Report No. 22527-ME Mexico Country Economic Memorandum: Challenges and Prospects for Tax Reform July 30, 2002 Colombia, Mexico and Venezuela Country Management Unit Latin America and the Caribbean Region Document of the World Bank ACRONYMS SHCP Secretaria de Hacienda y Credito Publico CIT Corporate Income Tax PIT Personal Income Tax OECD Organization for Economic Cooperation and Development GDP Gross National Product PEMEX Petr6leos Mexicanos NAFTA North American Free Trade Agreement DSH Derecho Sobre Hidrocarburos IRP Impuesto a los Rendiniientos Petroleros DEP Derecho a la Extracci6n de Petr6leo PEP PEMEX Exploration and Production IEPS Impuesto Especial Sobre Producci6n y Servicios Aplicado a la Enajenaci6n de Gasolina y Diesel ARE Aprovechamiento sobre Rendimientos Excedentes PROGRESA Programa de Educaci6n, Salud y Alimentaci6n SNCF Sistema Nacional de Coordinaci6n Fiscal ISAN Impuesto Sobre Autom6viles Nuevos SAT Sistema de Administraci6n Tributaria CHALLENGES AND PROSPECrS FOR TAX REFORM TABLE OF CONTENTS INTRODUCTION AND ACKNOWLEDGEMENTS EXECUTIVE SUMMARY i I. LEVEL AND STRUCTURE OF TAXATION IN MEXICO 1 II. CURRENT POLICIES AND OPTIONS FOR NATIONAL TAXES 3 Oil Taxation 4 Corporate Income (profit) Tax 6 Personal Income Tax 7 Value-Added Tax 8 Excises Summary Revenue Effects 10 III. ADMINISTRATION, COLLECTION AND ENFORCEMENT 14 IV. SUB-NATIONAL TAXES AND TRANSFERS 11 V. TAx REFORM IMPLEMENTATION STRATEGY 15 Annexes: 1. Mexico: An Evaluation of the Main Features of the Tax System 2. Mexico Oil Taxation 3. NAFTA and Mexico's Domestic Tax Policy 4. Efficiency Effects from Reforms to the Mexican Corporate Income Tax INTRODUCTION AND ACKNOWLEDGEMENTS The Mexican govemment is strongly committed to fiscal reform. In 2001 they proposed a comprehensive reform to Congress, which after lengthy debate passed some but not all of the measures. The purpose of this report is to inform the Board and management of the Bank about the issues for tax reform in Mexico, the progress made in 2001, and the challenges remaining. It also shares with the Mexican authorities and others some results of analysis done by the Bank. A number of people have contributed to this report in its various stages and deserve many thanks: Duanjie Chen, Thomas Courchen, Alberto Diaz-Cayeros, Robert Duval, Stephen Everhart, Michael Engleschalk, Michael Geller, Christian Y. Gonzalez, Omar Lopez-Escarpuli, Manuel Marfan, Jorge Martinez-Vasquez, Robert McNab, Charles P. McPherson, Kenneth Messere, Alma Rosa Moreno, Parthasarathi Shome, Elizabeth Toxtle, Jaime Vazquez-Caro, and Steven B. Webb (final task manager). Marcelo Giugale, Lead Economist, and Emnesto May, Sector Director, provided guidance for the task Olivier Lafourcade is the Country Department Director. Richard Bird and Charles McLure, the peer reviewers, provided valuable comments on earlier drafts but bear no responsibility for any remaining mistakes. The Secretaria de Hacienda y Credito Pulbico (SHCP) supported the preparation of this report with advice and data, for which the World Bank is very grateful. The calculations presented here were done by the World Bank team and are not the responsibility of SHCP. EXECUTIVE SUMMARY Since the late 1970s, Mexico has carried out a number of reforms for establishing an adequate and effective tax system. This report was drafted in 2001, before the last round of reforms, at the end of that year. Those latest reforms are described in the corresponding sections. They implement some of the agenda suggested in this report, but not all of it. Many of the instruments of modern tax policy are in place, including income and value-added taxation at the national level and property taxation at the local level. Income taxes are integrated between the corporate and personal levels, with indexation for the effect of inflation on net asset values. A corporate asset tax backs up the corporate income tax and is creditable against it. A VAT has replaced scattered and often competing sales taxes. Tax competition between states has been largely eliminated. Nevertheless, the reforms have not yet borne the expected fruit with: * Increased resources that could be allocated to provision of adequate social services and economic infrastructure for growth and poverty reduction; and * Improved accountability of the public sector to the taxpaying citizens. Over the last ten years it has become clear that Mexico needs greater non-oil tax revenues, sustainable in the long term, in order to provide for more public expenditure in areas such as poverty alleviation, health, education, and infrastructure, as well as to pay for the recent social security reform and banking sector support. Mexicans are not paying enough taxes to cover the public expenditure that they want, because they do not see enough connection between what they pay and what services they get and because of problems in the tax system itself, which are the focus of this report. What are the key problems with Mexico's tax and revenue system? * Administration has been the weakest part of Mexico's tax system. With weak and sometimes corrupt enforcement, evaders are rarely caught and even more rarely punished, especially large taxpayers. This seriously reduces revenue collection, makes tax burdens inequitable, and increases resistance to paying taxes that are seen as unfair. Poor information systems and the inconsistent-participation of the states in enforcement of the VAT has contributed to less consistent and weaker enforcement. Compliance with the tax code has been unnecessarily difficult due to the complexity of exemptions and special regimes and to the lack of a customer-service attitude in the revenue agency. Administrators have had wide latitude to interpret statutes and regulations, which leads to inconsistency, unfairness, citizen resentment, and confusion of incentives and creates opportunities for corruption. The weakness of tax administration has contributed to political resistance to broadening the tax base. The present administration is giving high priority to addressing these problems. * The system relies heavily on oil revenues, which are still about 30 percent of the total, depending on world prices. Inefficient taxation unnecessarily distorts the incentives of the sector. Even with more efficient taxes on the sector, a decline in potential and actual i oil revenues as a share of GDP will almost surely accompany the long-term growth of per capita GDP. So the rest of the economy will have to bear a larger tax burden as a share of GDP. * Various exemptions and special regimes erode the base of the most important taxes- VAT, corporate and personal income taxes. This not only reduces revenue, but distorts incentives away from efficient resource use and leads to a narrow and unfair distribution of tax burdens. Prior to 2002, the special corporate tax regimes for agriculture and trucking probably had the largest distorting effects and allowed the most evasion. The exemption of food and medicine from the VAT and (before 2002) the exclusion of fringe benefits from the personal income tax have the largest direct effect on revenue (besides weak enforcement). * Most tax decisions and the political consequences for them continue to be at the national level, while the delivery of services is increasingly devolved to subnational levels. Thus, the governments that spend those resources take little political responsibility for the taxes that raise the resources, and they have little incentive to expand the tax base in their territory. States have mainly the payroll tax and automobile taxes; and the municipalities have the property tax. While the states could do more, their taxes would fall far short of their spending mandates, even in the states with the best tax bases and with improved enforcement. Reform Options A variety of reforms could address these problems. In the menu of options in this report, a few seem most relevant, and they fall in three areas: national tax policy, administration, and intergovemmental fiscal relations. The central theme is to improve the revenue capacity, efficiency, and horizontal equity of the system by simplifying the laws, eliminating exemptions, and improving administration, facilitating compliance and tightening enforcement. With the revenue, the. government can increase resources for poverty alleviation and reduction through education and other social programs. Experience in many counties has shown that social spending, rather than heavily progressive taxation, is more effective in reducing poverty and income inequality. As improved policy and administration brings enough revenue to eliminate the fiscal deficit and meet the most pressing social needs, the government will have the option to lower the basic tax rates or to expand spending programs. The democratic process will need to choose between these options. National Tax Po/i. Raising more tax revenue cannot come from any major increases in the standard rates, for they are already in line with the major international comparators and Mexico does not want to lose out to intemational competition for skilled labor, capital, or retail sales or see transfer pricing used to attribute expenses to Mexico and transfer income to other countries. Almost all the policy reforms on the menu would broaden the tax base by eliminating exemptions and special regimes-so that revenue increases because more payers and transactions are subject to the same rates and enforcement is easier. For raising more revenue, and ultimately reducing fiscal dependence on the oil sector, the most important policy reform option in the short term is the inclusion of food, medicine, drinking water, books and travel, and their inputs, in the value-added tax (VAT). With such a change, upper income groups would pay most of the additional tax, since they spend most in these areas, although the burden would be heavier as a share of consumption on lower income groups. Exempting grains and legumes (including tortillas, rice and beans) would reduce the impact on the lowest income deciles by almost half, making the tax almost proportional to total consumption, while reducing the total revenue impact by only a third. Even with these exemptions, the elimination of zero rating on most food, plus travel, water, books, magazines and medicines would raise about 1.3 additional percentage points of GDP. The direct revenue effects of other policy reforms considered in 2001 would come to about one percent of GDP in total- including most fringe benefits in the personal income tax, ending the subsidy fot low salaries, applying the standard VAT rate in border areas, lowering the size threshold for firms to pay the VAT, and ending the special regimes of the corporate income tax (CIT) for agriculture, trucking, and publishing. Of these proposals, some important income-tax reforms were passed at the very end of 2001. Putting in place the standard calculation of income for agriculture and trucking, replacing a preferential cash-flow regime, was most significant reform for reducing distortions, raising revenue, and closing loopholes for evasion. The revised income-tax law also simplifies the method for factoring inflation in the calculation of net-worth and capital gains, phases out the highly preferential tax rate for the publishing industry, moves the corporate rate up to 35 percent for retained as well as distributed profits, institutes a simplified cash-flow regime for small firms, ends most of the government wage subsidy for the formal sector (with the incentive of a penalty tax if corporations do not pay this subsidy), includes some excessive fringe benefits in the income tax base, and brings all of a person's income into the same tax base (rather than separately taxing interest income, etc., often at a lower rate). These reductions of distortions will facilitate growth and further increase tax revenue, in the medium term. The Tax Reform Adjustment Loan ($303 mnillion, approved by the WQrld Bank Board in June 2002) supports these reforns. The distribution of the burden of taxes and benefits of public spending-and people's understanding of-it are crucial for winning political approval of tax reforms and for getting people to actually pay the taxes. Good enforcement of the legally required burdens of taxation are critical for achieving a public perception of fairness. The second critical point to remember is that most public spending is progressive: consumption of basic education, public health services, etc. is several times greater as a share of household income for poor families than for rich ones. The key to achieving this progressivity is assuring equal access and provision of services to poor families and assuring that a substantial share of revenue growth goes to such services. The PROGRESA program has established a reputation for being able to deliver income support to the rural poor, especially in the informal sector, and the government has renamed the program, OPORTUNIDADES, and expanded it to aid in urban areas as well. The proposed reforms would clearly improve the horizontal equity of the tax system-making more equal the tax burden on those at similar income levels. Increasing the sense of fairness among taxpayers is widely seen as critical for increasing voluntary compliance. Tax Administration. Improving administration to reduce evasion and facilitate compliance is the most important element of the strategy for improving the tax system. More resources need to go into catching tax evaders, especially those with large tax obligations, and they need to pay the penalties, -when caught. In a few areas the statutory penalties need to be increased, but usually enforcement of the existing ones would be an important improvement. The government should rationalize tax-payer identification numbers and make them comprehensive, collect more of the iii information readily available from payers (including the territorial location of income and transactions), and share it more efficiently between different branches of the system, including state tax administrations. The social security information, considered generally the best in Mexico, should be made more widely available to tax administrators. Common standards for interpreting the code should be more clearly established and widely published. A national tax tribunal should set the standards of interpretation of the federal tax code, including for the states enforcing the VAT. Paying taxes should be facilitated with the attitude of serving the client, including electronic filing and easier payment at banks. The fiscal reform recently approved by Congress has a tax administration component. Its objective is to increase the efficiency and effectiveness of tax collection through improvements on tax compliance and reduction of evasion of internal national taxes. This would be done by i) improving taxpayer assistance; ii) improving the management of human, physical and financial resources; iii) ensuring the sharing of information throughout the organization; iv) expanding information and communications technology; v) sinplifying and strengthening the compliance process; vi) strengthening auditing, control and enforcement; and vii) improving planning and quality control. The Tax Administration Institutional Development project aims to achieve these objectives and is financed in part with a $52 million technical assistance loan from the World Bank, approved June 2002. The impact of improved administration is difficult to quantify in advance, but the experience with the VAT.in the.early 1990s, when the government strengthened enforcement, shows the potential in Mexico to increase revenue substantially without raising rates. The possible revenue increases from improved administration should not be considered as an alternative to policy reforms, for the reforms to eliminate exemptions will contribute critically to the hoped-for advances in collection and enforcement. Subnational Taxation and Fiscal Relations. Rather than further expansion of transfers, such as the Fund for Strengthening Federal Entities, states should be given more tax authority and responsibility, commensurate with their recently expanded expenditure responsibilities. This has several aspects, which can be phased in: giving states legal claim to more tax bases within their borders, letting them set rates, and giving them clearer responsibilities for enforcement. Of the three main taxes, the VAT seems most amenable to partial devolution. This could be in the form of a state retail sales tax or a piggy-back VAT-in either case with the same base and administration as the federal VAT. There was serious discussion in 2001 of a measure to devolve to the states a 2% tax on final sales on the same base as the federal VAT, with that VAT rate dropping from 15% to 13% on final sales. Thus the consumption tax rate would stay at 15% for all stages, but the states would participate in the last stage. The federal government could best (continue to) handle the collection and distribution of revenue. The states would continue to play a role in enforcement and in promoting economic developmnent to expand their tax bases, with stronger incentives to support the overall system.' Such a coordinated reform would have worked best politically among 'the states 1. As the devolution demonstrates its effectiveness, a larger share of the VAT or parts of other taxes could be devolved, perhaps up to the point that baseline value of the devolution equals the value today of all participaciones. A final step, optional at a later date, would be to give states control of their part of the rate, moving it up or down from the original uniform percent. Only at this stage, when state rates might differ, would it be necessary to install a specific system for handling the tax on inter-state sales. Prior to that, the allocation of revenues for the states' share could be on the basis of estimates of final sales in each state. iv in conjunction with an expansion of the VAT base, to increase the sharable pie that was to be redistributed. That reform did not pass, but a symbolic gesture in the 2002 annual revenue law gave the states the authority to collect a sales tax, holding out the promise for further attention to the issue in the future. Devolution of tax authority, especially if it included a reduction of sharable federal taxes, as was discussed in 2001, would change the distribution of resources among states and would therefore require corresponding revisions to the transfer system. The federal government could not afford to transfer some of its tax authority without reducing transfers, such as eliminating the Fund for Strengthening Federal Entities or reducing the amount in the revenue-sharing pool. States with relatively small tax bases per capita would lose from such a trade-off and presumably would require a corresponding change in the remaining transfers, to make them more redistributive. Simulations in this report show how the combined revision could equalize benefits in a revenue-neutral projection. Since the devolution of tax bases should result in larger total revenue summed across all levels, because of the improved incentives, the combination of tax devolution and equalization transfers could make all states better off. Strategy for implementation. As indicated above, the reform strategy for the revenue system in Mexico has three main components-statutory policies, administration, and intergovernmental fiscal relations-and they will have the greatest chance of success if implemented in a coordinated way. * Reducing evasion and improving collection will bring more revenue and allow the improvements in the code to actually improve the incentive framework for the private sector. The biggest distortions and inequities at present come from the inconsistent apphication of the existing statutes. Reducing evasion and improving administration-for greater fairness as well as more revenue-will help gain popular acceptance for revisions to the code, and so they need to come at the beginning of the program. * Reform of the income tax at the end of 2001 was an important achievement, supported by the Tax Reform Adjustment Loan from the World Bank. Two important areas-the VAT and petroleum taxation-were not touched in the legislation in 2001 and thus remain on the agenda for future action. Also, some excise taxes were increased in 2001 to raise revenue and could be lowered again to reduce distortions, once other revenue sources become stronger. * State cooperation in the tax reform program will help improve enforcement, especially of the VAT, and thus needs to be an early priority. Information sharing would benefit the states' collection of the payroll tax. In the next steps, the federal government should devolve more tax authority to the states, especially piggy-backing a sales tax on the VAT. This would work best in conjunction with strengthening the code and administration at the national level. Implementation of tax reform will require strong political support, which in Mexico's competitive political environment will require publicizing the information showing progress toward the goal of everyone bearing a fair share of taxes, through the reduction of evasion, corruption and special exceptions. This could be developed into a fourth component of the fiscal reform strategy- a pact for progressive expenditure and transparency. It would be an explicit agreement with the people to use the majority of increased revenues to fund human development programs that benefit v the poor and to make clearer and more public disclosures of where the money comes from and where it goes. The government has already been moving in this direction for several years, and international experience indicates that it can help increase public faith in the government and thus help increase revenue collection and thereby the actual delivery of public services. vi CHALLENGES AND PROSPECTS FOR TAx REFORM To continue the modernization of its economy and reduction of poverty, Mexico will need to invest more in human resource development and public economic infrastructure, to complement investment in the private sector. Without overburdening or crowding out the private sector, which must remain as the main engine of growth, the public sector needs to increase its resources and reduce its dependence on the oil sector (World Bank 2000). Mexico has developed an impressive array of spending programs for reducing poverty through developing human resources and alleviating some of the most severe incidences of poverty, especially in rural areas (World Bank 1999). Although public spending on average is progressive- benefiting the poor more than the rich as a percent of their income-the per capita benefits of public spending are still greater for the rich in Mexico, as in most places. Thus, reforms to increase tax revenue should be accompanied by measures to increase benefits to the poor, more than enough to off-set any increase of the tax burden on the poor. International evidence indicates that once direct taxes on the poor are eliminated, as they are in Mexico with the virtual exemption of low income from the income tax, the best way to increase progressivity of the public sector as a whole is to increase the progressivity of public spending, not the progressivity of the remaining taxes. This report focuses on the questions of how to increase revenue and reduce distorting effects of taxes. If otherwise desirable measures would raise the tax burden on the poot, it considers how to phase these in, so that the compensating spending measures or adjustments in relative prices (and wages) will have time to take effect. While Mexican tax law has some features that could promote economic efficiency and attract domestic and foreign investment, to have these effects this report argues that the tax system needs to be made simpler and fairer by broadening the bases for the value-added tax (VAf), corporate income tax (CII) and personal income tax (PMI); that subnational governments should take responsibility for a larger share of taxes; and that tax collections should be increased by improving relations between the tax adrministration and actual taxpayers and at the same time mounting a more effective attack on tax evasion. (See also Gil Diaz and Thirsk 1997; World Bank, 1989; OECD, 1999; Diaz-Cayeros and McLure 2000). Reducing evasion would not only have a immediate effect on revenues but would also improve the fairness of the system, with longer term positive effects on public attitudes toward taxation and the public sector. I. LEVEL AND STRUCTURE OF TAXATION IN MEXICO The overall level and structure of taxes in Mexico has changed little over the last two decades (see Table 1), although the details of the separate taxes have gone through vatious changes, for better or for worse. Except for the early 1990s when vigorous enforcement efforts temporarily pushed up federal revenues, the revenue from any single tax has not vatied by more than one percent of GDP since 1980, and total revenue has not varied more than two percentage points. General government revenues are similar to those in Argentina and Colombia, six or seven percentage points higher than in Peru and Venezuela, and six or seven percentage points lower than in Brazil and Unrguay. Mexico's tax revenue shares, however, are lower than most middle-income Latin America countries and are inconsistent with the OECD-type of standards of living and ptovision of public services to which Mexico aspires. Federal tax revenues (excluding social security and PEMEX revenues) are only 10 or 11 percent of GDP, and subnational tax revenues are only about 1 percent.2 Mexico is now at the bottom of the OECD ranking order of tax revenues to GDP, and throughout the nineteen eighties and nineties the difference has increased. Table 1. Mexico: General Government Revenues (percent of GDP). 1980 1990 1995 2000 Total General Government Revenues 22.0 25.5 22.8 21.9 Total Federal Government Revenues 13 16.1 15.2 16 Federal Government Tax Revenues 9.4 11.1 9.3 10.7 Income Tax 4.8 4.9 4.0 4.7 Value Added Tax 2.1 3.5 2.8 3.5 Excises 0.8 1.3 1.3 1.5 Import Taxes 0.9 0.9 0.6 0.6 Others 0.8 0.5 0.6 0.4 Federal Government Non-Tax Revenues 3.6 5.0 5.9 5.3 Products, Services 0.2 0.4 0.5 0.1 Others 0.2 0.8 1.6 1.3 Duties 3.2 3.8 3.8 3.9 Of Which Hydrocarbon Duties 2.9 3.4 3.5 3.6 Other Duties 0.3 0.4 0.3 0.3 State-Owned Indust,jes 9.0 9.4 7.6 5.9 PEMEX 3.6 3.5 2.7 2.0 Other Industries 5.4 5.9 4.9 3.9 Subnational Own-Source Revenues 1.05 1.48 Taxes 0.43 0.54 Non-Tax Revenues 0.82 0.94 Sourc SHCP and World Bank calculations. Oil sector "non-tax" revenues are substantial and fluctuate widely-between 4.2 and 5.9 percent of GDP in the 1990s-and declining on average.3 The fluctuations are unavoidable because of international price fluctuations and could be partially mitigated through a well-designed stabilization fund. In the long-term, the share of oil revenues will inevitably decline if the per capita economic growth is to close the gap with the rest of the OECD. The gross revenue from oil seems likely to stay roughly constant in real terms and is unlikely to grow as fast as per capita incomes. Therefore, if public sector revenues are to keep up as a share of GDP, much less to grow as a share, then the non-oil revenues-taxes-will have to expand as a share of GDP. Income taxes-corporate and personal-are the most important source of non-oil revenue in Mexico, with the corporate income tax holding steady at around 2.5 percent of GDP and personal 2. This report does not discuss Social Security (available to workers in the formal private sector) because it has been separated from the budget and thus from the rest of fiscal policy. Collections cover the cost of participant's medical services, and each participant has his/her own contribution-defined retirement account. So Social Security is not one of the nation's immediate fiscal problems, nor part of their solution. 3. Revenues of PEMEX are counted separately, within the firm. The derechos on oil are, in effect, taxes that go to the federal government. 2 income taxes declining since 1980 from about three percent of GDP to about two percent. The increase of income-tax revenue should come from widening the base and improving administration, not raising rates. Income taxes are also the most important revenue source in Mexico's main competitor economy-the United States-which allows credit for taxes paid abroad. Thus, Mexico can strengthen its collections without discouraging foreign investment. As explained below the increase of income-tax revenue should come from widening the base and improving administration, not necessarily raising rates. Social security for participating private-sector workers has tumed into an involuntary personal saving plah, which effectively puts it out of the immediate fiscal concerns, as a revenue source or a liability. This is an important achievement of the 1990s. The remaining challenges include broadening the coverage of people in the private sector (as part of bringing the informal sector into the formal sector), deepening capital markets so that pension funds can have high returns with acceptable risk, and reforming the social security (pension) plans for the public sector, both to facilitate labor mobility and to prevent (or reduce) the build up of contingent liabilities of the public sector. The VAT and Excises are the growing portions of Mexico's tax revenue, now up to 3 and 2 percent of GDP, respectively. Consumption taxes are alleged to have a number of benefits- encouraging saving, fostering exports, and facilitating compliance, but many problems remain in the details, as noted below. The objective of fairness in sharing the tax burden argues against moving. toward too great a reliance on these taxes. Import taxes are low-less than a percent of GDP-and declining, as in most countries, due to ttade liberalization. To the extent that the decline of revenue results from lower rates, because of NAFTA or other reductions in rates, this is good, because increasing international integration improves the prospects for efficient growth. Sub-national taxes, mainly the state payroll tax and the municipal property tax (ptrdia), are very small-less than one percent of GDP. They cover less than 15 percent of the spending of sub- national governments, which has increased dramatically in the last decade. Those governments may never be able to raise enough revenue to fund all their programs, which include large-ticket items such as education and health, but they do need to raise more of their own revenue if they are to have proper fiscal incentives and be creditworthy. The overview provided above suggests that income taxes and value-added taxes have the best potential for growth in Mexico, as does taxation by sub-national governments. Whether sub- national taxation could grow by taking parts of the federal income or value added taxes is a difficult issue discussed later. The next section summarizes the major strengths and weaknesses of each of the major taxes and gives recommendations to improve them. II. CURRENT POLICIES AND OPTIONS FOR NATIONAL TAxES While each tax has its unique issues, two issues are especially noteworthy and common to all the taxes. First, redistribution. Experience in Latin America and elsewhere shows that taxes are a poor tool for redistributing income or wealth. Redistribution is done better through the spending side (transfers to the poor, primary education, health, etc.). In the treatment of each individual tax 3 below, there is a discussion of the expected distributional, as well revenue effects, but these discussions need to be set in a context of recognition that the fundamental improvement in distribution will come from using increased revenue to fund better services for the poor. Second, exemptions and exceptional regimes are the most problematic part of tax policy, leading to evasion, inequity and inefficiency, as well as revenue loss. Exclusions from the bases of the VAT, CIT and PIT, generally in response to the demands of pressure groups, create inequities and economic inefficiencies. They occur in most countries, but are excessive in Mexico. Because these measures have an effect siniilar to that of a budgeted subsidy and they cost the government resources, they are called "tax expenditures." They save some transaction costs by avoiding the overhead expense of passing through the tax collection and budget spending processes. But tax expenditures have serious disadvantages too. They are less transparent politically, easier to pass initially and harder to repeal, as they do not face annual review in the budget process. The incidence of tax expenditures is less clear than with budget expenditures and is usually quite different from the stated intention. While helping sick people, small farmers, or some other group of the poor is usually the rationale, the greatest benefits almost always accrue to high-income individuals who earn and spend the most, because they are subject to the highest tax rate and would otherwise pay the most tax. Thus, eliminating a tax expenditure almost always improves equity as well as reducing evasion and increasing efficiency and revenue. What began as an attempt to bring difficult-to-tax small enterprises and those in certain economic sectors, like agriculture and transport, into the tax net has become a complex mix of presumptive and cash-flow taxation, which results in uncertainty and more possibilities for tax evasion than previously existed.4 Oil Taxation Prior to 1994 oil taxation (essentially the taxation of PEMEX plus oil products), was negotiated between PEMEX and SHCP on the basis of the Government's budget requirements and PEMEX's perceived capacity to pay. In 1994 a new system was introduced, designed to explicitly tax PEMEX based primarily on profits, while maintaining Government revenues from the sector at levels equivalent to those negotiated in earlier years. The new system contained transition provisions, in particular the DSH (Derecho sobre Hidrocarbuns, see below), which allowed PEMEX time to prepare its accounts for income taxation and to safeguard the revenue-maintenance requirement. These provisions were to last only one year. Six years later they are still in effect, and PEMEX has not yet presented a single tax declaration according to the new regime. The explanation given is that PEMEX is still not in a position to provide the data required for income taxation. To date only one corporation within PEMEX is providing accounting statements. The existing system of oil taxation contains the following elements: * Income Tax. The Impiesto a los Rendimientos Petroleros RP) is an income tax, based on the same elements as Mexico's generally applicable corporate income tax. It is levied at a 35 percent rate on the net income of PEMEX and each of its subsidiaries. In practice it has no impact because, 4. This system was advocated as a simplified one by those promoting it as a political compromise. In the end, however, it became a Trojan horse that allowed policymakers to enlarge, year after year, the artificial definitions that have characterized the system in the last decade. 