiency of taxing the tourism sector, we apply a small increase in the tax rate to each of the 15 sectors at a time, keeping the other tax rates constant. 7 We then calculate and compare the associated MEB for each sector. There are alternative ways of calculating the excess burden, and Fullerton (1991) provides a summary of the different methods used. Excess burden is normally measured as the change in consumer surplus (oCS) minus the change in government revenue (oR). Fullerton argues that the key issue is about whether the compensating variation or the equivalent variation should be used to calculate oCS. However, Mayshar (1988) proves that for a strictly marginal change in tax, and hence in price, both measures are equivalent. In our case, the equivalent variation is used since it allows a convenient comparison of results across simulations because it uses a common benchmark reference point. Under our general equilibrium framework, the excess burden is simply the change in household welfare (oV) which accommodates both the consumption effect (oCS) via the change in demand and the revenue effect (oR) via transfers from the government. The MEB, ki, is often expressed per dollar of additional revenue and is hence given as the ratio of the change in household welfare

1188

WORLD DEVELOPMENT

sector j, and social welfare is increased by shifting revenue generation at the margin from taxing of i to that of j. The social welfare function follows the Bergson–Samuelson formulation, V = f(W1, W2, . . . , Wh), where Wh is the welfare level of each household group. Welfare for each household group is measured using the equivalent variation (EV) and is expressed as UN  UO ð2Þ W h ¼ EV h ¼ h O h I Oh ; Uh where U Oh , U Nh , and I Oh are the old utility, the new utility and the benchmark income level of household group h, respectively. We use Bentham’s additive unweighed functional form for the eight household groups such that V ¼ W 1 þ W 2 þ    þ W 8;

ð3Þ

Wh is made up mainly of factor income, savings, transfers, and consumer surplus. It also includes welfare of the corporate sector apportioned to each household group using the same proportion that capital income is allocated to each household group. A priori, we expect MEB to be high for sectors with a high tourism ratio. The tourism ratio is defined as the proportion of tourism consumption out of total final demand (excluding government consumption). Taxing tourism intensive sectors generates additional tax revenues at a relatively low domestic cost. By contrast, the more households consume from the taxed sector, the higher will be the burden inflicted on them following an increase in the

tax rate of that sector, that is, a higher reduction in the consumer surplus of domestic households and hence a lower MEB. The burden of a tax, however, is passed on to the consumers via consumption of not only good X but also by the consumption of goods from the sectors using intermediate inputs from sector X. If sector Y purchases most of its intermediate inputs from sector X and sector Y’s output accounts for a higher share in the total demand of domestic household than sector X, then taxing sector X can reduce welfare more than if the output of sector X was allocated solely to household consumption. 4. EVIDENCE ON THE EFFICIENCY EFFECTS OF INDIRECT TAXATION We carry out a set of sequential simulations, increasing sales tax rate in each sector by 0.1% at a time. It should be noted that Sugarcane and Government Services sectors are not included in the simulation. The output of the Sugarcane sector is mostly used as an intermediate input into the Sugar Milling sector and is not taxable. Similarly, Government Services mostly represents public goods and is not liable for sales tax. The extra tax revenue generated is assumed to be given back to households as transfers (in the same proportions as in the benchmark). The results are given in Table 3. The sectors are ranked in ascending order of ki, that is, with sectors which are more efficient to tax higher up the table. The initial tax rates and the tourism ratio are given in the adjacent

Table 3. MEB(ki) of 0.1% increase in sales tax Sectors Restaurants and hotels Transport and communication Foodcrops and fruits Mining and quarrying Other services Wholesale and retail trade Livestock, poultry, and fishing EPZ nontextiles EPZ textiles Electricity, gas, and water Other manufacturing Other agriculture Financial services Construction Sugar milling

Initial tax rate (%)

Tourism ratio

ki

Rank

2.616 2.510 0.000 0.000 3.640 3.542 0.000 3.968 2.962 2.952 2.240 0.000 4.676 3.343 3.998

95.4 61.4 0 0 12.3 2.7 0 0 0 0 12.1 0 0 0 0

0.9167 0.3819 0.0831 0.0412 0.0248 0.0116 0.0145 0.0301 0.0406 0.0630 0.0658 0.0688 0.0764 0.1377 1.1680

