Tax evasion and tax reform in a low income economy : general equilibrium estimates for Madagascar

Policy Research W mli') WORKING PAPERS Trade Policy Country Economics Department The World Bank June 1992 WPS 918 Tax Evasion and Tax Reform in a Low-lncome Economy General Equilibrium Estimates for Madagascar Jaime de Melo David Roland-Hoist and Mona Haddad If Madagascar moved toward a simpler, uniform tax structure, it could raise the same revenues it now raises - with less incentive for the tax evasion and smuggling now prevalent. Policy Recach Woiking Papors disscrninatche findings of woik in progress and enonougetho exchange of ideas among Bank staff and allodmitcedindevlopmnisTheucpapezs,dtbutedbydteRearchAdvioyStaffcyhenani esoftheauthon.rdflct oly heizviewsndh sould ba use and dtedaccingly.b Tfindings, itanatiors, and eof usions aretheauthoiown.heyshould not be attnbuted to the Wodd Banlk, ixs Beard of Dimos, itS mansgerncn, or any of itS memnber countrics  Policy Research Trade Policy WPS 918 This paper.- a product of the Tra e Policy Division, Country Economics Department - is the product of a foHow-up activity from a technical assistance mission to Madagascar in February 1989, carried out under the auspices of the joint UNDP/World Bank Trade Expansica Program. Copies of this paper are available free from the World Bank, 1818 H Street NW, Washington DC 20433. Please contact Dawn Ballantyne, room NIO-023, extension 37947 (June 1992, 34 pages). Madagascar's weak administrative system and Next they calculate the welfare gain that would complex tax structure (with many exemptions) result from less distortionary tax structures. have led to tax evasion and smuggling. De Melo, Simulation results suggest that the excess burden Roland-Holst, and Haddad compare of taxes would be greatly reduced if Madagascar Madagascar's fiscal system with that of other moved closer to a tax system with uniform rates low-income countries, noting its greater reliance across sectors and instruments. Relatively low on distortionary taxes. uniform taxes would raise the same revenue as the structure prevailing in 1988, and would Using a 10-sector model and general-equilibrium reduce incentives for tax evasion and smuggling. calculations, they estimate revenue losses from exemptions, tax evasion, and smuggling for three Assuming that the uniform tax (for import and important tax instruments: import duties, value- export duties, VAT, and excise taxes) would be added taxes, and excise taxes. Allowing forthe imposed only in sectors in which tax collecdon is agricultural and informal sectors to remain now positive, simulations suggest that a uniform exempt from taxation, they estimate that apply- tax rate of 6 percent across instruments would be ing published tax rates to the nonexempt sectors enough to raise the same revenues collected would raise tax revenue (from these three under the current structure. Moreover, lower instruments alone) from 6.4 percent to 15.1 bound estimates indicate that the excess burden percent of GDP at a relatively low welfare cost of taxation would be reduced by moving toward (0.4 percent of GDP), because such a move uniformity of about 5 percent of the tax base. would reduce dispersion across instnuments and within the import tariff structure. h e Policy Research WorDing PaperSeriesdisseminates thedings of workunder way in theBanr Anobjectiveof theseries is to get these findings out quickly, even if presantations are less than fully polished. The findings, interpretations, and conclusions in these papers do not necessarily represent of ficial Bank policy. Produced by the Policy Research Dissemination Center  Table of Contents 1. Introduction 1 2. A Description of the Malagasy Tax System and a Comparison with Other Countries 2 3. A General Equilibrium Tax Model 10 4. An Evaluation of Tax Revenue Loss 1S S. lhe Relative Efficiency of Alternative Tax Instruments 22 6. Conclusions 28 Refwrences 30 Appendix: The Madagascar Data Base 32 This research is a follow-up activity from a technical assistance mission to Madagascar in February 1989 carried out under the auspices of the joint UNDP/WB Trade Expansion Program. The views are those of the authors and should not be attributed to their respective affiliations. We thank Philippe le Houerou and a referee for helpful comments.  