(2PS 133~0 POLICY RESEARCH WORKING PAPER 1330 Revenue Uncertainty A temporaiy import surcharge may be the most effective and the Choice of Tax way to mobilie resourcesin Instru.ment during the Eastern Transition in Eastern Europe Delfin S. Go The World Bank Policy Research Departnent Public Economics Division August 1994 I POLicY RESEARCH WORKING PAPER 1330 Summary findings Go aramines the eroding tax base facing transitional revenue sources. To emphasize the transitory nature and economnies by employing a framework that allows risk reversibility of the policy recommendation, import tariffs factors in assessing tax instruments. should be implemented in the form of a temporary In an uncertain world, he asiks, which tax instruments uniform import surcharge. should be used? He examines Eastern Europe's revenue This conclusion seems to hold whether the problem, including the implications for public revenue of government formulates tax policy with correct or different causes of uncertainty-and investigates which incorrect expectations. But the choice of revenue target taxes are abettert at generating revenue. He defines a matters. All tax instruments will do almost equally well if 'betcer" tax as one that has greater stability in a risky the commonly used tax-to-GDP ratio is the target. But it environment (that is, less variation in generating a target is a misleading measure since the ratio does not reflect revenue) and has the least adverse impact on the the immense erosion of domestic tax bases in the economy (for example, on consumption). economy and how real revenue in absolute level may Go ernploys the framework to explain much of the actually be decreasing rapidly as a result. output and revenue fall in transitional economies. The The revenue decline and uncertainty can also be terms-of-trade shocks from the collapse of the CMEA viewed as a necessity toward downsizing the larg e state trade as well as the rigid but uncertain cconomic sector and in redirecting trade away from former responses in transitional economies are all important nonmarker partners. The results emphasize that restoring factors. revenue should never lead to maintaining subsidies The results of his model indicate that import tariffs are toward nonprofitable state enterprises or other public more effective than other traditional tax instrumcnts in spending no longer relevant in a market system. Doing so raising revenue, especially if real revenue is defined in will only lead to unreasonably high taxation. doilar terms (the price anchor). The contraction in No less important is moving assets out of collapsing domestic output and prices and the devaluation of the sectors, privatizing them, and rmaking them productive real exchange rare needed in the transition are significant again. reasons that favor imrports as a tax base over other This paper - a product of the Public Economics Division, Policy Research Department - is part of a larger effort in the department to develop tools for analyzing tax policy. Copies of the paper are available free from the World Bank, 1818 H StreetNW, Washington, DC 204331 Please contact CarlinaJones, room NIO-063, extension 37699 (37 pages). August 1994. T7he Poliy Research Working Paper Sories disseminates the findings of work n progress to cncwage thc chan of ies abosa development issues. An objectie of he sries is to get the fndings out qukly, even if the presentaiom are les th,an fidly poishe T7he papers cany the nantes of the authors and shold be used and citedaccordingly. The findings, interpretations, and condlusios are he atbhors' own and should nor be abibured to the World Bansk. its Execive Bod of Directors, or any of its member countrits Produced by the Policq Research Dissemination Center Comments Welcome Revenue Uncertainty and the Choice of Tax Instrument During the Transition In Eastern Europe Delfin S. Go' 1. Introduction As Eastern European economies undertake reforms on all fronts in their transition to market-based systems, their public revenues are not just falling, they are becoming more uncertain. The decline and increased uncertainty in revenue have several inter-related sources, including such factors as fundamental changes in the structure of the economies, steep output declines, difficulties in privatizing state entezprises, the collapse of traditional export markets, severe terms-of-trade shocks, nascent tax systems and 'Public Economics Division, Policy Research Department (PRDPE), the World Bank. This paper is part of PRDPE's effort to develop tools for analyzing tax policy. Thanks are due to Shantayanan Devarajan, Alan Gelb, and Richard Bird for their valuable comments, to Pekka Sinko for his able assistance in the numerical implementation of the framework, .md to Shankar Acharya and Christine Wallich for suggesting the topic. administations etc. (See, for example, Fischer and Gelb [1990], Gelb and Gray [1991], and Bruno [1993]). Inflation and accounting/amortization rules in the initial phase of reform also affect revenue generation through their impact on profits and profit taxes (see Shaffer [19931 for the case of Poland.) Compounding the revenue problem is the need to overhaul the entire tax system. However, introducing a market-based tax system takes time while governments critically need revenue to operate in the transition. In an uncertain world, which tax instrunents should be used? This paper examines the revenue problem in Eastern Europe, looks at the implications for public revenue of different causes of uncertainty, and investigates which taxes 'better' at generating revenue. While it is possible to raise revenue from alterative tax instrments, a 'better' tax is defined as one that has greater stability in a risky environment (i.e., lesser variation in revenue) and smaller adverse impact on the economy (e.g. on consumption). While firm estimates of revenue are impossible in economies undergoing fundamental transformation, it is important to understand how the transition will affect revenue and how much variation in tax intake each factor can generate. Efforts to appraise financial flows and economic performance of these countries are stymied by the lack of reliable data and the practical problem of modifying statistical information not immediately comparable to those in market economies (see Marer et. al. [1992]). Also, because the underlying economic relationships are changmg rapidly dunng the transition, parameters - such as the responsiveness of import demand and export supplies to terms-of-trade shocks - are inherently unstable and a range of possibilities may exist. In order to produce plausible illustrations of the revenue performance, this paper develops a simple analytical and accounting framework that imposes little data requirement but also incorporates the significant factors of change - such as unpredictable terms-of-trade shocks and shifting economic parameters. The analysis bnngs revenue uncertinty in greater focus much like a financial appraisal of a risky Go: Reve UncrtaiW 2 project. For each set of conditions identified, a risk assessment with a measure of the dispersion and likelihood is made regarding what would happen to public revenue or how high a tax rate would be needed among alternative instruments. Certainly, the usual way of measuring a single revenue potential for each new tax is inappropriate. Such an approach assumes a stable environment, which is clearly not the casc in Eastern Europe. In looking at tax policy in transitonal economies, point estimates of revenue of the most likely or the best and worst scenarios have a rather small chance of occurring. To determine the rate of taxation, e.g. a sales tax, what matters is not just the likely tax rate but the variance generated by the postulated uncertainties. Where conditions are very risky, the fiscal policy required (e.g. in choosing among alternative tax instruments) to achieve stabilization and other macro goals may be quite different from those cases in which revenue circumstances are more secure. Among recent studies of Eastern European transitional economies,2 very few look at temporary revenue measures. Holzman (1991] addressed public finance issues in a changing political enviromnent, including transitory tax measures. McKinnon [1991 & 1992] also emphasized the problem of generating revenue by looking at temporary measures, including distortionary taxes like import tariffs. This paper investigates whether such recommendations (import tanffs) are supported by a more systematic revenue analysis. The outline of the paper is as follows. In section i, the revenue performance in 5 Eastern European countries - Bulgaria, former Czechoslovalkia Hungary, Poland, and Romania - is briefly reviewed. Taking the former Czechoslovakia as a case in point, section III descnrbes the framework and the treatment of feues of a transitional economy and uncertainty of public revenue. In section IV, a few simulations are performed and interpreted. Section V offers some conclusions and suggestions for further research. 