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

ASSESSING INDIRECT TAX REFORM

columns. 9 The lines in this and later tables in italics show the tourism-related factors. The simulations for increasing the sales tax are applied to all sectors, including those that were not previously taxed. Although there is some tendency for sectors with initially low or zero rates to be ranked relatively high in terms of the ki value and for initially highly rated sector to have a lower ranking (and negative ki values), the relationship between the initial tax rate and the MEB score is mixed. Sectors with an initial zero sales tax rate are ranked third, fourth, seventh, and 12th out of the 15 sectors covered by the simulations. Contrast, for example, the rankings of Foodcrops and Fruits and Other Agriculture. Both have an initial zero tax rate, but the respective MEBs of raising the tax rate to 0.1% are +0.0831 for Foodcrops and Fruits and 0.0688 for Other Agriculture. Similarly, although an initially high taxed sector such as Financial Services has a low ranking (with ki < 0), there are other relatively high sectors (e.g., Other Services and Wholesale and Retail Trade) with much higher rankings (and with ki > 0). Interestingly, the two main tourism-related sectors (Restaurants and Hotels and Transport and Communication) have the largest (and positive) MEB score for this experiment, despite not having particularly low initial tax rates of taxation. Clearly, the important influence on this outcome is the high tourism ratio; the presence of substantial direct tourism consumption in these sectors allowing the generation of tax revenue at the margin without a corresponding decrease in social welfare. Indeed the initial tax rate must be below the optimal tax rate of taxation since the MEB values for these two sectors are +0.8951 (Restaurants and Hotels) and +0.358 (Transport and Communication). Thus, the results can be interpreted as identifying the relationship of the initial tax rate ðt0i Þ to the constrained optimal tax rate (~ti ). If initial tax rate is above the constrained optimal rate ðt0i > ~ti Þ then ki < 0, and if ðt0i > ~ti Þ then ki > 0. Table 3 shows therefore that there are six sectors for which tax rates could be separately raised to increase efficiency, and nine sectors where efficiency would be lowered if the tax rate for that sector was raised in isolation. (a) Efficiency and tourist arrivals The above simulations have been carried out under the assumption that tourist arrivals are endogenous in the model. In most previous

1189

tourism CGE analyses, tourist arrivals are assumed to be exogenously fixed, and tourist consumption is modeled like domestic demand with substitution and income effects only. In the latter method, for any given change in price, demand will change under the assumption that the representative tourist is in the country already. However, this abstracts from the role of tourism demand at a macrolevel, in particular tourist arrivals. An increase in the price of tourism products in the destination country, for instance, will reduce the demand for those products via two channels. One is through the normal substitution and income effect of the representative tourist. The second is through the reduction in the number of tourist arrivals in the country. Most studies that have estimated demand functions for international tourism for destination countries report that tourist arrivals are sensitive to changes in the ‘‘tourism price.’’ 10 When tourism demand is a function of tourist arrivals as well as the normal substitution and income effect, taxing the tourism sector might be expected to be less efficient than when tourism demand is based on the microdemand only. A given increase in the tax rate will increase the aggregate ‘‘tourism price’’ and decrease tourist arrivals. Consequently, tourism consumption will fall by more than due to microincome and substitution effects alone. This will result in a lower increase in government tax revenue and hence a lower MEB. On the other hand, as Hazari and Ng (1993) have pointed out, tourism consumption normally crowds out domestic private consumption. For example, in Mauritius, domestic residents find it difficult to buy fresh fish on the market since hotel chains buy most of them. In which case, a decrease in tourism consumption will also reduce this crowding-out effect and tend to raise welfare. The overall effect on MEB will depend on the relative magnitude of the tax revenue and ‘‘crowing-in’’ effects. To illustrate the implications of endogenizing tourist arrivals on the results, the simulations of increasing the sales tax rate for each sector at a time by 0.1% are carried out without endogenizing tourist arrivals, and the results are then compared with the results of the simulations where tourist arrivals are endogenous (as reported previously). The MEBs for the two set simulations are given in Table 4. It can be seen that, although the ranking is unchanged, the MEB with exogenous arrivals is higher for most sectors, and hence has a higher overall average

1190

WORLD DEVELOPMENT Table 4. MEB(ki) of 0.1% increase in sales tax—exogenous and endogenous arrivals

Sectors

Rank

ki Endogenous arrivals

Exogenous arrivals

Restaurants and hotels Transport and communication Foodcrops and fruits Mining and quarrying Other services Wholesale and retail trade Livestock, poultry, and fishing EPZ nontextiles EPZ textiles Electricity, gas, and water Other manufacturing Other agriculture Financial services Construction Sugar milling

0.9167 0.3819 0.0831 0.0412 0.0248 0.0116 0.0145 0.0301 0.0406 0.0630 0.0658 0.0688 0.0764 0.1377 1.1680