Tax Evasion and Tax Reform In a Low-Income Economy: General Equlibrium Estimates for Madagascar 1. Introduction Like many other poor countries with underdeveloped aministrative capacities, Madagascar experiences substantial smuggling. The motives for smuggling are several, ranging from the desire to survive in the face of a host of barriers created by government policies and the lack of infrastructure, to tax evasion. The reasons for (and effects of) smuggling in the African context are well described by Stolper and Deardorff (1990) who argue that smuggling is unlikely to involve any extra real costs of trading and note that trade along an East-West axis is natural whereas states are often organized along North-South axes where trading is more difficult to organize. They even go further and argue that under African conditions smuggling is likely to shift out the production possibility curve as it is likely to reduce pervasive domestic distortions. Nonetheless, it is no exaggeration that, notwithstanding the proliferation of tax exemptions, smuggling to avoid taxes is pervasive in low-income countries where tax structures are exceedingly complex and tax administration is notoriously weak. Madagascar is no exception. Exemptions and evasions lead to a revenue loss and discriminatory taxes lead to a welfare loss. In this paper, we take a first step towards analyzing potential losses from these two effects. We also examine the resource allocation and welfare effects of revenue neutral tax reform. Before embarling in an evaluation of potential tax loss, it is interesting to note the preoccupation of many developing country authorities with tax evasion. Again, Madagascar is no exception and has engaged the services of preshipment inspection (PSI) firms to verify that the quality and quantity of goods shipped meets contractual standards and that prices charged are within "reasonable" norms. A recent evaluation of Madagascar's use of PSI by Yeats (1991) comes to the conclusion that, by and large, it has not been either cost effective, or successfil in reducing capital flight or customs duty avoidance. These concerns result from scrutinizing Madagascar's relative import prices before and after PSI requirements  2 were adopted and suggest that the problem of tax "avoidance" remains a lively issue deserving of further analysis. In section 2, we describe briefly the Malagasy tax system and recent reforms aiming at streamlining the tax structure. We also compare the composition of government revenues with that prevailing in other low-income countries, noting that the Malagasy government raises a substantial proportion of its revenues from distortionary trade taxes. In section 3, we present the structure of a standard static general equilibrium model which we use to address the issues raised above. In section 4, we give rough calculations of tax revenues lost through tax evasion. In section 5, we report results of a standard analysis of tax reform in the absence of tax evasion and without a revenue constraint. The estimates in sections 4 and 5 are with a ten-sector model so as to allow for sufficient disaggregation to capture the distortionary effects of the Malagasy tax system. Conclusions follow in section 6. 2. A Description of the Malagasy Tax System and a Comparison with Other Countries As the majority of developing countries, Madagascar has relied heavily on trade taxes as the major source of fiscal revenue. At the time of writing Oate 1990), the tax system in Madagascar can be broadly described in terms of three major categories. First, domestic direct taxes which include income taxes, property taxes, and other taxes. Second, domestic indirect taxes which include the value added tax and the consumption tax. Third, international trade taxes which include export and import taxes. Table I summarizes the structure of the Malagasy tax system described in the text. 2.1 Domestic direct taxes Tax on company Drofits (IBS). Companies are imposed a tax on profits (Imp6t sur les Bendfices des Societds, IBS) which affects net profit from all origins, including activities abroad. Foreign  Table 1. The Malagasy Tax System Source Type Tax Rates 1988 revenue' and (% of tax revenue) D 45% for commerce 22.7 (6) 0 D Profits (IBS) 35% for other activities 22.2 (6) M I Individuals (IGR) Progressive max. rate: 45% E R Property (see text) 5.1 (1) S E Other (see text) 6.5 (2) T C I T C I Value-added (TUT) 15% except for exports, 40.3 (11) T N necessities and A D agricultural products x I E R S E Consumption tax (TC) 5% to 10% for most of 28.8 (8) C 300 products (see text) T T I Customs duty (DD) 5% to 45% (see text) 18.3 (5) R N A D D I Import tax (TI) 5% to 80% (see text) 73.1 (20) E R E Other import taxes (see text) 11.2 (3) T C A T Export taxes Specific tax on vanilla, 43.8 (12) X cloves, coffee (see text) E S Value-added (TUT) 15% on imports 54.1 (15) a. In billions of FMG.. Source: Authors' elaboration ffom Guillaumont et al. (1990).  