2Most studies concentrate on the practical issues of tax design without serious revenue analysis. An index of selected articles, papers and books on trasition economies is compiled in a special issue of Transiion, a newsletter about reforming economies published by the World Bank. Go: Rewnue Uncertainty 3 2. Revenue Performance in Eastern Europe 2.1. Public Revenue 120.. 10 I I - -o - Bulgaria , -- - ---- ao 70 FonnerCzechoslovakia -- _ > Hungary e So - Poland l -------- At,Romania 1986 1987 1988 1989 1990 1991 Year Figure 1: Revenue in Eastern Europe (in constant 1986 prices) Current revenues of governments in Eastern Europe, after adjusting for inflation, are falling rapidly (see Figure 1)- In some cases, revenues during this transition can only buy half of what they used to buy before the collapse of the centrally planned economiesw As percent of the peak revenue in 1988189, the revenue figures reported in recent years (expressed in constant 1986 prices) are only: 36.9 percent in Bulgaria (1992); 64.2 percent in the former Czechoslovakia (1992); 76.4 percent in Hungary (1991); 50.4 percent in Poland (1991); and 63.2 percent in Romania (1991). An irnportant reason for the decline in revenue is the collapse of output. The dismantling of the Council for Mutual Economic Assistance (CMEA) in Eastem Europe has led to the disappearance of traditional export markets, while die prices of inputs such as fuel and raw material imports have soared. The resulting extemal shocks has caused widespread output declines (see Rodrik [1993])_ Go: Reene Uncernainry 4 Moreover, while prices have been liberalized and the subsequent stabilization efforts have been impressive, the structurl change sought in the production system has been slow. The speed of adjustment is hampered by institutional and legal difficulties, such as the privatization of large enterprises, the introduction of property rights, new accounting system etc. A slower privatization and structural adjustment in turn implies that a large segment of supply is unable to respond efficiently to liberalized prices. Thus, the expanded tax bases promised by a more productive market economy have largely not been realized (see Sachs [19911 and Zou [1993]). Figure 2 reports the contraction of the general tax bases, as measured by the gross domestic product (GDP). Compared to the peak in 1989, real GDP in 1992 has shrunk on average by about a quarter: 37.5 percent in Bulgaria; 22.6 percent in the former Czechoslovakia; 14.1 percent in Hungary; 18. 1 percent in Poland; and 32.3 percent in Romania. To be sure, the extent of decline of output may be overstated because of measurement problems. It could be claimed that the pre-reform output was largely 120 5 00 - - - ----- Fm C l __ 0.. -8- Hungary\ go -a- Pola---nd -- 70 - - -v--- RomBulgania go 8 _ + Fanner Czechoslovakia ___ _. n ~~-[3 - Hungary' o ~~~-o - Poland v 70 _ Romania --- ----__ , 60 .,,, 1986 1987 1988 1989 1990 1991 1992 Year Figure 2: GDP in Easten Europe (in constant 1986 prices) Go: Revenue Uncenainry 5 inflated while the post-reform production is understated. For example, Berg [1993] estimated that real output decline in 1990 was more in the range of 7 to 8 percenr rather the official 12 percent in some countries (a factor of 0.67). Despite the correction however, this is still a sizable drop for a single year. 00 T ---- --------------l----- ----->- 30 _ -~- Romanla> IL~~~~~~~~~~ 1986 1987_ 198 198 1990_ _ ___1___991___ 40 .3. Bulgaria ._ _Yer 5.pee9 p1 i theformer Czechoslovakia 61.5 percent (9)o Polandb 30 .Ro1ana percen (1991) in Ro -- 1986 1987 1988 1989 1990 1991 Year Figure 3: Revenue Effort in Eastern Europe (revenue as percent of GDP) While rae dishntegrabon of ndeitonal revebaue bases accounted for much of the revenue problem, Vit did not explain ad . The revenue effort in each country, as measured by a siscple ratio of public reveise to GDP, suggested a significant decline in the overlal colection performanc ae broken upn Figure 3). The co:Rection rate fell rapidly in a few years: from 60.2 percent (19 to 34.2 percent (1992) in Bulgana; 57.6 percent Qi988) to 49.5 percent (1991) in the fonmer Czechoslovalia; 61.5 percent (1988) to 52.9 percent (1991) in H"gary; 48.0 percent (1988) to 29.2 percent (1991) in Poland; and 51.1 percent (1989) to 40.1 percent (1991) in Romania. The radical change in the underlying tax bases has engendered collection problems in several ways. Value added of services is nsing as thie distnbution system is decenttalized. The scale of enterprises is also srhdngknk as the service sector grows and as thie large socialized enterprises arc broken up and privatized. Go. Renue Unwrrtainw 6 The number of individual tax payers, including self-enployed entrepreneurs, are therefore multiplying. These shifts to services, smaller enterprises, and individual tax payers will make it generally harder for the governments to collect taxes. Hence, a significant factor is the organization of a new tax administration oriented towards collecting revenues from numerous private firms and individuals in the transitional economies. This is necessary because market-based taxes, unlike previous differentiated turnover taxes and discretionary profit remittances, need to be fixed, transparent, and enforceable on a decentralized market economy. Before the tax system become fully operative however, collection efforts suffer. Privatization, which encourages tax avoidance, also compounds the problem. An important source of revenue, profit remittances from state enterprises, is being replaced by direct taxes which have uncertain revenue prospects. Where profits used to be transferred completely to the states in the socialist regime, only a portion can be claimed with the new corporate tax. Moreover, profits are falling in the transition and the share of profits going to interest payments - generally not taxed - may be increasing. In the attempt to spur the growth of private firms, governments in Eastem Europe have also given generous tax exemptions, e.g. to foreign investors. The introduction of market-based indirect taxes also has required a shift away from very high turnover taxes to more reasonable and explicit tax rates. 2.2. Coglapse of the CMEA Trade The disintegration of exports to the CMEA region and its iinpact on output and public revenue is an important focus in the analysis. To model it correctly, the extent of decline is briefly reviewed here. Because of the problems of calculating the right exchange rates for the CMEA trade, it has been difficult Go: Revenue Uncenailny 7 to assess the recent historical export performance in Eastern Europe. Figure 4 shows exports in 5 countries using preliminary estimates in Marer et. al. [1992] from 1986 to 1990, extrapolated to 1992 using data from the IMF International Firance Statistics. These figures are very tentative and should be used with care, specially the data points of 1991-92. Nevertheless, it is evident that - there was almost a total collapse of exports in Romania and a huge drop in Bulgaria; a smaller decline is registered in the former Czechoslovakia registered; and slightly beter performance were reported in Poland and Hungary. 150 ________ __ x --4uba o- Bu-ad E : Forme rCzechoslovakbiav| - aL ~~50 - - Hungaiy __ is > Lb _- Poland \ _ ; s- ~~~~~~Romania\ 19e6 19i7 1988 1989 4M9 t99t ts.92 Year C ~ ~ ~ ~ ~ ~ ~ ~ 'ilgure 4: E xports in Easten Europe (in current U.S. dollas) Figure 5 displays the shares of CMEA trade in exports. The shares are generally and rapidly declining. The two countries baving difficulties switching away faom CMEA trade, Romania arxi Bulgaria, are also the ones experiencing the most difficulties in their total export performance. 'Me share of CMEA exports still fcll drmatically in the other countries: from 53% to about 5% in the case of Poland, 589% to 15% in Hungary, and 60% to 22% in the forFer Czechoslovakiar Go. h gevenue Uncererinn dc 100 - -o- Bulgaria _-_- Hungary J-.