4 as noted above, PEMEX has yet to come up with the accounting numbers that would enable it to comply. * Duy on Oil Extraction. The Derecho a la Extraccion de Petroleo (DEP) is a cash flow tax, levied on total revenues of PEMEX Exploration and Production (PEP), PEMEX's upstream subsidiary, less total approved expenditures in the year. It consists of three components: Ordinario (52.3 percent); Extraordinatio (25.5 percent); and Adicional (1.1 percent). The three components have different beneficiaties. These are, respectively: state governments (based on agreed formulas); the Federal Government; and cities from which oil is exported. * Petroleum Products Tax. The Impuesto Especial sobre Producciony Serviios Aplicado a la Enajenacion de Gasolinasy Diesel (IEPS) is a gasoline and diesel tax levied on consumers. The tax is calculated as the difference between the consumer price, fixed on an occasional basis by Government at essentially the same level for all regions of the country, and the producer price equal to Houston spot prices plus transport costs and a quality adjustment. Since Houston spot prices change daily, the IEPS also varies almost daily, while Mexico's consumer price remains fixed for much longer periods. These fluctuations in the IEPS are in the opposite direction from fluctuations in other oil revenues, so this tax helps to reduce the fluctuations in total oil-sector revenues.' The challenge is to assure that at least some of the medium-term movements in international oil prices are passed to consumers in order to give them the correct conservation incentives. This tax on petroleum consumption has beneficial environmental effects in reducing consumption. This, as well as the revenue generating capacity, would justify increases in the rate. * Gross Revenue Tax. The Derecho sobre Hidrocarburos (DSH) is levied on PEMEX at a 60.8 percent rate on the gross revenues from all of PEMEX's operations, upstream and downstream, plus the IEPS. The DSH was designed to have a short life, maintaining Government revenues during the transition to income-based taxation. It remains in effect, however, and is the most important feature of Mexico's oil taxation system. Both the DEP and the IEPS are creditable against the DSH, and whenever their sum is different from the calculated DSH obligation, the DEP is adjusted upwards or downwards to make them equal. * Price Cap. The Aprovechamiento sobre Rendimientos Excedentes (ARE) is levied on PEP whenever crude oil export prices exceed the price assumed for annual budget purposes. The ARE rate of 39.2 percent applies to the difference between the budget price and the export price and is additive to the basic 60.8 percent DSH rate, which means that effectively 100 percent of the difference between the export price and the budget price goes to the Governtnent/Hacienda. At present, Mexico's tax instruments for oil result in investment incentives that are exacty opposite to those for efficient development of the sector. The best tax instrument to promote an optimal allocation of resources would be a profit tax, like the dormant IRP (See Annex 2). Under profits-based taxation, a project which shows a positive pre-tax return will show a positive, albeit smaller, post-tax return, thus satisfying the important objective of not discouraging expansion in the underlying activity. Unfortunately, Mexico relies entirely on revenue rather than profits taxes in its upstream oil sector. 5. One convenience of the IEPS for fiscal management arises because the government can use an administrative decree to change the domestic target price and therefore the effective tax. (Changing an explicit tax rate requires an act of Congress). They use this for short-term fiscal management. 5 The profits-based character of the DEP and ISR notwithstanding, the DSH and ARE. are the measures actually in effect-both taxes based on gross revenue. Because of the insensitivity of these taxes to profit, production and projects with positive pre-tax margins or returns will not always have positive margins or returns post-tax. The higher the revenue tax, and/or the more modest the pre- tax return, the more likely that outcome becomes, with the result that already established,production (flowing oil) is not extended to socially desirable limits or is prematurely abandoned, and investment in desirable new production is not pursued. Economic modeling of Mexico's oil sector shows exactly that outcome. Many progressive oil producing countries have abandoned the use of revenue taxes. The average effective revenue or royalty tax rate world-wide is now in the 15 to 18 percent range. In sharp contrast, Mexico's 60.8 percent revenue tax is one of the highest in the world. As costs increase on flowing oil, Mexico's aggressive oil-revenue tax begins to take more than 100 percent of pre-tax cash margins, even where these are substantial as in Mexico at today's oil prices. For example, post-tax margins on Norte production and part of Sur appear to be negative, implying that cash operating costs cannot be recovered, let alone prior investment costs. Under normal commercial circumstances, such production would be suspended or abandoned. The DSH and ARE have even more adverse effects on incentives for investment in new projects. Only the very best projects will produce post-tax returns on investment above the usually required minimum threshold of 15 percent. Discouraging as they are, these results are probably an overly favorable portrayal of the impact of Mexico's tax system on its oil industry. They assume the DSH applies at its nominal 60.8 percent rate to upstream revenues. In fact, the effective rate is significantly higher. This is because the 60.8 percent obligation applies to all PEMEX revenues, both upstream and downstream. The downstream sector, whether in Mexico or elsewhere is incapable of supporting a revenue tax at this level or any level near to it. At the very best, it might be able to cover a 10 percent tax. This means that, to meet its overall tax obligation, PEMEX must be paying an effective tax on its upstream operations well above 60.8 percent of revenues, and a correspondingly higher percentage of pre-tax profits. PEMEX survives, nonetheless, because of the budget subsidies it receives. In order for the company to operate on a commercial basis, there needs to be separate accounting for each subsidiary as well a phase-out of both the budget subsidies and the high revenue taxes. Refom Options: What could be done to correct tax distortions in Mexico's oil sector? The strategy implicit in the current law (not being implemented) is to extract revenue from the sector as one would with efficient taxation on private corporations, thus opening the way for having PEMEX operate as a non-political business corporation. The tax on the production side of oil, including exports, would be on economic profit and not on gross revenue or output. This would mean implementing the IRP and dropping the DEPD, DSH and ARE. The rate for the IRP might need adjustmnent in order to be revenue neutral. PEMEX should be required to present proper and audited accounts to facilitate this. A modest royalty, say 10 to 12 percent of revenues, deductible for purposes of the IRP and applicable only to PEMEX's upstream operations (PEP), would provide early and dependable income from development and production operations without the severe disincentive effects created by the DSH. The royalty would collect a small percentage of the rents (and other profits) generated by upstream operations before application of an explicit rent tax. Because revenues from the IRP would be very sensitive to world oil prices, although perhaps slightly less so than the current regime, the government should consider (even more intensely) the 6 establishment of a stabilization fund. While further study would be needed, clearly its constitution should assure that short-tenn price rises, and even medium-tern ones, do not lead to unsustainable increases in spending. The reserves thus collected would cushion the effects of price declines. The copper fund in Chile and the coffee and petroleum funds in Colombia offer instructive examples. With a politically divided Congress in Mexico, a law and eventually a constitutional amendment could provide relatively strong protection of the reserves against raids to pay for ordinary spending. Revenue effect. The proposed reforms for-oil taxation could be revenue neutral, in tetms of the expected level of net revenue for the same oil production and world prices. It is important to evaluate the revenues of the sector net of the budget transfers to PEMEX, which are now negotiated annually, and compensate for inappropriate taxation, but which could be eliminated with the introduction of proper taxation. The package of reforms, especially the stabilization fund, would make net revenues less volatile in response to fluctuations in world prices. Also the reduced distortion of incentives for production and investment could lead to greater output and profit in the medium term, and hence more net revenue for the government. There would not be any substantial effect on the distribution of tax burden among income groups. Corporate income (profit) tax The corporate income tax (CIT) is the most important part of the income tax, which exceeds even the VAT in revenue importance. Total income taxes brought over 4.3 percent of GDP in the late 1990s, most of which was the CIT; this includes the tax on the self-employed, which most countries count as part of personal income tax revenue.6 This section discusses the general case of the CIT, which has a number of sound and sophisticated features, then discusses the exceptions and administration, which have undermined the CIT in practice, then discusses the links between the Mexican and the other North American CIT systems (OECD 1999; Gil-Diaz 1995; Annex 1), and finally, discusses the reform options and the recent measures approved by Congress. The general case of the CIT applies a rate of 35 percent to corporate income (profit), including net capital gains. The capital gains calculation is adjusted for inflation in principle, not only on the asset side, correcting asset values for inflation, but also on the debt side, counting as taxable profit the reduction in the real value of a firms debt (OECD 1999). If interest rates fully anticipated inflation, then that cost would have been deducted from income, so inflation would increase a firm's tax liability only if the interest rate on debt did not include full inflation compensation. The rules to implement the indexation principle are complex, however, so that the sophisticated objectives of the law are not fully realized, and compliance and enforcement become more costly. On retained earnings, prior to 2002 firms paid a reduced rate of 30 percent, with the remaining 5 percent deferred until the earnings were paid out. The corporate tax is integrated with the personal income tax, so that, at least in principle, individuals pay tax only on profits that have not been previously taxed at the corporate level. Also, from the point of view of the individual, profits are taxed at the same rate whether they are received via capital gains or distributed dividends. The unification of the personal and corporate income tax in Mexico is an important strength of the system-eliminating double taxation-which will show more clearly as other defects of the system are eliminated. Untdl recently a problem with the general regime itself was expensing of purchases 6. There is only one income tax law in Mexico, covering both physical and "moral" persons-individuals and corporations. For exposition and analysis, howevet, it is convenient to discuss the corporate and personal income tax issues separately, since the legal provisions and the underlying issues are largely distinct. 7 regardless of inventory changes; the immediate depreciation provisions allowed a substantial deferral of tax payments, but this was dropped in 1999. Since 1989 an asset (net worth) tax of 1.8 percent has complemented the CIT. It was introduced to deter evasion, such as through transfer pricing and overstating costs, and to assure at least a minimum degree of corporate taxation. It is creditable against the CIT. Since firms have an incentive to report investment expenses to reduce their taxable profits, information on net worth is relatively good. In its early years, an estimated additional 3.5 pesos came into the CIT for every peso collected by the asset tax (Gil-Diaz 1995). It yields the same revenue as the CIT when the rate of return on assets is 5.3 percent per annum, so firms making more than that (as any efficient firm would do on average) should pay at the higher CIT rate. Firms making less in the long term should close in any case. New firms do not have to pay the tax before operations start and for the first three years of operations. The measure seems effective as a minimum tax, as long as tax administrators do not accept the asset tax as a full substitute for the CIT properly due, which they sometimes do. Prior to 2002, there were three partly overlapping exceptional (and preferential) regimes to the general case of the CIT: i) Businesses in agriculture, agro-industry and trucking could choose a simplified cash flow accounting system, by which they pay tax only on the money taken out of the firms and not on the retained earnings. To have the desired effects, especially simplicity, cash-flow taxation has to be adopted for all sectors. Cash-flow taxation could make sense as part of a strategy to shift to taxing only consumption rather than all income (consumption plus saving). Applying it only to a few sectors, however, generated non-transparent micro-economic distortions and opportunities for evasion, for example, via manipulation of transfer pricing in multi-sectoral conglomerates!. ii) Firms in agriculture, forestry, fishing, or publication also enjoyed a special rate regime of half the standard rates - thus 17.5 percent on profits and 0.9 percent on net assets. The reduced rates also applied to the simplified cash-flow scheme. There was no economic justification for these sectoral preferences, and they created a lot of distortions, not least of which were manipulation of transfer prices and cross-sectoral mergers to reduce tax liabilities (as with the cash-flow regime). iii) There was a small-payers regime for companies with annual gross income below 2.2 million pesos, which simply taxes gross income at graduated rates ranging from 0.25 to 2.5 percent. This simplified compliance and enforcement for small firms, except for the graduated rates. During the 1990s these special regimes eroded the base of the CIT. Many of these tax expenditures were the consequence of political pressures and were granted without measuring the revenue losses incurred nor considering how much they increase complexity in the tax system. Not only did they narrow the statutory tax base, but also provided an opportunity for tax avoidance. The cash-flow system for agricultute and trucking had some attractions of simplicity, but it was much more generous than the standard regime, creating opportunities and incentives for evasion, 7. Zodrow and McLure 1991. Cash-flow taxation also has the practical drawback that, at least for now, it is not creditable against US corporate taxes, the way a traditional CIT would be. In effect then, cash-flow taxation creates a protective barrier against competition from US firms. 8 especially in the trucking industry, which is now largely integrated with other sectors of the economy. Opinions differ on how widely the combined use of the general provisions, the simplified regime, the consolidation of profits and losses, and tax treaty provisions contributed to tax avoidance. Unquestionably, the provisions were open to abuse and were abused. The situation was particularly bad because of the absence of qualified staff in the tax administration to understand, let alone challenge, what the corporate accountants declared. (he quality of tax administration in Mexico is addressed later on). Mexico's corporate taxation has become more harmonized with that of NAFTA and especially that of its major trading and investment partner, the United States (See Annex 3 and 4). Still much remains to be done; some measures Mexico should take on its own in any case, and others will need joint agreements. The US federal CIT rates are similar (35 percent federal rate, but various state rates complicate the comparison). Both countries use a worldwide basis for taxing corporate income and foreign tax credits to avoid double taxation by different countries. The Mexican tax system is more advanced than the US in some ways: in its system for integrating personal and corporate income taxes to avoid double taxation and in the indexation for inflation. These policies should continue, although with refinements (usually simplifications), as with the inflation indexation. In other areas, Mexico should move toward the US practice, for Mexico has much looser rules for transfers of losses among corporations, and more special incentive regimes. There should also be a harmonization of the withholding rates for the cross-border payments of dividends, interest, and royalties, in the tax treatment of leasing, and in the customs duties on investment goods, although neither country's system (nor Canada's) has obvious advantages in these areas. Recent Rejonms. The most important reforms to the corporate income tax that took effect in 2002 make it more uniform across sectors, thus reducing distortions and evasion. As a group they are expected to increase revenue by only about half a percent of GDP in the short run; some measures will reduce revenue and others will bring more. The medium-term revenue effects will be stronger because the reform reduces economic distortions and facilitates administration. The reform includes the trucking and agriculture sectors in a general system for calculating income, ending the special cash-flow regimes that favored them so heavily and required them to pay the income tax only if financial resources were withdrawn from the company. From now on, taxes in those sectors will be payable on income minus expenditures and deductions, which do not include financial investments and only include the present value of the straight-line depreciation of fixed investment. The 50 percent reduction in tax rate for publishing is being phased out over four years; the lower rate for agriculture remains. With these and the other changes, the inter-sector dispersion of marginal effective tax rates is reduced from 5.5 to 2.4 percent. (See Annex 4). The elimination of the negative effective marginal rate for trucking is especially important, since it was such a strong incentive for tax evasion through various tricks of corporate organization and pricing. The fiscal regime that will now apply to trucking and agriculture has clearer standard-depreciation and expensing rules, which will facilitate compliance by firms and enforcement by the Government. The reform expands the standard rate of 35 percent to all profits, eliminating the special rate of 30 percent for reinvested profits. Part of the rationale (the good part) for the special cash-flow regime for agriculture and trucking was the difficulty that small operators have in documenting all their business costs. To address this concemn, the reforms allow simplified tax calculations and reporting in several ways. The trucking and agriculture may report on a simplified cash basis, but the deduction for investment 9 is only the present value of future capital depreciation, not the whole value that they had in the cash- flow regime. Thete is a new simplified regime, without sectoral discimination for individuals with entrepreneurial activities under 4 million pesos gross annual revenue, allowing them to pay tax on a cash-basis definition of income. For individual entrepreneurs whose income is lower (under 1.5 nillion pesos gross revenue) and who sell only to the general public, the reform modified the Small Taxpayer Regime, so that it allows payment of a flat-rate presumptive tax to the federal government-1 percent on gross revenue-which is simpler than the graduated rates previously charged. The reforms are designed to reduce the diffetences in effective tax rates and thus to deter taxpayers from inappropriately shifting from the general system to the simplified cash-basis regime. Also, requiring that the payments for purchases from firms belonging to the simplified tegime will be deducted only after the payment is made (in accotd with the cash-basis regime) maintains the fiscal symmetry and improves the timing of the revenue impact for the government. The expected gains in revenue and efficiency from better compliance and enforcement, due to simplicity, was felt to outweigh the losses from the lower legal tate. Making this happen will depend on actual itaprovements in the Servicio de Administracidn Tributaria (SAT), the institution in charge of collecting federal taxes. The one percent federal rate on gross income of individuals with entrepreneurial activities is a lower rate on average than the progressive rates previously charged by the federal govemment, but the revenue law (Ly de Ingirsos) for 2002 gives the state governments the right to put a tax of up to 5 percent on the net income of all individuals with entrepreneurial and professional activities whose revenue does not exceed 4 million pesos. For individuals who contribute in the Small Taxpayer Regime, the states could charge a rate up to 2 percent on gross income. The federal government has collected very little from these groups of potential taxpayers, and the authorities believe that bringing in the state governments, who know this populace better, will improve enforcement and collections. If this happens, it would be a creative and important way to strengthen the states' fiscal authority, although first the federal governtnent needs to put into perrnanent law the option for states to introduce the these taxes. Distribution effects. The incidence of the corporate income tax is notoriously difficult to discern, but we can identify certain trends. First, additional tax that is collected from foreign-owned firms will just reduce their home-country tax lhability to the extent that they have not yet used up their opportunity to take foreign tax credits at home. For instance, the burden of Mexican taxation on US firms up to 35 percent (the US rate) is exported to the US. On Mexican firms the tax is shared among capital owners, workers, landowners, and consumers in complex general equilibrium way. When actual collection rates are first changed, the burden falls on capital owners, making the initial impact very progressive, with the burden falling mostly on households in the top 30 percentiles of the income distribution. As the input and product prices adjust, however, the tax becomes much less progressive, if at all. Making the system more uniform would reduce horizontal inequity- making different effective rates of incidence on persons with similar incomes and similar situations. Personal Income Tax The personal income tax (PIT) is the most progressive part of Mexico's tax system, but exemptions, the salary subsidy in the formal sector (cr6dito alsa/aio), and uneven administration make it unequal horizontaLly, and inefficient for raising revenue. Revenues have declined as a share of GDP over the past two decades. Rates have been reduced and the tax bases narrowed by the introduction of new concessions, such as limits on taxation of fringe benefits, which resulted in a 10 more complex and unmanageable PIT. This contrasts with the trend in most other OECD countries, which have increased their income tax revenues (PIT plus CIT) during the last two decades by widening the tax bases, even while reducing the rates. About 10 percent of total remuneration consists of fringe benefits that are currently exempted, see Table 2.8 The increase of revenue from including these benefits in the tax base would-be less than 10 percent because the percent of excluded remuneration was lower in the upper income brackets where the rate is higher. The PIT is progressive because the regular (positive) PIT starts at a low rate of 3 percent and moved up to 40 percent, until the 2002 refotms. The brackets are indexed twice annually for inflation. On income from agriculture, fisheries, and book editing (until 2002), the rates are one half of the regular rates (consistent with the CIT). There is no double taxation of dividends, for shareholders can take as a credit the tax paid by corporations they own, including via stock ownership. Persons with exclusively agricultural incomes below 20 times the local minimum wage pay no PIT. On incomes in the range of the salary credit (credito al salaio) and the fringe-benefit subsidy, the effective marginal rate is lower, as described next, which makes the PIT even more progressive, at least in the initial impact. The salary credit part of Mexico's PIT, along with the subsidy for those whose fringe benefits are low or absent, was until 2002 a subsidy for the wages of low-paid workers in the formal sector. These subsidies were generous at the lower levels and phased out gradually, so that the effective marginal rate is reduced. For workers in the lower wage ranges, the subsidy exceeded the regular tax, resulting in an actual outlay from the government. The salary subsidy goes with the job, not the worker (or her or his low-income household), so households with more than one worker can get more than one subsidy. The subsidy comes without regard for the size of the household or its other wage and non-wages sources of income, which results in additional inequities and dilution of the program as an antipoverty measure.9 While the initial impression and impact of this subsidy are very progressive, although not equal horizontally, the main effect of the salary subsidy may be to attract low income workers into the formal sector, rather than to raise their take-home wage. Once wages adjust in the medium term, most of the benefit accrue to employers, who could pay lower (pre-subsidy) wages as a result of the credit, because there are so many low wage workers in the rural and informal sectors and thus the supply of such workers to the formal sector is very elastic. The salary subsidy cost the government considerable revenue, at least 0.2 percent of GDP and perhaps as much as 0.6 percent, depending on assumptions, and it greatly complicates the administration of the PIT, without policy justification. The salary credit is alleged to provide aid to the poor employed in the formal sector, but they are far from the poorest members of society and are no more deserving than other moderately poor persons in the informal sector. Recipients of the salary subsidy (and the fringe benefit subsidy) are certainly less deserving than the very poor, who are rarely in the formal sector, although many of the rural poor benefit from the PROGRESA (now OPOTUNIDADES), Contigo, and ProCampo programs. Around 10 percent of the total wages and 8. Fringe benefits include: motor vehicles, accommodation, low interest loans for housing or other purposes, discounts on goods or services, gifts to employees, educafion expenses of employee or dependants, contributions to pension or super-annuation schemes, medical and dental insurance, life insurance, canteen facilities-or meal vouchers, and leisure facilities. 9. Mexico is tie only OECD country that actually pays a fringe benefit subsidy for low and middle wage earners, and it includes in the base for the positive income tax a far narrower range of fringe benefits than most countres. Mexico is also the only country with a salaty subsidy for low paid employees that does not take account of the family situation and non-wage income. salaries are not taxed at all because of the special treatment of benefits given to the employees in addition to their salary and wages. Such simplifications could help increase revenue, because concentrating personal income taxes on the middle- and upper-income groups would make tax administration more manageable, and little revenue would be lost from the lower brackets. Table 2. Fringe Benefits to Salaried Workers. Marginal Tax Rate Total Fringe Benefits Exempted Fringe Benefitsa (Percent of total remuneration) (Percent of total remuneration) 3 26.0 17.0 10 24.2 14.6 17 29.0 14.0 25 29.1 12.4 32 29.3 11.4 33 29.6 9.7 34 28.3 7.9 35 19.6 2.9 Total 25.8 9.7 Note a Not all Fringe Benefits are exempted and there are some limits to the value of the exemptions. Source: SHCP using a sample of special fiscal declarations of 1998 and the National Survey of Employment. Recent Reforms. The main reforms to the personal income tax, taking effect starting in 2002, consolidate the taxable income of individuals from all sources and charge more similar rates to people with similar incomes. Consolidation addresses the previous problem that persons receiving total incomes that should be taxed at a high rate could receive income from various sources, including interest income, some of which were taxed at lower rates. Tax on interest income in excess of the return needed to compensate for inflation will continue to be withheld, and credited against the tax owed on consolidated income when real-interest income exceeds 100,000 pesos annually. This reform largely eliminates the previous distortion of the effective tax rate on equity versus debt financing of corporate investment. The reforms introduce taxation of real interest earned on government bonds and of capital gains from some stock sales.'0 The new income tax law partly addresses the problem of tax-exempt fringe benefits through a change in the corporate section, which limits the deduction (from taxable profits) of the cost of fringe benefits for non- union employees to the lesser of one minimum wage or 10 percent of taxable income. This approach, simnilar to that of Hungary, will eliminate the incentive for firms to evade taxes by compensating their executive and professional employees with benefits like apartments, cars, and vacation facilities. The reforms simplify the personal part of the tax code by reducing the number of tax brackets from ten to five by 2005 (starting with eight rates in 2002 and dropping one rate bracket each year during the transition) and make the top rates consistent with the corporate tax, induding the decline of the top rate from 35 to 32 percent by 2005. The reforms largely eliminated the government subsidy for salaries of workers in the formal sector, and add a tax incentive for firms to 10. Capital gains taxes will be levied on stock market transactions involving companies that have less than 35 percent of their stock traded on the Mexican exchange and that have been public for less that 5 years. This intends to deter evasion through schemes involving artificial trading in dosely held stocks. 12 continue paying the subsidy themselves, in the form of a penalty tax if they stop, in order to protect the workers. As noted above, the low-income wages in the long run will be determined by productivity (labor demand) and the wage levels in the informal sector (supply). Distnbution effects. The PIT is a very progressive tax and it would remain so, since the reform has little effect on low-income groups. Within the middle and upper income groups that pay positive PIT, the inclusion of fringe benefits would bring the most revenue from the upper income groups where the value of fringe benefits is greatest, although for the middle income groups the increase would be a larger share of income. Phasing out the cutrent form of the salary subsidy would have little effect on distribution, since firms face tax subsidies to keep paying the subsidy (and the government pays if they do not). Furthermore, the government is using some of the fiscal savings to fund redistributive spending programs, with very progressive effects. Value-Added Tax The VAT is currently the third largest revenue source for the Government, after petroleum and income taxes, and holds the most potential for increasing revenues. The VAT in Mexico has a standard rate of 15 percent, which is lower than VAT rates in Canada and in other Latin American countries, but higher than sales taxes in the US Even before one considers enforcement problems, the standard rate does not apply to many activities; the base includes only about 55 percent of total consumption. 11 Small and medium enterprises, up to annual sales of MxP 1 million have the option to be exempt; there is zero rating for most food and medication.'2 Furthermore, the rate for taxable items is only 10 percent in the broadly defined border areas. The exceptions to the standard rate and poor enforcement seriously undermine the revenue generation of the VAT, for it yields only about what would be expected from a rate of 5 percent on total consumption, as shown in Table 3. This 32 percent effectiveness rate as percentage of statutory rate is far lower than any other OECD country. The low effectiveness in Mexico only partly reflects administrative problems. The zero rating for food cuts out a larger portion of total consumption expenditures in Mexico than a similar exclusion would in other OECD countries, because Mexico is poorer. Italy, where there are also collection problems, is the next lowest at 51 percent effectiveness; the OECD average is around 70 percent. In addition to the most setious problem of evasion, there are four main problems with the VAT policy in Mexico: exemptions that erode the base, a lower rate of 10 percent near the border, a high threshold for firms to have to participate, and enforcement delegated to the states. 