4 companies are taxed only on profits realized in Madagascar. Collectives are not subject to this tax since the partners are subject to the tax on individuals, The tax rate on profits became uniform in 1983 and was fixed at 45 percent. Before that, companies were imposed different taxes depending on their economic activities. However, in 1987, in order to encourage industrial activity the following preferential rates were applied: 40 percent for industrial enterprises, 35 percent for agricultural enterprises, and 45 percent for commercial enterprises. Besides creating administrative difficulties, the adoption of differential rates across activities gave an incentive to arbitrage by shifting towards activities with a lower tax rate. Effective January 1st, 1989, the tax rates were once again set at 45 percent for commercial activities and at 35 percent for agriculture, industry, mining, hotels, and transport. Moreover, the fixed portion of the minimum tax levy was no longer differentiated according to the legal form of the company and was set at FMG 400,000 in all cases. On the other hand, the variable portion of the minimum levy was raised from 0.1 to 0.5 percent of sales revenue. This description of the frequent changes in the structure of the IBS shows the difficulties the Malagasy authorities have had with settling on a satisfactory tax on company profits. Also, the numerous exemptions and differences in tax rates must have provided strong incentives for arbitraging across tax categories if not for outright evasion since the tax rates are quite high. Taxes on individuals (IGR). The personal income of individuals was subject to two progressive taxes: a tax on wages and salaries (nmp6t sur les Revenus Salariaux et Assimiles, IRSA) and a tax on non-wage income (ImpOt sur les Revenus Non Salariaux, IRNS). The IRSA is independent of any revenue from sources other than wages and salaries, all of which are subject to the IRNS. There was no general complementary tax on revenues from all sources until recently. Effective January 1, 1989, a major reform on the personal income tax system was introduced. For the IRSA, it involved a marginal tax structure with 9 'tranches' and a maximum marginal rate of 40 percent beginning at FMG 500,000 per month. For the IRNS, the reform involved a marginal tax  5 structure with 8 'tranches' and a maxinum marginal rate of 50 percent beginning at FMG 5 million per annum. In 1990, a new reform introduced a general tax on revenue (Imp6t GenEral sur le Revenu, IGR), which was still strongly progressiv.e and with a maximum marginal rate of 45 percent. The IGR, although not purposely set by the authorities to decrease the average tax rate, would eliminate the regressive elements that the previous system incorporated. The total revenue, independently of its origin, would constitute the criterion on the ability to pay taxes, so that some of the differentiations and injustices hidden in the old system would be eliminated. Taxes on proper. Taxes on property include (i) taxes on real estate which include a land tax levied annually on the estimated productive value of land based on the type of crop used, a tax on buildings levied on the rental value of buildings, and a surtax on buildings; (ii) death and gift duties which are levied on the net value of property causa mortis or inter vivos; and (iii) property transfer duties which are levied on sale, lease or exchange of property. Oher taxes on income. Other taxes on income include taxes on capital income for which the tax rates vary from 45 percent on dividends to 25 percent on other profits distributed by companies, and a tax of 15 percent on transfers abroad. Collection of the direct taxes described above is low. Administrative capacity of central and local tax authorities is weak. Furthermore, the absence of an accounting system for taxpayers worsens the situation, since they do not have verifiable accounts. Indeed, eighty percent of the approximately 10,000 businesses subject to IRSA are taxed on a presumptive basis. Underestimation of income is therefore a widespread phenomenon.  6 2.2 Domestic Indirect taxes The Malagasy domestic indirect tax system is based on two major types of taxes: a value-added tax (Taxe Unique sur les Transactions, TUT) and a consumption tax (Taxe I la Consommation, TC). Value-added tax (TUT). The TUT is a tax on value added which has a fixed rate. This rate was changed in 1983 from 10 percent to 15 percent. The TUT is applied to all sectors involved in local production including the services sector, and to imports. It excludes the necessity consumption goods, agricultural products and exports. The TUT is a major source of fiscal revenues. It represented on average 10 percent of the fiscal revenues between 1981 and 1988. Moreover, the TUT is not distortionary between sources, although exemptions between activities can be viewed as distortionary. Consum2foitn tax (TC). The second major tax on goods and services is the consumption tax. The TC covers more than 300 products including many inputs. A large number of goods which are excluded from the TUT are subject to the TC. The TC comprises multiple rates which depend on the type of product. It ranges from 5 percent to 500 percent, although most products are taxed at 5 percent or 10 percent. [Other indirect taxes include taxes on insurance premium and motor vehicle but represent a negligible share of the fiscal revenue (0.9 percent in 1988)]. 