- Poland lolal ffi < , _ _ ~~~~-&- Romania-} a' 20- - ~ ~ -Foie zehulvlci 1968 19i17 1968 1989 1990 1991i 1992 Your Figure 5: Share of CMEA Trade in the Exports of Eastern Europe 3. The Framework Two elements are important in the analysis. First, given the severe data constraint in Eastern Europe, the framework (a tax model) needs to be simple in the sense that it makes use only of available national income accounts and revenue infornation, while capturing at t.e same time the minimum, essential features of an open economy. The model employed in this paper is a modified version of the 1-2-3 general equilibrium model presented in Devarajan, Lewis and Robinson [1990 and 1993]. It is a simple one- country, two-activity and three-commodity model. In the original setup, a single open economy engages in the production of two goods, a pure domestic good and a pure export good, with a constant elasticity of transformation between the two. The domestic good also competes with impors in the domestic market Go: Revenue Uzceriainry 9 but with an elasticity of substitution that is finite, The modified version adds features of a transitional economy (e.g. treatment of CMEA exports) and various taxes into the model.3 Second, to highlight revenue uncertainty and introduce risk analysis in the simulations, the framework permits the sampling of random variables pertaining to key parameters or exogenous shocks. The Monte Carlo method (as well as the Latin Hypercube technique) is employed as in Go and Sinko [1994].4 Using this approach, alternative tax instruments and their revenue impact, encompassing many "what-if" scenarios, can be compared regarding the range of tax rates required - What is the most likely rate needed? What is the dispersion or variance? What are the relative welfare effects of different taxes? And what tax policy covers an acceptable risk? The salient features of the framework are discussed below. To carry out the model, data from any Eastern European country may be used. Mainly because of data availability, a Czechoslovakia-like economy with 1989 as the base year is employed for illustrations (see Table 1). However, the conclusions are drawn and applied to transitional economies in general. A list of equations appears in the appendix. A reference such as 'A2' means equation '2' in the list. 3.1. Production and CMEA Exports To account for the importance of CMEA trade in the framework, gross output in the economy, GDP, is defined as the sum of exports to the CMEA region, exports to the rest of the world excluding 3Denizer and Gelb (1992) also use a simple general equilibrium framework with dualistic (rural-urban) structure to look at transformnation issues in Mongolia. 4Go and Sinko [1994J implements Monte Carlo sampling method in a simple tax model of an open economy (Sri Lanka). A discussion of the statistical inference and confidence set in the context of CGE models and simulations when parameters are uncertain can be found in Abdelkhalek and Dufour (1993). Go: Revenue Uncerwainot 10 CMEA trade, and domestic goods (equation AI). CMEA exports, taken as exogenous, are imnportnt risk factors subject to uncertain market loss or price collapse. Their preduction has a set price and a fixed- coefficient technology of labor and capital, reflecting its non-market nature (equation A2-A3). Net output, defined as GDP Ier CMEA exports, is a CES function of primary inputs and is dependent on output and input prices (equation A4-A6). The production of net output must also be allocated between domestic goods and non-traditional exports (rest of the world) based on a constant elasticity of transformation (equation A8-A10). In the setup, a fail in CMEA exports releases labor and will lead to a fall in real wage in order to maintain employment (equation A7). Prices of net output and domestic goods will fall. Net output shifts outward, absorbing excess labor and increasing non-traditional exports (Figure 6). The extent of the shift and how much new exports will grow depend on how much of the freed resources are reallocated, the relative price of domestic goods and exports, and the CET elasticity. However, even if net output increases, GDP may still fall because the insalled capital used in the production of CMEA exports is fixed. D~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Errawj Figure 6: CET Transfonnation Between Domestic Goods and Other Exports Go: Rewene Uncertainry 11 Redistributing that capital mimnics the effects of privatizing assets, i.e. making them productive again.' Alternatively, it is possible to postulate that some of the lost output is prolonged and supported (hoarded?) by government out of social concerns, thus indirectly subsidizing wage earners and socialized enterprises with the prospects of a runaway fiscal deficit. 3.2. Taxes and The Rest of the Model Tax revenue consists of the major types of direct and indirect taxes - domnestic indirect taxes (sales/exciseNAT), import tariffs, export duties, payroll tax, incone tax on capital and odter income. The rest of the framework follows the standard 1-2-3 design. Imperfect substitution characterizes the competition between domestic goods and imports. This is reflected in the CES (Armington) function between domestic goods and imports (equation Al1-A13). The CMEA shock on the import side may be addressed by a rise in import prices. Personal income is the total of factor income (net of tax) from labor and capital plus transfers from the government and abroad (equation AM4). A simple Keynesian-type consumption function defines consumption at market prices (inclusive of sales tax) as a fixed proportion of personal income after tax (equation A15). Current account balance (equation A16) is the residual of imports less exports at world prices, adjusted for grants and remittances from abroad. Domestic demand (equation A 17) consists of private consumption, investment, and govemment consumption. Public savings are endogenously determined in the government budget (equation A19) as the balance of tax revenues (equation A18) plus foreign grants (exogenous) less government consumption (exogenous) and tansfer to the household (exogenous). By Walras' law, the savings-investment identity (I ! Sp + Sg + Sf) is implied slf data are available, output can also be distinguished in terms of socialized and private enterprises so that the effects of privatization can be looked at more directly. It will not change the qualitative results of the paper however. Go: Revenue Unceriainty 12 by the above equations. Foreign savings (Sf = eB) are presently fixed, so that the model is savings-driven.6 3.3. Risk Factors Each uncertainty is defned by a normal probability distribution, N(p,a,min, max), with mean IA and standard deviation a.? The distribution is truncated with a minimum (mnu) and a maximum (nmax) liiit to rule out extreme and infeasible values. Several sources of risk can be defined in the framework: 1) Key parameters like the trade elasticities: A high CET elasticity, for examnple, corresponds to a more responsive and greater capacity to export when relative puice are more attractive. The elasticity of substitution between domestic goods and imports is another potential risk factor. A value of iess than one implies that the real exchange rate will depreciate during an import price shock so that additional exports are encouraged to pay for tie more expensive imports. This is usually the case in developing countries and may be the case in transitional economies (see Devarajan et. al. [1993].) 2) Exogenous shocks like the collapse of the CMEA markets and the rise in import prices as described above. 3) Adjustment cost in moving assets from the declining sector to the rest of the economy.' 6 The nominal exchahg-. rate (fixed at 1.00), hence implicidy the world prices of foreign goods, is the numeraire in the model. 7Other probability distributions, such as log-normal, uniform, poisson, beta etc., are also possible. 