0.9140 0.4012 0.1100 0.0703 0.0604 0.0431 0.0187 0.0042 0.0031 0.0263 0.0266 0.0319 0.0397 0.0950 1.0215

Average

0.0137

0.0252

of 0.0252 as compared to 0.0137 for the endogenous case. The exceptions are the Restaurant and Hotel and Transport and Communication sectors. The higher MEB in the endogenous case for the latter two sectors means that the increase in welfare caused by the rise in domestic consumption due to the reduction in the crowding-out effect offsets the reduction in welfare caused by a smaller increase in government revenue. 5. ALLOWING FOR EQUITY EFFECTS Following the Feldstein (1972) model, we can assert that taxing tourism tends to improve income distribution because tourism products are generally luxury products and/or because they are generally consumed by individuals from the higher income brackets. The factors determining the relative distributional effect of taxing different sectors are complex, however, as Mieszkowski (1969) argues, the distributional effect of taxation under a general equilibrium setting depends on the composition of the use and of the source of income of the different household groups. Generally, if poorer households have a larger proportion of their income spent on consumption from a given sector A than richer households, then taxing this sector will tend to lead to a deterioration in income distribution. But the distributional effect will

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

also depend on the income effects of the tax change. This effect is fashioned by the factor income changes and income transfers induced by the tax change. Transfer income is mainly generated in the current context by the government sector. The way transfer income is modeled merits comment. The government closure rule is formulated in this model such that all tax revenue collected is redistributed to the household sector as transfer income in the same proportions as in the benchmark data. Thus, increases in transfer income induced by tax changes tend to favor poorer household groups. This is because a higher proportion of existing transfers go to poorer household groups. Although a given amount of tax revenue is distributed in the same way to the household groups irrespective of which sector is taxed, the amount of tax revenue generated depends on the sector being taxed. Larger sectors tend to generate more revenue for a given increase in the tax rate. With transfers tending to favor the poorer household groups, the greater the revenue generated the greater improvement in income distribution. 11 For this formulation of transfer income, tax revenue increases induced by tax changes must increase equity. The adjusted MEB approach extends the MEB approach by attaching different weights to the eight household groups. It has the advantage that it considers both the source and use of income effect on the distribution of welfare. We

ASSESSING INDIRECT TAX REFORM

assign different weights to every household’s welfare in the social welfare function, while computing social welfare such that a given increase in the welfare of the poorest household and of the richest household are no longer regarded as being socially equivalent. The social welfare function thus takes the following form: SW ¼

8 X

bh  W h ;

ð4Þ

h

where the bh are the welfare weights for household group h, and are given by 12 bh ¼

ðM 1 =M h Þe ; N

ð5Þ

N is simply coefficient and is P8a normalization P8 e equal to h ðM 1 =M h Þ so that h bh ¼ 1. M1 and Mh are the average income level of the poorest household group and household group h, respectively. Thus bh can be regarded as representing the marginal social benefit of one unit of income to household group h relative to a unit to the poorest household group. e is a positive number and can be interpreted as the degree of inequality aversion. If we are not concerned about inequality at all, then e = 0 such that bh = 1 for all h and Eqn. (4) reduces to Eqn. (3). If the modeler has an aversion to inequality, then e will be greater than

1191

zero such that bh < 1 for h 5 1 and becomes smaller for richer household groups. We are thus giving a higher weight to poorer household groups. For a value of e = 1, for example, the weight of the poorest household will be twice the weight of the household group with twice the income of the poorest household group. We calculate the weights for values of e of 0, 1, 2 and 5 from the average income level of the eight household groups in the benchmark data. 13 The weights are normalized such that they sum to one across the household groups. To assess the relative equity effect of taxing the tourism sector, we increase the sales tax by 0.1% for each sector at a time, and estimate the corresponding value of the adjusted MEB for the above values of e. The results are given in Table 5. The adjusted MEBs for e = 0 in Table 5 are the same as those reported in Table 4 where no weights are used. Table 5 reveals that the MEBs tend to be higher when e > 0 than when e = 0. For moderate equity consideration, MEBs increase for higher values of e, that is, as we give more weight to the poorer household group social welfare increases for a given tax increase. However, in the extreme equity consideration case, the MEB value is lower than for the moderate equity case, although it is higher than the no equity concern case. These patterns can be observed for most

Table 5. Equity adjusted MEB for increase in sales tax by 0.1% Consideration for equity None