2.3 Taxes on foreign trade Import taxes in Madagascar serve two purposes: to protect the local industry against imports, and to raise fiscal revenue. Prior to 1988, restraints applied on imports included quantitative restrictions (QRs), customs duties, import taxes, consumption surcharges, and special import surcharges. After the fiscal reform of 1990, import duties were reduced to a customs duty, a fiscal duty on import, a value- added tax on imports (see above), and import duties on petroleum products.  7 Quantitative restrictions (QRs) were imposed in Madagascar mainly to overcome the shortage of foreign exchange. Moreover, imports of goods for which local production could satisfy demand were prohibited. Q3RS were completely eliminated in 1988 and 1989. Customs duly (DD). The customs duty (Droit de Douane I l'entree, DD) has seven different rates (0, 5, 10, 15, 20, 35, 45 percent) applied to the c.i.f. value of imports. Import tax (TI). The import tax (taxe I l'importation, TI) is levied mostly on the c.i.f. value of imports or on physical volumes for selected goods. On January 1, 1988, the tariff reform introduced a simplified tariff structure which reduced the minimum number of brackets from 69 to 16, with a maximum rate of 80 percent and a minimum rate of 5 percent. (For certain products, a temporai3 surcharge of 30 percent was introduced to ease the transition period.) Effective January 1, 1989, the minimum duty was raised to 10 percent, except for some products such as fertilizers, pesticides, and pharmaceutical products. In addition, the temporary surcharge was cut to 10 percent. The ultimate objective of the tariff reform is to put into place a simplified tariff structure with rates ranging from 10 percent to 50 percent. Other taxes on imports. The TUT and eventually the TC also affect imports as well as domestic production. The TC is applied to the c.i.f. value of imports, while the rate of the TUT is fixed at 15 percent and is applied to the c.i.f. value of imports inclusive of the DD, the TI, and the TC. A stamp duty (droit de timbre douanier) of 1 percent is levied on the perceived taxes (DD, TI, and the TC). Export taxes. In order to encourage exports and reduce reliance on export duties and taxes. the government eliminated export taxes on all goods in 1987, except for vanilla, coffee, and cloves. In 1988, the export duty on cloves was set at the specific rate of FMG 1 10/kg, and the one on coffee was set at the specific rate of FMG 19/kg. Furthermore, an export surcharge is levied on coffee, cloves, and  8 vanilla in addition to the export duty. It consists of a rate of 10 percent for coffee, US$1I per kg. for vanilla, and 15 percent for cloves. 2.4 A Comparison with Other Countrles It is apparent from the above description that Madagascar has a very complex tax structure. This is not uncommon among developing countries which have been heavily influenced by their colonial legacy. The emphasis on progressive income taxes, a cascaded structure of indirect taxes, a schedular system for direct taxes, and a proliferation of exceptions encourage too many arbitrages that erode the tax base. Also, the resulting system would appear to be far too complex for the country's administrative capabilities as the recent simplifications in the tax system suggest. In his recent review of tax reforms in several developing countries, Thirsk (1990) has noted a general across-the-board move towards a streamlining of tax rates, abolishment of exemptions so as to bring transparency to the tax system and remove the opportunities for arbitrating across tax rates and tax categories. Before turning to a quantitative analysis of the likely effects of such a tax reform, we compare briefly Madagascar's structure of tax revenues with that of other low-income countries. Comparisons are reported in table 2. The comparisons in table 2a siggest two observations. First, even among low- income countries, Madagascar's tax revenues (as a share of GDP) are a third lower than in other developing countries. From the description of the tax structure in section 2, this certainly cannot be due to low tax rates. Rather it must be a combination of tax exemptions and tax evasion. Second is the unusually high share of trade taxes in total revenues. The relatively high share of trade taxes in total tax revenue reflects a combination factors. First, is a weak administrative capability that must have reflected itself more strongly in the application of ta- rates and exemptions for domestic taxes. Second, is the important share of coffee, vanilla and cloves (two-thirds of agricultural exports). For vanilla and cloves, Madagascar is likely to have monopoly power in world markets. Hence there is an argument for taxation  9 Table 2. Tax Revenues (2a) A Comparison with Other Low Income Countries (1986-88 average) Tax Revenue Income Tax/ Domestic International (IR)/GDP TR Tax/TR Tax/TR Low income countriese 16.3 22.4 27.3 30.9 Madagascar 11.8 14.3 28.3b 55.6 (2b) Madagascar: Budgetary Revenue (1988)9 Foreign Budgetary Income Other Trade Tax on Goods & Profits 200.5 96.8d' 51.49 11.8 a. Low-income countries: sample of 36 countries with 1980 income per capita below $500. Average values for 1986-88. Source: Faini and de Melo (1991) table 2. b. Includes taxes on goods and services and taxes on property. c. Billion FMG. Source: Guillaumont et al. (1990) and authors' calculations. d. Includes monopoly profits tax (24.8 billion). e. Includes profits (22.7 billion) and wage tax (14.4 billion).  