'Another possible risk factor is the colection problem associated with the introduction of market-base taxes during the transition. This factor will create a wedge between statutory and effective taxes but may not add to the analysis. For this reason, it is not included in the simulations. It should be understood that tax rates in the simulations refer to collection rates. Go: Revenue Uncerainty 13 4. Simulations 4.1. Trade Shocks The first experiment deals with two uncertain trade shocks - a sharp decline in CMEA exports and an increase in import prices. CMEA exports are expected to decrease from 21 percent of GDP at factor prices to 10 percent - a market loss of about half the original size. This market loss may vary by a standard deviation of 0.025 (1/4 of the expected value). Its range has arn upper limit equivalent to the base year level (not permitted to rise) and a lower bound equivalent to 5 percent of GDP, i.e., N(0. 10, 0.025, 0.05,; 0.21). Import prices (base year = 0.95) are expected to rise by 30% with a maximum and minimum gain of 47% and 21%, respectively, i.e., N(1.24, 0.05, 1.15, 1.40). These shocks are comparable to those observed in the Section 2.2. In addition to the trade shocks, economic responses as defined by the trade and output elasticities also vary. The parameters chosen are generally on the low side, reflecting the economic rigidity of a transitional economy in reallocating resources: the elasticity of transformation between domestic goods and exports is assumed to be N(0.50, 0.02, 0.30, 0.90); the elasticity of substitution between domestic goods and imports is defined as N(0.80, 0.02, 0.50, 1.10); and, the elasticity of substitution between labor and capital in net output is N(0.60, 0.02, 0.30, 0.90). Monte Carlo sampling is used to draw good distributions of the risk factors. About 300 iterations and simulations are made. In each iteration, values of the shocks and parameters are sampled randomly from the above probability distributions and a solution is computed. The results for key variables are reported in Table 2, which shows their expected levels, standard deviations, and various percentile values. The experiment mimics the collapse of CMEA trade in a changing enviromnent. The pattern of Go: Revenue Uncertaingy 14 Simulated Decline of Output 0.08 - - - - - - - - -AI p 921 0.70 _ _,_ _ - - EtcedGP - - - - - - B B0 XL0 2 ------- 0 86 s8 90 92 94 96 98 GDP in a Changing Economy (Base Year = 100) Flgure 7: Impact of Trade Shocks on Output output and revenue decline in Eastern Europe is duplicated in Figure 7 and 8, respectively. Real output is expected to decline by about 8 % GA). Tax revenue, on the other hand, is expected to fall by 22 % (u).9 The effects of uncertainty in the outcomes are reflected in several ways. A straightforward statistical interpretation indicate that the risk of output deviating ± 2.2 percentage points {one standard deviation) around the mean is 68.3 percent. The chance of output declining anywhere between 8 to 14 percent (bottom half of the distribution) is 50 percent. The average deviation of the revenue fall is higher than output. The likelihood of revenue falling by ± 4.2 percentage points (la) around the expectation is about two-third and by ± 8.4 percentage points (2a) is about 95%. There is an even risk that tax revenue rnay fall from 22 to 31 percent. The scenario above assumes that real wage is flexible and labor is free to move from the collapsing sector to the rest of the economy. This assumption leads to the following - net output will increase by 7.8 percent (p) ± 1.6 percent (a) while exports to non-CMEA region will rise by as much 21.8 percent (U) 'Tax revenue is reported in real terms. See Section 4.2, Case I, for definition and discussion. Go: Revenue Uncertainy 15 Simulated Decline in Tax Revenue 0.14 S. 8 -- - - - - - - - - - - - R 68 73 78 83 8a Tax Revenue in a Changing Economy (Base Year = 100) Figure 8: Impact of Trade Shocks on Revenue ± 4.4 percent (a)- If the assumption is to be violated, the rise in net output and other exports may be reversed; the fall in GDP and tax revenue will be much more than reported. The scenario assumes conservatively that exports to the CMEA region will shrink by half. If the collapse of the CMEA trade is near total, the expected fall in real GDP can reach 18% in magnitude. Consumption and investment will contract on average by 16% and 49%, respectively. Such a general breakdown of tax bases can lead to a 38% fall in tax revenue. Hence, raising revenue remains a critical problem in transitional economies. 4.2. Which Tax to Raise? What if the government tries to raise or maintain revenue? Which tax instrument is better in an uncertain world? The once-off systemic change in Eastem Europe wil eventually require a new, market- based tax system. However, introducing a market system is a slow process and not all reforms can be Go: Revenue Uncerwiaty 16 implemented at once due to legislative, political, administrative, and other institutional bottlenecks. Because of the serious revenue problem in the transition, governments have to view and sequence prospective reforms by their short-term impact on revenue as well. Abstracting from the institutional issues and assuming that the current revenue structure can be converted to market-based taxes without problems, this section examines three types of taxes - domestic indirect tax (e.g. sales/excise/VAT), import tariff, and capital income tax - under three reasonable possibilities. In the first case, the govermnent tries to maintain revenue by fine tuning individual taxes in response to the shocks. In the second case, the government has to set tax policy in advance and guesses correcdy the distributions of the shocks and parameters; however, it uses an altemative but commonly-used revenue target, the ax-to-GDP ratio. In the third case, the government guesses wrong. Case I: The Government Maintains Revenue In an economy experiencing severe trade shocks, what is the appropriate revenue target (R)? The open-economy framework used here takes the nominal exchange rate (er) as the price anchor and defines real reveeme as Rler. Thus, maintaining revenue or equal yield means keeping this ratto constant (at the base year level). Since the model deals only with relative prices and keeps er fixed (i.e., no nominal inflation), maintaining revenue should be interpreted to mean that revenue is keeping pace with nominal devaluations of the exchange rate. To compare taxes, a broad indication of welfare is aggregate consumption. This is because the framework is a static one (no dynamic decision on savings) and all income (after fixed rates of sales tax and savings) is consumed on one aggregate good. Investnent, which is driven by available savings, is also affected by the uncertain pubiic revenue. The main findings (presented as IL ± a) regarding the distribution of tax revenue, consumption, and investment are Go: Reeue Uncertainy 17 summarized below:"0 Case I: Maimtaing Revenue Tax Instrument Ave. Tax Rate Consumption Inveshnent 1) Dom. Indirect Tax (22.0)' 39.2 ± 4.1 77.5 ± 3.3 80.9 ± 2.6 2) Import Tariff (6.0)' 31.1 ± 6.4 83.4 ± 2.2 71.6± 3.5 3) Capital Income Tax (32.0)' 72.8 ± 9.8 73.9 ± 4.5 85.9 ± 2.8 Base-Year Level 100 100 Wagures in pauhasis refer to tax rates in die base year. The tax rate needed is quite high for each instrument. For example, the expected rate of domestic indirect taxation is 39.2 percent (an increase of 78.2 percent); 31.1 percent for import tariffs (an increase of 4.9 times from the low base); and about 73 percent for capital income tax (an increase of 2.3 times). The variation around the expected rate is substantial, ranging from 4.1 to 9.8 percentage points. The highest variance is in the capital income tax, which also has the highest exected tax rate. To be able to cover 95 percent of the revenue contingencies, the average rate of corporate tax will have to be raised by two standard deviation, which, at the tax rate of 92.4 percent however, is extremely high. Import taiffs has the next highest variance but the lowest expected rate. Even at one standard deviation higher than the mean, the average import tariff (at 37.5 percent) is lower than the expected rate of domestic indirect taxes. Normnally, one would expect that domestic indirect taxes would be lower because of the broader base in the domestic market. Here, a 40 percent sales/VAT rate, or the 47 to 50 tax needed for a 95 percent likelihood, is very high indeed. I"Table 3 to 5 in the appendix present the reslts in greater details. Go: . Revnue Uncnory 18 The effects on constuption has a similar story. Import tariffs cause the least decline in consumption, followed by domestic indirect taxes, and capital income taxes." The impact on investment, on the other hand, is the reverse of consumption