11. Subsecretario de Ingresos Manuel Ramos Francia in Refonma, August 21st , 2000. 12. Exempt operations do not collect and pay VAT on what they sell, but they do pay VAT on taxable inputs. So their consumers effectively pay the VAT only on value added in stages prior to the exempted one. This makes sense, for instance, with micro retail operations, where collection would be prohibitively difficult. With zero-rated items, on the other hand, the final consumer does not pay any VAT even indirectly, because the seller does not collect VAT and gets a rebate of the VAT on any taxed inputs. This is a more significant tax benefit than an exemption and entails administrative costs. 13 Table 3. Comparing Effectiveness of Value Added Tax Collection, 1997. :Value added Tax . . Effective VAT rate in Revnue i pecen o Standard rate Effective VAT Reveues percent of Percent Rate, percent per cent of standard GDP Percent Rate ,percent rate A B B/A Japan2 1.8 5.0 3.1 89 Germany 6.6 15.0 11.5 77 France 7.9 20.6 14.7 71 Italy3 5.7 19.0 9.8 52 United Kingdom 6.9 17.5 10.8 62 Canada2 2.5 7.0 4.3 61 Mexico 3.1 15.0 4.7 32 OECD average 6.7 17.7 12.5 73 G7 average 5.3 14.4 9.0 69 EU average 7.3 19.4 14.0 71 Notear 3. The effective VAT rate is VAT revenues divided by the base (i.e. consumption excluding VAT). 3. Fiscal year basis (Q2/97-QI/98). 3. The standard VAT rate was raised from 19 to 20 per cent on 1 October 1997. Sourre. OECD 1999 Economic Surv9 ofMexico; Bank staff calculation. Extensive zero-rating-for food, medications, water, transportation, books and magazines- reduces the VAT revenues by almost 40 percent and is the largest policy deficiency. The zero rating requires rebating substantial amounts of tax revenues to traders, complicates the system, and opens the way for tax evasion, but keeping an exemption for basic foods would greatly mitigate that effect. For instance, Canada and some US states exclude food from the VAT or sales tax, which makes the tax more progressive. To have an exemption that achieves a similar degree of progressivity relative to its income distribution, a poorer country like Mexico should only exclude from taxation a basic basket of foods whose consumption is concentrated in the lowest income deciles. Ending these zero-rating provisions would increase the burden as a share of income more for the poorer two deciles than for the upper two deciles. To address equity concerns, discussed further below, increased income subsidies to the poor through the OPORTUNIDADES and other programs should accompany the end of zero rating. This compensation should be fiscally feasible while still leaving net revenue gains, since one-third of the additional revenue would come from the two highest income deciles of taxpayers (they get 33 percent of the implicit subsidy today from the zero rating), while less than 9 percent of the additional revenue would come from the two lowest income deciles of the population. The second problem with the VAT is that a rate of only 10 percent, instead of the standard 15 percent, is charged in the broadly defined border areas, including all the territory of the states of Baja California, Baja California Sur, and Quintana Roo. Since the markets in the border areas are estimated to have around one-tenth of the total market, the lower rate represents a substantial revenue loss. In addition, the present 10 percent rate for border states complicates the system, provides opportunities for tax evasion, and distorts internal competition within Mexico." This measure was designed to protect border merchants against smugglers, but a well-enforced VAT would increase the international competitive position of Mexican producers, including those near 13. No other OECD country has special low rates for border areas (with two or three very limnited exceptions), even though rates of general consumption taxes vary considerably between neighboring countries-for example, Canada/US, Denmark/Germany, and France/Spain. 14 the border, because public services in the US are supported mostly with non-VAT-type taxes that cannot be rebated for exports. Thus, ending the differential rate, making it the same throughout Mexico, would have important benefits, especially to reduce evasion and simplify administration. Even reducing the scope of "border areas" to those in true competition with neighboring markets, as proposed in 2001 would reduce distortion and increase collection. The third problem with Mexico's VAT policy is the unusually high threshold below which traders can opt out. At 1 million pesos (about US $100,000) the Mexican threshold is much higher than that of other OECD countries, with the exception of Japan."4 To bring more medium size taxpayers into the regular VAT regime, the threshold should be substantially reduced. Enterprises (including those in the agricultural sector) with receipts of 500,000 pesos or even 300,000 pesos, should have to register and pay the VAT. The threshold could drop in phases over two or three years, which would help SAT to cope with the increasing documentation that would result. -The fourth problematic feature of Mexico's VAT is that the states participate in its administration and make uneven enforcement efforts. They are often given responsibility and authority for enforcement Cfiscafi!aci6n) with certain firms-in a process negotiated with the federal SAT-and are allowed to keep the fines that they collect as a result of their efforts. The logic of this arrangement is that the state authorities know the local situation better and therefore will be better able to identify and prosecute evaders. And the incentive to the states is the extra revenue they collect, which they are allowed to keep and is substantial in some cases. The Federal Government should also benefit as stronger enforcement, with penalties for evasion, makes more firms willing to comply with the regular federal collection. Actual results have not lived up to these expectations. States vary in their interpretation of the tax law and in the vigor of their enforcement, because of varied capacity of local officials, possible corruption, and perhaps even tactics to attract business investment. This creates geographic variation in the effective tax. Also, the reduced federal government involvement in enforcement deprives it of information that would help it to improve and coordinate enforcement efforts not only of the VAT but also of excises and the corporate and personal income taxes. Since direct devolution. to states of a share of the VAT is a possible route for reforming the Pacto Fiscal in the future, as discussed below, states should continue some involvement in enforcement. But clearly the federal government needs to get more involved in enforcement, at least to assure common standards and to collect and share data. More payment information should be collected to allow better determination of the location of value-adding activity, something that will help improve enforcement of state taxes like the payroll tax, as well as other federal taxes, like corporate income. Pofify Refrn Opmions. The most important reforms of the VAT would be to improve enforcement and to remove the preferential rate in border areas and the zero rating for non-basic food, medicines, magazines, books and travel. Including food would be the most important for raising more revenue, with books and magazines, and travel next (Possible devolution of the VAT to the states in discussed below in the section on subnational taxes and transfers). Lowering the threshold for the size of firms from MxP 1 million to MxP 500,000 would bring in some additional revenue as administration and enforcement improved but more importantly would reduce the opportunity for larger enterprises to evade taxation by using subsidiaries. To move to a uniform 14. Comparative thresholds for other countries to the nearest thousand US dollars are Japan 270, UK 75, Canada 23, Germany and New Zealand 21, France 14, Portugal 13, Greece 7, Italy 3. (OECD 1999 and T. Dalsgaard and M. Kawagoe, 2000). 15 VAT system, as practiced in most other OECD countries, would increase not only VAT revenues but also corporate income tax revenues in the long term, because of the additional accounting information that would become available. Strengthening enforcement and coordination with the states is clearly crucial, to make the tax paid correspond to the tax in the law. Revenue effects. Compared to the current VAT revenues of about 3.5 percent of GDP, the contemplated expansion of the base could increase revenues by over two percentage points of GDP from direct effects, with further indirect effects as well. The most important effect would be from eliminating zero rating, except for exports. Household expenditures on unprocessed food, medicines, water, books and travel were about 62 percent of the expenditures on items currently taxed, according to the 1998 Household Expenditure Survey, and if one excludes cereals, grains, vegetables, and legumes-tortillas, rice and beans-the increased coverage is still 42 percent of currently taxed expenditures. So, at the maximum one could expect that broadening the base would raise revenues by about these percentages."5 Since about 10 percent of total sales are in border areas, bringing the rate there from 10 to 15 percent (an increase of one-haLf) would raise total VAT revenues by about 5 percent. In combination, the maximum increase of revenue (currently about 3.5 percent of GDP) from these two policy changes would range from 1.5 to 2.2 percent of GDP, depending on the extent of foods covered. We lack information to make a good estimation of the effect of lowering the threshold of firm size. Improved enforcement could add even more revenue, and the elimination of most zero-rating would make collection and enforcement easier, with effects discussed below. Distnbutional effects. The main effect on equity would come from the end of zero rating of food."6 As mentioned earlier, most of the spending on these items, and thus most of the burden from taxing them, is in the upper income groups, which is shown in Figure 1 as the vertical distance between the top and bottom lines. But taxing these items would fall heavier as a share of income in the lower income brackets, and indeed would make the VAT a slightly regressive tax (higher percent in lower income brackets than in lower ones), rather than a slightly progressive tax that it is today. Compare the black bars (all inclusive VAI) with the dark-gray or blue bars (current VAT) in Figure 2.17 If the VAT were extended to all food except cereals, grains, vegetables and legumes-tortillas, rice and beans-then it would be roughly the same as a share of income across all groups except the richest, as shown by the light-gray or green bars. To address equity concerns from the end of zero rating, the government could increased income subsidies to the poor through Seifores (for those in the pension system) and the OPORTUNIDADES (formerly PROGRESA) program, especially with an expansion to urban areas. 15. Since the determination of which farm inputs went to which products would be too difficult, exemption of certain food items is a more realistic transition policy than selective zero-rating, with the result that the government would still get the tax on farm inputs for exempted items, and consumers would bear that hidden burden. 16. Water and especially medicines are a much smaller shares of spending than food at all income levels and are more nearly equal shares of spending across income levels. Compared with the poor, upper income groups spend a much higher share of income on, travel books and magazines, although the absolute share is still small. 17. Bank staff calculations from 1998 household expenditure survey. Looking at the VAT burden as a share of income is important for obtaining these results. If one looks at the share of expenditure, which is much closer to income for poorer groups than for richer ones, then the current VAT looks much more progressive and even the all- inclusive VAT is about proportional to income. The calculation here uses net income reported in the household survey, net of taxes and including income subsidies. For assessment of a changes in one tax this procedure seems appropriate. 16 Figures 1. VAT Revenue According to Income Decile. 9000000 8000000 , 7000000. O 6000000 5000000 1X 4000000/ 3000000O-": 2000000 1000000 .* - . 0 0 2 4 6 8 10 12 Income Deciles ---*-VAT act. -O---Ref. VAT - -* --Ret. VAT wle cereals and veg. Figure 2. Theoretical VAT Burden as a Share of Income by Deciles. 0.12 0.1 - 0.08 -. _ _ _ _ _ _ 0.0211111 1 2 3 4 5 S 7 8 9 10 Income Deciles *VAT act. *Ret. VAT *lot. VAT Wle cereals and vag. lowest income highest income 17 Excises In Mexico the Federal Government collects excise tax on gasoline, tobacco, and alcohol. Revenue from excises other than gasoline (discussed above in the oil section) has roughly doubled over the last two decades, exceeding two percent of GDP by the end of the 1990s. Unlike most places, Mexican excise taxes are levied on an ad valorem basis, rather than by physical unit, which is the common practice. The government introduced the ad valorem rates in 1980 to protect excise yields at a time when inflation was high and increasing rapidly. Nevertheless, ad valorem taxation has many defects and now that inflation is relatively low, there is a strong case for following the practice of other OECD countries in basing taxation on physical units, for example, the volume of hydrocarbon oil and the weight of tobacco."8 Taxation of alcoholic beverages according to the strength of the alcohol was already enacted in 1999. The Mexican excise taxes also differ from other OECD countries in applying a VAT-type credit mechanism each time the products are sold, instead of applying a final tax when the goods are imported or leave the domestic factory. The original purpose of this feature was to combat tax evasion. Since a few firms controlled the market from the production to retail stage, they put much of the mark-up at the post-import or post ex-factory stage and thus could reduce the taxable value at the earlier stage. (Taxing by physical volume, not ad valorem, would remove any advantage from the VAT-type mechanism). Mexico has recently returned to the standard OECD practice of taxing alcohol at the production or import stage, due to the emergence of many small wholesale and retail producers of alcoholic beverages, but it has not yet done so with the oil and tobacco excises. Recent Reforms. The reforms to excise taxes are different than those proposed above. The new Law for Special Taxes for Goods and Services raised the rates on cigarettes, to raise revenue and further discourage a socially undesirable activity. The tax on soft drinks using fructose, a protectionist measure for the domestic sugar-cane industry, was passed by Congress but has been temporarily suspended. The income law of 2002 also put a luxury tax of five percent on top of the regular 15 percent VAT on luxury goods and services, such as restaurants, luxury foods, expensive automotives and accessories, sport boats, and recreational services. While these areas of consumption are not socially undesirable, putting an extra tax on them is a progressive way to raise the increased revenue that the Government needs to continue paying for its program, most of which are in the social sectors and thus are pro-poor. Further Reform Options. The excise taxes could be simplified in two main ways. First, ad valorem rates need to be converted into rates per physical unit, with provision for automatic and regular indexation for overall inflation. Second, the VAT-type credits for intermediate sales should be eliminated and taxes should be collected at the production or import stages. While these reforms would not have direct effects on revenue or distribution, they would facilitate administration and reduce distortion, both with indirect effects to increase revenue in the medium term. 18. In the case of cigarettes there is a case for partial ad valorem treatment. 18 III. ADMINISTRATION, COLLECTION AND ENFORCEMENT Tax administration is the weakest link between the often sophisticated objectives of Mexican tax policy and the actual tax revenue collected (Gil Diaz 1995). A study in 1996 indicated that the evasion of the VAT in Mexico was about 60 percent (CIDE 2000). While the VAT is perhaps the most serious area of evasion, the income taxes also fall far short bf their potential; total revenues (PIT and CIT) are only 7 percent of corporate profits and less than 5 percent of total wages, salaries and profits (INEGI and World Bank staff calculations). Existing evidence indicates five key problems: a) evasion is too easy-rarely discovered and even more rarely punished; b) filing and payment, even by those willing to pay, are still too difficult; c) the tax code is too complex; d) the application of the code by those who adlminister it is inconsistent across taxpayers and across time; and e) the system collects insufficient information and fails to use it well. Evasion is the biggest problem with taxes in Mexico, because there is too little effort to stop it. Taxpayer coverage is insufficient. A national survey carried out toward the end of 2001 reported that global coverage reached 83 percent -81 percent among individuals and 95 percent among firms. Nevertheless, in addition to the 7.1 million taxpayers registered in December 2001, there are approximately 1.9 million unregistered potential taxpayers. Of the registered taxpayers, 5.8 million must present tax returns, and of these only 3.9 million present them on time, thus producing a second gap. Not enough resources go to enforcement, particularly catching people and firms that have not filed or have not reported all their sales (VAT) or income. A relatively low percentage of the unpaid amounts identified through audits are actually collected. Of the court cases filed against taxpayers who fail to fulfill- their obligations, many of those submitted for administrative adjudication fail because of procedural errors, faulty notifications, and inadequate case preparation; there is a similar pattern for judicial cases. Of all the audit actions, only 30 percent lead to increased payment in the mid 1990s, compared with about 75 percent in other similar countries, and the most effective type of audit, home visits, are extremely -rare (Casanegra and others. 1997, pp. 2, 7). Human and technical resources should thus be shifted from collection to enforcement 'and be given training in methods of data gathering and investigation. If necessary, more tax auditors should be hired. Furthermore, sanctions are rarely imposed. The government perhaps should eventually stiffen sanctions, but the first priority is to apply the ones existing to any offender, even against those with political connections. While not paying taxes in Mexico is too easy, actually paying them is still too difficult. A well-functioning tax system requires that most people comply voluntarily with it, but in Mexico the payer often finds it difficult to file in a way that meets the rules. The introduction of the Tax Administration Service (SAT) in the late 1990s has brought some improvements, such as getting most payments at banks, but has still not achieved an institutional culture with the objective of customer service. Efforts to make sure that the filed returns are correct and acceptable have too often gone to the extreme of being a barrier to payment. Setting standards for prompt and helpful customer service, and training employees to meet these objectives, is an obvious priority. Establishing a voluntary compliance program, where returns are automatically accepted and then spot checked on a random basis, is another one. Both steps should be complemented by the expansion of the programs of payment at banks and filing and paying by computer. The complexity of the tax code contributes to the problems with compliance and enforcement. Officials and honest tax payers are often equally unsure of what is required. Evaders 19 take advantage of ambiguities and contradictions to reduce their payment and avoid prosecution. During the 1990s, parts of the tax code originally designed to simplify collection for particular classes of taxpayers became, in practice, open-ended legal constructions that year-after-year have increased the difficulties of tax administration. The code lends itself to tax planning combinations that allow real profits to be transferred to companies eligible for the special regimes. For example, the increase in the allowance for conglomerates to consolidate tax liabilities has also increased the possibilities of tax avoidance through a) using standard consolidation provisions to buy up loss- making enterprises for the sole purpose of reducing or eliminating taxable profits, and b) combining the benefits of consolidation with those of double-tax treaties to reduce tax liability. For example, in 1991 the special cash-flow taxation in the agricultural and transport sectors replaced the previous presumptive taxation, as a way to bring small traders into the tax net. Both presumptive and cash flow taxation have their particular disadvantages, and both enable taxpayers who are subject also to the general regime to arrange their affairs to reduce or eliminate their taxes. Such difficulties indicate that the overall simplification of the tax code, as described in the previous section, especially the elimination of privileges, is critical to improve the quality of tax administration in Mexico. In terms of consistency of application, Mexico exemplifies to an extreme the adage that "tax administration is tax policy" (Bird and Casanegra 1992). Bureaucrats in the tax administration, even some lower level staff, often have wide discretion to interpret the law and regulations, thereby determining the effective tax rate and base. The situation encourages corruption among officials under the pretext of misinterpretation. It also has rendered both tax inspectors and taxpayers unsure of what is the proper amount of tax payable. When administration is delegated to the states, as with car taxes and the VAT enforcement, the problem becomes even worse. It is thus imperative to publish national standards for tax implementation and train tax administrators to follow those standards. To assure that tax administration meets national standards, most OECD countries have an independent tribunal to which taxpayers can appeal and Mexico should follow this example, as it has started to do with the Tax Court. The independent tribunal generally does not include the tax authorities but rather is composed of professionals with tax expertise-accountants, lawyers, judges, and the like-appointed by the government. Finally, Mexico's tax administration lacks adequate information and does not use well what it has. Enforcement is weak in part because it does not seek and use information beyond what the filer submits for the tax in question. There has not been a unified system of taxpayer numbers, and tax filers are not required to report the tax ID numbers for all persons and firms involved (workers, suppliers, and business customers). Current programs seek to address both problems. Indeed, information reporting requirements were reduced in the 1990s. The electronic data management systems are not all compatible, even within SAT, or capable of sharing data. Tax authorities should therefore increase data-reporting requirements for major transactions, including the ID numbers of all taxpayers whom they pay or by whom they are paid. The computer systems should be fixed to enable efficient sharing of information, including between authorities at different levels. Enforcement should use information that is available from other taxes and fees; for example, data from the VAT and automobile registration (tenenca) could indicate unreported corporate and personal income; and data from personal and corporate income could reveal property for which no property tax was being paid. The goal of reducing evasion needs to be at the core of efforts to improve the informnation systems. Revenue effects. The revenue effects from improved administration are harder to predict than the effects of policy changes, holding enforcement constant. Historical and international 20 comparisons provide some indications. As mentioned earlier, the revenues from stronger enforcement of the VAT in 1992-94 with a 10 percent rate, compared with revenues with a 15 percent rate in 1998-99, suggest that improved enforcement could raise VAT revenues by 25 percent. This is conservative compared with the CIDE estirnate of 60 percent evasion (reducing evasion to 40 percent would raise revenues by 50 percent) and with the international comparison of revenue shares in Table 3. For the income taxes, revenues in the early 1990s peaked at about 20 percent higher than in recent years, and CIDE estimates of the potential for increasing income tax revenues suggest imply a possible 25 percent increase due to better enforcetent.'9 The revenue increases from improved administration should not be considered as an alternative to policy reforms, for the reforms to eliminate exemptions will contribute to the hoped- for advances in collection and enforcement. If Mexico decides that the government does not need as much revenue as obtained with the combination of policy and administrative measures, then it should cut the standard tax rates, after eliminating the exceptions and as the administrative improvements are actually bringing in more revenue. Tax Administration Reforms. The Bank, with a proposed Tax Administration Institutional Development Loan, is supporting the Government in strengthening and implementing the institutional development program of the SAT. The project has seven components: a. eSAT and taxpayer senices, to strengthen SAT's on-going efforts to provide better taxpayer service and increase the transparency of SAT. The project will further simplify taxpayers' filing of returns by employing more advanced electronic means, including all web-based activities under SATs e-business strategy. b. Management of human, financial and physical resources, to help SAT establish an Integrated Resource Management System for planning and managing its manpower, financial and physical resources: assessing SATs staff, documenting the results in a Human Resource Management System, upgrading training, and introducing performance management systems. c. Information hub, to improve the communication, sharing and management of information within SAT. A Chief Information Officer will take charge of policy, coordination, quality assurance and security of information, and wiU implement a series of databases that will be optimized for access and security and will automatically receive data from the operational databases in the various areas. d. Technology services, to develop an ICT strategy aligned with SAT's business strategy, to provide technical assistance for the development and implementation of an Internal Service Policy for ICT services, and to upgrade equipment to improve performance. e. Compliance prvcess, to improve SATs units for Collections and Large Taxpayers, to develop a management information system and a Standard Notification System (with performance indicators and a control system). The component would also finance increased coverage of the Taxpayer Registry and the creation of the single taxpayer account. f. Compliance contml and enfornement, to integrate into a "Compliance Control and Enforcement" process all the elements of SATs enforcement, auditing, legal and other functions directed at problems of evasion and non-payment. g. Planning and .Qualipy Control, to provide training, seminars and technical assistance on planning and performance indicators to senior and middle managers in order to carry forward the 19. Comparing the collection of income tax in Mexico with that in the US, which collects 11 percent of GDP with the same top rates, CIDE estimates that the combination of broadening the base and improving administration in Mexico could close about half of the gap, raising an additional 3 percent of GDP (CIDE 2000). 21 strategic planning exercise and to improve it as a means to achieve the necessary, but previously missing, coordination among administrative units. IV. SUB-NATIONAL TAXES AND FISCAL TRANSFERS Sub-national governments in Mexico, as noted earlier, collect very litde revenue either as a share of GDP (about 1 percent) or as a share of their own expenditures (about 10 percent). This imbalance is inconsistent with the reality that has emerged in the past ten years: political and constitutional autonomy of the states, opposition control of the federal congress (and of many state congresses), major spending responsibilities in the hands of states, and substantial subnational autonomy to borrow. Redressing that fiscal imbalance is more than a technical tax matter, as it will implicitly define the kind of federal country Mexico wants to be (Courchene and Diaz-Cayeros 2000). Much of the imbalance has a long historical background. The taxes and fees that are the main source of revenue in Mexico are now clearly allocated between levels of government, so that there is virtually no overlap vertically between levels and only minimal tax competition horizontally between states. These are important achievements for making taxes more efficient. This was not always so, as counterproductive tax wars were common in the first half of the 20'h century. A series of tax agreements from 1947 to 1980 stopped those wars and centralized taxation much more than in most federations. Although there has been much fiscal decentralization during the last decade, none of it has affected taxation, and taxes remain as centralized as they were twenty years ago. Indeed, the increase of transfers in the 1990s has decreased the attention of subnational governments to raising own revenues, which have declined in many cases. The constitution does not say much about taxes, only that taxes and fees on natural resources (including oil) are exclusively for the federal government. The law of the Sistema Naaonal de Coordinacidn Fiscal (SNCF) allocates most of the taxes and was set largely in its present form in 1980. Income taxes, excise taxes, and the VAT are exclusively federal, although an agreement delegated to the states the enforcement powers (fiscak!atci6n) for the VAT and the right to keep the arrears and fines they collect. Two other federal taxes, the sales tax for new cars (ISAN) and the annual registration fee (tenencia) for cars less than .10 years old are "shared" 100 percent with the states, effectively giving them the right to collect and enforce the taxes. Since they are federal taxes, however, people pay them in any state they choose, based on convenience, irrespective of location of residence. The annual fees for older cars is fully allocated to the states. The main source of own revenue for most states is the payroll tax. Most states charge one or two percent (one state charges four percent). States depend on federal transfers for 85 to 95 percent of their revenue. Municipalities depend almost as heavily on transfers.' Their main source of municipal own revenue is the property tax (predia). They also collect various fees, the most 20. The federal district, with an elected executive since 1997 and with characteristics of both a state and a municipality, collects predial and payroll taxes, raising enough to pay for over half of its outlays. This is far more than other states and demonstrates the untapped potential for revenue raising in at least the more economically advanced states. 22 important of which is for water, although this could be counted as a revenue of the water 21 The centralization of Mexican taxes, as culminated in the SNCF described above, represented a bargain in which the states and municipalities gained a share of 20 percent in most federal revenues-oil royalties, income tax, excises, and VAT, but not import duties. This amounts to almost half of the federal transfers received by states, and these transfers (pati ip acones) are administered by the Subsecretaria de Ingresos (in SHCP). Most of the rest of federal transfers are earmarked for sectors and not linked to taxes, and the Subseetada de Egresos administers them, in conjunction with the sectoral ministries. The allocation of partiipaiones between the states initially followed the distribution of tax revenues that the states relinquished under the 1980 Pacto Fiscal, but since then has become somewhat more on a per capita basis and thus redistributive. The allocation formula since 1994 distributes 45.17 percent of the participaciones according to the previous year's allocation plus an adjustment to reward relatively strong tax collection (or to punish weak collection) by the state in the previous year. This encourages states (although not very strongly) to administer taxes better and to foster economic development that expands the tax base. Another 45.17 percent of the partiapaciones is distributed according to population numbers (the progressive part). The rest is distributed by the inverse of the two other criteria, effectively giving more aid to small poor states. Of the parlicipaciones that come to the states, 20 percent is automatically passed on to the municipalities by formulas set by the states. The typical formula is similar to the federal formula for distribution of general revenue sharing among states: 40 percent is linked to the revenue collection (property tax and water fees) in the municipality, 40 percent is proportional to population, and 20 percent provides compensation (to small poor municipalities) that rate low on the other two criteria. In the past, the federal government would also make extraordinary transfers to states and municipalities (until the late 1990s) or it would do federal investment projects with strongly local benefits. These were often large, larger than the sub-national government's own revenue collection, and their allocation was highly politicaL Also, borrowing was usually done with a federal guarantee, which might turn into a federal bailout and effectively a partial grant.' So a mayor or governor in need of resources might try to get more federal money or borrow from soft sources, rather than try to raise more taxes. Indeed the decision to raise local taxes seriously was often simultaneous with the decision to break with the party in office at the federal level (Diaz, Magaloni and Weingast 2000). In more recent years, however, the situation has changed dramatically and perhaps permanently. Discretionary federal transfers are no longer available. The allocation of federal investment is mostly disclosed, and is potentially susceptible to congressional oversight and control. New financial regulations have eliminated the federal guarantee of state borrowing and made it more costly for banks to lend to uncreditworthy states (Giugale and Webb 2000). 