10 for taxation on both revenue and welfare grounds. Apart from export tax revenues, Madagascar's source of budgetary revenues is fairly similar to that of other low income countries. The detailed figures of budgetary revenues for Madagascar in table 2b for 1988 reveal another characteristic of tax structures in low income countries: a distortionary tax structure across markets and activities. Trade taxes discriminate across markets, and profits and wage taxes discriminate against investment and employment. We retain two conclusions from this brief look at the Malagasy fiscal system: a complex tax structure that yields relatively low revenues, and hence a suggestion of tax evasion; and a distorted tax structure that discriminates against trade and agricultural activities. In the following, we attempt to quantify both the potential loss in revenue from evasion and the distortionary costs of taxation while recognizing that the Malagasy administrative tax ability is weak. 3. A General Equilibrium Tax Model We now describe briefly the general equilibrium tax model we shall use to evaluate the welfare and resource allocation effects of tax reform. The features of the model are standard to computable general equilibrium (CGE) models, except for the inclusion of the various taxes which reflect the Malagasy tax system. We therefore describe briefly the model using a one-sector formulation to save on notation.' The empirical application is with a ten-sector model calibrated to 1988 data whose aggregation is described in table 4. Consumer behavior is represented by a linear expenditure system. The resulting demand functions (equation 1) are derived from the maximization of the Stone-Geary utility indicator. These 1 Except for the treatment of technology for intermediate demand and the treatment of taxes, the model's structure is quite similar to the one presented in de Melo and Tarr (1992, Chapter 3). For an introductory presentation to CGE models, see Dervis, de Melo and Robinson (1982, Chapters 5 and 6).  11 Table 3A. A One-Sector Tax Model Consumer Behavior: C = LES (PQ. ) (1) Production Technology: X = CES (LD' KD' V; 0) (2) V=aX (3) Factor Demands: LD= CES(r,$) (4) RD Domestic Demand and AllocatioL. of Traded Goods: Q = CES (DD, M; a) (5) DD =CES(PD PY a) (6) X = CET (Ds,E; t) (7) Ds = CET(P,,Piz) (8) Foreign Commodity Supply and Demand Functions: _mHM (9) II, Hi or E=Dg(PE;,.) (10) Domestic Prices: PvA| = PX (I - t) - a PQ ( 1 P, X = PD D, D PE (12) P Q = (1 + tQ + tVA PVA) [PD Ds PMJ (13)  12 Table 3A. A One-Sector Tax Model (continued) Foreign Commodity Prices: PM= (I + t,M) Hfie (14) Pe = (I - t E) (15) Market Equilibria: DD = VD +CD (16) DD = Ds (17) LD = (18) KD = S (19) Foreign Trade Constraint eB = IIUM - H,E (20) Income and Government Revenue: Y = (1 - tl) WD + (1 - tr) r KD + YG + eB (21) YG =t PX X + tQ PQ Q + tVA PVA PQ Q . tM M M e + tI ,Ee tL * t,rK (22) Numeraire: PD a 1 (23) Notes: A bar ovef a variable indicates an exogenous variable.  13 Table 3 (continued): Variables and parameters in the one sector tax model DD domestic demand for domestic goods PD domestic disposable (after tax) purchaser prices of domestic goods Y domestic income C personal consumption (composite) X gross domestic output LD demand for labor KD demand for capital by sector V total intermediate use (composite) Ls aggregate labor supply (exogenous) w average wage rate Ks aggregate capital supply r rental rate on capital Q composite good for domestic demand M imports Pm domestic currency price of imports Ds domestic production for domestic use E exports Pl domestic currency price of exports Px producer price of domestic output PQ purchaser price of composite domestic demand PvA price of value added YO government income e exchange rate B exogenous net foreign borrowing .r, exogenous world price of imports xit exogenous world price of exports (except when indicated) tVA value added tax rate (base is total value added) tk excise or sales tax (base is domestic sales) tm import tariff rate (base is domestic import demand) to export tax rate (base is exports) tx indirect or monopoly tax (base is monopoly revenue) tL tax on labor income tk tax on capital income StruOural and policy parameters * eONsticity of substitution between labor and capital in domestic production a intermediate use coefficient a elasticity of substitution between domestic