because, in part, with less consumption more savings are invested. The effects on revenue and consumption seem to suggest that import tariffs are good temporary tax measures. Such a recommnendation is made by McKinnon [199 1. If collection problems are allowed to raise the nominal rates,'2 McKinnon's suggestion of a cascading tariff with a top rate of 100 percent, to be gradually brought down to a uniform and low rate over 5 years, seems within range and reasonable. The pattern of results points to tax issues peculiar to transitional economies. Two factors are critical: the trade shocks and the choice of Armington elasticity. The collapse of the CMEA exports reduces the base of domestic taxation by its deflationary impact on output, domestic prices, and factor prices. Tile import price shock when combined with an Anmington elasticity of less than one (because of its plausibility in transitional economies) also tends to reinforce this effect. Figure 9, for example, ilustrates the impact of a 10% import price shock on the rate of change in domestic prices, PD., with altemative Annmington elasticity, a, and CET elasticity, a, in a 1-2-3 framework.'3 The f'all in domestic prices relative to world prices is an indication that the real exhange rate must fall in the transition so that output switching can occur and non-CMEA exports can be pushed while keeping the current account balance from deteriorating. Under the circumstances, import tariffs are favored to maintain real revenue "Although strictly not comparable, the average consumption in simulation 1 (the no tax case) is 88.3 with a standard deviation of 1.4. 12[t is possible to add less than 100 percent collection rates for various taxes in the framework. This may change the nominal rates of taxes but will not affect the results. '3Disregarding xaxes, the underlying relationship is 45D = * ' (.-1) , where is the change in import prices. See Devarajan, Lewis, and Robinson (1990 & 1993) for the algebraic derivation and discussion. Go: Revenue Uncertainty 19 5~~~~~~~~~~~~. Figure 9: Import Price Shock, Trade Elasticities, and Domestic Prices defined in dollar terms (price anchor). The traditional corporate tax base, i.e., profits of state-owned enterprises, is shrinkdng so rapidly that any attempt to restore revenue from this source will require excessive taxation. Likewise, the fall in output and domestic prices dininishes the revenue potential of domestic indirect taxes. Moreover, import tariffs provide at the margin some protection of factor income from the contractionary impact of trade shocks; hence, it will allow for greater consumption relative to domestic taxes, which either reduce consumption directly (direct taxes) or indirectly by increasing the cost of consumption (domestic indirect taxes). The above result is striking in the sense that, witlout bringing in the issue of administrative ease, imports tariffs are still favored because of the type of shocks and trade elasticities postulated in the transition. Note that this is not an argument for imposing high and permanent import tariffs, which have become synonymous with excessive protection, inefficient production, and rent-seeking behavior. In the Go: Reeme Uncerraincy 20 long run, other factors become important: market orientation may finally take hold and supply, operating more responsively to price signals, can expand rapidly. In that situation, domestic taxes will tend to have a broader base and greater efficiency in generating revenue. Until the immense erosion of domestic tax bases are reversed, however, irnport tariffs a :1 play a vital role. Case II: The Government Guesses Correctly In this experiment, the govermnent needs to prepare tax legislation in advance but does so with correct expectations. Tax policy, which can only be defined once and at a single rate, is set at the expected values of the shocks and parameters. However, there are risks around the expected shocks and parameters. The issue is whether such variation may generate effects that favor certain taxes: some taxes may be more stable (e.g. in revenue); the expected revenue may be different whet risks are introduced (i.e., the relationships are highly non-linear.) In addition, the government looks at another revenue target, the tax- to-GDP ratio, and plans to raise the ratio by 5 percentage points from 59.2 percent in the base year to 64.2 percent. Using expected values of the shocks and parameters as defined in Section 4.1, the tax rate required are as foliows: domestic indirect taxes need to be changed from 22.0% in the base year to 25.7 %; importtriffs,from6.0% tol3.5%; capitalincometax, from31.7%to40.3%. Fixingeachnew tax ruxe in turn, the risk factors are introduced altogether and simulations are conducted repeatedly using Monte Carlo sampling. The results, shown below, indicate that consumption is slighdy higher in the import-tariff case. Hence, investment is also less with reduced savings when compared to the other two tax instruments. However, the differences are numerically small. This is also true the revenue target - the tax/GDP ratio. The results are generally stable when the Monte Carlo sampling is increased from 300 to 500 iterations. Go: Revenwe UwxeriaintY 21 Thus, when the government formulates its fiscal policy with correct expectations, it appears that it does not matter which tax instnrment is chosen. Case 13: The Government Guesses Correctly Tax Instrument TaxzGDP Consumption Invesament 1) Dom. Indirect Tax 64.69 ± 0.15 85.71 ± 1.36 69.38 ± 4.38 2) Import Tariff 64.25 ± 0.19 86.58 ± 1.36 67.89 ± 4.23 3) Capital Income Tax 64.23 ± 0.18 85.21 ± 1.37 69.98 ± 4.44 Memo: Base-Year Level 59.24 100 100 However, the tax-to-GDP ratio is a misleading revenue target when the tax base is shrinkdng. In fact, real revenue as measured by Rev/er is (82.8 ± 4.1) for domestic indirect taxes, (85.0 ± 4.6) for import tariffs, and (82..8 ± 4.4) for capital income tax. While the increased tax efforts as measured by the tax-to-GDP ratios are attained, real revenue levels in terms of Rev/er are significantly below the base- year 100 but higher than the 78.14 in the no tax policy case in Section 4.1. Moreover, import tariffs still appear better in generating real revenue. On the other hand, if Revler is the target instead and held fixed, the tax-to-GDP ratios will vary. Hence, it does matter which revenue target is used. Case ll: The Government is Overly Optimistic Achieving a particular tax-to-GDP ratio or a real revenue target as in Case II requires absolutely correct expectations - not just of real GDP, but also about the external shocks, the responsiveness of the Go. Revenue Uncertainly 22 economy, and how they may affect relative prices (e.g. the real exchange rates between domestic and foreign goods), and various components of GDP that utimately determine revenue generation by individual tax instruments. This is often a heroic assumption. Governments in Eastern Europe are confronted by pressing reforms in all fronts and are just learning GDP accounting properly. What if simplistic and wrong assumptions are made in the budgetary preparations? Do import tariffs minimize the downside risks relative to other taxes? Take the case in which the government anticipates the external shocks correctly but the economy is less responsive that what government originally expected. In particular, Case II is repeated: the governmet fonnulates the same tax policy given the same expectations about the shocks and parameters; however, except for the Armington elasticity, all expectations are correct; the correct Armington elasticity is only half of what is anticipated and has greater standard deviation. i.e., a is changed from N(0.80, 0.02, 0.50, 1.10) to N(0.40, 0.05, 0.25, 0.90). The results are sununarized as follows: Case m: The Government is Overly Optimistic Tax Instrument Tax/GDP Rev/er Consumption Investment 1) Dom. Indirect Tax 65.21 ± 0.29 72.51 ± 4.84 83.99 ± 1.64 65.12 ± 5.25 2) Import Tariff 64.23 ± 0.20 74.62 ± 4.81 84.51 ± 1.78 64.26 ± 5.38 3) Capital Income Tax 65.28 ± 0.32 72.08 ± 4.80 83.49 ± 1.61 65.83 ± 5.30 Memo: Base-Year Level 59.24 100 100 100 The tax-to-GDP ratios are similar to those in Case II. The results suggest that the target of raising the ratio by 5 percentage points is sdll achievable and appear comparable for altemative tax instruments. Go: Revenie Uncerairnt 23 As in Case II however, the tax-to-GDP ratios do not reflect the underlying and