21. In many cases the water tariff is a flat fee for connection, not a metered rate based on consumption. The water sector has many problems, extending beyond the scope of this report, but two important ones are the frequent failure to collect water tariffs and the frequent failure to pass the collected tariffs along to the state and national entities that provide the water. 22. Up to 1996, the guarantee was explicit, but the subnational debt crisis of 1995 (part of the general economic crisis) led officials to end the guarantees. Banks were unwilling to lend to states without guarantees, so the federal govertument agreed to accept mandates from states to withhold debt service from their revenue-sharing participaciones, which restored an implicit federal guarantee. 23 The total of own tax resources plus transfers (but not counting federal investment) is relatively equal per capita across states. Although states with similar levels of poverty have sometimes strking differences in total resources per capita, the governments in poorest states have about the same level of per capita resources on average as the middle income, and only slightly less on average than the least poor states. This is because the states with higher GDP per capita tend to get more resources per capita from own revenue and participaciones (Ramo 28), but this is offset with higher aportaciones (Ramo 33) per capita to the poorer states. See Figure 3, where the states are ranked with the states with the highest marginality index of poverty on the left and the states with the least poverty on the right. The recent federal government measures-especially ending extraordinary transfers and tightening the regulatory environment for subnational borrowing-should give the states more incentive to collect taxes, because now each state's creditworthiness and access to capital depend significantly on raising its own revenue. The results remain to be seen. These federal policies need to continue, but more is needed as well. States and municipalities need technical help to develop their ability to collect taxes, not only with training but also by accessing federal data bases, such as that of social security, which is one of the best, and the corporate asset tax. 24 Figure 3. Subnational Resources fot Each State in 2000. 8,000 - fOwn Revenues U Parficipaciones O Excise Taxes El Aportaciones * Fondo de Apoyo p 7,000- 6,000- 5,000.- p 0~~~~~~ r4,000-- C 3,000 -- a p 1 2,000 - - - - Sourre- SHCP data and World Bank calculations. A growing reaction to the new political and economic context is that many states, typically those with the strongest economic growth and increased need for public services, are demanding more authority to impose taxes. As the motivation and technical capacity of the states are reaching an adequate level, the federal government needs to devolve some tax bases to the states. Some excise taxes are one possibility. A larger step, eventually requiring a revision of the Pacto Fiscal, would be sharing a tax like the VAT between the federal and state levels. The sharing could go in stages: first, setting up the reporting and information system to know the location where consumption occurs; second, giving each state a fixed percentage of the VAT corresponding to the calculation of consumption within its borders, like 2 percentage points of the 15 percent total rate (with 13 percent still going to the fedetal level); and third, possibly much later, going to a full piggy- backing system where the states can set their own rate of VAT or retail sales tax for sales within the state. The VAT and the retail sales tax (RST) are both consumption-based taxes, but they are collected in different ways. Administrative considerations, therefore, dictate whether and how they can be efficiently devolved to states. It is difficult for states separately to adninister theit own VATs or RSTs, especially if there is also a federal VAT and even more so if the bases differ. On the other hand, it is readily feasible for the states to have their own VATs or RSTs if they have the same base as the federal VAT and if the same national administration (federal jointly with states) administers the taxes at both levels. Box 1 explains why. For more details see McLure 2000, Varsano 2000, and Diaz and McLure 2000. For those stages to be feasible, there would need to be some revision of the VAT/RST administration, making federal-state coordination even doser than at present. 25 Box I. Devolving Consumption Taxes: VAT or RST Taxinsg ronusdon. A consumption-based tax applies only to sales to households; it does not burden sales to business. An ideal VAT is collected on virtually all sales by registered traders, except those for export, without regard to whether the sale is to. a household or unregistered trader or to another registered trader. Taxation of business purchases is eliminated by allowing registered traders to take "input credit" for taxes paid on purchases against tax due on sales. Thus the only tax that is not eliminated by input credits is that paid by households and unregistered traders. An ideal RST achieves directly the objective of exempting sales to business; it applies only to sales to households (and unregistered traders) and exempts al exports and sales to registered traders. The administrative problem is how to achieve this objective, without opening the door to evasion by households claiming to make business purchases. It would be sensible to condition exemption for sales to business on presentation of a business taxpayer identification number to the seller, so that, in principle, only purchases that would be eligible for input credits under the VAT would be exempt under the RST. With devolution, the federal tax administrations would, of course, need to verify that purchases by registered traders are for a legitimate business purpose, since only those are eligible for exemption (RST) or input credit (VAT). The weakness of the RST is that the buyer merely needs to lie to the seller to get the exemption. By comparison, under the VAT the buyer must pay the tax and then lie to the tax administration in claiming credits for purchases made for personal use.' Devoving taxes. Allowing states to have their own VATs or their own RSTs (not both) is straightforward administratively if there is a federal VAT, as in Mexico, and if the states' taxes have the same base as the federal VAT and use the same administration. Such an arrangement is called piggy-backing, because the state tax rides on the back of the federal. If the rates for state VATs or RSTs are the same for all states, which would facilitate the initial phase of transition, then the revenue going to each state wil be the same with VAT or RST, allocated on the basis of the vendor location for sales to consumers and non-registered traders. The tax would generate no incentive to buy outside one's own state. Thus with either system, the administration would initially not need to collect a compensating VAT (see below) nor to find out the location of domestic purchasers (except for direct interstate sales, as by mail order)2, only whether the purchaser was a registered trader or not, which is already done for the VAT. Location by state of each sale to a consumer or non-registered trader would be the only additional information needed, and with equal rates in all states, there would be little incentive to misreport. The standard advantages of a VAT administration, noted above, would pass along to both types of piggy-back state taxes, as long as they are not too high relative to the federal VAT rate. If or when the states are allowed to and have different RST or VAT rates, as would be the expectation under current proposals, some differences between the systems emerge. With either tax, final sales would migrate to states with lower rates, so a national floor might be necessary. Administration would be still manageable, as Icing as the tax bases and administration are the same as the federal VAT. With the state RST, sales to registered businesses would be taxed only by the federal VAT, so buyers would have the usual incentive to misrepresent themselves as registered traders, in order to avoid the RST. But this would be more difficult with collection of the federal VAT at the same time. Enforcers could check that the purchases of inputs as registered traders matched in appropriate ratios with their volume of sales reported and taxed by the at the next stage. For example, a simple consumer who used a phony business to avoid the RST would show-up obviously, because he would be collecting no VAT on further sales. With the state VAT, sellers would have to collect information on whether a domestic sale to registered traders was to someone in-state or out-of-state and charge perhaps different rates for these sales-the own-state VAT or the compensating tax on interstate sales. (Varsano 2000 and McLure 2000 offer details on how the compensating tax would reduce the problem of evasion with multiple state VATs). There could be problems with excess credits of traders in some states, but these would not be large if there were a national administration. Avoiding damngrs dimnfication. If the base for the state RSTs or VATs and the federal VAT were not substantially identical, vendors would need to distinguish between four possible combinations of tax treatment, in addition to the distinction between sales to registered traders and other sales for purposes of the RST: taxable under both federal and state taxes, exempt under both, and taxable under one, but not the other. See Diaz and McLure 2000 for details. Separate administration of state RSTs (from the federal VAT) would expose them to all the problems of 26 evasion that have led most countries to move to the VAT. Revenues would probably fall precipitously unless states raised the RST rate substantially above whatever percentage points of VAT were devolved to them; and even then revenue might fall. Thus, any devolution of the VAT to states (as VAT or RSI) should keep a consistent national base and administration. Box Notes: Special thanks to Charles McLure for his inputs and advice to this box. 1. Too much is made of the cross-checking that is made possible by the VAT-credits taken by registered purchasers must be shown on invoices issued by vendors, and thus subject to tax. This is probably less important than confirming that purchases are for legitimate business use 2. This an important problem in the United States and Canada and will grow in importance as the economy of Mexico develops. Note that this discussion pertains directly only to remote selling sales of tangible products sent to the customer across state lines. It does not concern cross-border shopping, in which the customer buys a product across the counter in a state where she or he does not live and takes it home. The latter transactions surely will be taxed in the state where the sale is made (the state of origin), and not that of the state of destination, except in rare cases, for example, where goods must be registered in the state to be used there, such as automobiles, boats and planes. Interstate sales of digital content over the Internet can be handled in the same way as interstate sales of tangible products, if the vendor knows the location of the buyer. Where vendors do not know the location of buyers, it might be appropriate to subject remote sales to households and unregistered traders to a uniform tax and divide the resulting revenue among the states in proportion to estimated consumption in the various states. This is almost certainly a small problem. The more difficult problems in this area involve sales originating outside the country. It is extremely difficult to tax a transaction in digital content between an unknown customer and a foreign seller (perhaps located in a tax haven), especially if payment is made in untraceable money, like credit cards. There is no ready solution to it. If the federal government no longer puts its VAT revenue (say, from the 13 percent federal rate, if 2 percent were given to the states) into the pool for parnicipaciones, states with weak tax bases would lose to an unacceptable degree unless there were also a revision of the parficipacones formula to compensate them. So renegotiating the transfer formulas would probably go hand-in-hand with renegotiating the allocation of tax bases. Designing such a revision will be technically and politically difficult, but possible (Courchene and Diaz-Cayeros 2000). The devolution of taxing authority should improve efficiency and motivation for tax effort, generating an increase in total revenue, so that a compensation scheme could make all states and the federal government financially better off than before, even without raising rates. In 2001, discussions on expanding the VAT base also included the idea of giving each state a fixed percentage of the VAT (probably as a retail sale tax) corresponding to the calculation of final consumption within its borders. States could charge 2 percentage points on final sales, for instance, and the federal government would reduce the VAT on final sales from 15 to 13 percent. The federal government, on its side, would not like to reduce its VAT rate without some compensating reduction of partiacpadones, and in the 2001 discussions it wanted a full compensation, so that the change would be revenue neutral relative to the estimation of their position with an expansion of the VAT base. The coalition to pass such a proposal it did not hold together in 2001, but it remains a possibility, and Figure 4 shows the simulation of winners and losers.t 23. Note that Tabasco is a big loser in the simulation, because of the extraordinary level of participaciones per capita it receives now, and Chiapas loses slightly. The federal government would, of course, compensate these two states to avert any absolute loss. States are ranked as in Figure 3, according to the index of marginality. 27 Figure 4. Simulation 1. Winners and Losers from 2001 ProposaL 800 600 P 400 o 200 C -400 aP -600 -800 a -1000 -1200 Sourre: World Bank Calculations. The Revenue Law for 2002 gave the states the right to tax consumption at the point of final sales to consumers at up to a 3 percent rate (with no reduction of the federal VAT) and to tax the income of individual entrepreneurs with medium and low levels of income. However, none of the states has taken advantage of this right. The states insist that the devolution of the tax should be done in a regular law, with effect for more than one year, and they do not want to have partiaipaaones cut. Figure 5 shows the disttibution of the state taxes and revenue sharing that would result from devolution to the states of a sales tax, with no change in partiapadones. In this simulation, the aggregate revenue from the devolved tax is based on the historical revenue from the VAT, with a horizontal distribution across states based on their relative GDP. As one can see from Figure 6, with the winners and losers, everyone wins, although not surprisingly the poorer states tend to win less. We cannot do a simulation of devolving an income surtax to the states, because we do not know how much is currently collected even in the aggregate from the small enterprises that would be paying the tax, but the geographic distribution (but not the levels) would look similar if the basis for distribution were state GDP. 28 Figure 5. Simulation 2. Allowing States to have a 3% Sales Tax, without affecting participaciones. * Own Revenues * Participaciones 0 Excise Taxes El Aportaciones * Fondo de Apoyo * 3% Sale. 7,000 p 4 _,000 a3,D Source Wotld Bank Calculations. Figure 6. Simulation 2. Winners and Losers. 1200. p a1,000 p 6,000 aI 2.000 Sourre: World Bank Calculations. 29 V. TAx REFORM IMPLEMENTATION STRATEGY The previous sections have identified reforms needed in both the tax regime and its administration. Proper grouping and sequencing will let the reforms work better technically and will make them more feasible politically. A strategic plan for Mexico's tax reform would have three parts. 1. Improve the tax law by reducing or eliminating tax expenditures and improving the methods of taxing petroleum and other excises. As in many other countries, but perhaps of more than usual importance in Mexico, taxation is a game of opening, closing and recreating loopholes. Entrenched political forces have, with a few short interruptions, perpetuated many privileges that reflect the structure of power in Mexico. International experience indicates that a comprehensive plan usually has more success-putting together a consensus among most major interests that they will all accept an end to their special treatment. Mexico largely achieved this for the income tax at the end of 2001; improving indirect taxes and energy taxation should be the next focal point, and probably will. 2. Imtprve administration and reduce evasion by making compliance easier for those who are paying and making enforcement more thorough for those who are not. This will require better information systems, collecting more information on each payer and potential payer, and linking the information from different tax bases and other sources. It will also require more autonomy and better incentives and working conditions for the tax agency and its staff. An independent tax tribunal to handle appeals should monitor the decisions of the tax agency and ensure the substance and appearance of fairness. Political impetus must come from the top. Some powerful people will have to pay more, and suffer penalties if they do not. If this does not happen, and does not happen visibly, no one will take tax reform seriously. Mexico had a period of more serious enforcement in the early 1990s, showing that it can be done. Making compliance easier and enforcement more forceful will reinforce each other, and can even lead to a virtuous cycle if social norms change and people who pay exert political pressure to make others pay as well. The Tax Administration Institution Development project addresses this issue. 3. Improve the tax effort and authority of sub-national governments. The last decade has witnessed rapid decentralization of spending authority and transfer of resources from the center. The federal government has little legal authority to monitor the use of these resources or to control directly the amount the states borrow with transfers as collateral. So there are limited opportunities for the federal government to exercise the role of principal over the states as agents. The taxpayer- citizens are the other potential principal over the states as agents, and they will come into this role more seriously if the states receive greater tax authority in order to fund their programs. In other words, sub-national taxation will strengthen sub-national democracy and accountability. Some progress is possible imnmediately in all three areas, but full results in each of them will come sequentially. Improving the tax code first is essential to make substantial progress with compliance and enforcement. Improving policy and administration at the national level will be necessary before substantial devolution of taxes to the sub-national governments can take root, because the most promising ways of devolution would be to share the VAT, PIT, or excises with the states, and eventually to let the states choose what rate they would piggy-back onto the federal tax. This would work well only if the corresponding federal tax is working well. 30 REFERENCES Bird, Richard and Milka Casanegra de Jantscher (1992), Improving Tax Administration in Developing Countries, International Monetary Fund, Washington, D.C. Carey D., K. Gordon, and P. Thalnann, (1999), "Tax Reforn in Switzerland," OECD Economic Departnent Working Papers, No. 222, pp. 55. Casanegra de Jantscher, Milka, Paulo dos Santos, Julio Escolano and Patricio Castro (1997). "Mexico: Fortalecimiento de la Administraci6n Tributaria Federal." Fiscal Affairs Departmnent, International Monetary Fund. Washington, DC. Casanegra de Jantscher, Milka, Anthony Pellechio, Julio Escolano, and Paul Bernd Spahn (1995), Mexico: Strrngthening the Fiscal System for Growth and Stability, Fiscal Affairs Department, International Monetary Fund. Washington, D.C. CIDE (Centro de Investigacion y Docencia Economnica) (2000). Los Impuestos en Mexico: Quieny Como se Pagan? CIDE, Mexico City. Courchene, Thomas and Alberto Diaz-Cayeros (2000), "Transfers and the Nature of the Mexican Federation," in Giugale and Webb, eds. (see below). Dalsgaard T. and M. Kawagoe (2000), "The Tax System in Japan: A Need for Comprehensive Reform," OECD Economic Department Working Papers, No. 231, pp. 87. Diaz-Cayeros, Alberto; Beatriz Magaloni and Barry R. Weingast (2000), 'Democratization and the Economy in Mexico: Equilibrium (PRI) Hegemony and its Demise." Stanford University, Processed. Diaz-Cayersos, Alberto and Charles E. McLure, Jr. (2000), "Tax Assignment," in Giugale and Webb, eds. (see below). Gil Diaz, Francisco (1995), " Fiscal policy and Tax Administration: The Experience of Mexico," in TaxAdministration Reform in Latin America, Inter-American Development Bank, Washington D.C. Gil Diaz, Francisco and Wayne Thirsk (1997), "Mexico's Protracted Tax Reform", in Tax Reform in Developing Countries, World Bank, Regions and Sectoral Studies. Giugale, Marcelo and Steven B. Webb, ed., (2000), Achievements and Challenges of Fiscal Decentra/itafion, Lessonsffrom Mexico. World Bank, Washington, DC McLure, Charles E., Jr. (2000) "Implementing Subnational Value Added Taxes on Internal Trade: The Compensating VAT (CVAT )," International Tax and Puiblic Finance Vol. 7, No. 6 (December): 723-40. OECD 1999, Economic Survey of Mexico, OECD, Paris. 31 Varsano, Ricardo (2000). "Subnational Taxation and the Treatment of Interstate Trade in Brazil: Problems and a Proposed Solution," in Burki and Perry, et al ed., DecentraliZation and Accountabiliy of the Pubc Sector, Proceeding of the Annual World Bank Conference on Development in Latin America and the Caribbean 1999. World Bank, Washington, DC. World Bank (1989), Mexico: Tax reform for Efficient Growth, Volume I - Main Report No. 8097-ME (November), Washington DC. World Bank (1999), Government Programs and Poety in Mexico, Report No. 19214-ME , Washington DC. World Bank (2000), Mexico: Fiscal Sustainabiliy, Report No. 20236-ME , Washington DC. Zodrow, George R. and Charles E. McLure, Jr. (1991). "Implementing Direct Consumption Taxes in Developing Countries," Tax Law Reuiew 46 (4): 405-487. 32 Annex I Al ANNEX 1 Mexico: An Evaluation of the Main Features of the Tax System A. Tax Revenue Developments in Mexico * 1989-1993 * 1994-1998 * 1999 B. Revenue Adequacy and International Comparisons * Tax levels and the Budget Deficits C. Buoyancy of Revenues D. Revenue Stability E. A Suggested Reform Package * Feasibility and Strategy * Content of the Tax Reforn 33 Annex I A Tax Revenue Developments in Mexico' Al. Although ostensibly a federal system, Mexico remains a highly centralized country from a tax revenue viewpoint. The share of federal revenue in general government revenues has declined slightly over the last two decades from close to 88 percent in the early 1980s to 83 percent in 1998. Some of the difference went to social security agencies and to subnational government own revenues. In 1998, subnational government own revenues represented less than 6 percent of general government revenues. A1.2 Of the Federal Government taxes, the combined income taxes (personal income tax (PIT) and corporate income tax (CIT), including the gross asset tax) are the most important source of revenue. In 1998, income taxes accounted for 31 percent of total federal revenues. The second most important source of federal revenues is the VAT, which in 1998 represented 22 percent of total federal revenues. The third most important source of federal revenues is hydrocarbon duties paid by Pemex. A very close fourth source of revenues is excise taxes, which have increased significantly over the last two decades. The relative importance of taxes on international trade has decreased as expected from trade liberalization before and after NAFTA to 4 percent of federal revenues in 1998. A1.3 The relative importance of income taxes has fluctuated significantly over the last two decades. These oscillations reflect frequent changes in tax policy, the high sensitivity of income taxes to the business cycle, fluctuations in revenues from petroleum, the significant hike in privatization proceeds in the early 1990s, possibly the effects of inflation and different moods of tax administration with respect to controlling evasion. 2 Income taxes were as low as 23 percent of all federal revenues in 1987 and as high as 36 percent in 1980-81. The relative importance of the VAT has also fluctuated widely during the last two decades. This fluctuation reflects, among other things, the increase in collection in 1990 after the federal tax administration took over collection from the regional administrations and the several rate changes like the drop in 1992 in the general rate from 15 to 10 percent and its increase again during the 1995 recession. A1.4 Between 1994 and 1999, there was a change in the tax mix. While, in 1994, the income tax comprised 45 percent of tax revenue, VAT 24 percent, and the tax on production and services (EEPS) -importantly gasoline and diesel-17 percent, by 1999, the proportions had changed to 41 percent, 29 percent and 20 percent respectively. To 1. Tables Al. I to A 1.4 provide the data on which this section comments. 2. According to Gil Diaz 1995, the CIT showed decreasing revenues during the 1980-87 period in part due to the asymmetric inflation adjustments to assets and liabilities in the midst of high inflation and the Oliveira-Tanzi effect of inflation on tax collections. With the full inflation adjustment to tax bases introduced in 1997 and the more frequent filing and adjustment for both liabilities and assets, the effect supposedly disappeared thereafter. For the PIT the lack of indexation until 1988 worked to produce more revenues, but after indexation this advantage was lost. On the whole, Mexico's tax system does not appear to have gained or lost due to inflation over the last two decades. The regression of real revenues from income taxes on the inflation rate is not statistically significant after controlling for serial correlation and real GDP. 34 Annex 1 some extent, this reflected the increase in the general VAT rate from 10 percent to 15 percent. Al.5 Policy changes during three different periods, 1989-1993, 1994-1998 and 1999 explain the evolution of tax collection in the 1990s. Between 1989-93 an important fiscal reform effort was launched. It comprised both revenue policies and an unprecedented administrative effort to organize, automate and monitor and control tax obligations. The second period (1994-1998) starts at the end of the Salinas de Gortari administration in 1994 with the weakening of revenue performance as a result of the introduction of a number of new tax expenditures and relaxation of tax controls. Since early 1999 there has been an effort to recover revenue looses that emerged during the second period of the 1995 crises. A1.6 As percent of GDP, federal government tax revenues both in 1994 and in 1989 were 11.3 in spite of important re-structuring of the composition by tax due to important reforms in the main taxes. These reforms included major amendments to the income tax that completed the adjustment for inflation of CIT and PIT, eliminated presumptive calculations done for privileged sectors paying under the special Tax Bases regime (BETs) and the small payers regime.3 1989-1993 A1.7 During this first period, despite the loss of revenue through generalized inflation adjustments of the PIT schedule, made in 1987, rates of PIT were lowered from 42% in 1986 to 40% in 1987 and to 34% at the end of 1994. The revenue loss from such tax rate reduction was not perceptible at the time, perhaps because of such other factors as (a) the introduction of a gross assets tax (impuesto al activo), (b) the conclusion of the amortization of losses linked to the high inflation periods in the past, and (c) the windfall gain -in terns of tax collection-linked to the negative Tanzi effect. Also, there were positive results of administrative efforts geared to improve taxpayer compliance. A1.8 Corporate income taxes (CIT) decreased from 2.52% of GDP in 1989 to 2.36% in 1993, while total income tax (CIT plus PIT) increased from 4.72% of GDP in 1989, to 5.13% in 1994. It is clear that this increase came from PIT and not CIT. Two factors were responsible for changes in VAT revenues during this period: (a) the federal tax administration took back from the states its full administration; and, (b) the application of a single 10% rate. The general rate was reduced from 15% and the lower border states rate that was established in 1980 was increased from 6%. In addition, the collection of this tax benefited from two additional elements: the across the board price adjustments of goods and services sold by public enterprises including Pemex and the increase in VAT collection by the customs as the result of the opening of the economy. As a percent of GDP, VAT decreased by only 3.1% in 1989 to 2.7% in 1994, in spite of the 33% reduction of the general rate. 3. In practice, the BETs allowed complete sectors of taxpayers to remain out of the tax circuit while permitting them to deal with other taxpayers by issuing receipts to back-up deductions. 35 Annex I A1.9 Custom collections showed an important improvement compared to earlier periods. (customs revenues increased from 0.7% of GDP in 1989 to 0.9% in 1994, and reached a peak of 1.1% in 1992 --the highest collection level during the last 20 years. A 1.10 Finally, a rather small tax that contributed to the increase is the Motor Vehicle tax (Tenencia). As a result of increasing rates and broadening its bases revenues increased from 0.07% of GDP to 0.22% in 1994. 1994-1998 A1.1 1 The second period started at the end of the Salinas de Gortari period when ad hoc tax reductions were granted. Revenue losses were compounded by the 1995 economic crisis and led the Government to increase VAT general rate to 15%. In contrast to the earlier rate change (reduction) when revenue decreased only slightly, VAT collection barely increased in spite of the 50% rate increase. The severe financial restriction during 1995 and 1996 practically stopped business from delivering the VAT collected. VAT was kept as their cash flow. The economic situation improved in 1997 when VAT collection started again to increase as percent of GDP. Nonetheless, collection did not catch up with the earlier years. As percent of GDP, VAT increased from 2.8% in 1995 to 3.3% in 1999. A1.12 For a number of different reasons the ratio of income tax revenues fell from 5.13% in 1994 to 3.85% in 1996. In addition to the losses caused by the crisis, the cost of the negative tax introduced in1994 which was further increased in 1995 and 1996. Income taxes on salaries fell from 2.29% in 1994 to 1.83% in 1999, in spite of a recovery in real wages. Part of the loss may, therefore, be attributable to the negative income tax schemes. In addition, the observed behavior of tax collection during this period responds to a widespread deterioration of tax compliance induced by the 1995 crisis accompanied by a reorganized tax administration team that was not able to tackle the problem. 