Moderate

Extreme

e=0

e=1

e=2

e=5

Foodcrops and fruits Livestock, poultry, and fishing Other agriculture Mining and quarrying Sugar milling EPZ textiles EPZ nontextiles Other manufacturing Electricity, gas, and water Construction Wholesale and retail trade Restaurants and hotels Transport and communication Financial services Other services

0.08307 0.01447 0.06875 0.04122 1.16802 0.04060 0.03006 0.06577 0.06297 0.13768 0.01163 0.91668 0.38186 0.07637 0.02479

0.29274 0.21434 0.24080 0.38367 0.83418 0.33005 0.25643 0.28931 0.18755 0.29692 0.36110 0.92351 0.65158 0.39279 0.36157

0.33426 0.29660 0.20222 0.43703 0.28563 0.41908 0.34190 0.37589 0.28249 0.39165 0.41274 0.71277 0.59022 0.46248 0.40950

0.26771 0.26621 0.00716 0.33929 0.16486 0.33585 0.30728 0.31828 0.26101 0.35140 0.31302 0.40021 0.37585 0.33780 0.30803

Average

0.01370

0.28988

0.35888

0.29026

1192

WORLD DEVELOPMENT Table 6. Ranking of equity adjusted MEB for different values of e Tourism ratio

Consideration for equity None

Foodcrops and fruits Livestock, poultry, and fishing Other agriculture Mining and quarrying Sugar milling EPZ textiles EPZ nontextiles Other manufacturing Electricity, gas, and water Construction Wholesale and retail trade Restaurants and hotels Transport and communication Financial services Other services

0 0 0 0 0 0 0 12.1 0 0 2.7 95.4 61.4 0 12.3

sectors, and two important points emerge here. First, because MEB tends to be higher for e > 0 than e = 0, it quite surprisingly indicates that the increase in the tax rate is less detrimental to the poorer household groups than it is for the richer groups. The main reason behind this is that a higher proportion of transfer income goes to the poorer household groups, and this tends to outweigh the larger consumer surplus loss inflicted on them by the higher tax rate.

Moderate

Extreme

e=0

e=1

e=2

e=5

3 7 12 4 15 9 8 11 10 14 6 1 2 13 5

9 13 12 4 15 7 11 10 14 8 6 1 2 3 5

11 12 14 4 15 5 10 9 13 8 6 1 2 3 7

12 11 15 4 14 5 8 7 13 3 9 1 2 6 10

Secondly, because the MEB for the extreme equity case is less than the MEB for the moderate equity case, this indicates that, within the lower income groups, the very poor household groups suffer more or gain less than the less poor household groups. The rankings of the MEB for different values of e are given in Table 6. It can be noticed that the ranks for e = 0 differ considerably with the ranks for e > 0. However, once the

Table 7. Sensitivity analysis of the government closure formulation Transfer income e=2

Foodcrops and fruits Livestock, poultry, and fishing Other agriculture Mining and quarrying Sugar milling EPZ textiles EPZ nontextiles Other manufacturing Electricity, gas, and water Construction Wholesale and retail trade Restaurants and hotels Transport and communication Financial services Other services

e=5

Original

Neutral

Original

Neutral

11 12 14 4 15 5 10 9 13 8 6 1 2 3 7

7 8 4 2 10 6 5 12 13 14 9 1 3 15 11

11 12 15 4 14 6 9 7 13 3 8 1 2 5 10

6 7 9 1 10 5 4 13 14 12 8 2 3 15 11

ASSESSING INDIRECT TAX REFORM

inequality aversion parameter is above zero, the ranks of the MEB are strongly correlated, showing that they do not change significantly across e for nonzero e. Three conclusions can be drawn from these results. First, they indicate that distributional characteristics matter in determining the MEB. Second, it shows that the exact specification of the distributional weights could be less important to policy makers than the broad indication of whether they are concerned at all with distributional concerns or not. Finally, the incorporation of distributional consideration does not alter the earlier conclusions about the relative efficiency or desirability of marginal tax increases in the tourist and nontourist related sectors since Restaurant and Hotel and Transport and Communication are still ranked first and second, respectively. In light of the previous discussion on the government closure rule used, we carry out a sensitivity analysis to check the robustness of the equity-adjusted MEB results. Given that the debt or public spending options also have specific income distribution effects, we use an income neutral budget balancing mechanism such that the relative income of all household groups are fixed. The additional tax revenue generated by each experiment is now transferred to households via reduced income taxation, such that the ratios of household incomes are not altered by the experiments. By allocating the tax revenue in this income in a neutral manner for each experiment, we are able to measure the equity effects only of the changes in product and factor prices caused by the change in the sectoral tax rate. The rankings are given in Table 7 alongside the rankings in the original transfer income formulation. The results are given for the inequality aversion parameter set at 2 and 5. It can be seen that apart from the Financial Services sector, the ranking does not change significantly, and the Restaurants and Hotels and Transport and Communication sectors are still ranked on top in the table. 6. CONCLUSIONS CGE models are well suited to investigating questions about the optimality of existing tax structures and the appropriateness of tax reforms. Here we investigate indirect taxation of the tourism sector in a tourist-dependent developing country, recognizing that tourist arrivals