severe changes in the economy. Real revenues in fact are now all lower than in Case 11. Morever, import taxes still do better. Consumption levels are also down but less with imports. Investment, driven by savings, have the revcrse story like in Case II. In all variables, the risks or standard deviations are higher. Another source of possible error is the government's forecasts about the external shocks. For example, if die collapse of the CMEA exports and the rise in import prices turn out to be higher, e.g. near 80% and 60%, respectivcly (instead of the projected 50% and 30%, respectively) and everything else the same as above, real revenue will fa ll to much lower levels but with less decline in the import-tariff case: 59.1 in the domestic indirect tax case; 61.4, import tariffs; 59.2, capital income tax. 4.3. Downsizing the State Sector The simulations so far sidestrack some fundamental questions - Whether public expenditures in transitional economies are already 'optimal'? Whether they should be at all Financed by high taxes? And whether other measures, such as privatization, will help? These issues are examined next. The extremely high tax-to-GDP ratios resulting from keeping the current revenue structure instandy confirm that tax policy should never be divorced from the expenditure fimction of government. As the former socialist countries move towards market-oriented economies, public expenditures also need to be examined, cut back, and restructured. Any attempt to finance current levels is not sustainable and will only lead to reasonably high taxes, as demonstated above. One single item like government transfers and subsidies, for example, can take up 45 % of gross value added at factor prices (see Table 1.) Just eliminating a third of the revemne demanded for that item alone, even while giving up non-tax revenue, will reduce the required tarff rate in Case I from 31.1 % to a much lower 85%. Go: Ramnm Unrainia 24 Required Import Tariffs 0.12 I 0.1-~~~~~~~~~~~~~Expected Vatue R 0.06 O 0- .0 -- - - - I--- L --- --- --- -- 0 B 0.04 --- -- -- I ------ I= 0.02 .--___-- 1__ 14 is 24 29 34 39 Tariff Rate (in percent) Figure 10: Expected hnport Tariff Rate with Privatization No less important is to move assets out of the collapsing sector, privatizing and makng them productive again (see, for example, Blanchard et. at. [1991]). The next experiment shows that even if there are substantial but uncertain adjustment costs in the order of 25 % with a standard deviation of 10 %, the benefits of 'privatizing' assets quickly are still substntial (at least the portion associated with the output loss). Tne resulting expansion of the market base will reduce the imposition of additional taxes. 1" Looking at the tariff instrument again, Figure 10 shows the distribution of the tax rate with such a privatization scheme; the average rate falls to 27.3 percent. It should be recalled that the stardtin point of this result is the expecttion that the CMEA market will shrink by only 50 percent. a very conservative scenario. If the loss is total and if, in addition, a significant portion of output from socialized enterprises cannot be sold in the domestic market, then the gains of privatization will considerably be more. Since 14It is entirely possible that privatization will not lead to an expnasion of the tax base initially because of collection and evasion problems. Such possibility reemphasizes the need for temporay tax measures like an import surcharge. Go: Revenwe Uncerainy 25 privatization is intricately associated with the reduction of govermment subsidies, their combined effects on public finance further make the issue of reducing the state sector critical. 5. Conclusions The purpose of this paper is to formulate a framework for making revenue assessments and for examining altemative tax instruments in a changing environment. The model is used to explain much of the output and revenue faU in transitional economies. The terns-of-trade shocks from the collapse of the CMEA trade as well as the rigid but uncertain economic responses in transitional economies are all important factors. nsofar as a new tax system will take time to institute and revenue is critical in the transition, the results indicate that import tariffs are more effective ;ian other traditional tax instruments in raising revenue, especially if real revenue is defined in dollars terms (the price anchor). The contraction in domestic output and prices and the devaluation of the real exchange rate needed in the transition are significant reasons that favor imports as a tax base over other revenue sources. To emphasize the ransitory nature and reversibility of the policy recommendation, import tariffs should be implemented in the form of a temporary, uniform, import surcharge. The above conclusion seems to hold whether the government formulates tax policy with correct or incorrect expectations. The choice of revenue target, however, matters. All tax instruments will do almost equally well if the commonly used tax-to-GDP ratio is the target. It is a misleading measure, however, since the ratiD does not reflect the immense erosion of domestic tax bases in the economy and how real revenue in absolute level may actually be decreasing rapidly as a result. The revenue problem and uncertainty can also be viewed as a necessity towards downsizing the Go: Revenue Jncerraby 26 large state sector from the previous control economy and as a consequence of redirecting trade away from former non-market partners. Hence, the results emphasized that restoring revenue should never lead to maintaining subsidies towards non-profitable state enterprises or other public expenditures no longer relevant in a market system. Doing so will only lead to unreasonably high taxation. The outcome hinted at the importance of privatization in expanding the tax base and reorienting or cutting public expenditures. Furthermore, policy uncertainty, i.e., the likelihood of success (or reversal) and the credibility of reform, although not examined here, is an important risk hctor by itself. Some of these issues are best investigated firther in a dynamic stochastic framework. Future research will examine policy uncertainty, privatization, and the likelihood of refbrm success as determinants and risk facwrs in investment and growth. Go: Rvemre Uncertainty 27 References Abdelkhakek, Touhami and Jean-Marie Dufour (1993). Infdrence statistique pour les mod6les de simulation et les modeles calculables d'equilibre g6n6ral. C.R.D.E. et Departement de sciences 6comiques, Universitd de Montr6al (Novembre). Berg, Andrew (1993). Measurement and mismeasurement of economic activity during transition to the market, in: Mario I. Blejer, Guilermo A. Calvo, Fabrizio Coricelli, and Alan H. Gelb, eds., Eastern Europe in Transition: From Recession to Growth, World Bank Discussion Papers No. 196. Blanchard, Olivier, Rudiger Dornbush, Paul Krugman, Richard Layard, and Lawrence Summers (1990). Reform in Eastern Europe (the WIDER Report). Helsinki: World Institute for Development Economics Research. Bruno, Michael (1993). Stablization and reform in Eastem Europe: preliminary evaluation, in: Mario I. Blejer et. al., (see Berg [I993] above). Denizer, Devdet and Alan Gelb (1992). Mongolia - privatization and system transformation in an isolated economy, World Bank working paper No. 1063. Devarajan, Shantayanan, Jeffrey Lewis, and Sherman Robinson (1993). External shocks, purchasing power parity and the equilibrium real exchange rate. The World Bank Economic Review, vol.7, Ino. 1, 45- 63. Devarajan, Shantayanan, Jeffrey Lewis, and Shemian Robinson (1990). Policy lessons from trade-focused, two-sector models. Jounal of Policy Modeling, 12(24): 625-57. Fischer, Stanley and Alan Gelb (1990). Issues in socialist economy refurm, World Bank PRD Working Paper No. 565. Gelb, Alan and Cheryl W. Gray (1991). The transfomation of Economies in Central and Ehstem Europe - issues, progress, and prospects. Policy and research series, No. 17, the World Bank. Go, Delfin and Il'ekka Sinko (1994). Terms-ol-trade shocks and uncertainty in a small open economy: Monte-Carlo sirnulations in a general equilbrium model. Forthcoming. Holzman, Robert (1991). Tax reform in transition: central policy issues. Paper presented in the conference, public finance in a changing political environenmt, International Institute of Public Finance, 47th Congress, Leningrad, 26-29 August. Marer, Paul, Janos