1999 A1.13 It was not until 1999 that income tax revenues recovered to 4.6% as the result of a tax revenue reform in 1998 that closed or reduced some observed loopholes in the system. In particular, it curbed benefits from fiscal consolidation, eliminated immediate deduction of capital investments, redefined income for insurance companies and eliminated the exemption to maquiladoras with permanent establishments in Mexico. B Revenue Adequacy and International Comparisons Al.14 One measure of fiscal adequacy in any country is whether sufficient revenues are generated to meet the desired level of expenditures. Failure to generate a sufficient amount of revenues can be attributed to either an unrealistic level of expenditures or inadequate revenue performance. Previous studies have noted that the level of public expenditure in Mexico is too low for the physical infrastructure and human capital needs of the country. Given this consensus, one of Mexico's tax system's most important problems is its inability to yield adequate revenues to finance basic public sector goods and services. This statement does not represent any normative judgment of the fiscal 36 Annex I performance of Mexico, there is no absolute scale against which one can assess the right size of a country's relative size for public sector. The share of government in GDP reflects, among other things, the collective preferences of a country for public goods and services vis-a-vis private consumption. Clearly, from an economic standpoint, these preferences cannot be judged right or wrong. It is rather a positive statement: that the current level of revenues is not adequate to support all the expenditure programs that are desired. That is why the traditional, but normative "tax effort" is not used; "tax effort" is a misleading metaphorical synonym for tax levels which governments and society may or may not wish to increase. A1.15 One way to approach the question of revenue adequacy in Mexico is to ask whether the country's tax level approximates to that of other countries with a similar level of development and economic characteristics. Although there is no ultimate way to establish how high taxes should be in a country, international comparisons may provide a first indication of whether an adequate level of public services is being provided, or whether the level of public infrastructure is less than that required for the country's development. Comparing Mexico's tax level with that of other countries with similar levels of development also provides a lead on how much tax levels can increase in Mexico without getting out of line with potentially competitor countries for foreign direct investment. A1.16 As noted in the core report, tax revenues in Mexico have lagged behind those in all other OECD countries and, as Table Al.5 shows, behind nearly all Latin American countries. However, simply making comparisons of the ratio of tax revenues to GDP across countries may be misleading because the ability to collect taxes across countries may differ due to the availability of tax handles and, the overall level of development. Therefore, it is useful in making international comparisons to provide controls for these differences in ability to collect taxes. To this end, regression analysis is used to estimate the average capacity to collect taxes for a sample of countries, controlling for GDP per capita and other proxies for the ability to collect taxes. These regressions are then used to predict the tax level effort that on average would be expected given the per capita income and other relevant characteristics of any given country. A1.17 More specifically, in the regression analysis approach, the dependent variable is taxable capacity as measured by the ratio of tax collections to GDP. This measure of taxable capacity is regressed with a variety of proxies for the tax bases, which are the independent variables in the regression analysis. From each estimated equation, a predicted value of the tax collection to GDP ratio is obtained, that is, the amount the country could collect if it reached an "average" tax level. The effective level is then defined as the ratio of actual to predicted tax level. By calculating the ratio of actual to predicted tax collections, a cleaner measure of can be derived since there is a control on the impact of changes in the general tax base over time. A1.18 There are many different specifications that can be used in estimating tax levels across countries. Here three different models that have often been used in the past in this 37 Annex 1 kind of analysis are followed.4 Each of the models allows the derivation of an index for, ranking countries according to their effective level. These models are Model 1: T =a+b Y+cXm Model2: T =a+bMine+cAg Model3: T =a+bY+cXm+dAg, where, T = ratio of tax collections to GDP Y = per capita GDP Xm = ratio of the sum of exports plus imports to GDP Mine = share of mineral and fuel exports in GDP Ag = share of agriculture in GDP A1.19 As indicated above, the independent variables attempt to capture the effects of differences in overall economic structure on the ability to raise taxes. In particular, a higher level of per capita GDP proxies a generally greater ability to collect taxes. Similarly, a larger external sector (imports plus exports) and a larger share of mineral and fuel exports in GDP also proxy a greater ability to collect taxes. On the other hand, the larger the share of the agricultural sector in GDP, the lower the ability to collect taxes. The data used for the estimation are from the International Monetary Fund's Government Finance Statistics Yearbook 1999 and the World Bank's World Development Indicators 2000. The models are estimated for a sample of 32 developing countries from 1990 through 1996. By estimating tax levels over time, inferences are less likely to be drawn from the impact of special conditions in a single year. The results in Table A1.6 show that regardless of the model specified, Mexico ranks consistently among the bottom third of countries in the sample in terms of the level of tax. According to the three models, Mexico's tax level is 82 percent, 63 percent, and 70 percent of the international average.5 Based on these estimates, a "normal" ratio of total tax revenues to GDP for Mexico during this period would have been between 12.75 and 16.75 percent of GDP as opposed to the actual 10.5 percent. Tax Levels and the Budget Deficit A 1.20 An alternative approach to assessing Mexico's tax levels is to examine the amount and persistence of the budget deficit. The basic premise in this approach is that a budget 4. See, for example, Bahl et al. 1996. 5. Note that this international average is exclusively based on the countries included in the sample. The sample does not contain any high-income countries. 38 Annex I deficit that is high and persistent could be taken as valid evidence that tax levels are too low for the expenditure needs of the country. A1.21 The actual budget deficit in Mexico was very volatile during the 1980s and 1990s. In the first decade, budget deficits over 10 percent of GDP were common. During the 1990s, however, the bud get deficit never exceeded 4 percent of GDP and was frequently under 1 percent of GDP. A1.22 The sustained deficits during the 1980s reflect a transition period, when oil revenues were falling and neither spending nor taxes had adjusted enough to fill the gap. The budget imbalance during the 1980s reached 10 percent of GDP and closing the gap purely from the revenue side would have required practically doubling the existing level of tax . During the 1990s Mexico substantially reduced its fiscal deficit, partly by more revenue effort but mostly by cutting spending, especially for infrastructure. These cuts were costly for economic development, as discussed in the World Bank report, Mexico's Fiscal Sustainability 2000. C Buoyancy of Revenues A1.23 Another important property of a tax system is its ability to generate automatic growth in fiscal revenues over time. The adequate rate of growth depends on the expenditure goals of government. A natural benchmark for dynamic performance of a tax system over time is its ability to grow at the same rate as GDP. Tax revenues increase over time either because tax bases grow with the economy or because changes in the tax laws either broaden tax bases or increase tax rates, or because of better enforcement of an existing tax structure. When only the first effect is present the ability to grow is measured by the elasticity and when all effects can be present the ability to grow is measured by the buoyancy. A1.24 Revenue adequacy also needs to be understood from a dynamic point of view, as the tax system should collect more as the whole economy grows, because the demand for public services is very likely to expand as the economy grows. If tax revenues grow automatically with the entire economy, a balanced budget can be maintained without needing repeatedly to introduce new taxes or raise the rates of existing ones. A1.25 Formally,- the elasticity is defined as the ratio of the proportional change in revenues to the proportional change in the tax base. It is often the case that GDP is used in lieu of the tax base. If the elasticity is greater than one, the government is able to expand the provision of goods and services as the economy grows and is even able to reduce taxes by lowering tax rates, for example. Conversely, if the elasticity is less than one, the government will struggle to keep up with the services demanded by a larger 6. Though, as in most countries, published deficits may be subject to massaging according to whether the government wishes to increase or decrease tax levels. 7. The elasticity property is more important because it gives government the ability to respond to increases in the demand for public services without having to interfere continuously with the tax system with ad hoc revenue raising measures. Elasticities are more difficult to estimate because it is necessary to control for the impact of changes in the tax structure and changes in enforcement. 39 Annex 1 economy and to avoid budget deficits will have to introduce new taxes or increase the existing ones by some combination of rate increases, base broadening and improved collection. A1.26 A correct measure of tax elasticities requires the observation of changes in tax revenue arising exclusively from changes in the tax base. Often the observed changes in tax revenues are also the result of changes in the structure of taxes, such as tax rates or the definition of the tax base, or also changes in the tax administration, such as a stricter enforcement of the tax laws. When it is not possible to disentangle all these different effects, the ratio of the proportional change in tax revenues to the proportional change in the tax base of GDP is known as the buoyancy of the tax to differentiate it from the stricter concept of elasticity. Although some attempt has been made here to control for the impact of discretionary changes in the tax structure, the values reported should be interpreted as buoyancy coefficients rather than elasticities. A1.27 This section examines the buoyancy of the Mexican tax system using two similar approaches. First, the year-to-year buoyancy for overall revenues is calculated for each separate revenue source with respect to GDP. In the second case, regression analysis is used to estimate the average buoyancy of tax revenues over the last two decades. Al.28 The average year-to-year buoyancy of total general government revenues with respect to GDP is 0.93 (Table A1.7) for the period 1980-1999. For federal government revenues, which, as opposed to general government revenues, do not include social security contributions and subnational government tax revenues, the average year-to-year buoyancy for the period is 0.88. This buoyancy is a composite of the behavior of federal tax revenues, with an average buoyancy for the period of 1.2, and federal non-tax revenues, with an average buoyancy of 0.71. As shown in Table A1.7, the year-to-year buoyancy for individual sources of revenues show both the impact of changes in the economic environment, including recessions,8 changes in the price of oil, and changes in tax structure, such as the lowering of the general VAT rate in 1992.9 The buoyancy coefficients for the separate taxes exhibit several interesting patterns. For the last three years (1997-1999) the buoyancy of all the major taxes (income taxes, VAT and excises) has been significantly greater than unity, meaning that these tax revenues have increased more than proportionally to' GDP. This may reflect both the recovery from quite low performance in the previous years and also that the current tax structure has a built-in elasticity that can generate an increasing tax levels if no discretionary tax measures are adopted to offset this pattern. The periods 1993-1997 and 1989-1992 are both periods of low buoyancy, the first because of bad economic conditions and the latter because of discretionary changes in the tax structure. But apart from a few exceptions, the major taxes did not show significant buoyancy in the 1981-1986 period. 8. During the 1980-1999 period Mexico has suffered four recessions 1982-83, 1986-87, a mild recession in 1988, and a sharp but short recession in 1995. 9. Some of the buoyancy coefficients in Table A1.7 are negative. This is due to a fall of revenues in nominal values from year to year. For example, non-tax revenues declined sharply from 83 billion pesos in 1992 to 51 billion in 1993. 40 Annex I A1.29 Turning to the second approach for estimating the buoyancy of the tax system, regression analysis is used to derive the average buoyancy of the tax system over the period 1980-1999, the natural logarithm of the revenue series being regressed on the natural logarithm of the GDP series. The resulting coefficient for GDP provides an estimate of the average buoyancy of the revenue series over the period 1981-1999.10 These results are reported in Table A1.8. The most significant finding is that the estimated buoyancy for most revenue aggregates including general government total revenues, federal government total revenues, and federal tax revenues are unitary or very close. The same is true for the major individual taxes. The only significant divergence is for hydrocarbon duties which show a buoyancy over the entire period of 0.88. In short, Mexico's tax system, after many structural reforms, improvements in tax administration, and significant turns in the business cycle, managed to yield increases in tax revenues that just kept pace with GDP, neither faster nor slower. This should not be a surprising result given that the ratio of tax revenues to GDP has remained quite constant over the period. Of course, these results support (although do not prove) the hypothesis that Mexico's policymakers steered the system over the last two decades in order to maintain the same aggregate share of taxation. D Revenue Stability A1.30 Another significant feature of tax systems is their relative revenue stability over time. The coefficient of variation"1 is used to examine how the different sources of revenues vary relative to their mean over the 1980 to 1999 period (Table Al.9). As the coefficient of variation increases, the relative dispersion or variability of the series increases. Several results in Table Al.9. are noteworthy. All aggregate measures of government revenues show a high degree of volatility over the entire period with the coefficient of variation for federal revenues of 1.16. Federal tax revenues show more volatility, with a coefficient of variation of 1.22, than federal non-tax revenues, with a coefficient of 1.11. Correspondingly, the most important taxes (income, VAT, and excises) show more volatility over the period than revenues form hydrocarbon duties. Volatility of all tax sources was higher, often more than double, during the decade of 1980s than the decade of the 1990s, for the most part due to differences in inflation during the two periods. A1.31 Revenue instability at the federal level of government is a positive feature if revenues move with the business cycle so that they expand more than proportionally during expansions and contract more than proportionally during contractions. In this case, the tax system works as a built-in stabilizer and helps moderate the swings in the real economy caused by the business cycle. Revenue stability, or lack of volatility, is a 10. This analysis was refined by using dummy variables as proxies for major changes in the structure of the taxes and also by substituting GDP by variables that could match more closely match the actual tax base of the revenue source in question. For example, for the Value Added Tax (VAT), dummy variables are introduced for each of the major structural changes in the VAT and GDP is replaced by private consumption. Since the resulting estimated coefficient for buoyancy did not differ much from those estimated with a simple regression and GDP as the tax base, only these latter results are discussed. 11. The coefficient of variation is defined as the ratio of the standard deviation of the series to its mean value. 41 Annex I desired characteristic at the subnational level, since state and local governments have in general less ability, and this is true too in Mexico, to borrow during business activity contractions and many of the services provided at this level (such as education) require a steady flow of funds. However, as shown in Table A1.9, subnational taxes in Mexico do not appear to be less volatile than federal revenues for the period (1990-99) for which subnational data are available. A 1.32 Has the volatility of federal revenues in Mexico worked as a built-in stabilizer to moderate the business cycle? This question has been studied recently (World Bank, 2000) and the results of the analysis are summarized here. Whether tax revenues have moved with the cycle or against the cycle is illuminated by the correlation coefficient between the revenue series and the cyclical component of real GDP.12 Over the 1980-1999 period total federal revenues were not tightly correlated with real GDP. The correlation coefficient was only 0.2. This result is the composite of two elements. First, tax revenues moved positively with the cycle as expected since most tax bases are positively related to GDP. But, second, non-tax revenues- (which include most importantly hydrocarbon duties) were approximately acyclical, meaning they did not show any pattem vis-a-vis GDP. Therefore, even though most tax revenues tend to work as built-in stabilizers, their impact on the business cycle is relatively weak in Mexico. One of the reasons for this weakness is that these taxes represent only a fraction of public sector revenues. Between one-third and one-fourth of federal government revenues still come from petroleum and the fluctuations of these revenues was independent of the business cycle. It may be added to the observations in World Bank 2000, that there is also evidence that discretionary tax policy has been pro-cyclical. Often during the past 20 years, Mexico's government has decreased tax rates and taken other measures that reduce revenues during the expansionary phase of the business cycle (such as the reduction of income tax rates in 1989 and 1994) and has increased rates or taken other measures that increased revenues during the recession part of the business cycle (such- as the increase in the general VAT rate during the sharp downturn in 1995 associated with the peso crisis). This behavior fits well, of course, with the hypothesis that the overriding objective, explicit or implicit, of Mexico's tax policy has been to maintain a constant tax level.13 A1.33 Mexico appears to have a pattern found in many other Latin American countries when pursuing pro-cyclical discretionary tax policy (tax rates are increased during recessions and decreased during expansions).14 This behavior has a rationale, however. Latin American countries, like many other developing countries, have large fluctuations in tax bases (larger anyway than those in developed countries). Compensating for these 12. The latter can be interpreted as the deviation of GDP from its long-term path. 13. During the last two decades one can find counterexamples to the assertion that discretionary changes in tax policy have been pro-cyclical. But in fact, it appears that the Mexican Government over the past 20 years never set out upon tax reform with the explicit objective of increasing revenues as a percentage of GDP. The main objectives of these reforms were efficiency and occasionally equity. In some cases the objective was clearly one of increasing revenues as when the contribution from the petroleum sector decreased in the late 1980s) but this increase in revenues was only to bring actual tax levels back to the average level and not to increase it as a percentage of GDP. 14. This pattern is common to other developing countries but not among OECD countries. See Gavin and Perotti 1997. 15. See Talvi and V6gh 2000. 42 Annex I large fluctuations would require obtaining large budget surpluses during expansions and large budget deficits during contractions. Politically, it becomes too hard to run big surpluses in times of plenty given the various competing demands for expenditures rather than reducing national debt. Therefore governments may proceed to lowering tax rates rather than seeing the resources "wasted" on unnecessary public expenditures.'6 E A Suggested Reform Package Feasibility and Strategy A1.34 Mexico has had too many and too frequent tax reforms. As argued above, continuously tinkering with the tax systems makes tax administration more difficult, increases compliance costs, and creates uncertainty for investors. The next round of tax reform should be thought out carefully and the resulting tax system left in place for the longest period of time possible. A1.35 To be successful, the tax reform package will need to gain consensus among the main sections of the community including support at the highest level of government, tax administrators, the business sector, labor unions, and the public at large. This will require small group discussions, presentations to large groups and publicity campaigns. A recent study of the political factors determining the success of Mexico's reform in 1988-1994 (Berensztein, 1998) emphasized several factors including the opportune use of an economic crisis and the formation of a coalition in support of the reform while preventing the forrnation of other coalitions to oppose it. But the late 1980s reform was revenue neutral while it pursued visible gains in efficiency. Currently, there is no apparent economic crisis, and the new tax reform, while pursuing gains in efficiency and equity, would have as its main goal a significant increase in the level of tax effort. In the past, Mexico's business elites appear to have been successful in blocking any permanent increases in the overall level of tax effort.17 What may change now to make a revenue- increasing reforrn feasible? Perhaps the answer is to make the benefits of increased government services clear to decisive majority coalitions. The political strategy needs to capitalize on the growing national consensus on the need to increase public services. A1.36 To be successful, the tax reform package will also need to be comprehensive, covering all aspects of the tax system, including policy and administration. A comprehensive reform may be the only politically viable way to address the difficult aspects of the current tax system involving privileges and special treatments. Taking a comprehensive approach also facilitates other desirable objectives of tax reform besides revenue adequacy. 8 Similarly, tax administration reform should accompany tax policy 16. Allegedly, this was the budget philosophy publicly espoused by Domingo Cavallo, Minister of Finance in Argentina during 1991-96. (Talvi and V6gh, 2000). 17. See, for example, Tonell and Esquivel 1995. 18. Fiscal systems are interconnected systems. It is incorrect, for example, to consider the progressivity or regressivity of single taxes. Getting rid of a certain degree of regressivity in indirect taxation may be quite costly in terms of administration costs or other objectives, but this regressivity can be simply redressed in some other 43 Annex I reform and a lot of care should be taken to ease problems for tax administration arising from tax policy. Content of the Tax Reform Al.37 There are some general objectives that should guide the tax policy reform together with the modernization programs for the SAT and customs. The general objectives of the comprehensive reform should include: 1. Revenue adequacy: Revenue adequacy should be of paramount importance in this reform because the share of the public sector in GDP is currently insufficient to provide the increased level of public services desired by government. How much to increase revenue effort is fundamentally a political question. 2. Elastic revenues over time: Taxes should be designed so that the government's share in GDP at least keeps pace with the growth of income in the entire economy. Elastic revenues are needed to keep government away from introducing ad hoc new tax measures to simply keep up with ordinary increases in the demand for public services. 3. Simplicity: The tax structure should be simplified -reducing complex inflation indexation and eliminating exemptions and special treatment. 4. Fair distribution of tax burdens: The new tax structure should attempt to increase the progressivity of the tax system, bearing in mind that the most effective progressivity must take place from the expenditure side of the budget. The tax system should also increase its horizontal equity by eliminating special treatments and privileges and by significantly improving tax compliance. 5. Reduction in distortions: The tax reformn should increase the efficiency of the tax system by minimizing current distortions and providing a more level playing field. This will mean getting rid of special regimes under the CIT and eliminating special treatment for certain sources of income under the PIT and certain forms of consumption under the VAT. 6. The modernization of tax administration: The rationalization of taxes and tax administration should have the goal of increasing the public's confidence in the tax system and its perception of fairness. Taxpayers are more likely to oppose a system that seems unreasonable or capricious and that they believe is unfair. Unfairness includes arbitrary enforcement by reaching only taxpayers subject to withholding. areas of the tax system. What, should matter is the performance of the entire tax system, not what happens with individual taxes. 44 Annex I Table A1.1. General Government Revenues as a Percentage of GDP. 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 Total Revenues 17A1 17.43 17.88 19.78 18.52 18.71 17.76 18.98 18.32 19.20 18.98 21.89 22.29 19.35 19.50 18.91 18.91 19.08 17.04 16.48 Total Federal Revenues 15.30 15.27 15.63 17.79 16.88 16.86 16.00 17.06 16.34 16.43 15.93 18.69 18.70 15.51 15.52 15.25 15.68 15.84 14.17 14.52 TaxRevenues 10.91 10.64 9.98 10.41 10.46 10.37 11.53 10.92 11.53 11.30 10.90 10.71 11.26 11.40 11.29 9.27 9.03 9.82 10.51 11.21 lncome Taxes 5.52 5.53 4.74 4.06 4.12 4.06 4.25 3.96 4.68 4.72 4.46 4.53 5.15 5.51 5.13 4.01 3.88 4.25 4.41 4.62 ValueAddedTax 2.68 2.59 2.21 3.03 3.20 3.12 3.15 3.24 3.37 3.10 3.60 3.43 2.71 2.64 2.71 2.82 2.88 3.07 3.12 3.27 Excises 1.07 1.02 1.81 2.39 2.24 2.16 2.76 2.51 2.62 2.30 1.52 1.34 1.62 1.54 1.97 1.35 1.19 1.43 1.99 2.29 Import Taxes 0.99 1.09 0.84 0.52 0.52 0.69 0.87 0.81 0.44 0.73 0.86 1.04 1.14 1.01 0.89 0.61 0.59 0.57 0.56 0.60 Others' 0.65 0.40 037 0.40 0.38 0.34 0.49 0.39 0.43 0.45 0.45 . 0.37 0.65 0.70 0.58 0.49 0.49 0.50 0.44 0.44 Non-TaxRevenues2 4.38 4.63 5.66 7.38 6.42 6.49 4.47 6.14 4.81 5.13 5.03 7.98 7.44 4.11 4.23 5.98 6.65 6.02 3.66 3.31 Products, Services 0.18 0.34 0.24 0.28 0.15 0.06 0.05 0.07 0.08 0.08 0.26 0.26 0.34 0.14 0.24 0.48 0.29 0.27 0.31 0.17 Others3 0.16 0.17 0.46 0.24 0.10 0.20 0.24 0.35 0.93 1.38 0.75 4.01 3.65 0.81 1.49 1.28 1.57 1.39 0.61 0.68 Duties4 4.04 4.13 4.96 6.86 6.17 6.23 4.18 5.72 3.80 3.66 4.02 3.70 3.45 3.17 2.49 4.22 4.79 4.37 2.75 2.46 Of Which: HydrocarbonDuties 3.66 3.82 4.67 6.49 5.75 5.73 3.69 5.21 3.21 3.23 3.51 3.27 3.04 2.78 2.19 3.93 4.51 4.08 2.32 2.11 Other Duties 0.38 0.31 0.29 0.37 0.42 0.50 0.49 0.51 0.58 0.43 0.51 0.44 0.40 0.39 0.31 0.28 0.29 0.29 0.42 0.36 Socal Secrity 2.94 2.92 3.20 2.67 2.41 2.58 2.50 2.29 2.36 2A2 2.57 2.74 2.90 3.24 3.44 3.10 2.94 3.06 3.28 3.41 Contributions 2.45 2.41 2.47 1.99 1.96 2.17 2.14 2.11 2.14 2.20 2.32 2.39 2.72 2.98 3.15 2.68 2.54 2.70 2.88 2.97 Non-Tax Revenues 0.49 0.51 0.73 0.68 0.45 0.41 0.36 0.18 0.21 0.22 0.24 0.35 0.18 0.26 0.29 0.42 0.40 0.36 0.39 0.44 Subnatlonal .. .- .. .. .. .. .. .. .. 0.71 0.88 0.88 1.16 1.17 1.19 1.11 0.89 1.07 0.96 Taxes .. .. .. .. .. .. .. .. .. 0.20 0.27 0.33 0.40 0.42 0.40 0.36 0.35 0.41 0.37 Non-Tax Revenues7 .. .. .. .. .. .. .. .. .. 0.52 0.61 0.55 0.76 0.76 0.78 0.76 0.55 0.66 0.59 GDP (Billion Pesos) 4.5 6.1 9.8 17.9 29.5 47.4 79.2 193.3 416.3 548.9 738.9 949.1 1125 1256 1420 1837 2504 3179 3847 4623 Source: Secretarfa de Hacienda y Cr6dito Psiblico, Banco de M6xico, and World Bank Staff Calculations. Tota federal revenues include compensated operations, which are revenues for social security institutions and expenditures for the Federal Government. Compensated operations are excluded from the calculations presented in this table. ' Includes export taxes, some payroll taxes, accessories revenues, etc. After 1992, data for accessories revenues are reported in other tax revenues. Prior to 1992, the data series was reported separately. For consistency, prior to 1992, the data for accessories is added to other revenues. 2 In 1988 includes revenues from zero-coupon bonds. In 1991,1992 and 1994 includes revenues from the sale of public enterprises 3Encludes revenues from palaties, surcharges, privatizaion proceeds. etc. 4Prior to 1991, revenues from PEMEX were reported as a separate revenue line item. Prior to 1991, the data in this table are the reportd data for revenues from PEMEX. 5 For 1977 to 1991, the data series did not contain sufficient detail to identify the social security contributions of the Federal Government and these contributions are included as others of social security. From 1992, the contributions of the Federal Government are identified and are classified as social security contributions. 6 Subnational revenue data is only available from 1989-1998. 7 Non-tax revenues include fees, public services (including Productos and Aprovechamientos), and other revenues. 45 Annex 1 Table A1.2. General Government Revenues as Percentage of GDP. General Government Total Revenues General Government Tax Revenues 1980 1985 1990 19971 1980 1985 1990 19972 Mexico' 18.48 19.29 19.34 19.05 16.69 17.81 16.92 16.50 Argentina 18.77 18.72 16.68 20.88 16.26 16.26 15.15 18.99 Brazil 29.90 33.80 41.88 42.33 24.24 22.01 27.53 29.29 Chile 4 33.60 30.09 .. 24.63 26.27 22.23 .. 20.44 CostaRica 18.60 21.48 23.80 27.72 17.20 19.18 20.15 24.05 Dominican Rep. 14.64 11.17 12.23 15.35 11.16 10.17 10.89 13.91 Panama 25.96 26.20 26.34 26.67 19.02 19.04 18.27 18.47 Paraguay 11.25 10.16 12.44 14.40 10.06 8.42 9.24 9.27 Peru 5 17.07 14.99 10.55 16.91 15.77 13.14 9.50 14.29 Canada 38.99 40.22 45.37 45.02 32.44 33.53 36.83 36.96 United States 31.81 32.53 33.51 36.71 26.84 26.30 27.18 29.22 Source: 1999 Government Finance Statistics, International Monetary Fund, 2000 World Development Indicators, The World Barnk, and World Bank Staff Calculations. ' The reported data for Brazil and Panama are for 1994. For Costa Rica and the Dominican Republic, the reported data are for 1996. For Paraguay, the reported data are for 1993. 2 The reported data for Brazil and Panama are for 1994. For Costa Rica and the Dominican Republic, the reported data are for 1996. For Paraguay, the reported data are for 1993. 3 General government revenues include federal govemment compensated operations, which are revenues for social security institutions and expendimres for the Federal Government. The reported data include compensated operations. Subnational government data are not reported for 1989-1991. The reported data for 1980 and 1985 are for consolidated central government operations only. Data on subnational govemment operations are not reported prior to 1990. 46 Annex 1 Table A1.3. Composition of General Government Revenue.' (As percent of General Government Revenues) 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 General Governrene 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 t00 100 100 100 100 Total Federal Revenues 87.86 87.60 87.43 89.97 91.15 90.13 90.08 89.88 89.17 85.60 83.94 85.39 83.91 80.16 79.56 80.64 82.93 83.02 83.19 88.11 Tax Revenues 62.68 61.03 55.80 52.65 56.49 55.41 64.90 57.53 62.94 58.87 57.44 48.93 50.54 58.91 57.88 49.02 47.74 51.46 61.68 68.03 Income Taxes 31.70 31.75 26.52 20.55 22.23 21.69 23.92 20.87 25.52 24.59 23.51 20.70 23.10 28.48 26.32 21.22 20.53 22.27 25.86 28.00 Value Added Tax 15.42 14.86 12.39 15.34 17.27 16.67 17.76 17.08 18.39 16.15 18.99 15.66 12.14 13.63 13.91 14.91 15.23 16.11 18.29 19.85 Excises 6.14 5.85 10.11 12.10 12.12 11.54 15.55 13.25 14.27 11.99 7.99 6.12 7.25 7.95 10.09 7.11 6.27 7.48 .11.69 13.89 Import Taxes 5.69 6.26 4.71 2.63 2.82 3.69 4.89 4.26 2.38 3.80 4.55 4.77 5.14 5.22 4.59 3.21 3.14 2.98 3.28 3.63 Others3 3.73 2.31 2.07 2.03 2.05 1.82 2.77 2.07 2.37 2.34 2.39 1.68 2.91 3.62 2.97 2.58 2.57 2.61 2.56 2.65 Non-Tax Revenues4' 25.17 26.57 31.63 37.32 34.65 34.72 25.17 32.35 26.23 26.73 26.50 36.46 33.37 21.26 21.69 31.62 35.19 31.56 21.51 20.08 Products, Services 1.05 1.94 1.33 1.44 0.80 0.33 0.27 0.38 0.42 0.44 1.37 1.20 1.52 0.73 1.24 2.53 1.51 1.39 1.83 1.02 Others5 0.91 0.95 2.56 1.21 0.52 1.09 1.36 1.82 5.09 7.21 3.95 18.34 16.38 4.16 7.66 6.77 8.32 7.27 3.56 4.12 Duties6 23.21 23.68 27.74 34.67 33.33 33.30 23.55 30.14 20.72 19.08 21.18 16.92 15.47 16.37 12.79 22.32 25.36 22.90 16.11 14.94 Of Which: Hydrocarbon Duties 21.03 21.90 26.13 32.80 31.05 30.64 20.78 27.45 17.53 16.83 18.51 14.93 13.65 14.35 11.20 20.81 23.84 21.36 13.64 12.78 Other Duties 2.18 1.77 1.61 1.87 2.28 2.66 2.76 2.70 3.18 2.25 2.67 1.99 1.82 2.02 1.59 1.51 1.52 1.54 2.48 2.16 16.87 16.77 17.90 13.48 13.01 13.79 14.05 12.08 12.86 12.59 13.52 12.54 13.02 16.72 17.65 16.40 15.58 16.06 19.24 20.70 Contributions 14.08 13.85 13.81 10.06 10.59 11.60 12.02 11.11 11.69 11.46 12.24 10.93 12.20 15.39 16.13 14.17 13.45 14.17 16.92 18.04 Non-TaxRevenues 2.79 2.92 4.09 3.42 2.42 2.18 2.03 0.97 1.17 1.13 1.28 1.61 0.82 1.34 1.51 2.23 2.13 1.88 2.32 2.66 Subnational .. .. .. .. .. .. .. .. .. 3.71 4.64 4.02 5.20 6.07 6.08 5.89 4.71 5.62 5.63 Taxes .. .. .. .. .. .. .. .. .. 1.02 1.40 1.50 1.80 2.16 2.07 1.90 1.83 2.16 2.14 Non-Tax Revenues' .. .. .. .. .. .. .. .. .. 2.69 3.24 2.52 3.40 3.91 4.00 4.00 2.88 3.46 3.49 Source: Secretarfa de Hacienda y Ciddito Pdblico, Banco de Mexico, and World Bank Staff Calculations. Total federal revenues include compensated operations, which are revenues for social security institutions and expenditures for the Federal Govermment ' The reported data may not equal 100% due to the exclusion of compensated operations from the calculations. 