1193

should be viewed as endogenously determined alongside the tax burden that falls on tourists. This is an innovation in the study since tourist arrivals are often modeled in the literature as being exogenously given. In order to give practical policy relevance to the work, we do not explicitly seek to identify the ‘‘optimal’’ tax structure. Rather we examine the effects of piecemeal, marginal reforms of the indirect taxes, exploring the sensitivity of the results to whether efficiency only or efficiency combined with equity considerations are incorporated in the objective function and to whether tourism arrivals are endogenously or exogenously determined. Although the optimal structure of indirect taxes is not identified, the variation in the estimated MEB associated with adjustments to each sector’s sales tax rates indicate that the current structure of indirect taxation in Mauritius is not optimal (unless there are positive social costs and/or benefits that are not being captured by the model and for which the current structure of taxation is optimally adjusted for). From the variation in the MEBs (which tend to be both positive and negative), we can identify more or less efficient means of raising additional tax revenue. For most specifications of the social welfare function and of tourist arrivals used in this study, there are some sectors that are shown to be currently overtaxed (i.e., with negative MEBs) and some that are undertaxed (i.e., with positive MEBs). Sugar milling falls in the first category, for example, and the tourism-related sectors (Restaurants and Hotel and Transport and Communication) fall in the latter category. Indeed tourism-related sectors under all specifications receive the highest MEB estimate (and positive in all cases). An increase in the sales tax rates on these sectors would generate an additional unit of tax revenue and increase social welfare in the process. The tourism ratio (proportion of tourist consumption out of total consumption) is sufficiently large in the economy for indirect taxation to generate additional revenue at a relatively low domestic welfare loss. However, the high tourism ratio also means that the authorities need to be very conscious of the impact of tourism taxation on tourist arrivals. It was important therefore that the conclusion about the relative efficiency of tourism taxation is subjected to an examination of the sensitivity of the finding to endogenizing tourist arrivals. The finding is found to be robust.

1194

WORLD DEVELOPMENT

The results suggest that Mauritius should be able to increase tax rates on tourism and reduce rates on other sectors, while increasing welfare without reducing tax revenue. The results also suggest that taxing tourism related sectors can potentially have positive income distribution effects. In designing an ‘‘optimal’’ reform program, one would also need to consider whether partial (i.e., sector specific) or comprehensive reforms were feasible or desirable. In the context of the current modeling, it makes a difference in formulating the appropriate level of taxation of tourism whether we take the tax rates on other nontourism sectors as fixed at their current rates or not. Such an exercise in ‘‘optimal’’ tax setting is not undertaken here,

though our results give guidance as to the direction of tax policy reform. Further considerations would also have to be taken into account for a complete tax reform. Macroeconomic impacts such as effects on the balance of trade, employment, inflation, investment and consumption level, and sectoral impacts should also be taken into consideration. Another important factor is administration costs, including collection costs and enforcement costs. For instance, discriminatory taxes such as taxes on tourists only may have higher collection costs than nondiscriminatory taxes. As Ahmad and Stern (1991) argued, administration cost can be as important as tax rates in determining the effectiveness of a reform.

NOTES 1. See, for example, Thirsk (1997).

2. Sugar and Textile are the other major tradable goods sectors in the economy. Growth in the latter was associated with major income growth and sectoral change in the two decades up to the 20th century (see Milner & Wright, 1998).

3. Fiscal dependence on trade taxes has declined in Mauritius, like many other developing countries, as a result of income growth and trade policy reform (Greenaway & Milner, 1991).

4. The GAMS/MPSGE code of the model is available upon request.

5. Government consumption is predominantly from the government services sector itself and consists mainly of public goods. However, since we do not have enough data on the valuation of public goods, we treat government consumption as a standard demand rather than as public goods.

6. Marginal tax rates are preferred to average tax rates in principles although estimating the marginal tax rate is problematic and does not usually capture the complexity of the actual tax code.