Arvay, John O'Connor, Martin Schrenk, Daniel Swanson (1992). Hsoricalyplanned economies - a guide to the data, a World Bank publication. McKinnon, Ronald I. (1991). The order of econonic liberaizaton -financial control in the transitiom .o Go: Rnwzue Uncertainty 28 a market economy, John Hopkins University Press, Baltimore. McKinnon, Ronald I. (1992). Taxation, money, and credit in a liberalizing socialist economy, Economics of Planning, (U.K.), 25(1):97-112. Rodrik, Dani (1993). Making sense of the Soviet trade shock in Eastern Europe: a framework and some estimates, in: Mario 1. Blejer et. at. (See Berg [1993] above). Sachs, Jeffrey (1991). Accelerating privatization in Eastern Europe: the case of Pcoand, in: Proceedings of the World Bank annual conference on development economics, 1991, supplemert to the World Bank Economic Review and the World Bank Research Observer, 15-30. Schaffer, Mark (1993). The enterprise sector and emergence of the Polish fiscal crisis, 1990-91, World Bank working paper No. 1195. World Bank (1991). Transition. Special issue - index of selected articles, new working papers and books. Zou, Heng-Fu (1993). On the dynamics of privatization, Polciy Research Department, the World Bank. Go: Reweme Unrertainy 29 Bil. Kcs. Output=1 Si1. Kcs. Output = 1 1 NatIonal Accounts 3 Fiscal Account Output (Value Added) 840.46 1.00 Revenue 484.10 0.58 Wages 479.06 0.57 NonTax 37.90 0.05 Current Expenditure 443.90 0.53 GDP at market prices 977.76 1.16 Goods & Services 253.94 0.30 Private Consumption 390.14 0.46 Interest Payments 22.07 0.03 Public Consumption 253.94 0.30 Transfers & Subsidies 167.93 0.20 Investment 292.00 0.35 Capital Expenditure 61.60 0.07 Exports 371.25 0.44 Fiscal Balance -21.40 -0.03 Imports 329.57 0.39 2 Tax Revenue 4 Balance of Payments Sales & Excise Tax 125.80 0.15 Exports - Imports 41.68 0.05 Import Tariffs 11.50 0.01 Net Prafits & Dividends 0.00 0.00 Export Duties 0.00 0.00 Interest Paymrents 0.00 0.00 Payroll Tax 107.20 0.13 Net Private Transfers 0.00 0.00 Personal Income Tax 59.00 0.07 Net Offidal Transfers -14.91 -0.02 Capital Income Tax 137.00 0.16 Current Account Balance 26.77 0.03 Total 440.50 0.52 Extemal Debt 324.52 0.39 Debt Service Payments 50.03 0.06 Table 1: Data of the Former Czechoslovakia, 1989 Go: Rvemae Uncertainty 30 Table 2: Terms-of-Trade Shocks Slmulatlon Sta2433.0s 5%f.ls4 Year = .00%J IterNtions: 300 Variabl0: GDP Exports Domestl0 Imports N.t Consump- Inv0st- Tax Pothea) Good Output tlon mant Revenue Minnum= 86.80% 109.35% 100.63% 56.03% 102.91% 84.70% 54.95% 68.83% Poximum- 97.7E% 133.12% 103.69% 73.98% 105.96% 91.93% 76.77% 90.71% Mean=- 92.12% 121.83% 102.24% 62.09% 107.60% 88,27% 65.82% 78.15% Std D90aton 2.17% 4.37% 0.43% 3.70% 1.56% 1.375% 462l% 4.17% Variance - 0.05% 0.19% 0109% 60.14% 0.02% 8.05% 05189% 0.17% skownessr 94.07.% -11.62% *41.49% 11.34% -26.86% 7.2% -7.88% 75.6ff kurtosh - 298.57% 276.29% 327.32% 2601.11% 275.40% 278.18% 255.80% 278.351% Percentile Valuos UPerc= 9 88.39% 114.35% 101.24% 56.70% 104.96% 86.03% 59.13% 71.38% 1OPerc - 89.23% 116.409% 101.69% 57.89% 1M7.75% 86.39% 60.02% 72.84% 16Perc= 89.79% 117.37% 101.79% 56.910% 106.01% 86.815% 61.15% 73.84% 2OPerG- 90.19% 118.24% 101.85% 59.54% 1068.7% 87.05% 62.06% 74.38% 2SPerc= 90.52% _ 123.95% 101.96% 60.03% 108.64% 87.35% 62.95% 74.97% 3OPerjc= 90.91% 119.62% 102.04% 60.53% 106.88% 87.52% 63.B1% 75.69% 3SPerc - 91.30% 120.20% 102.09% 60.98% 107.10% 87.70% 64.10% 76.51% 4OPerc= 91.59% 120.66% 102.15% 61.33% 107.33% 87.82% 64.53% 76.99% 45Pere- 91.88% 121.27% 102.24% 62.01% 107.34% 88.04% 75.10% 77.41% 50Aerc = 92,22%f 121.93% 102,28% 82.55% 107.75% 38.31%e 65.73% 77.88% 65perc = 92.40% 122.83%6 102.33% 63.07% 107.95% 838.50% 66.57% 78.47% 60Pe,rc- 92.71 % 123,03% 102.40% 83.66ff 108.14% 88.63% 67.01 % 79.06% 65Perc - 92.999% 123.52% 102.43% 64.18% lOB.35% 8B.83% 67.45% 79.53% 70Perc - 93.38% 124.26% i02.47S6 64,73% 108.56% 89.08% 68.42% 80.34% 7SPOer - _ M64% 125.13% 102.53% 65.36% 1 OB.80% 89.24% 68.94% _ 80.82% BOPerc = 93.B9% 1256.4% 102.81 % B5.93% 109.04% 89.52% 691.70% 81.649% 8bBerc s 94.40% 128.33% 102.88% 66.69% 109.34% 89.65% 70.11% 82.43% 90Perc = 94.92% 127.39% 102.76% 67.26% 109.68% 80.05% 71.25% 83.55% 95Petr- 95,61% 128.97% 102.88% 69.14% 110.12% 90.51 % 72.72% 85.28% Paga 31 Table 3: Raising Revenue from Sales Tax (Revenue-Neutral Simulation) Simulation Statist/cs (Base Year - 100% except Sales Taxi Iterations: 300 Varable: GDP Nat Exports Domestic Imports Consump- Invest- PD Sales Tax Output Good tion ment Rats MlnImum- 86.83% 10.67% 109.03% 100,34% 52.66% 69.o5% 73.34% 60.13% 28.05% Mexsmumn- 98.01% 129.89% 132.59% 103.34% 74.39% 87,39% 85.97% 85.07% 50.16% Mean 92.12% 120.72% 121.80% 102.27% 62.71% 77.46% 80.88% 70.69% 39.25% Std Deviation = 2.14% 3.78% 4.29% 0.44% 3.66% 3.32% 2.57% 4.58% 4.15% Variance- 0.05% 0.14% 0.18% 0.00% 0.13% 0.11% 0.07% 0.21% 0.17% ISkowness- -2.71% -19.92% -15.62% 56.69% 3.52% 8.34% 30.10% 33.75% 3.60% Kurfosls = 261.41% 279.59% 274.53% 382.59% 289.08% 277.06% 273.B6% 292.56% 267.23% Percentile Values SPare- 88.4696 114.15% 114.64% 101.49% 56.82% 71.97% 76.24% 63.15% 32.35% 10Perc= 89.23% 115.55% 116.26% 101.69% 57.83% 73.17% 77.57% 64.89% 33.80% 75Perc = 89.S2% 116.76% 116.98% 101.82% 58.84% 74.01% 78.18% 65.92% 34.83% 2OPera - 90.26% 117.55% 118.10% 101.90% 59.63% 74.38% 78.67% 66.98% 35.79% 25Perc = 90.70% 118.08% 118.87% 102.00% 60.01% 75.22% 7S.12% 67.56% 36.27% 3OPerc= 90.93% 118.61% 119.41% 102.06% 80.53% 75.60% 79.40% 68.07% 36.90% 35Perc= 91.28% 119.43% 120.29% 102.15% 81.23% 76.15% 79.84% 68.91% 37.65% 4OPerc- 91.55% 119.98% 120.91% 102.21% 61.79% 76.45% 80.29% 69.20% 39.18% 4SPerc = 91.87% 120.52% 121.45% 102.27% 62.20% 76.99% 80.63% 69.66% 38.64% SOPero- 92.11% 121.04% 121.97% 102.30% 62.71% 77.50% 81.03% 70.18% 39.37% 5SPerc= 92.43% 121.45% 122.51% 102.37% 63.31% 77.76% 81.32% 70.76% 39.73% 6OPerc= 92.68% 121.85% 122.86% 102.41% 63.64% 77.98% 81.68% 71.50% 40,23% 6SPerc- 93.06% 122.08% 123.50% 102.45% 64.26% 78.65% 82.05% 72.13% 40.92% 7OPerc m 93.37% 122.80% 124.03% 102.50% 84.90% 79.43% 82.43% 73.08% 41.35% 75Pero= 93.58% 123.21% 124.72% 102.56% 66.24% 79.92% 82.71% 73.79% 42.01% 80Pero = 93.95% 123.79% 125.54% 102.64% 65.96% 80.25% 83.02% 74.43% 42.52% BSPerc= 94.40% 124.55% 126.06% 102.71% 68.38% 7 80.98% 83.60% 75.45% 43.88% 9OPerc - 94.91% 125.41% 127.46% 102.80% 67.08% 81.88% 84.25% 77.03% 44.61% 9SPerc = 95.59% 126.77% 128.74% 102.93% 68.53% 82.73% 84.98% 78.69% 46.37% Page 32 Table 4: Raising Revenue from Import Tariffs (Revenue-Neutral Simulation) Simulaion Statlaslcs 103se Year - 100% except Tariff Rate) Iterations: 300 Variable: GDP Not Output Exports Domestle Imports Consump- Invest- Tariff Goods tion ment Rate Minimum- 89.12% 103.28% 107.68% 101.81% 51.08% 78.18% 63.41% 15.23% MaXImum= 97.62% 110.99% 124.25% 108.72% 70.77% 88.90% 79.59% 48.39% Mean=- 93.15% 107.74% 116.42% 104.64% 60.35% 83.41% 71.63% 31.07 % SWjDevatZon= 1.75% 1.65% 3.17% 0.96% 4.0B% 2.22% 3.49% 6.37% Vrlrancea= 0.03% 0.02% 0.10% 0.0196 0.16% 0.05% 0.12% 0.41% Skewness- -4.35% .19.30% -4.12% -31.78% 16.34% 2.72% .15.55% 16.98% lKursoshJ- 248.29% 259.58% 258.70% 277.07% 251.08% 250.25% 244.64% 250.E0% Percenlile Values 5Parc= 90.03% 105.04% 111.27% 102.95% 63.59% 79.69% 6S.44% 21.32% lOPerc- S0.75% 106.64% 112.20% 103.26% 55.13% 80.57% E6E.56% 22.70% 75Porc- 91.17% 106.03% 112.84% 103.66% 66.84% 81.12% 67.68% 24.05% 2OPerc= 91.59% 106.38% 113,45% 103.77% 56.87% 81.40% 68.66% 25.22% 25Perc= 91.88% 106.63% 114.05% 103.98% 57.33% 81.77% 69.21% 26.37% 3OPerc- 92.19% 106.89% 114.80% 104.17% 57.91% 82.12% 69.72% 27.32% 35Perc- 92.f0% 107.19% 115.17% 104.34% 58.43% 82.44% 70.27% 28.47% 4oPetc- 92.73% 107.40% 116.711% 104.48% 59.26% 82.83% 70.72% 29.00% 45Perc= 92.96% 107.60% 116.05% 104.62% 59.60% 83.12% 71.23% 29.78% 6OPeor- 93.16% 107.78% 116.37% 104.68% 60.09% 83.44% 71.70% 30.81% 65Perc- s3.38% 107.97% 116.88% 104.80% 60.62% 83.67% 72.11% 31.41% 60Pe,c= 93.63% 109.13% 117.19% 104.92% 61.32% 84.01% 72.46% 32.62% 65Porc= 93.86% 108.41% 117.65% 105.05% 61.84% 84.25% 73.19% 33.26% 7OPorc- 94.12% 108.E3% 118.02% 105.18% 6i.35ff 84.60% 73.92%6 34.42% 75Pero- 94.43% 108.83% 118.58% 105.32% 63.18% 84.92% 74.37% 35.25% 8OPere- 94.71% 109.07% 119.21% 105.45% 63.93% 85.30% 74.64% 36.63% p6re = 95.06% 109.45% 119.88% 105.55% 64.83% 85.91% 75.41% 37.80% 19OPorc- 95.46% 109.74% 120.81% 105.8B% 65.69% 8B.50% 76.04% 39.73% 9SPerc - 95.89% 110.24% 121.42% 106.14% 67.19% 87.03% 76.95% 