2 Data for 1980-1988 and 1999 do not reflect subnational government revenues. 3Includes export taxes, some payroll taxes, accessores revenues, etc. After 1992, data for accessories revenues are reported in other tax revenues. Prior to 1992, the data series was reported separately. For consistency, prior to 1992, the data for accessories is added to other revenues. 4 In 1988 includes revenues from zero-coupon bonds. In 1991,1992 and 1994 includes revenues from the sale of public enterprises 5Includes revenues from penalties, surcharges, privatization proceeds, etc. 6 Prior to 1991, data on hydrocarbon revenues were not reported and revenues from PEMEX were reported as a separate revenue line item. Prior to 1991, the data in this table are the reported data for revenues from PEMEX. 7 For 1977 to 1991, the data series did not contain sufficient detail to identify the social security contributions of the Federal Government and these contributions are included as others of social security. From 1992, the contributions of the Federal Government are identified and are classified as social security contributions. Non-tax revenues include fees, public services (including Productos and Aprovechamientos), and other revenues. 47 Annex I Table A1.4. Composition of Fedetal Government Revenues. (As percent of Federal Government Revenues) 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 Federal Goverrenent 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 Tax Revenues 71.35 69.67 63.82 58.52 61.98 61.48 72.05 64.01 70.58 68.77 68.43 57.30 60.24 73.48 72.75 60.79 57.57 61.98 74.15 77.21 Income Taxes 36.08 36.24 30.33 22.84 24.39 24.06 26.56 23.22 28.62 28.72 28.01 24.24 27.53 35.53 33.08 26.31 24.75 26.83 31.09 31.78 Value Added Tax 17.55 16.97 14.17 .17.05 18.95 18.50 19.71 19.00 20.62 18.87 22.63 18.34 14.47 17.00 17.49 18.49 18.37 19.41 21.99 22.53 Excises 6.99 6.68 11.56 13.45 13.30 12.80 17.27 14.74 16.01 14.00 9.52 7.17 8.64 9.92 12.68 8.82 7.56 9.01 14.05 15.76 Import Taxes 6.48 7.14 5.39 2.92 3.09 4.09 5.43 4.75 2.67 4.44 5.42 5.59 6.12 6.51 5.77 3.98 3.78 3.59 3.94 4.12 Others2 4.25 2.63 2.37 2.25 2.25 2.02 3.08 2.30 2.66 2.74 2.85 1.97 3.47 4.52 3.73 3.20 3.10 3.14 3.08 3.01 Non-Tax Revenues3 28.65 30.33 36.18 41.48 38.02 38.52 27.95 35.99 29.42 31.23 31.57 42.70 39.76 26.52 27.25 39.21 42.43 38.02 25.85 22.79 Products, Services 1.20 2.22 1.52 1.60 0.88 0.36 0.30 0.43 0.48 0.51 1.64 1.40 1.81 0.91 1.55 3.14 1.82 1.68 2.20 1.16 Others4 1.04 1.08 2.92 1.34 0.57 1.21 1.5t 2.02 5.71 8.42 4.70 21.48 19.52 5.19 9.63 8.39 10.03 8.76 4.28 4.67 Duties5 26.42 27.03 31.73 38.54 36.57 36.95 26.14 33.54 23.23 22.29 25.23 19.82 18.43 20.42 16.08 27.67 30.58 27.59 19.37 16.95 Of Which: Hydrocarbon Duties 23.94 25.00 29.89 36.46 34.07 33.99 23.07 30.54 19.66 19.66 22.05 17.48 16.27 17.89 14.08 25.80 28.74 25.73 16.39 14.50 Other Duties 2.48 2.03 1.85 2.08 2.50 2.95 3.07 3.00 3.57 2.63 3.18 2.34 2.16 2.53 1.99 1.87 1.83 1.86 2.98 2.45 71.35 69.67 63.82 58.52 61.98 61.48 72.05 64.01 70.58 68.77 68.43 57.30 60.24 73.48 72.75 60.79 57.57 61.98 74.15 77.21 Total Federal Revenues (as of GDP) 15.30 15.27 15.63 17.79 16.88 16.86 16.00 17.06 16.34 16.43 15.93 18.69 18.70 15.51 15.52 15.25 15.68 15.84 14.17 14.52 Source: Secretarla de Hacienda y Cr6dito P6blico, Banco de Mexico, and World Bank Staff Calculations. Total federal revenues include compensated operations, which are revenues for social security institutions and expenditures for the Federal Government. ' The reported data may not equal 100% due to the exclusion of compensated operations from the calculations. 2 Includes export taxes, some payroll taxes, accessories revenues, etc. After 1992, data for accessories revenues are reported in other tax revenues. Prior to 1992, the data series was reported separately. For consistency, prior to 1992, the data for accessories is added to other revenues. 3 In 1988 includes revenues from zero-coupon bonds. In 1991,1992 and 1994 includes revenues from the sale of public enterprises 4 Includes revenues from penalties, surcharges, privatization proceeds, etc. 5 Prior to 1991, data on hydrocarbon revenues were not reported and revenues from PEMEX were reported as a separate revenue line item. Prior to 1991, the data in this table are the reported data for revenues from PEMEX. 48 Annex 1 Table A1.5. Government Revenues as Percentage of GDP. 1 General Central Taxes on Social Domestic Taxes on Non-tax Government Government income, security taxes taxes on international revenue Total Total profit, and goods and trade Revenue Revenue capital gains services 1980 1997 1980 1997 1980 1997 1980 1997 1980 1997 1980 1997 1980 1997 Mexico2 18.48 19.05 15.06 17.93 5.15 4.25 2.18 3.06 7.58 6.96 1.05 0.57 1.12 2.51 Argentina3 18.77 20.88 15.62 12.26 .. 1.67 2.60 3.35 2.60 4.99 1.34 0.93 5.21 1.06 Brazil 29.90 42.33 22.63 31.01 2.42 3.65 5.66 8.22 7.27 6.57 1.62 0.52 4.85 6.85 Chile 33.60 24.63 32.77 22.79 5.64 4.03 5.55 1.39 11.47 10.45 1.39 1.87 6.38 3.77 Costa Rica 18.60 27.72 17.80 26.73 2.44 2.85 5.14 7.26 5.41 10.68 3.36 2.25 1.04 3.24 Dominican Rep. 14.64 15.35 14.39 15.23 2.75 2.49 0.55 0.62 3.07 5.09 4.45 5.55 3.18 1.18 Panama 25.96 26.67 25.33 18.61 5.38 5.21 5.37 4.93 4.24 0.67 2.62 0.10 6.75 6.91 Paraguay 11.25 14.40 10.74 14.08 1.63 1.46 1.41 .. 1.90 5.03 2.66 1.75 0.94 5.00 Peru 17.07 16.91 17.07 15.74 4.42 3.29 6.29 1.32 6.35 7.80 4.62 1.35 1.31 1.71 Canada 38.99 45.02 18.48 20.28 9.71 10.39 1.92 3.69 3.07 3.53 1.30 0.38 2.52 2.30 United States 31.81 36.71 20.16 21.43 11.40 11.76 5.69 6.84 0.89 0.72 0.28 0.24 1.66 1.48 Sources: Govermnent Fuiance Statistics Yearbook, International Monetary Fund (1999) and Global Development Network database (2000), The World Bank. ' Actual data for Brazil is 1994, Canada 1995, Costa Rica and the Dominican Republic 1996, and Paraguay 1993. 2 General government revenues include Federal Government compensated operations, which are revenues for social security institutions and expenditures for the Federal Government. 3 Data for taxes on international trade is for 1981. 49 Annex 1 Table A1.6. Estimated Tax Level Rankings and Indices (1990-1996). Country Ranking Index I Country Ranking Index 2 Country Ranking Index 3 Mexico 21 82.725 Mexico 27 62.84 Mexico 27 70.09 Hungary 1 216.894 Lesotho 1 235.28 Hungary 1 183.36 Lesotho 2 185.648 Hungary 2 184.41 Romania 2 180.01 Uruguay 3 175.467 Romania 3 169.70 Uruguay 3 174.38 Romania 4 167.857 Bulgaria 4 155.73 Lesotho 4 155.37 South Africa 5 154.631 Estonia 5 141.05 Bulgaria 5 141.85 Bulgaria 6 144.605 Uruguay 6 140.43 Chile 6 135.58 Estonia 7 136.331 Tunisia 7 127.98 Costa Rica 7 132.59 Tunisia 8 125.879 Costa Rica 8 121.43 South Africa 8 125.35 CostaRica 9 114.532 South Africa 9 120.03 Tunisia 9 115.26 Chile 10 107.784 SriLanka 10 115.35 Estonia 10 113.35 SouthKorea 11 102.801 Zambia 11 106.69 Sri Lanka 11 106.52 Sri Lanka 12 94.276 Malaysia 12 101.94 Cameroon 12 103.69 Zambia 13 93.069 Mauritius 13 93.56 South Korea 13 102.30 Indonesia 14 90.951 Chile 14 92.69 Mauritius 14 90.20 Mauritius 15 89.469 Indonesia 15 91.22 Burkina Faso 15 88.06 Venezuela 16 88.949 Panama 16 87.26 Turkey 16 86.96 Thailand 17 88.924 Pakistan 17 85.40 Malaysia 17 85.95 Malaysia 18 84.768 Thailand 18 85.24 Pakistan 18 85.190 Turkey 19 83.764 South Korea 19 75.68 Indonesia 19 84.96 Argentina 20 82.769 Turkey 20 74.99 Zambia 20 84.25 Panama 22 81.096 BurkinaFaso 21 74.13 Venezuela 21 83.52 Pakistan 23 77.936 Republic of Yemen 22 73.79 Bhutan 22 83.21 Peru 24 77.258 Cameroon 23 73.71 India 23 82.41 Dominican Republic 25 69.709 Dominican Republic 24 73.06 Thailand 24 80.92 Bolivia 26 63.063 Venezuela 25 72.55 Argentina 25 77.19 India 27 60.011 India 26 69.27 Argentina 26 76.45 Cameroon 28 57.98 Bolivia 28 59.74 Dominican Repub 28 64.34 Republic of Yemen 29 56.88 Peru 29 58.51 Republic of Yemen 29 62.97 Burkina Faso 30 54.325 Bhutan 30 56.35 Bolivia 30 54.10 Bahrain 31 32.862 Argentina 31 55.02 Peru 31 53.52 Bhutan 32 30.255 Bahrain 32 34.02 Bahrain 32 42.51 Source: Government Finance Statistics Yearbook (1999);1ntemational Monetary Fund, World Development Indicators CD-ROM (1999), World Bank; and World Bank StaffCalculations. l Index I is equal to the ratio of actual collections divided by predicted collections, where predicted collections are based upon the regression of the ratio of tax revenues to GDP on per capita GDP and the sum of imports and exports divided by GDP. Index 2 is calculated in the same method except that predicted collections are based upon the regression of the ratio of tax revenues to GDP on the share of fuel and mineral exports in GDP and the share of agriculture in GDP. Index 3 is calculated using the regression of tax collections to GDP on per capita GDP, the sum of exports and imports divided by GDP, and the share of agriculture in GDP. 50 Table A1.7. Year-to-Year Buoyancy of Revenue Sources. 1 2 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 Mean Total Revenues 1.00 1.05 1.16 0.87 1.02 0.90 1.06 0.96 1.17 0.96 1.56 1.11 -0.29 1.07 0.88 1.00 1.04 0.41 0.82 0.93 Total Federl Revenues 0.99 1.05 1.20 0.90 1.00 0.90 1.06 0.95 1.02 0.90 1.62 1.00 -0.70 1.01 0.93 1.09 1.04 0.42 1.13 1.00 Tax Revenues 0.92 0.87 1.07 1.01 0.98 1.20 0.95 1.06 0.93 0.88 0.93 1.30 1.11 0.92 0.24 0.91 1.35 1.35 1.35 1.02 Income Taxes 1.01 0.68 0.75 1.02 0.97 1.09 0.93 1.19 1.03 0.81 1.06 1.74 1.61 0.42 0.04 0.89 1.37 1.19 1.25 1.00 Value Added Tax 0.89 0.67 1.47 1.10 0.95 1.02 1.03 1.05 0.70 1.49 0.80 -0.39 0.77 1.23 1.15 1.07 1.27 1.07 1.26 0.98 Excises 0.85 2.07 1.42 0.88 0.92 1.44 0.91 1.05 0.54 -0.41 0.51 2.09 0.55 2.98 -0.48 0.60 1.76 2.70 1.75 1.16 Import Taxes 1.30 0.46 0.20 1.01 1.55 1.42 0.93 0.21 2.73 1.55 1.74 1.53 -0.14 0.02 -0.51 0.93 0.83 0.90 1.37 0.95 Others3 -0.53 0.83 1.12 0.89 0.78 1.65 0.77 1.12 1.12 1.03' 0.16 4.15 1.71 -0.56 0.33 0.99 1.09 0.31 1.01 0.95 Non-Tax Revenues4 1.17 1.40 1.40 0.73 1.02 0.28 1.29 0.70 1.23 0.93 2.74 0.59 -4.31 1.23 2.29 1.34 0.58 -1.60 0.45 0.71 Products, Services 2.77 0.25 1.28 -0.32 -0.87 0.51 1.39 1.07 1.29 4.15 1.02 2.47 -6.67 5.22 3.44 -0.68 0.70 1.84 -2.33 0.89 Others5 1.12 2.74 -0.08 -0.82 2.34 1.31 1.32 1.93 2.36 -1.06 5.99 0.44 -11.0 5.78 0.40 1.64 0.47 -3.24 1.60 0.43 Duties6 1.07 1.37 1.48 0.80 1.02 0.23 1.29 0.48 0.87 1.31 0.67 0.58 0.23 -0.95 2.91 1.40 0.61 -1.43 0.41 0.50 Of Which: Hydrocarbon Duties 1.13 1.40 1.49 0.77 0.99 0.15 1.31 0.39 1.02 1.28 0.71 0.58 0.16 -0.95 3.12 1.43 0.58 -1.93 0.47 0.74 Other Duties 0.36 0.86 1.37 1.25 1.33 0.97 1.04 1.15 -0.09 1.52 0.40 0.56 0.70 -0.92 0.68 1.03 1.10 2.83 0.07 0.85 Social Security 0.98 1.18 0.71 0.80 1.14 0.94 0.91 1.03 1.09 1.20 1.26 1.33 1.98 1.50 0.60 0.83 1.16 1.35 1.22 1.12 Contributions 0.95 1.05 0.65 0.97 1.20 0.97 0.99 1.02 1.10 1.18 1.12 1.74 1.82 1.45 0.38 0.83 1.25 1.33 1.17 1.11 Non-Tax Revenues 1.15 1.71 0.87 0.18 0.81 0.76 0.26 1.18 1.03 1.38 2.42 -2.81 4.09 .2.04 2.33 0.84 0.54 1.48 1.57 1.15 Subnational' .. .. .. .. .. .. .. .. .. 1.69 1.00 2.58 1.11 1.07 0.76 0.28 1.76 0.42 .. 1.19 Taxes .. .. .. .. .. .. .. .. .. 1.99 1.82 2.15 1.38 0.73 0.53 0.88 1.72 0.37 .. 1.29 Non-Tax Revenues .. .. - .. .. .. .. .. .. 1.56 0.57 2.82 0.97 1.26 0.88 -0.05 1.78 0.45 .. 1.14 Source: Secretarfa de Hacienda y Credito Publico, Banco de Mdxico, and World Bank Staff Calculations. Not adjusted for discretionary changes. ' The Year to Year Buoyancy of variable X at time t is given by ((X(t) - X(t-1))/((X(t) + X(t-1))/2) / ((GDP(t) - GDP(t-1))/((GDP(t) + GDP(t-1))/2). 2 Total federal revenues include compensated operations, which are revenues for social security institutions and expenditures for the Federal Govermment. Compensated operations are excluded from the calculations presented in this table. 3Includes export taxes, some payroll taxes, accessories revenues, etc. After 1992, data for accessories revenues are reported in other tax revenues. Prior to 1992, the data series was reported separately. For consistency, prior to 1992, the data for accessories is added to other revenues. 4 In 1988 includes revenues from zero-coupon bonds. In 1991,1992 and 1994 includes revenues from the sale of public enterprises 5Includes revenues from penalties, surcharges, privatization proceeds, etc. 6 Prior to 1991, data on hydrocarbon revenues were not reported and revenues from PEMEX were reported as a separate revenue line item. Prior to 1991, the data in this table are the reported data for revenues from PEMEX. 7 For 1977 to 1991, the data series did not contain sufficient-detail to identify the social security contributions of the Federal Government and these contributions are included as others of social security. From 1992, the contributions of the Federai Government are identified and are classified as social security contributions. 8 Subnational revenue data is only available from 1989-1998. 51 Table A1.8. Average Estimated Buoyancy of Revenue Sources (1980 -1999).' Total Revenues2 0.99 Total Federal Revenues 0.99 Tax Revenues 1.00 Income Taxes 1.01 Value Added Tax 1.02 Excises 0.94 Import Taxes 0.98 Others3 1.05 Non-Tax Revenues 0.96 Products, Services 1.08 Others 1.34 Duties4 0.89 Of Which: Hydrocarbon Duties 0.88 Other Duties 1.01 Social Security 1.04 Contributions5 1.06 Non-Tax Revenues 0.91 Source: Secretarfa de Hacienda y Credito Publico, Banco de Mexico, and World Bank Staff Calculations. Total Federal Revenues include compensated operations, which are revenues to social security institutions and expenditures for the Federal Government. ' The average estimated buoyancy is equal to the coefficient for the natural logarithm of GDP when the natural logarithm of the revenue source is regressed on a constant, the natural logarithm of GDP, and an autoregressive term of order 1. 2 Total revenues include compensated operations that are those transactions that are considered revenues for social security institutions and are expenditures of the Federal Government. 3 After 1992, revenue data for accessories are reported in other tax revenues. Prior to 1992, the data were reported separately. For consistency, prior to 1992, the data series for accessories is added to other tax revenues and is not reported as a separate item. 4 Prior to 1991, data on hydrocarbon revenues were not reported and revenues from PEMEX were reported as a separate revenue line item. Prior to 1991, the data in this table are the reported data for revenues from PEMEX. 5 For 1977 to 1991, the data series did not contain sufficient detail to identify the social security contributions of the Federal Government and these contributions were included in the non-tax revenue series for social security. From 1992, the contributions of the Federal Government are identified and as reported as social security contributions. 52 Annex 1 Table A1.9. Coefficient of Variation of Major Revenue Sources (Millions of Current Pesos).' 1980-1999 1980-1989 1990-1999 Standard Mean Coefficient of Standard Mean Coefficient of Standard Mean Coefficient of Deviation Variation Deviation Variation Deviation Variation Total Revenues 242090.10 210861.44 1.15 36738.77 25386.80 1.45 214315.50 396336.08 0.54 Total Federal Revenues 205056.19 176833.50 1.16 31855.61 22316.06 1.43 186269.53 331350.93 0.56 Tax Revenues 146618.83 119785.32 1.22 22117.53 15225.33 1.45 143521.58 224345.31 0.64 Income Taxes 61667.29 51308.32 1.20 9131.10 6130.99 1.49 58389.52 96485.65 0.61 Value Added Tax 43642.99 34867.22 1.25 6220.06 4327.08 1.44 43700.10 65407.36 0.67 Excises 27635.85 20224.81 1.37 4708.63 3296.87 1.43 30876.86 37152.74 0.83 hnport Taxes 8344.24 7834.59 1.07 1274.02 884.54 1.44 6166.46 14784.65 0.42 OtherS2 6430.27 5550.38 1.16 862.48 585.85 1.47 5637.11 10514.92 0.54 Non-Tax Revenues3 63407.50 57048.18 1.11 9772.64 7090.74 1.38 53349.21 107005.62 0.50 Products, Services 3782.42 2936.75 1.29 155.19 114.12 1.36 3531.85 5759.38 0.61 Others4 16476.62 14515.79 1.14 2528.04 1257.20 2.01 13269.76 27774.38 0.48 Duties' 45722.75 39595.64 1.15 7256.40 5719.42 1.27 42550.57 73471.86 0.58 Ojf Which: Hydrocarbon Duties 41468.58 35451.53 1.17 6323.31 5052.58 1.25 39202.36 65850.49 0.60 Other Duties 4947.94 4144.11 1.19 959.21 666.84 1.44 4888.21 7621.37 0.64 So-clial Security 45920.01 35579.90 1.29 4623.01 3251.32 1.42 45909.59 67908.48 0.68 ContributionS6 40170.59 31446.84 1.28 4229.80 2922.13 1.45 39757.10 59971.55 0.66 Non-Tax Revenues 5812.86 4133.06 1.41 395.58 329.19 1.20 6246.77 7936.93 0.79 Subnafo 11057.80 17720.48 0.62 .. .. .. 10539.77 19254.46 0.55 Taxes 4375.77 6405.69 0.68 .. .. .. 4194.51 6998.00 0.60 Non-Tax Revenues 6710.18 11314.79 0.59 .. .. .. 6378.00 12256.46 0.52 Source: Secretarfa de Hacienda y Credito Piblico, Banco de Mexico, and World Bank Staff Calculations. Total federal revenues include compensated operations, which are revenues for social security institutions and expenditures for the Federal Government Compensated operations are excluded from the calculations presented in this table. ' The coefficient of variation is determined by dividing the standard deviation of the series by its mean. 2 Includes export taxes, some payroll taxes, accessories revenues, etc. After 1992, data for accessories revenues are reported in other tax revenues. Prior to 1992, the data series was reported separately. For consistency, prior to 1992, the data for accessories is added to other revenues. 3 In 1988 includes revenues from zer-coupon bonds. In 1991,1992 and 1994 includes revenues from the sale of public enterprises 4 Includes revenues from penalties, surcharges, privatization proceeds, etc. ' Prior to 1991, revenues from PEMEX were reported as a separate revenue line item. Prior to 1991, the data in this table are the reported data for revenues from PEMEX. 6 For 1977 to 1991, the data series did not contain sufficient detail to identify the social security contributions of the Federal Government and these contributions are included as others of social security. From 1992, the contributions of the Federal Government are identified and are classified as soial security contributions. 7 Subnational revenue data is only available from 1989-1998. 53 Annex I Annex 2 ANNEX 2 Mexico Oil Taxation 1.- Introduction A2.1 Mexico's oil sector is a major contributor to tax revenues. At the same time, the performance of this key sector is heavily influenced by the tax system. 2.- Taxation Objectives A2.2 In designing or critiquing any tax system, it is essential to clearly specify the objectives which the system is meant to satisfy and/or against which it is to be judged. Key objectives in the Mexican context include: * Efficient development and extraction of the resource; * Efficient capture of oil rents for the state; * Creation of an accountable, commercially viable PEMEX; and * Consumption of oil products in an efficient and environmentally sustainable way. A2.3 The annex provides a brief description of Mexico's existing tax system, evaluates its performance against the objectives listed above, and makes a number of recommendations for its reform. 3.- Mexico's Existing System Of Oil Taxation A.- Background A2.4 Prior to 1994, oil taxation, which essentially equated to the taxation of Pemex, plus product taxation, was negotiated between Pemex and the Hacienda on the basis of the Government's budget requirements and Pemex's perceived capacity to pay. In 1994 a new system was introduced. The new system was meant to introduce explicit taxation of Pemex based primarily on profits; while maintaining Government revenues from the sector at levels equivalent to those negotiated in earlier years. A2.5 The new system contained transition provisions, especially the DSH (see below), designed to allow Pemex time to prepare its accounts for income taxation and to safeguard the 55 Annex 2 revenue maintenance requirement. These provisions were to last only one year. Six years later they are still in effect. The explanation given is that Pemex is still not in a position to provide the data required for income taxation. B.- Design of the Existing System A2.6 The existing system of oil taxation contains the following elements: * Income Tax. The Impuesto a los Rendimientos Petroleros (IRP) is an income tax, based on the same elements as Mexico's generally applicable corporate income tax. It is levied at a 35% rate on Pemex and each of its subsidiaries. In practice it has no impact because, as noted above, Pemex has yet to come up with the accounting numbers that would enable it to comply. * Duty on Oil Extraction. The Derecho a la Extraccion de Petroleo (DEP) is a cash flow tax, levied on total revenues of Pemex Exploration and Production (PEP), Pemex's upstream subsidiary, less total approved expenditures in the year. It consists of three components: Ordinario (52.3%); Extraordinario (25.5%); and Adicional (1.1%). The three components have different beneficiaries. These are, respectively: state governments (based on agreed formulas); the Federal Government; and cities from which oil is exported. * Petroleum Products Tax. The Impuesto Especial sobre Produccion y Servicios Aplicado a la Enajenacion de Gasolinas y Diesel (IEPS) is a gasoline and diesel tax levied on consumers. The tax is calculated as the difference between the consumer price, fixed on an occasional basis by Government at essentially the same level for all regions of the country, and the producer price equal to Houston spot prices plus transport costs and a quality adjustment. Since Houston spot prices change on an almost daily basis, while Mexico's consumer price remains fixed for much longer periods, the IEPS is variable. * Gross Revenue Tax. The Derecho sobre Hidrocarburos (DSH) is levied on Pemex at a 60.8% rate on the gross revenues from all of Pemex's operations, upstream and downstream, plus the IEPS. The DSH was designed to have a short life, maintaining Government revenues during the transition to income based taxation. It remains in effect, however, and is the most characteristic feature of Mexico's oil taxation system. Both the DEP and the IEPS are creditable against the DSH, and whenever their sum is different from the calculated DSH obligation, the DEP is adjusted upwards or downwards to make them equal. * Price Cap. The Aprovechamiento sobre Rendimientos Excedentes (ARE) is levied on PEP whenever crude oil export prices exceed the price assumed for annual budget purposes. The year 2000 budget price is US$15/barrel. The export price for the so- called "Mexican Basket" is currently above US$20/barrel. The ARE rate of 39.2% applies to the difference between the budget price and the export price and is additive to the basic 60.8% DSH rate, which means that effectively 100% of the difference between the export price and the budget price goes to the Government/Hacienda. 56 Annex 2 C.- Tax Admuinistration A2.7 Several agencies are involved in the administration of oil taxes or related activities, as follows: * Pemex Tax Department. Pemex is responsible for payment of all of the above taxes. The IEPS, although a consumer tax, is collected and paid in by Pemex Refining. Pemex maintains a large, sophisticated Tax Department which is responsible for calculating taxes, tax management, tax payment (daily for most taxes, with a monthly reconciliation), tax policy and tax negotiations. * Hacienda. A small, experienced team within Hacienda is responsible for tax policy, disputes and tax/budget related negotiations with Pemex. * SAT. SAT focuses on compliance. A 4 to 5 person team manages the relationship with Pemex. It is generally accepted that SAT does not have the capacity to conduct a comprehensive tax audit of Pemex. Only two partial audits have been completed to date, the most important of which relates to Pemex salaries and benefits. * Comptroller's Office. In contrast to the Hacienda and SAT, the Comptroller's Office maintains a very large group to watch over Pemex. Comptroller's audits the metering and measurement of production, the invoicing of and payment for crude oil exports, and procedures used for the domestic and international procurement of goods and services by Pemex. 4.- Critique Of The Existing System A2.8 Mexico's current oil taxation system appears to be deeply flawed with respect to achieving the objectives listed under Section B above. In particular: A.- Efficiency of Development and Extraction A2.9 Both Mexico's oil reserve base and its pre-tax economics provide assurances of the enormous potential of the sector to support the country's economic and social aspirations: Resource Base A2.1O Mexico is blessed with a very substantial oil reserve base, both in absolute and relative terms. Current oil reserves are estimated at 28.5 billion barrels, or 13 % of world reserves outside the Middle East. Production of crude oil, at approximately 3 million barrels per day, is equivalent to 7% of world production outside the Middle East. Mexico produces a variety of crude oils of differing quality. Heavy crudes, less favored by the market, dominate and the Mexican Basket typically sells at a discount of around 20% relative to the nearby West Texas Intermediate (WTI) marker. However, these quality disadvantages are offset by Mexico's proximity to major export markets and the attractive cost conditions which apply to a high percentage of Mexico's production and reserve base. It is estimated that close to 80% of flowing oil and a possibly higher percentage of reserves can be produced at a cost of US$5/barrel (US$2.25 for operating costs and another US$2.25 for capital costs). 57 Annex 2 Pre-Tax Economics A2.11 At a Mexican Basket price of US$20/barrel, these conditions translate into very favorable pre-tax economics. Pre-tax cash (i.e. before depreciation) margins on flowing oil range from an estimated low of US$14.00/barrel in the high cost Norte region, to US$17.00/barrel in the prolific low cost Marin area. Margins for the Sur region range somewhere in between. Pre-tax rates of return calculated for investments in new oil projects varied from a low of 26% in the Norte region, to highis in excess of 70% in the Marin area. If 15% is considered to be the threshold required to justify investment, i.e., recover costs including the cost of capital, then clearly these projects are all worthy of development on a pre-tax basis. Under normal circumstances, however, investors base their decisions on post- rather than pre-tax returns. If it is to achieve a socially desirable allocation of resources, the tax system, to the maximum extent possible, should ensure that all projects with acceptable pre-tax margins or returns show acceptable, albeit reduced, post-tax returns. The broad range of activity that such a system would encourage would be not only desirable in its own right, but also attractive from a fiscal point of view in that it would expand the tax base. A2.12 In fact, Mexico's choice of tax instruments has resulted in investor incentives which are exactly opposite to those required for efficient development of the sector: Choice of Tax Instruments A2.13 The best tax instrument to promote an optimal allocation of resources is a profits tax. Under profits-based taxation, a project which shows a positive pre-tax return will show a positive, albeit smaller, post-tax return, thus satisfying the important objective of encouraging a broad range of activity. Unfortunately, Mexico relies entirely on revenue rather than profits taxes in its upstream oil sector. The profits-based character of the DEP and ISR notwithstanding, it is the DSH and the ARE that control, and both of these taxes are based squarely on revenue. Because of the insensitivity of these taxes to profit, it is quite possible that production or projects with positive pre-tax margins or returns will not have positive margins or returns post-tax. The higher the revenue tax, and/or the more modest the pre-tax return, the more likely this outcome becomes, with the result that already established production (flowing oil) is not extended to socially desirable limits or is prematurely abandoned, and investment in desirable new production is not pursued. Impact on Incentives A2.14 This is precisely what economic modeling shows in the Mexico case. Many of the more progressive oil producing countries have abandoned the use of revenue taxes. The average effective revenue or royalty tax rate world-wide is now in the 15% to 18% range. In sharp contrast Mexico's 60.8% revenue tax is the highest in the world, with the possible exception of Syria. As costs increase on flowing oil, Mexico's aggressive revenue tax begins to take more than 100% of pre-tax cash margins, even where these are substantial as in Mexico at today's oil prices. For example, post-tax margins on Norte production and part of Sur appear to be negative, implying cash operating costs cannot be recovered, let alone prior investment costs. Under normal commercial circumstances, such production, although socially desirable, would be suspended or abandoned. When to comes to investment in new projects, the adverse influence of the DSH and ARE on incentives is even more pronounced. Only the very best of the Marin 58 Annex 2 projects will produce post-tax returns on investment above the required minimum threshold of 15%. A2.15 Discouraging as they are, these results are probably a generous portrayal of the impact of the Mexican tax system. They assume the DSH applies at its nominal 60.8% rate to upstream revenues. In fact, the effective rate is significantly higher. This is because the 60.8% obligation applies to all Pemex revenues, both upstream and downstream. The downstream sector, whether in Mexico or elsewhere is incapable of supporting a revenue tax at his level or any level near to it. At the very best, it might be able to cover a 10% tax. This means that, to meet its overall tax obligation, Pemex must be paying an effective tax on its upstream operations well above 60.8% of revenues, and a correspondingly higher percentage of pre-tax profits. Thus the negative impact of the current system on upstream development and extraction incentives must be greater than that suggested by the limited economic modeling reported on here. B.- Efficient Capture and Management of Rents A2.16 Efficient capture by the taxation system of the rents generated by oil development and extraction depends on the ability of the tax system to identify and target rents and to collect a high but non-confiscatory share of those rents, positively correlated with the level of rents. Mexico's tax system does not score well in either respect: Identification of Rents A2.17 Economic rents are the excess of a project's pre-tax benefits over cost, including the cost of capital. The most important component of Mexico's tax system, the DSH, neither seeks to identify rents nor targets them for collection. The ARE does have some characteristics of a rent tax insofar as the excess price on which it is levied can be expected to generate rents. The ARE is considerably less significant than the DSH, however, and fails to recognize other important determinants of rent such as costs. Collection of Rents A2.18 As suggested above, Mexico's tax system, more often than not, may collect more than 100% of a project's economic rent, either reducing Pemex's equity if the investment has already been made, or discouraging investment in the case of a potential project. Even under the conservative assumptions made above with respect to the impact of the tax system, Mexico's typical share of project rents places it at the top end of the league table of major oil producing countries. As to the desired positive correlation between government share and rents, the DSH works sharply against this. Simple revenue taxes are notoriously regressive -insensitive to profit, they take a lower percentage from the more profitable projects than they take from less profitable projects. In the former case, potential tax revenues may be " left on the table" (although this seems unlikely given Mexico's high effective rates of tax); in the latter case, the tax system, under normal circumstances, would act to limit the range of commercially viable projects and erode the tax base. A2.19 The fundamental point under this heading is that the Government, except for the limited contribution of the ARE, is not attempting to collect rents through the tax system. Rather, it expects to ensure their capture through ownership of Pemex. This comes at a considerable cost, however. The failure of the tax system to focus on rents in an explicit way and to distinguish 59 Annex 2 them from required returns on investment confuses debate and the development of policy on incentives and efficiency, and on another major issue, namely the management of rents: Management of Rents A2.20 Historically, oil sector rents have swung widely from one period to the next as a result of price volatility. These swings create both positive and negative windfalls which can have very substantial impacts on a country's economic performance depending on the importance of the oil sector. In Mexico, where oil accounts for roughly 30% of budget revenues, their influence has been considerable, not only on overall economic performance, but also on the taxation of non-oil sectors, which tends to be correlated inversely to the movement of oil rents.' Recognizing the destabilizing potential of oil windfalls, a number of other major oil producing countries have established special funds or procedures to actively manage them. Mexico recently announced its intention to create an oil windfall stabilization fund, but as yet there has been no follow-through. C.