7. Two out of the 17 sectors are not included in the simulations. The first one is the Sugarcane sector because 99% of the output is used as intermediate inputs

in the Sugar Milling sector, and the second one is the Government Services sector because the output represents mainly public goods.

8. This is not to be confused by the convention that a high MEB means a high dead weight loss. This is applicable when the sign of MEB is always negative and we ignore the sign so that a bigger value means a higher dead weight loss. In our case, since MEB can be positive as well, its sign is important, and thus a high MEB implies either a higher increase or a lower reduction in welfare.

9. The initial sales tax rates are generated from the data as total tax paid divided by total sales.

10. The ‘‘tourism price’’ is usually proxied using the Consumer Price Index. See Sinclair (1998) for a review of the evidence.

11. Taylor and Lysy (1979) and Bourguignon et al. (1989) demonstrate how the choice of closure rules affects the distributional effect of policy changes.

12. This formulation of the weights is commonly used in empirical studies in taxation. See Ahmad and Stern (1991) for such applications. 13. While some empirical studies, such as Ahmad and Stern (1991), use a range of values of 0–5 for e, other studies, such as King (1983), use a range of 0–10.

ASSESSING INDIRECT TAX REFORM

1195

REFERENCES Adams, P. D., & Parmenter, B. P. (1995). An applied general equilibrium analysis of the economic effects of tourism in a quite small, quite open economy. Applied Economics, 27, 985–994. Ahmad, E., & Stern, N. (1991). The theory and practice of tax reform in developing countries. Cambridge: Cambridge University Press. Bird, R. (1992). Taxing tourism in developing countries. World Development, 20, 1145–1158. Blake, A. (2000). The economic effects of tourism in Spain. Discussion Paper Series 2000/2, Tourism and Travel Research Institute, University of Nottingham. Bourguignon, F., Branson, W. H., & de Melo, J. (1989). Adjustment and equity in developing countries; A new approach. Paris: OECD Development Center. Dervis, K., de Melo, J., & Robinson, S. (1982). General equilibrium models for development policy. Cambridge: Cambridge University Press. Dieke, P. U. C. (1993). Tourism in the Gambia: Some issues in development policy. World Development, 21, 277–289. Durbarry, R. (2002). The economic contribution of tourism in Mauritius. Annals of Tourism Research, 29, 862–865. Feldstein, M. (1972). Distributional equity and the optimal structure of public prices. American Economic Review, 62, 32–36. Feltenstein, A. (1984a). Money and bonds in a disaggregated economy. In H. Scarf & J. Shoven (Eds.), Applied general equilibrium analysis. Cambridge: Cambridge University Press. Feltenstein, A. (1984b). An intertemporal general equilibrium analysis of financial crowding out: Theory with application to Australia. Journal of Public Economics, 31, 79–104. Fullerton, D. (1991). Reconciling recent estimates of the marginal welfare cost of taxation. American Economic Review, 81, 302–308. Gooroochurn, N. (2002). Computable general equilibrium modelling of tourism taxation: The case of Mauritius. Unpublished Ph.D. Dissertation, University of Nottingham. Gooroochurn, N. (2004). The economy-wide effects of tourism taxation in a distorted economy: A general equilibrium analysis. Discussion Paper Series 2004/1, Tourism and Travel Research Institute, University of Nottingham. Gray, P. (1987). The role of tourism in economic development. In J. Millet (Ed.), The role of tourism in development, Institute of National Affairs Discussion Paper No. 89.

Greenaway, D., & Milner, C. R. (1991). Fiscal dependence on trade taxes and trade policy reform. Journal of Development Studies, 27, 95–132. Hazari, B., & Ng, A. (1993). An analysis of tourists’ consumption of non-traded goods and services on the welfare of the domestic consumers. International Review of Economics and Finance, 2, 43–58. Hertel, T. W. (1997). Global trade analysis: Modelling and applications. New York: Cambridge University Press. King, M. (1983). Welfare analysis of tax reforms using household data. Journal of Public Economics, 21, 183–214. Mayshar, J. (1988). On the measures of excess burden and their application. Paper No. 199, The Hebrew University of Jerusalem. Mieszkowski, P. (1969). Tax incidence theory: The effects of taxes on distribution of income. Journal of Economic Literature, 7, 1103–1124. Milner, C. R., & Wright, P. (1998). Modelling labour market adjustment to trade liberalisation in an industrialising economy. Economic Journal, 108, 509–528. Robinson, S., Naude, A. Y., Ojeda, R. H., Lewis, J. D., & Devarajan, S. (1999). From stylized models: Building multisector CGE models for policy analysis. North American Journal of Economics and Finance, 10, 5–38. Sinclair, M. T. (1998). Tourism and economic development: A survey. Journal of Development Studies, 34, 1–51. Taylor, L. (1983). Structuralist macroeconomics: Applicable models for the Third World. New York: Basic Books. Taylor, L. (1989). Stabilisation and growth in developing countries: A structuralist approach. Zurich, Switzerland: Harwood Academic Publishers. Taylor, L., & Lysy, F. J. (1979). Vanishing income redistributions: Keynesian clues about model surprises in the short-run. Journal of Development Economics, 6, 11–29. .Thirsk, W. (1997). Tax reform in developing countries. Washington, DC: World Bank.