42.48% Page 33 Table 5: Raising Revenue from Capital Income Tax (Revenue-Neutral Simulation) Simuladron Statistics (Base Year - 100% except CapitalIncome Tax - tkl Iterations: 300 Variabl: GDP Exports Domeatlo Imports Not Consump- Invest- PD tk Good Output tion ment Minfmum= 86.95% 109.95% 100.71% 53.02% 103.15% 63.44% 76.72% 60.51% 48.69% MAxlmum= 97.51% 131,65% 103.18% 73.24% 110.95% 85.14% 92.27% 83.99% 95.19% Mean- 92.11% 121.82% 102.24% 62.82% 107.68% 73.93% 85.8E% 70.58% 72.75% Std DaviatIon_ 2.17% 4.39% 0.40% 3.81% 1.55% 4.47% 2.77% 4.78% 9.80% Variance = 0.05% 0.19% 0.00% 0.15% 0.02% 0.20% 0.08% 0.23% 0.98% Skewness- *6.17% .12.58% *59.99% .1.45% .25.07% 0.04% -41.01% 27.83% 7.02% Kurtoul = 253.63% 2M3.52% 382.34% 278.36% 271.75% 260.68% 310.90% 276.97% 258.04% Percentile Values 5Perc= 88.39% 114.32% 101,58% 56.62% 105.01% 86.07% 80.87% 63.09% 56.73% lOPerc_ 89.13% 116.27% 101.71% 57.82% 105.62% 87.78% 82.24% 64.13% 60.29% 16Perc= 89.74% 117.10% 101.85% 58.85% 106.03% 68.85% 83.00% 65.11% 62.04% 2OPerc 90.20% 118.15% 101.92% 69.58% 106.36% 70.03% 83.53% 66.44% 64.26% 2SPew= 980.87% 118.89% ¶02.00% 60.29% 106.64% 70.95% 84.01% 67.17% 66.07% 3OPe,a= 90.94% 119.47% 102.08% 60.81 % 105.89% 71.49% 84.36% 67.95% 67.05% 35Perc= 91.27% 120.19% 102.13% 61.34% 107.12%' 72.27% 85.04% 68.70% 68.44% 40Po,c= 91.60% 1 20.58% 102.16% 61.E6% 107.33% 72.76% 85.42% 69.06% 69.81% 45Perc- 91.86% 121.19% 102.23% 62.24% 107.53% 73.23% 95.79% 69.66s% 70.83% 6OPerc- 92.10% 121.99% 102.28% 62.71% 107.74% 74.09% 86.21% 70.40% 72.44% 65Perc = 92.46% 122.56% 102.32% 63.39% 107.93% 74.68% 86.50% 71.02% 73.78% 6OPerc- 92.70% 122.99% 102.35% 63.72% 108.14% 76.19% 86.78% 71.75% 74.88% 66Perc - 93,06% 123.65% 102.44% 64.46% 108.34% 75.91% 87.08% 72.44% 76.31% 7OPerc- 93.34% 124.12% 102.47% 64.96% 108.56% 76.43% 67.38% 73.07% 78.12% 75Percx 93.60% 124.93% 102.52% 65.49% 108.80% 76.81% 87.81% 73.65% 79.27% OOPere_ 94.03% 125.46% 102.57% 61.95% 109.05% 77.62% 88.24% 74.12% 81.40% 86'perc- - g4.35% 126.67% 102.62% 66.97% 109.32% 78.52% 88.69% 75.69% 83.38% 9OPerc= 94.85% 127.71% 102.73% 67.63% 109.67% 79.58% 89.21% 76.65% 85.82% 96Pe,re 95.58% 129.19% 102.80% 69.95% 110.17% 91.51 90.29% 79.11% 89.00% Page 34 Appendix A.l. Equations of the Model (1) GDP (7) Labor Market Y Y Px X + Pl1+a)t Lx + LF u L (2) CMEA Exports k (8) CET Transformation Ir 9(1+a) = miz(aLL4, UEKO X = t,[ EPt + (3) Cost of (9) EJD Rabo Pt (1 + = PL LE + rK E = _ _E_+P, (4) Net Output X (10) Net Output Pnce x =~~~~~~~~~~~~~~~~ x =a PC + P-1P Z) 9px =PEX~ + PD°X (5) Factor Demand in X (11) Supply of Goods K,_ pi, V,L LK [(1-I3;J1 Q = 4M ,pP + (I-PP)D (6) Cost of X (12) M/D Ratio L K, = __rPq p = p r + r- D Lp LX rx D [(1-PPMJ Go: Reenue Uncerainiy 35 (13) Supply Pnce (18) Tax Revenue Rep =tm hcm Mer D M + te(P/z+P4) + PQ Q P Q =PM + t PL L + tk rr Kr + tk rfKC + ty Y (14) Personal Income (19) Govemment Budget Y, = Px(l-t)LS + (rA=+rK)(I-tk) S = Rcv - PQ(1+erG +flP +flgr - TR PQ - uPzg + FT er (15) Personal Consumpton (20) Prce Indices .YN(1_N') Pr =(I (l)Px X APC I = PQ(1 +ts) PM h '(1 b)r p =E PE Il+tC (16) Current Account Balance p = E B = SCMM - E E -1E-FT -RE (21) Capital Stocks (17) Domestc Demand K, =KO4) Q = Ct + I + G K = KE + (I-O)( KO^ - K) Go: Rae Unceawrbs 36 A.2. Parameters & Scalars A.3. Variables Uq - shift parameter In 0 CN - consumption at - shift parameter In CET ax - shift parameter In X D domestic good Pq share parameter In Q , - share parameter In CET E - export good 13. - share parameter In X B - current account balance I - investment CS - expenditure share 4 - exports to the CMEA KI - capital in region er - exchange rate Kx - capital in X FT - foreign transfers to government G - govemment consumption Le - labor demand in E KOt - base year capital in t KOx - base year capital in X Lx - labor demand in X L - labor supply A - weight of CMEA exports M - import good pq - exponent in 0 pt - exponent in CET P - price of CMEA exports p, - exponent in X RE - remittances from abroad to PE - price of other exports households Rev - tax revenue PM - import price u - production support from govemment Po - domestic price a - adjustment cost te - export duty P, - wage tk - tax on capital income XI - paroll tax Po - supply price tn - import duty ts - saleslexcise tax Px - net output price ty - tax on personal income TR - transfers from govemment to Py - GDP deflator households nk - world price of CMEA exports Q - supply rE - world price of other exports nM - world price of imports r, - return to K, rx - return to Kx Sg - govemment savings X - net output Y - GDP YH - personal income Go: Renu Uncerainiy 37 Policy Research Working Paper Serles Contact Titlo Author Date for paper WPS1319 The Financial System and Public Asil DemirgOg-Kunt July 1994 B. Moore Entorpriso Roform: Concopis and Ross Levine 35261 Casos WPS1320 Capital Structures In Dovoloping Asli DemirgO-Kunt July 1994 B. Moore Countries: Evidence from Ton Vojislav Maksimovic 35261 Countries WPS1321 Institutions and the East Asian Jose Edgardo Carnpos July 1994 B. Moore Miracle: Asymmetric Information, Donald Uen 35261 Rent-Seeking, and the Deliberation Council WPS1322 Reducing Rogulatory Barriers to Barbara Richard July 1994 M. Dhokai Private-Sector Participation in Latin Thelma Triche 33970 America's Water and Sanitation Servicos WPS1323 Energy Pricing and Air Pollulion: Gunnar S. Eskeland July 1994 C. Jones Econometric Evidence from Emmanuel Jimenez 37699 Manufacturing in Chile and Indonesia Lili Liu WPS1324 Voucher Funds in Transitional Robert E. Anderson July 1994 F. Hatab Economies: The Czech and Slovak 35835 Experience WPS1325 The Economics of Research and Anwar Shah July 1994 C. Jones Development: How Research and 37699 Development Capital Affects Production and Markets and Is Affected by Tax Incentives WPS1326 Banks, Capital Markets, and Gerhard Pohl July 1994 L. Hovsepian Corporate Govemance: Lessons Stijn Claessens 37297 from Russia for Eastem Europe WPS1327 Is the Debt Crisis History? Recent Michael Dooley July 1994 S. King-Watson Private Capital Inflows to Developing Eduardo Femandez-Arias 31047 Countries Kenneth Kletzer WPS1328 The Use of New York Cotton Futures Panos Varangis July 1994 0. Gustafson Conrarcts to Hedge Cotton Price Risk Elton Thigpen 33714 in Developing Countries Sudhakar Satyanarayan WPS1329 The Regulation and Supervision of David H. Scott August 1994 K. Waelti Domestic Financial Conglomerates 37655 WPS1330 Revenue Uncertainty and the Choice Delfin S. Go August 1994 C. Jones of Tax Instrument during the Transition 37699 in Eastem Europe Policy Research Working Paper Series Contact Title Author Date for paper WPS1306 Capital Flows and Long-Term Ibrahim A. Elbadawl June 1994 R. Martin Equilibrium Real Exchange Rates Raimundo Solo 39065 In Chie WPS1307 How Taxation Affects Foreign Direct Joosung Jun June 1994 S. King-Walson Invesiment (Country Specific Evidence) 31047 WPS130O Ownership and Corporate Control in Brian Pinto June 1994 M. Karm-Cheong Poland: Why State Firns Defied the Sweder van Wijnbergen 39618 Odds WPS1309 Is Demand for Polluting Goods Gunnar S. Eskeland Ju1ie 1994 C. Jones Manageable? An Econometric Study Tarhan N. Feyzioglu 37 699 of Car Ownership and Use in Mexico WPS1310 China's Economic Reforms: Pointers Justin Yifu Lin June 1994 C. Spooner for Other Economies in Transition Fang Cai 30464 Zhou U WPS1311 The Supply Response to Exchange Muslapha Rouis June 1994 J. Schwartz Rate Reform in Sub-Saharan Africa Weshah Razzak 32250 (Empirical Evidence) Carlos Mollinedo WPS1312 The New Wave of Private Capital Eduardo Femandez-Arias June 1994 R. Tutt Inflows: Push or Pull? 31047 WPS1313 New Estimates of Total Factor Vikram Nehru June 1994 M. Coleridge- Productivity Growth for Developing Ashok Dhareshwar Taylor and Industrial Countries 33704 WPS1314 The Significance of the "Europe Bartiomiej Kaminski June 1994 M. Patefna Agreements" for Central European 37947 Industrial Exports WPS1315 Global Tradable Carbon Permits, Bjom Larsen June 1994 C. Jones Participation Incentives, and Anwar Shah 37754 Transfers WPS1316 Preserving the CFA Zone: Macroeco- Shanlayanan Devarajan June 1994 C.Jones nomic Coordination After the Michael Walton 37699 Devaluation WPS1317 Estimating the Efficiency Gains Jeremy Bulow July 1994 R.Vo of Debt Restructuring Kenneth Rogaff 33722 Ning S- Zhu WPS131 B Exchange-Rate-Based Stabilization Miguel A. Kiguel July 1994 M. Divino in Argentina and Chile: A Fresh Look Nissan Liviatan 33739