- Accountability, and Commercial Viability of Pemex. A2.21 Existing fiscal arrangements make it unnecessarily difficult for PEMEX to develop into an accountable, and commercially viable enterprise. The effect of the tax system is to put Pemex on a very "short chain". Projects that should be profitable under modern efficient taxation are more often than not unprofitable for Pemex. Proceeding with investment in such cases would normally result in an erosion of equity. Pemex can only continue in operation because what it loses through the tax system, it recovers through the budget process. This is ostensibly an annual process. In practice, it turns out to be a running negotiation between Pemex and the Hacienda, especially when, during the year, oil prices move significantly above or below the price on which the budget is based. A2.22 This state of affairs has a number of negative consequences. Two of the most important are a loss of accountability and good governance, and reduction of the contribution of the oil sector to Mexico's overall economy. A system which forces investment decisions to taken on the basis of closed door, non-transparent negotiations is certain to cloud accountability, politicize decision-making and even raise corruption risks. Efficient decisions will only be made to the extent these negotiations can compensate, and not overcompensate, for the significant distortions to incentives created by the tax system, a highly unlikely outcome. The practical potential of the sector to contribute to the overall economy will be further inhibited by the limits on Pemex's ability to plan for even the medium term under the existing system. A2.23 It would be a mistake to interpret Pemex's ability to borrow on international capital markets as evidence contrary to the remarks made here. Lenders to Pemex are almost certainly looking past the corporation's ability to generate a positive cash flow to the Government, the 100% owner of Pemex, and /or Pemex's crude oil exports to ensure repayment. D.- Efficiency of Consumption in Petroleum Product Markets A2.24 The IEPS system of petroleum product (gasoline and diesel) taxation suffers from several drawbacks in terms of achieving economically and environmentally efficient consumption: 1. One possible explanation of this inverse relationship is the likely incentive to be more generous with politically motivated tax breaks at a time when overall tax revenues, thanks to oil, are high. 60 Annex 2 Prices Are Non-responsive to Market Conditions A2.25 Rather than allowing market conditions to be reflected in consumer prices, the IEPS increases or decreases as a function of decreases or increases in underlying product market conditions, while the administered product price remains constant. This not only discourages efficiency in the consumption of products, but also distorts supply incentives, even to the extent of causing losses for Pemex when costs of providing product to consumers rise above the administered price. This deficiency is likely to become more pronounced, the longer the gaps between administered price determiinations. Prices Are Non-responsive to Regional Cost Differentials A2.26 Because product prices are set at the same level for all regions of Mexico, the IEPS varies inversely with the cost of supplying each region, resulting in the same inefficiencies in consumption and supply as are caused by the infrequent adjustment of prices to overall market conditions. Instead of benefiting from their proximity, regions close to refining centers or import points effectively are required to subsidize more remote consuming areas. High Administrative Costs A2.27 The variability IEPS with market conditions and regional cost differentials, plus the need to keep close track of all components of product cost in order to calculate the tax, makes for administrative complexity and high administrative costs, which only add to the economic costs resulting from distorted incentives. Tax-inclusive Prices Inadequately Reflect Environmental Costs A2.28 Prices of petroleum products frequently fail to reflect the costs which consumption of those products imposes on society in terms of harm to the environment and health. While prices in Mexico under the JEPS system have on average been comparable to prices in the neighboring North America, they have been significantly (roughly 50%) below average tax-inclusive prices in Europe, where environmental concerns run high and contribute to higher taxes on the consumption of petroleum products. 5.- Recommendations For Reform. A2.29 The forgoing gives rise to a number of preliminary recommendations for reform of the Mexican oil taxation system as applied to Pemex and to petroleum products. A.- Taxation of Pemex. A2.30 For efficiency of development and production, clarity in the definition and capture of rents, effective rent management, and a commercially viable and independent Pemex, the following alternative to the existing system merits consideration: Corporate Income Tax A2.31 This should be the 35% income tax generally applied to all enterprises in Mexico, or its equivalent, the IRP, described above but made ineffective by the current DSH formula. This tax 61 Annex 2 would be levied on all Pemex operations,. both upstream and downstream, and unrelated to rents. Its application would create, as a starting point, a level playing field between Pemex and all other enterprises, public and private, in the Mexican economy. Royalty A2.32 A modest royalty, say 10% to 12% of revenues, deductible for purposes of the IRP and applicable only to Pemex's upstream operations (PEP), would provide early and dependable income from development and production operations without the severe disincentive effects created by the DSH. The royalty would collect a small percentage of the rents generated by upstream operations before application of an explicit rent tax. Rent Tax A2.33 Focussed explicitly on rents, this tax should apply only to PEP, where rents are generated, and only after recovery of costs and a reasonable return on investment. A variety of options exist for structuring such a tax and these should properly discussed and researched before a final selection is made. Choices relate especially to the tradeoffs between economic and administrative efficiency, e.g., should the allowed return on investment be project -specific (economic efficiency), or based on the enterprises's overall undepreciated assets (administrative efficiency). The tax in any event should be positively correlated with the measured level of rents so as to introduce progressivity to the tax system. When combined with the IRP and a royalty, the rent tax might produce a rate of government take, measured appropriately as a percentage of pre-tax net present value (NPV), in the range of 50% to 85% depending on achieved profitability. Rent Management A2.34 This has more to do with the management of the proceeds of taxation than with the tax system itself. If the tax system identifies rents, as is advocated here, it would be best to channel the proceeds of their taxation to a special account before disbursement rather than simply consolidating them with revenues from general taxation. This will allow explicit consideration of their appropriate use. A variety of options exist, including: allocation of a certain percentage to long term investments for future generations; another percentage to offshore investments to hedge foreign exchange exposures; and a percentage to address current development needs. Consolidating rents with the rest of budget revenues without special recognition encourages the pro-cyclical expenditure patterns and contra-cyclical non-oil tax collection observed in Mexico to date. These topics will be dealt with in greater detail in a separate paper on oil windfalls. A2.35 A reformed upstream taxation system, by improving development incentives, expanding the tax base and collecting rents efficiently, would have the potential to significantly and positively impact both production and tax revenues. Under the existing tax regime, without Mexico's compensating budget system, it would be impossible to finance new development, and production could be expected to decline at about 7% per annum from its present level of approximately 3.1 million barrels per day to 1.8 million barrels per day by the year 2010. Associated annual tax revenues might decline from US$16.0 billion to US$9.5 billion. A modern tax system such as that described above should be capable of attracting investment sufficient to cause production to increase to as much as 5.3 million barrels per day by 2010, with estimated tax revenues of US$25.0 billion. These are very major differences. Of course, the Government is not going to allow the 7% decline case to occur. It may come up with a budget 62 Annex 2 for Pemex which is adequate to maintain or even slightly increase current levels of production and tax revenue. However, it is unlikely to be able to come up with the funds needed to achieve the full upside potential. The investment required over the 10 year period would be on the order of US$40 billion. Under tax reform, Pemex should be able to come up with part of this sum itself, through retained earnings and enhanced borrowing capacity based on improved earnings prospects. The sheer scale of the potentially viable investment under tax reform also suggests the possible merits of expanding participation in the upstream sector to include domestic and foreign private sector investors. Finally, it is important to keep in mind that reform would bring rewards not just in terms of incentive-related impacts on production volumes and fiscal revenues, but also in terms of increased accountability and transparency in Pemex's operations, which in turn should have a further positive impacts on volumes and revenues. B.- Petroleum Product Taxation. A2.36 With respect to petroleum product taxation, efficiencies in consumption, supply and administration could be restored and environmental gains achieved through: Excise or Ad Valorem Taxation A2.37 The administered price element of the IEPS system should be replaced by fixed specific tax per liter or flat ad valorem tax, either of which would allow market signals (changes in import parity prices) to filter through to consumers and suppliers. Regional Price Variation A2.38 Regional prices should be allowed to vary to reflect differing costs of delivery. Because of the risk of monopoly abuse in remote, poorly served areas, some regulatory oversight will be required. Addition of Environmental Premium A2.39 Comparison with other countries suggests there is room for Mexico to consider incorporation of an environmental premium into its petroleum product taxes. This could be expected to have the combined impact of reducing consumption while raising revenue. 6.- Obstacles To Reform/Transition Issues A2.40 Changes of the kind proposed here are bound to present difficulties, albeit offset by the benefits they bring. Principal concerns regarding these reform recommendations are likely to relate to: Maintenance of Revenues A2.41 Maintenance of short term revenues is a classic concern whenever fiscal reform is discussed. In the upstream sector, tax revenues, excessive now, must decline under reform. However, the net impact on the budget need not be negative as revenue declines will be partially offset by reduced budget allocations to Pemex and by dividends. In the medium to longer term, tax revenues should grow as a result of expanded activity and an expanded tax base. As to petroleum products, replacement of the EEPS need result in no decrease and may even increase revenues. 63 Annex 2 Tax Administration /Cost Control A2.42 Recommendations made here would involve a massive shift away from revenue taxation towards profits taxation. Effective administration of profits taxes requires the collection, presentation and audit of a sometimes bewildering variety of costs on a timely basis. Where tax administration is weak, profits taxes may offer the taxpayer considerable scope for tax evasion through misrepresentation or inflation of costs. While a move to profits taxation in practice may require some upgrading of institutional capacity in Mexico, the risks of a change are reduced by the experiences gained through, admittedly pro forma, assessment of the IRP and DEP, and by the cost containment roles played by the Controller of the Federation and by legal requirements for competitive bidding in contracts. At the same time, there is clearly a need to parallel any tax reform program with a careful review of existing administrative capacity and plans for its enhancement. 7.- Summary And Conclusions. A2.43 This paper leaves a considerable number of issues still to be addressed and its findings and recommendations, summarized below, are necessarily preliminary and incomplete. Comments and corrections are welcome. * Existing taxes create serious inefficiencies in both the upstream oil sector and in petroleum products markets. * Both the level and structure of upstream taxes are in need of major revision. The level needs to be reduced to extend the margin of commercially viable production. The structure needs to shift dramatically away from revenue to profits based taxation, with explicit taxation of rents. * Rents from oil production should be singled out for transparent, professional management with an eye to regulating their impact on the overall economy. * Tax reform has the potential to produce significant gains in oil production and tax revenues. Over the next decade, production could increase by as much as 3.5 million barrels per day over levels that would result under the existing tax system and normal commercial behavior, and by an estimated 2.2 million barrels per day over today's levels. Corresponding increases in annual tax revenues would be US$15.5 & US$9.0 billion. * Taxes on Pemex do not really constitute a tax system at all, but rather represent an opening position in a time consuming, resource intensive process of negotiation between Pemex and the Hacienda on taxation and Pemex's budget. This process has serious negative implications for accountability and governance in the sector. * The IEPS system of product taxation is excessively cumbersome and puts an undue administrative burden on Pemex, in addition to the inefficiencies in consumption and supply which it encourages. 64 Annex 2 * The transition to a tax reform will present challenges, but these can be overcome. Particular care should be taken to ensure that the institutional capacity is in place to administer any new system. 65 Annex 2 66 Annex 3 A3 ANNEX 3 NAFTA and Mexico's Domestic Tax Policy INTRODUCTION I. THE ECONOMIC IMPACT OF NAFTA HI. NAFrA AND MEXICO'S TAX POLICY Im. CONCLUSIONS AND RECOMMENDATIONS APPENDIX I APPENDIX 2 67 Annex 3 Introduction A3.1 The North American Free Trade Agreement (NAFTA) was signed in December 1992 and came into effect January 1, 1994. By most accounts NAFTA has had a significant effect on Mexico's economy and institutions. The consideration of tax reform in Mexico requires an evaluation of the role of NAFITA in Mexico's economy and how Mexico's tax system may affect its trade with its NAFTA partners, Canada and the US, and equally important, how tax reform may affect the cross-border investment flows from those two countries into Mexico. Clearly, no good tax reform in Mexico can ignore the role of NAFTA. However, it is not entirely obvious how NAFTA concerns should shape or constrain Mexico's tax policy reform. This report addresses that basic question. The report first reviews some of the evidence on the economic impact of NAFTA by focusing mainly on the evolution of foreign trade and foreign direct investment (FDI) flows in Mexico. The report then considers the differences between Mexico's tax system and those of Canada and the US and the consequences of these differences. We conclude with a consideration of the main implications for tax reform in Mexico. I. The Economic Impact of NAFTA. Mexico's standing within NAFTA: A3.2 Mexico's share of North America's GDP is the smallest. As of 1998, Mexico represented 4.3 percent of NAFTA or North America's GDP as opposed to Canada's 6.5 percent and the United States' 89.2 percent. Over the last two decades, Mexico's share in North America's GDP was at its highest in 1981 (8.4 percent) and at its lowest in 1986 (2.7 percent). Overall, as shown in Figure A3.1, Mexico's share has fluctuated up and down and no definite trend has emerged. Impact on cross-border trade: A3.3 The openness of the Mexican economy increased dramatically over the last decade. The sum of exports and imports as a proportion of GDP rose from 35 percent in 1991 to 62 percent in 1999. Over this period, exports in US dollar terms have grown by 165 percent or at an average annual rate of 14 percent per year (Table A3. 1). Over the last decade, also, Mexico has become much less dependent on oil for its export revenues. In 1991 oil exports still represented 19 percent of all Mexico's exports. By 1999 this share had fallen to 7 percent (Table A3.2). While oil exports in US dollars remained basically at the same level over the last 10 years, non-oil exports took off, especially after 1994. A3.4 The high rate of growth on exports over the past decade has been uneven (Table A3.3). The merchandise trade by type of industry shows that exports by manufacturing industries tripled in US dollars from 1991 to 1998, while the value of exports in the extractive industries decreased and in agriculture and forestry increased more moderately. Over the decade, exports from manufacturing went from representing 76 percent of total exports in 68 Annex 3 1991 to representing 91 percent in 1998. Within the manufacturing sector, the best export performers over the 1994-1998 period were "textile, apparel and leather" with an export growth of 202 percent and "metallic products, machinery and equipment" with an export growth of 111 percent. Many other manufacturing industries had export increases in 1994- 1998 of over 80 percent (Table A3.4). The most significant of all of these increases was in the "metallic products, machinery and equipment," mostly the auto industry, which represented 64 percent of all Mexico' exports in 1998, up from 48 percent in 1991. A3.5 This explosion of manufacturing exports has been accompanied by a significant growth of imported intermediate inputs, linking many of the fastest growing areas of Mexican imports to the demand for Mexican exports rather than to fluctuations in Mexican domestic demand.' As shown in Table A3.5, total imports by manufacturing industries have remained dominant. They represented 93 percent of all imports in 1998 and many of the fastest growing import sectors were also among the fastest growing export sectors over the 1991-1998 period. A3.6 A significant share of the export growth has come from the export assembly plants or maquiladora sector. By 1998, the maquiladora sector represented 45 percent of all exports, up from 37 percent in 1991. Over the 1994-98 period, exports from maquiladoras grew by 102 percent by comparison to 86 percent of non-maquiladora exports. However, some manufacturing industries traditionally not in the maquiladora sector, such as "metallic products, machinery and equipment," mostly the auto industry, grew even at faster rates than the maquiladora sector (Table A3.4). A3.7 The impressive performance of the maquiladora sector in production, total employment, and salaries is documented in Tables A3.6 and A3.7. From 1990 to 1997, total output in constant 1993 pesos tripled and value added doubled.2 Over the same period of time, the total number of workers in factories and other locations doubled, and the annual pay per worker in current pesos more than tripled. As shown in the last row of Table A3.7, by far the largest gains on all counts of the maquiladora sector took place from 1994 onwards. A3.8 Another significant feature of the growth in Mexico's exports is that geographically it has been highly concentrated in exports to the US. This is clear from Figure A3.2, where we see that the growth in exports to the US closely mirrors the growth of Mexico's exports, at the same time exports to Canada and the Rest of the World have increased at much slower rates. From 1989 to 1999, Mexico's total exports in US dollars went from US $ 35 billion to US $ 137 billion and Mexico's exports to the US went from US $ 28 billion to US $ 121 billion (Table A3.8). By comparison, Mexico's exports to Canada, the second largest trading partner for Mexico, went from US $ 277 million in 1989 to US $ 2.3 billion in 1999. This is also an impressive increase, but Mexico's exports to Canada represent less that 2 percent of those to the US. Despite their large size, Mexico's exports to the US have increased over 133 percent from 1994 to 1999. Exports to other countries have increased faster over the past five years, but no other country comes even remotely close to the relative size played by the US. The 1. See Hinojosa Ojeda et al. 2000 2. Interestingly, at the time of the NAFIA signing it was expected that Mexico would suffer the biggest losses in the maquiladora sector because of the loss in competitive advantage since all finms would thereafter face the same tax and tariff regulations See Bulmer-Thomas et al. 1994. The role of NAFTA is discussed below. 69 Annex 3 share of exports to the US in Mexico's total exports went from 81 percent in 1989Ato 90 percent in 1999. In a distant second place was Canada, which represented less than 2 percent of Mexico's total exports in 1999. The economic integration of Mexico with the US economy is two-sided. Mexico's total imports in 1999 were US $ 142 billion, of which US $ 105 billion were imports from the US. Mexico's imports from Canada were under US $ 3 billion (Table A3.9).3 Not surprisingly, imports from the US and Canada grew fast over the 1989-1999 period. Imports from the US represented 74 percent of all of Mexico's imports in 1999. Canada at 2 percent was behind Germany, Japan and South Korea. A3.9 What has been the impact of NAFTA on the tremendous growth in Mexico's exports? This impact is difficult to disentangle for several reasons. First, Mexico's foreign trade and foreign investment regime liberalization, through the reduction in tariffs and quantitative restrictions or quotas, started in the mid l980s. Other important institutional breakthroughs in Mexico, including joining GATT, the liberal Foreign Investment Law of 1989 (which reversed Mexico's. previous restrictive policies toward foreign investment), and the elimination of foreign exchange controls in 1991, had their effect on trade before 1994 when NAFTA came into being (Figure A3.3). The time series for exports makes it clear that an increasing integration of North America's market through cross-border trade was already occurring in the late 1980s and early 1990s prior to NAFTA. Therefore, NAFTA may have continued and consolidated trends that already existed but it did not necessarily represent a fundamental shift in the growth of exports. A3.10 Second, as we have reviewed above, the growth in exports to the US and Canada, especially to the US, did take off sharply in 1994 -95. Again, however, there were other factors which may have had as much, or more, of an impact on the growth of exports than NAFTA itself.4 The most important of these factors were the sharp devaluation of the peso in December 1994 and the 1995-96 recession, which pushed exporters and traditionally non- exporters alike to seek sustained demand in the export markets. What is clear is that NAFTA. facilitated the acceleration of exports by providing increased access to the US and Canadian markets and by providing safety and certainty to US investors in Mexico. This also meant that because of NAFTA, quite likely Mexico's recession was much less pronounced than it would have been without NAFTA. 3. By comparison, Mexico's imports in 1999 from Germany and Japan were US $ 5 billion each, and those from Korea were about the same as from Canada. 4. See, for example Krueger 1999 and Hinojosa Ojeda et al. 2000. 5. Hinojosa Ojeda et al. 2000 argue that statistically the lower tariffs due to NAFIA can explain only a small portion of the increase in Mexico's exports since 1994, and that a much lager role should be attributed to the collapse of the peso and the subsequent recovery and to the on-going bi-national integration between Mexico and the US Hinojosa Ojeda et al. also argue that exports for the commodities liberalized by NAFTA actually grew more slowly than those commodities that were not (either because they were already liberalized before NAFTA, liberalized by other means or not at all liberalized). 70 Annex 3 Impact on cross-border investmentflows: A3.11 Even though there may be some controversy as to the net impact of NAFTA on cross- border trade vis-a-vis other factors, the impact of NAFTA on cross-border investment flows appears to be much clearer. Total foreign direct investment into Mexico took off with the liberalization reforms of the mid 1980s (Table A3.10). After the approval of the Foreign Investment Law in 1989, FDI into Mexico further shifted up totaling between US $ 3 billion and US $ 4 billion per annum. After NAFTA came into force in 1994, FDI again experienced a significant increase reaching over US $ 10 billion per annum in 1994 and 1997 and not less than US $ 7 billion in 1995-96.6 The exception was 1998 when total FDI was only US $ 4 billion. This latter dip was in the aftermath of the East Asian crisis in 1997 and the Russian crisis in 1998. A3.12 FDI in Mexico has been dominated by the US since the early 1980s (Table A3.1 1). The share of US FDI in total FDI was 66 percent in 1980 and 71 percent in 1998. Because of the lumpiness of many large FDI projects, the shares of the US, and other important home countries, such as the UK, Germany, and Japan, for FDI in Mexico have fluctuated over the years, but the US continues to be the dominant presence. On the other hand, Canada represented less than 3 percent of all FDI in Mexico for 1998. Regional effects: A3.13 NAFIA has also had an impact on the distribution of economic activity through the location of FDI, especially in the maquiladora sector. Before economic liberalization in the mid 1980s, the import-substitution economic activity was located mainly in Mexico DF and some other areas in the center of the country. With economic liberalization and the growth of the maquiladora sector, a considerable share of economic activity shifted to the northern border states. After NAFTA it would appear that the economic activity, including the location of the maquiladora has spread to some extent throughout the country. For example, the state of Puebla had 146 new maquiladoras started from 1994 to 1998, almost double those started in the states of Sonora and Coahuila, both northern border states, during the same period (Figure A3.4). However, there is no indication that NAFTA has been able to narrow the divide between the poor South and the rich North.7 If any thing, the impact of NAFTA on production, employment, and exports has been more pronounced in the northern border states of Mexico than elsewhere in the country. This has helped to keep wages at a higher level in 8 the area, especially after the devaluation and macroeconomic adjustment of 1994-95. 6. Although Mexico's 1994 crisis and devaluation and ensuing recession in 1995-96 can be argued to have had an independent impact (from NAFrA) on Mexico's exports, it is less likely that these same events have had a significant independent pull on FDI. However, see Trigueros 2000 for a more skeptical view. See also Rubin and Alexander 1995 for an early review of FDI issues in NAFrA 7. See Bulmer-Thomas et al. 1994. 8. See Davila Flores 2000 and Katz 2000. 71 Annex 3 Impact on Tax Bases: A3.14 The most relevant question, but also the hardest to answer, is whether NAFTA has had any discemable impact on tax bases or the ability to collect taxes in general. This information would be valuable to policymakers in order to adapt the country's tax structure to the new economic environment created by NAFTA. We have seen, however, that even some of the more direct effects of NAFTA, such as the impact on cross-border trade, are difficult to disentangle. So we have no expectations of being able to identify the direct impact of NAFTA on tax bases and tax revenues. However, we can observe that NAFTA has likely contributed to several changing trends in the composition of GDP. A3.15 Since the late 1980s, the time of the most important tax reforms, there has been significant economic growth. But some of this growth may not have translated into growth of the main tax bases. Value added from the service sector increased from 50 percent of GDP in 1986 to 61 percent of GDP in 1998. Over the same period, value added in the industrial sector decreased from 32 percent to 26 percent. Actually, the impact of NAFTA may have been to slow down the decline of industrial value added, specifically in manufacturing, vis-a- vis the value added in services. The implication of these trends is that it is easier to collect both income taxes (CIT and withheld PIT) and VAT from large manufacturing enterprises than from more fragmented service-oriented finms. A3.16 The service sector in Mexico includes the difficult-to-tax "banking and financial institutions" which grew from 8 percent of GDP in 1986 to 13 percent of GDP in 1998. The service sector also includes "transport and communications," part of which is more lightly taxed through the special regime in the CIT, and which grew from 7 percent of GDP in 1986 to 10 percent of GDP in 1998. Finally, the service sector also includes "public administration and defense" which are not taxed, and grew from 16 percent of GDP in 1986 21 percent of GDP in 1998 (Table A3.12). However, not all trends in the composition of GDP have narrowed the tax base or made tax collection more difficult. In particular, the share of agriculture, a sector that is traditionally hard to tax and which is also preferentially treated under the VAT and the CIT, decreased from 9 percent of GDP in 1986 to 5 percent in 1998. Another positive trend has been, as mentioned above, the growth of FDI and export oriented firms in manufacturing. The downward trend of the contribution of manufacturing to GDP was reversed after NAFTA in 1994. However, this advantage for tax base growth is offset by the fact that this sector tends to be dominated by the maquiladoras, which traditionally pay lower taxes, and by firms with foreign ownership and large Mexican enterprises, which have been able to escape taxation in varying degrees through the use of consolidation. A3.17 How should the tax system adapt to these major trends? There are no easy answers, but these findings seem to strengthen the case for the elimination of special tax treatment of some service sectors such as transport, the elimination or restriction of consolidation for large fimns, foreign and or domestic, the strengthening of the taxation of services under the VAT, and in particular for more effective ways to tax the banking and financial sectors. 72 Annex 3 Intangible BeneJits of NAFTA A3.18 The impact of NAFTA on the Mexican economy does lies not only in the increased trade and investment flows, but also in numerous intangible benefits that should yield more efficient use of economic resources and sustained economic growth in the years to come. Numerous NAFTA observers have emphasized the relevance of these intangible benefits from NAFTA including the following:9 * NAFTA has served as a commitment device to force reforms and some of the reform process has been extended to other sectors of the economy (Tornell and Esquivel, 1995; Blomstrom and Kokko, 1997). * NAFITA has produced major advances in areas such as government procurement, intellectual property rights, and conflict resolution with binding investor-state arbitration (Blomstrom and Kokko, 1997). * The foreign competition introduced by NAFTA, in turn, has induced significant gains in productivity in the Mexican economy (Trigueros, 2000). * NAIFTA has contributed to more stable and mature economic and political relations between Mexico and the US and Canada, helped to open Mexico to the world, and brought certainty and stability in the international arena (Fernandez de Castro, 2000; von Wobeser, 2000; and Krugman, 1993). * NAFTA has served as an insurance device to foreign investors against policy reversals not only by Mexico but also by the US, as exemplified by the fact that the explosion in Mexico's exports to the US after the 1994 peso devaluation was not followed by US retaliation. In addition, NATErA very likely played a role in the US support to prevent a default Mexico in 1995 (Fernandez, 1997; Studer, 2000). A3.19 However, not all is well after NAFTA. The question is how broadly and deeply the export-led growth has benefited the rest of society. World Bank 2000 takes a pessimistic view. In this view, the maquiladoras and the large exporting and foreign-owned firms may be creating an enclave that is not integrated with the rest of the economy. The export sector has been an engine of growth, but much less successful as a vehicle for equitable growth. [. NAFTA and Mexico's Tax Policy. NAFTA is about cross-border trade and investmentflows and not about taxpolicy: A3.20 The fundamental objective of NAFrA is to achieve trade and investment liberalization within the three member countries.'0 The goal of achieving the free flow of goods is pursued by imposing a nondiscrirnination rule (granting the trade partners the same treatment provided to nationals), and by removing over time (for up to 15 years) the existing tariffs." The goal of freeing investment flows is pursued by requiring each member to provide investors and 9. But see Fernandez 1997 for a critical view on some of these intangible benefits. 10. See, for example, McDaniel 1994 for a good summary of the issues. 11. Still extensive rules of origin apply for trade within NAFTA and a few economic sectors are exempted completely from the removal of tariffs. 73 Annex 3 investments from the other two countries the same treatment provided to its own nationals in all aspects of the investment process (from acquisition to management to disposition of investments). 