APPENDIX A See Tables 8–10. (See overleaf )

1196

Table 8. An aggregated SAM for Mauritius, 1997, Rs. million Expenditures

Recipients Production activities A

Production activites

Capital

Institutions Households

Government

Corporate sector

Institutions

Capital

Households

Government

B

C

D

E

A Intermediate inputs 36,205.1

Consumption 36,075.9 Consumption 10,428

Export 52,573.5

Consolidated capital account

Total

H

I

Gdfcf + invtories 16,186.3

B Compen. to empl. 34,679.0 C Opening surplus 40,622.2 D

40,622.2

Wages 34,679

E Ind taxes/ other/sub 6,617.6 F

Entrepreneurship 19,341

151,469.1

Transfers 4,672

Operating surplus 201

Dir/ind tax + other 6104.3

Gross opr. surp 21,080

Net transfer 45 Imports 15,824.8

Consolidated capital H account I

151,468.9

34,679.0

Rest of the world G Imports of inputs 33,345.2

Total

Corporate Rest of sector the world F G

34,679.0 40,622.0

Interest, Net dividend transfers 3,460 2,402 Corp tax + other 1,332

Net interest Net 253 interest 45 Net transfers 12 Net factor Imports income 374 1,783.5

Savings 6,504

Savings 435

Savings 16,212

Trade deficit 21,66.9

64,554.0

15,800.0

21,423.0

58,926.0

64,554.0

1,545.1

15,800.0

21,423.0 Imports 7,586.5

58,926.0

25,317.9

25,318.0

WORLD DEVELOPMENT

Factors Labor

Factors Labor

ASSESSING INDIRECT TAX REFORM

1197

Table 9. List of equations used in the model Price equations World price of imports World price of exports Composite commodity price Marketed output price Intermediate input price Value added price Aggregate price index Quantity equations Value added Composite demand Domestic output Demand for labor Demand for capital Export demand Import demand Tourist Arrivals Income equations Household income Government income Labor income Unincorporated capital income Incorporated capital income Expenditure equations Household total expenditure Government expenditure Household consumption Intermediate consumption Tourism total consumption Representative tourist consumption Equilibrium equations Aggregate demand Labor market equilibrium Capital market equilibrium Government savings Household savings Total savings Investment-saving equilibrium

PM i ð1 þ tM i ÞER PEi PWEi ¼ ð1 þ tEi ÞER PM i M i þ PDi Di PX i ¼ Xi PEi Ei þ PDi Di PQi ¼ P Qi PN i ¼ j wij PX i PVAi ¼ PQi ð1  tSi Þ  PN i P PINDEX ¼ i windexQ i  PX i PWM i ¼

h i va va va va va ri =ri 1 rva va ri 1=ri i 1=ri þ ð1  ava VAi ¼ /va i ÞK i i ai Li h im m m m m ri =ri 1 rm m ri 1=ri i 1=ri X i ¼ /m þ ð1  am i ÞDi i ai M i h ie e re 1=rei re 1=rei ri =ri 1 Qi ¼ /ei aei Ei i þ ð1  aei ÞDi i  va VA rvai  rvai 1 a P DDLi ¼ /va VAi i Li i P va  VA ri  va rvai 1 ð1  ava i ÞP i K DDi ¼ /i VAi K P  rei  e Ei PEi 1  aei ri ¼ Di PDi aei  rmi  rmi Mi PM i 1  am i ¼ Di PDi am i   price PINDEX eT TA ¼   TA0  ER YH h ¼ ½wdist h TL þ ½kdisth TKU  þ tdisth ½TGH þ TCH þ TWH  ðERÞ þ TOH  þ iddisth IDRH EG ¼ CG þ IDEG þ SUB þ TGH þ TGW  ER þ TGO þ SAV G P L TL ¼ i P Li P TKU ¼ i ji P K K i P TKI ¼ jC i ð1  ji ÞP K K i EH h ¼ CH h þ THW h  ER þ THOh þ THGh þ IDEH h þ INCTAX h þ PROPTAXh þ SAV H h G E ¼ CG þ IDEG þ SUB þ TGH þ TGW  ER þ TGO þ SAV G P kH Y HDIS CH h ¼ i ih h PX i P CN i ¼ ni aij Qj CTi ¼ TA · CTUi ki  Y T1  ER CTU i ¼ PQi Qi ¼ CNi + CHi + CGi + CTi + CIi + CSi + Ei P DDLi ¼ L Pi K i DDi ¼ K SAVG ¼ YGEG  