and imported goods T elasticity of transformation between domestic and exported goods tD indirect tax rate on domestic sector production tM import tariff rate ts export subsidy rate r, elasticity of foreign export demand  14 demand functions allow for non-unitary income elasticities of demand and non-zero cross-price elasticities of demand between domestically-produced and foreign-produced consumption goods. The production technology is constant returns to scale and involves intermediates and two primary factors, capital and labor, which are mobile between sectors and are both in fixed supply. The functional form describing the production technology is a constant elasticity of substitution (CES) function to represent capital-labor substitution and substitution between domestic and foreign intermediates (equation 2), and a Leontief function between intermediates (as a whole) and gross output (equation 3). Atomistic firms maximize profits independently and are price-takers in factor and product markets. The resulting factor demands are given in equation 4. The treatment of foreign trade recognizes that in an economy like Madagascar, domestic and foreign-produced goods are poor substitutes. Likewise, goods sold abroad and goods sold in the domestic market are imperfect substitutes. This formulation is known as the national product differentiation assumption. This assumption gives rise to the composite good aggregation functions in equations S and 7. By assuming that demanders (suppliers) minimize (maximize) the cost (revenue) of purchasing (selling) a given quantity of composite good Q(X), gives rise to the first order conditions in equations 6 and 8. Imports are in perfectly elastic supply (equation 9), but foreign export demand may not be perfectly elastic (equation 9) to reflect the possibility that Madagascar may have monopoly power for its principal agricultural commodity exports (vanilla, cloves).2 The following five equations describe prices and the various wedges introduced by the Malagasy tax system. The equations describing domestic prices (equations 11, 12, 13) result from the application of Euler's theorem to the linear homogenous functions describing technology choice and goods allocation across domestic and foreign markets. Three wedges are introduced: the "monopoly profits' tax, t., 2 For a description of the implications of this external closure on the shape of the domestic offer curve, see de Melo and Robinson (1989).  15 which is applied to all sales;3 the sales tax (TUT), tQ, which is applied to all sales on the domestic market; and the value-added tax, tVA, which is applied to imports and to domestic value-added. The commodity tax structure is completed by tariffs on imports (equation 14) and by taxes on exports (equation 5). The next set of equations (16, 17, 18, 19) describes the conditions for equilibrium in the goods and factor markets. The model includes a foreign trade constraint (equation 20) so that the equilibrium real exchange rate is determined endogenously. To facilitate the welfare interpretation of tax reforms, all government revenue, YO (equation 22) is returned to the representative consumer in lump-sum fashion (equation 21). Because of the linear homogeneity of all demand and supply functions, only relative prices can be determined. Hence the need to select a numdraire (equation 23).' 4. An Evaluation of Tax Revenue Less In this section we use the model to estimate the potential government revenue loss through exemptions and various forms of tax evasion (smuggling, bribery, etc.). Table 4 describes the ten-sector sectoral aggregation and tax revenues by tax instrument. The disaggregation into ten sectors was deemed the minimum one to capture the incidence of the main tax instruments used in Madagascar. The structure of the economy in table 4 and the values of the endogenous variables correspond to a 'base" solution of the model. This calibrated base simulation replicates the actual disaggregated flows in the Malagasy economy in 1988.5 Three quarters of exports originate in agriculture which is also, by far, the largest 3 In the numerical application, this tax is only applied to sales of agricultural products. 4 By Walras law one of the equations in the model is redundant. However, for expositional purposes all equations describing the model are included in table 3. 5 How we updated the 1984 input-output table and reconciled it with national accounts, foreign trade figures and budgetary revenue figures is discussed in an appendix available upon request. It would have been desirable to disaggregate agriculture into export cash crops and other (mainly for domestic use) agriculture. However, the input-output table did not make this distinction so further disaggregation was not possible.  Shams Aca 4Adc_b Emeq Pw_ced Food UL Inui HaVy b1hufty OL Mmaf. Trnpoet Cnmmwe Snies No-mahet Ws Vahe Sa") ~~~~~~~~~~~ (1)3) (4) mS (6 (7) m (9) (!