H H  YH SAV H 1  tIT h h  PROPTAX h  mpsh h ¼ W H C G SAV ¼ SAV þ SAV þ SAV þ SAV  ER TINV ¼ SAV

1198

WORLD DEVELOPMENT Table 10. List of variables used in the model

Endogenous variables (Due to space consideration, the equation of some variables listed below have not been provided above) CNi CHi CIi CINSTKi CORTAX CTi Di EC EG EH h Ei ER EW GAMTAX HOTTAX IDEC IDEG IDEH IDRC IDRG IDRH INCTAX Ki Li Mi P INV i PNi PQi PINDEX PROPTAX PVAi PMi PXi Qi SALTAX SAV SAVG SAV H h TA TARIFF TCH TCO TGH TGO THG THO TID TINV TKG TKI TKU TL

Intermediate demand Household consumption Investment demand Inventory stock demand Corporate tax Total tourism consumption Domestically consumed output Expenditure of Corporate sector Total government expenditure Total household expenditure Volume of export Exchange rate Rest of the world expenditure Gambling tax Hotel and restaurant tax Interest/dividend paid by Corporate sector Interest/dividend paid by Government Interest/dividend paid by Household Interest/dividend received by Corporate sector Interest/dividend received by Government Interest/dividend received by Household Income tax Total demand for capital by sector Total demand for labor by sector Volume of import Price of a unit of capital Price of composite intermediate goods Domestic price of composite good Aggregate price index Property tax Value-added price Price of imports Domestic price of composite good Composite domestic output Sales tax Total savings Government savings Household savings Tourist arrivals Import duties Transfer income from Corporate to Household Other transfers paid by Corporate sector Transfer income from Government to Household Other transfers paid by Government Transfers from Household to Government Other transfers paid by Household Total endowment of interest/dividend Total investment Incorporated capital to Government Incorporated capital to Corporate sector Unincorporated capital to Household Total labor income to Household

ASSESSING INDIRECT TAX REFORM

1199

Table 10—continued Endogenous variables (Due to space consideration, the equation of some variables listed below have not been provided above) TOC TOG TOH TTO VAi Xi YC YG YH h Y HDIS h YW PEi

Other transfers to Corporate sector Other transfer income to Government Other transfer income to Household Total endowment of transfers Value Added Component Composite good Income of the Corporate sector Total Government income Total income of Household Total disposable income of Household Rest of the world (World) income Price of exports

Exogenous variables CG IDRW K L PWEi PWM i SAV W TGW INSTK IDEW TGW THW TWC TWG TWH YT

Government consumption Interest/dividend received by World Supply of capital Supply of labor World price of export World price of imports Foreign savings Transfer income from Government to World Endowment of inventory Interest/dividend paid by World Transfer income from Government to World Transfer income from Household to World Transfer income from World to Corporate sector Transfer income from World to Government Transfer income from World to Household Tourism total income

Parameters /i ai ri aij mij wij kh mps ti H g wdist kdist tdist iddist ji jc sub otr otp

Shift parameter in production function Share parameter in production function Elasticity of substitution Input–output coefficient for domestic inputs Input–output coefficient for domestic inputs Input–output coefficient for composite inputs Share parameter in utility function Marginal propensity to save Tax rate Shift parameter in macrotourism demand Price elasticity of demand of exports Proportion of wages to household h Proportion of return to capital to household h Proportion of transfers to household h Proportion of interest/dividend to household h Proportion of unincorporated capital income Proportion of incorporated capital to corporate sector Rate of subsidy to producers Proportion of ‘‘other transfer’’ from income Proportion of ‘‘other transfer’’ from total ‘‘other transfer’’ (continued next page)

1200

WORLD DEVELOPMENT Table 10—continued

Parameters idp windexQ

Proportion of interest/dividend from total interest/dividend Weights in construction of PINDEX