12 A3.21 Quite clearly, NAFTA is not about tax policy coordination among the member countries. NAFTA lets the member states freely develop their domestic tax policies and relies on the bilateral treaties to coordinate any problems that may arise. The only article in NAFTA dealing with tax policy issues (article 2103) states that "nothing in NAFTA shall apply to taxation measures" and that "nothing in NAFTA shall affect the rights and obligations of any of the three countries under any tax convention."'3 In the case of any contradictions between the tax treaties and NAFrA, the former are supposed to prevail. A3.22 The build-up to NAFTA led Mexico to conclude comprehensive bilateral tax treaties with Canada, which became effective January 1992, and with the United States, which became effective January 1994. The three treaties (including the Canada-US treaty) apply to all income taxes imposed by the federal governments. The two treaties with Mexico also specify the coverage of Mexico's asset tax. All three treaties contain anti-discrimination provisions which apply to additional taxes. These provisions ensure that national taxes placed on goods and services do not discriminate against foreign goods and services in favor of domestic ones. Of course, the three treaties accept the different treatments of capital income among the three member countries and nothing is done to address the impact of these differences on cross-border investment flows.1 In addition, the three countries have agreements for information sharing to simplify the tasks of the tax administration and improve tax enforcement. Differences in tax regimes within NAFTA and their implications A3.23 Of course, the tax systems of Mexico, the US, and Canada differ in some ways and are similar in other ways. One main difference is the level of overall tax effort in the three countries. In 1997, general government tax revenues represented 37 percent of GDP in Canada, 29 percent in the US, and less than 17 percent in Mexico. Clearly, the tax systems in the three countries are used to pursue different objectives, including the level of services to be provided through the public sector. There are also differences in tax structure. For example, Mexico and Canada have a national VAT while the US does not. Other differences include rates and base definition for income taxes, social security taxes and excises. The US and Canada have wider social security programs and use payroll taxes more heavily than Mexico 12. In particular, a country may not impose minimum levels of equity to be held by its nationals, nor require senior management to be of a particular nationality, nor impose performance criteria, such as exporting a given percentage of production. However, the majority of the board of directors may be required to be of a particular nationality. 13. Three other articles in NAFTA touch upon tax issues. There is a general nondiscrimination provision, extended to state and local governments, in article 301, accompanied by the prohibition against using discrimninatory taxes on exports (article 314). In addition, article 604 has several provisions on energy taxes. Tax-like barriers to trade such as customs duties, anti-dumping and countervailing duties, and importation fees are not considered "taxation measures" according to article 2107. 14. See Cockfield 1998 for a discussion of how the three bilateral tax treaties coordinate the tax treatment of cross-border flows in trade and investment. 74 Annex 3 does. On the other hand, all three countries during the 1980s introduced similar reforms for income taxes by cutting rates, broadening bases and reducing tax incentives. The most significant round of reforms was after the 1986 US tax reform, to some extent followed by both Canada and Mexico.15 A3.24 Which differences in tax structures matter within the context of NAFTA? Or, which differences in tax structures have the potential of negatively affecting trade and the cross- border flow of investment funds? Few of the differences in tax systems in the three countries are likely to affect trade and cross-border investment flows. For example, differences in personal income taxation do not count for much because NAFTA does not provide for the free mobility of labor.16 Other tax differences with the potential to distort trade patterns, such indirect taxes and differences in corporate income taxes, in reality do not because exchange rates offset the impact of differences in uniform taxes.17 A3.25 The differences in the tax systems that are of relevance in the context of NAFIA are those with the potential to distort cross-border investment patterns. The definition of taxable income and tax rates in each of the countries may impact the mobility and final allocation of investment resources. These effects should come primarily from differences in the CIT, but also from differences in property taxes and because of the gross asset tax in Mexico. The most important differences in the CIT across the three countries include:'8 * different withholding rates imposed by the three bilateral treaties on cross- border payments of parent/subsidiary dividends, portfolio dividends, interest, and royalties (See Table A3.13), * different systems of mitigating double taxation employed by the three countries (worldwide taxation by the US and Mexico versus the exemption of territorial systems by Canada), * differing levels of integration between the CIT and PIT, with the use of dividend credit in Canada, dividend exclusion in Mexico, and the classical system with no integration at all in the US, * the potential for over and under-taxation caused by the lack of agreement on source rules for different categories of income and deductions, * different tax subsidies used in each country to encourage the development of particular economic activities, 15. The pressure on Canada and Mexico has been to narrow differences, mostly for the corporate income tax (CIT), with US taxes in order to continue to offer an attractive environment to highly mobile capital. The CIT in Canada and the US have converged considerably over the years (Boadway and Bruce, 1992) and so has the CIT in Mexico with that in the US (Gordon and Ley, 1994). But, as reviewed below, significant differences remain in the CITs of the three countries. 16. See Gordon and Ley 1994. 17. Gravelle 1986 shows that direct effects of corporate income taxes are offset in the aggregate by an adjustment in the exchange rate. 18. See for example McDaniel 1994 and Gordon and Ley 1994. 75 Annex 3 * differences in the tax treatment of leasing (Mexico does not allow the use of leasing agreements to transfer depreciation allowances from one firm to the other, while the US and Canada do), * differences in indexing for inflation (full indexing of assets and liabilities in Mexico and not in the U.S and Canada), * differences in the treatment of inventories (expensing of purchases in Mexico versus traditional LIFO/FIFO treatment in the US and Canada), * differences in depreciation allowances for fixed assets, * differences in capital import duties (both Canada and the US exempt the import duty on capital goods but Mexico only does that for exporters), * differences in the transfer of losses among enterprises through purchases and other means (which are much more restricted in Mexico vis-a-vis the US and Canada). A3.26 The differences in tax regimes clearly can lead to the distortion of investment decisions on how much to invest, in what economic activity and in what country.19 The differences in tax regimes may also lead to tax arbitrage (i.e., corporations attempting to gain tax benefits offered by one country without any changes in their real economic activities.20 The gross asset tax in Mexico plays a particular role in tax arbitrage between the three countries. Since the gross asset tax can reduce the CIT to zero on reported income, US and Canadian multinationals have an incentive to transfer income to Mexico via transfer pricing or other means. A3.27 The basic case for reforminrg the tax structures of the three countries within NAFTA is that overall consumer welfare (in the three countries) would be maximized if current distortions to the cross-border investment flows were eliminated. The urgency to carry out these reforms is that existing distortions are expected to get more pronounced as cross-border activity continues to increase.21 However, it will be important to know how significant these distortions may be.' The next section discusses the estimates of marginal effective rates of taxation (MERT) on new investment in the three countries, building on the analysis in Annex 4. Two qualifications are necessary to the welfare loss argument. First, capital is unlikely to bear the burden of the tax distortions induced on investment activities. Factors with less mobility including labor and, of course, land, are more likely to bear that burden. Second, the differences in business costs may not always lead to distortions (changes in investment behavior). Taxpayers may not change location if they derive additional benefits from higher government expenditures in a particular location or if there exist pure rents that firms enjoy in reference to a particular location. A3.28 There is one final potential implication of differences in tax regimes of member countries in a free trade area, negative tax competition. This has been an important concern, 19. For example, because intermediate inputs are treated more favorably in Mexico, firms with substantial inventories may want to locate there; or because of Mexico's restrictions on the transfer of losses, firms with tax losses may want to locate in the US or Canada. See Gordon and Ley 1994 for other examples. 20. For example, profits are moved from a high to a low tax rate jurisdiction via transfer pricing, or because of the existing differences in the treatment of leasing between companies in the US and Mexico. 2i. See McDaniel 1994 or Cockfield 1998. 76 Annex 3 for example, among European Union officials.22 The traditional concern about tax competition is that it may lead to a "race to the bottom." By continuously lowering tax burdens on capital income, every country may find that its revenues are insufficient to cover 23 all needed expenditures. However, there is no evidence of harmful tax competition within NAFTA, or if there is tax competition, that it has led to undesired lower tax revenues. However, as pointed out above, there has been a process of convergence in tax rates and the broad definition of the base for corporate income taxes, with Canada and Mexico following the lead of the US.24 A Marginal Effective Tax Rate Analysis of Mexico's Corporate Tax System within the NAFTA: A3.29 This section provides a marginal effective tax rate (METR) analysis on the Mexican corporate tax. system in comparison with those in Canada and the United States. A summary of the corporate tax systems in these three countries is presented in Appendix 1, and an explanation of the impact of non-tax parameters on the marginal effective tax rate in Appendix 2. The simulations of the effective tax rate on capital are carried out for multinational firms from each of the three NAFTA countries investing in the other two NAFTA countries. The simulation covers only manufacturing and service sectors, which are the focus of foreign direct investment. Assuming that multinational firms in these sectors are generally large, this simulation does not include any special tax treatment for small taxpayers. A3.30 The main results of the simulations are presented in Table A3.14 with METRs on foreign capital investment in Mexico, Canada and the US respectively. In each of these three countries as a host, the other two are simulated as foreign investors. The first two panels (1A and IB) in Table A3.14 are for Mexico as a host and foreigners as non-exporters and exporters respectively. As described in Appendix 1, the special tax benefit enjoyed by exporters is the import duty exemption for inputs including capital goods, which is not available to the non-exporters. Panels 2 and 3 of Table A3.14 are for Canada and the US as host country respectively. It should be noticed that both Canada and the US exempt the import duty on capital goods. A3.31 There are mainly four findings from the simulations: 22. See Weiner 2000 for a recent discussion of issues in tax competition within the European Union, including the Code of Conduct introduced in 1997 with measures against harmful tax competition.. 23. Tax competition may have advantages intra-nationally by keeping subnational governments more efficient (McLure, 1986 ). However, these benefits are much less likely to arise internationally among countries in a free trade area. 24. Given the relative size of the US economy vis-a-vis its partners in NAFIA, any tax competition will be necessarily one-sided. The differences in relative size imply that the US tax system will always have a disproportionate effect on capital movement within NAFTA and that the US is less likely to be affected by the tax policies of its NAFTA partners (Cockfield, 1998). 25. A description of the methodology used in the computations of the METRs was provided in Appendix 1 of the companion report, "Main Features of Mexico's Tax System." 77 Annex 3 A3.32 First, when the import duty is exemrted, Mexico appeared to be the lowest taxed country among the three NAFTA members. 6 This is well justified by its low CIT rate compared with the other two (i.e., 35 percent versus around 40 percent in Canada and the US). However, when we look at the case for non-exporters (Panel IA, Table A3.14), Mexico's tax advantage disappears in the manufacturing industry and withers in the services sector compared to the case for the US as a host for foreign investors. A3.33 Second, Canada appears to be the highest taxed country for foreign investors within NAFTA. This is also evident due to its high CIT rate (36 percent for manufacturing and 43 percent for services sector27). The other factors contributing to the high METR in Canada include the provincial capital tax rate (about 0.3 percent) and the FIFO accounting method required for the tax purpose.28 A3.34 Third, Canadian and Mexican investors appear to be at a disadvantage, when they invest in each other's country, compared with their US cousins (Panels IA, lB and 2, Table A3.14). This is mainly because they both have a better treaty with the US but not with each other. That is, the withholding tax on repatriated dividends is higher between Canada and Mexico (10 percent) compared to that between each of them with the(5 percent). A3.35 Fourth, in any given host country, the marginal effective tax rate borne by foreign investors differs from each other. This is a combined result of the given home country's tax system and the bilateral treaty between the home and host countries. More specifically, a foreign investor from a country with higher CIT rate could benefit more from the interest deduction and hence incur a lower financing cost of capital brought to the host country.29 Furthermore, a higher withholding tax rate could cause a higher financing cost for capital brought by the foreign investors from home. For example, Panel 2, shows that the US investors incurred a lower METR in Canada compared with their Mexican counterparts. Similarly, the Canadians incurred a lower METR in the US compared with their Mexican counterparts (Panel 3, Table A3. 14). In both cases, the lower CIT rate in Mexico reduces the tax benefits from the interest deductibility for the Mexican investors. In the case where Canada is the host country (Panel 2, Table A3.14), the higher withholding tax on dividends 26. Similar results have been found in previous research. Chen and McKenzie 1997 estimated METRs for investment in capital employed in manufacturing and services undertaken by domestic investors in the G7 countries (which include Canada and the US), plus Mexico and Hong Kong. In the manufacturing sector, Mexico's domestic investors for large firms face the lowest METR after Hong Kong. The METR in Mexico is 16.5 (while in Hong Kong it is 11.9). By comparison, these rates were 25.5 for Canada and 21.5 for the US. In the case of services, Mexico's METR is slightly higher at 17.7 (versus Hong Kong 3.7). For services, the METR in Canada is 32.2 and in the US 19.9. In a previous study Iqbal 1994, using a cash-flow model, also found tax burdens in Mexico to be more competitive than those in Canada and the US. 27. For simplicity, we use Ontario's CIT rate (i.e., 13.5% for manufacturing and 15.5% for other sectors) representing the provincial CIT rate in Canada. 28. As explained in Appendix 2, the FIFO accounting method could cause inflated taxable income and hence pump up the METR. This impact can be significant when a rather high capital share has to be allocated on inventory such as often happens to the manufacturing sector compared with the services sector. 29. It should be noticed that we are aware of the restriction which could be imposed by the US interest allocation rule on the interest deductibility for the US multinationals at home. For simplicity, our simulation for the US multinationals includes only the case of "excessive limit for foreign tax credit," in which the US multinationals do not face the restriction on the interest deduction (for the tax purpose) at home. 78 Annex 3 between Canada and Mexico (10 percent compared to 5 percent with the US) further increase Mexico's tax disadvantage compared with the US. When we look at the case where Mexico is the host country (Panels 1A and lB, Table A3.14), we are unable to draw such a clear-cut conclusion. This is true in particular for the services sector where not only the Canadian CIT rate is higher but also Mexico's withholding tax rate for dividends repatriated to Canada are higher than to the US. Obviously, the effect of the higher withholding tax on the dividends to Canada appears to more than offset the effect of the higher CIT rate (for the interest deduction) in Canada. How does Mexico compare and what needs to be done? A3.36 In summary, given the existing differences in the taxation of capital income within NAFTA, Mexico does well in being competitive for attracting cross-border investment flows. Mexico could do better if it were to follow the US and Canada in exempting from import duty all capital imports for both exporters and non-exporters. Although Canadian foreign direct investment flows into Mexico are not large, bringing the current withholding tax rate on repatriated dividends between Mexico and Canada from 10 to 5 percent (the latter is the rate between Canada and the US and Mexico and the US) would increase Mexico's attractiveness to Canadian investors vis-a-vis the US. A3.37 The differences in METRs estimates could encourage investment to move to Mexico because of its lower rates even if Mexico has lower before-tax rates of return for those investment activities. If this were the case, the overall pool of capital available in the three countries would be used less efficiently, or other words, the overall level of output for NAFTA would be lower. But clearly, a more efficient allocation of resources within NAFTA would not necessarily mean that Mexico would become better off. At any rate, it would not be possible for Mexico alone to eliminate existing distortions in cross-border investment flows. In addition, the welfare losses arising from these distortions are not likely to be large. Note that the METRs are only an approximation of the manner in which the current tax systems favor or discourage investment relative to other countries, but they do not provide an estimate of the actual welfare losses.30 III Conclusions and Recommendations A3.38 NAFrA has so far had a very significant impact on Mexico's economy. Even though the tremendous increase in exports since 1994 can be partly explained by other factors, mainly the devaluation of the peso in December 1994 and the pressure to export that followed with the recession of 1995-96, NAFTA also appears to have played a significant role in the sustained increase in the level of exports. The positive impact of NAFTA on the sharp increase in cross-border investment flows is much less controversial. Mexico's dependence on oil exports in the past has been shed for a strong export oriented manufacturing sector based not only on the maquiladora sector but also on the general economy fueled by sustained foreign direct investment, largely from the US NAFTA has also had a variety of positive 30. In an early estimate, Brown et al. 1992 concluded that NAFTA could add around 0.1 percent to US real income and around 4 percent to Mexico's real income. We would expect the distortions to investment flows be a fraction of those gains. 79 Annex 3 intangible effects on the modernization and opening of Mexico's economic and political institutions. A3.39 Mexico's profound economic transformation over the last decade has also had important effects on tax bases, and quite likely on the ability of the government to collect taxes. The relative importance of agriculture in GDP has declined sharply. The relative roles in GDP of some types of manufacturing have held steady, but most of them have also declined in importance. On the other side of the coin, the relative roles of the service sector and public administration in GDP have increased. These changes in economic structure and tax bases call for the adaptation of the tax structure to a service and manufacturing-export oriented economy. A3.40 Joining NAFTA has enhanced the potential effects of Mexico's tax structure on trade and, more importantly, on cross-border investment flows. The obvious significant implication of NAFTA for Mexico is that a traditional constraint for tax policy reform has become more binding. No reform proposals should now be considered without an explicit analysis of how they may affect Mexico's standing in NAFTA, in particular how new measures may affect cross-border trade and investment flows into Mexico from Canada. A3.41 The existing differences in the tax treatment of capital income among the three NAIFTA members translate at times into quite different METRs. This has the potential of distorting cross-border capital flows. The existing differences in taxes may also lead to tax arbitrage as multinational firms take advantage of national tax differentials in their financial planning. But so far there is no evidence that these differences in tax structure are motivated by tax competition or that tax competition has produced revenues losses for Mexico or other NAIFTA members. A3.42 What ought to be done, if any thing about the existing differences in CIT regimes? The first option is to do nothing. These differences may be justified because they reflect the different objectives of the governments in the three countries. Note that this issue is not only about the level of overall tax effort or what share of GDP should be channeled through the public sector, but also about how to raise those funds. After all, tax policy typically pursues quite different objectives from those of trade policy, including maintaining different types and amounts of public goods and services, as well as degrees and patterns of income distribution.31 Maintaining sovereignty over tax policy also allows policy makers to neutralize other sources of economic distortion or encourage activities that are considered important at a national level. A3.43 Doing nothing has the cost of not fully exploiting the potential gains from trade and from an efficient allocation of investment resources. But it is not clear that Mexico is harmed by many of the current differences in tax systems. Because, in general, Mexico's METRs are lower than those in the US and Canada too much capital may be invested in Mexico vis-a-vis the US and Canada. Other existing differences in CIT structure also benefit Mexico. In the case of Mexico's gross asset tax, not only does it facilitate tax enforcement domestically, but US and Canadian parent companies have an incentive to shift income to their Mexican 31. See, for example, Bird 1994. 80 Annex 3 subsidiaries to convert the asset tax into an income tax which then becomes creditable in their home countries. A3.44 A second possibility is to attempt to bring the CIT systems in the three countries closer together. A concrete proposal is for the three countries to adopt a trilateral tax treaty, which would incorporate the formulary taxation of the unitary enterprises operating in more than one member country. 32 A3.45 What takes away from any momentum for moving the tax system of the three countries closer together is that at the present time there is little information available on the welfare costs imposed by the existing tax differences for Mexico, the other two countries, or for the trade block as a whole. The information there is would seem to point in the direction of small additional benefits to be gained form more coordination or uniformity of their tax structures. Under these circumstances, only a very weak case can be made for the three countries to relinquish some control over their tax policies to gain closer coordination in their tax treatment of capital income. This is not to say that NAFTA has already brought an erosion of real government control over certain aspects of taxation, especially for Mexico and Canada. A3.46 Even if the three countries were to move their CIT systems closer together the question is in what direction they should move. Given the very junior status of Mexico in the NAFTA partnership it is quite unlikely that Canada would move closer to Mexico's CIT structure, even if in many respects Mexico's CIT may be a priori more appealing. Of course, the other option would be for Mexico to move closer to the CIT structure in the US and Canada. But, that may suggest that Mexico give up the indexation of the CIT for inflation or the integration of the PIT and the CIT to avoid the double taxation of dividends. That would not seem right either. In short, there are no weighty reasons from a NAFTA perspective for Mexico to undertake fundamental changes in its tax structure. The new wave of tax reform should concentrate on the objectives of raising revenues, simplifying the tax structure, and increasing the efficiency and overall equity of the tax system. 32. See, for example, McDaniel 1994. 81 Annex 3 APPENDIX 1 NAFrA BUSINESS TAX PROVISIONS BY COUNTRY: AN OVERVIEW A3.47 This appendix this appendix provides an overview of business taxation in each of three NAIFTA member countries: Canada, Mexico, and the United States. The business taxation means taxes that may affect business activities, particularly the real capital investment. The major business taxes include capital taxes, and transaction taxes on business inputs. The capital taxes in our context include the corporate income tax, personal income taxes on investment income,33 and the property tax on immovable properties. The description presented in this appendix is based on the publication of International Bureau of Fiscal Documentation, the 1999 Worldwide Corporate Tax Guide published by Ernst & Young, and recent issues of Tax Note International. A3.48 Table A1 summarizes the main features of each country's corporate tax system. I Canada A3.49 Corporations resident in Canada are taxed on their worldwide income from all sources including income from business or property and net taxable capital gains. The capital tax provisions A3.50 The corporate income tax rate. The corporate income tax is levied at both federal and provincial level. The general federal CIT rate is 29% including the 4% surtax; however, manufacturing industries, pay a lower rate of 22%. The provincial CIT is not deductible for federal CIT purposes and the rates range from 8.9% to 17%. Some provinces also impose lower rates on manufacturing sector. The combined CIT rate, based on the industrial structure among provinces, is about 43% for the services sector and 35% for the manufacturing sector. A3.51 The tax depreciation rule. The tax depreciation is based on the declining balance and varies by capital asset classified for the tax depreciation purpose. The average depreciation rate for buildings is 5% for manufacturing and 6% for services; the rate for machinery is 38% and 31%.respectively. 34 A3.52 Capital Taxes. There are two types of capital taxes in Canada. At the federal level, a large corporate tax is imposed at a rate of 0.225% on paid-up capital in excess of $10 million. This tax, however, is creditable against the corporate surtax. At the provincial level, five provinces including British Columbia, Alberta, Manitoba, Ontario and Quebec also impose a tax on capital. The weighted-average of provincial capital tax rates is about 0.36%. 33. As illustrated by the effective tax rate analysis, taxes on any personal investment income could affect the cost of capital investment through financing. 34. These rates are our estimates based on the Canadian capital structure by industry. 82 Annex 3 A3.53 Inventory accounting method. In Canada, only the first-in-first-out (FIFO) method is allowed in inventory accounting for the income tax purpose. A3.54 Loss carry-overs. Business losses may be carried back for three years or forward for 7 years. A3.55 The withholding tax rate on dividends. There is no withholding tax on dividends distributed from the after-tax profits. Dividends paid by a Canadian company to a Canadian resident individual are generally taxable, but the individual also receives a tax credit because the income has already been taxed within the corporation. Dividends paid to a non-resident shareholder (e.g., a foreign multinational firm) are subject to a withholding tax. According to the Canadian bilateral treaties, the withholding tax rate on dividends paid to an US firm is 5% and that to a Mexican firm 10% assuming the recipients hold at least 10% of the voting shares of the payer. A3.56 The property tax. In Canada, the tax base and rate vary widely by locality, and there is no average estimate available. The transaction taxes A3.57 The main transaction tax that affects the capital investment in Canada is the provincial sales taxes applied to some capital goods. According to the Mintz Report, the effective sales tax rate on capital goods is about 1.7% for the manufacturing and 3.4% for the services sector. H Mexico A3.58 Mexico adopts certain rules regarding inflation adjustment: The adjustment factor is the proportional difference in the consumer price index between the starting month and the ending month of a given period. The income tax law recognizes the effects of inflation on the following items and transactions: depreciation of fixed assets, cost on sales of fixed assets, sales of capital stock (shares), monetary gains and losses, and tax loss carried forward. The capital tax provisions A3.59 The corporate income tax rate. The corporate income tax has been increased from 34% to 35% in 1999. There appears to be a tax deferral of 5 percentage points until the dividends are effectively distributed to shareholders. The taxable income for a residential corporation is its worldwide income from all sources, while that for a non-residential corporation is its income derived from its Mexican source. A3.60 Minimum tax on net assets. There is a minimum tax of 1.8% on the net assets of corporations, which provides a credit for the CIT payable. Any minimum tax paid in excess of income tax for any tax year may be carried forward 10 years or back three years to offset CIT liabilities or CIT paid. More specifically, in the case of carrying back the minimum tax credit, a refund of tax paid in the last 10 years (-IBFD) up to that credit (and adjusted for inflation) is permitted. 83 Annex 3 A3.61 The tax depreciation rule. The tax depreciation is based on the straight-line method. Depreciation is computed on original cost of fixed assets, with the amount of depreciation indexed for inflation as measured by price indices. The maximum annual deprecation rates are set by law. Our reading of the Official scheme of depreciation and amortization (IBFD 1999) indicates the following rates for annual depreciation allowance: 5% for buildings used by all sectors and 10 - 25% for machinery and equipment. More specifically, the annual allowance for machinery and equipment is 10% for manufacturing, public utility, trade and other services, 12% for transportation and storage, 20% for communication and 25% for agriculture, forestry and construction. A3.62 Inventory accounting method. For inventory valuation, the basic requirement is the adjustment for inflation, which is, in effect, equivalent to the average cost method. A3.63 Loss carry-overs. Business losses may be carried forward for 10 years. A3.64 The withholding tax rate on dividends. There was no withholding tax on dividends distributed from the after-tax profits until 1999. Under the new tax laws effective January 1999, the dividends paid out of the after-tax profits must first be grossed up by the factor of 1.5385 and then subject to a withholding tax of 5%. As a result, the effective withholding tax rate is 7.7%. The old regime, if the distributing corporation does not have sufficient accumulation in its "net tax profit" account to cover the dividend, then the dividends are taxed at the corporate level at the CIT rate of 34%. In this case, dividends distributed to foreigners subject to the lower of the treaty rate and the CIT rate. In its treaty with the US, the withholding tax is 5% or 10% with the lower rate applicable to the receipt owning at least of 10% of the payer (E&Y). A3.65 The property tax. The property tax is levied at the municipal level. As a result, the tax rate varies by location. In the Federal District, the tax rate ranges from 0.131% to 0.647%. The transaction taxes A3.66 The VAT is levied at a general rate of 15% with a lower rate of 10% in border regions. There is also a real estate acquisition tax, levied at the local or state level, on the market value of the transferred property. The approximate rate is 3.3%. The payroll taxes A3.67 The social security contribution (covering pension, unemployment insurance, health insurance, etc.) is levied on salaries up to a specified amount. A housing fund is also payable by the employer at 5% on salaries with a ceiling. Furthermore, the federal district and states levy a payroll tax on the total remuneration for dependent personal services at a rate up to 2%. The resultant gross rate payable by an employer is above 20% and that by the employee is 4%. 35. The maximum taxable amount is defined by specific times the minimum salaries, which varies from 15 to 25 minimum salaries depending on the category of contribution. The maximum amount will be set at 25 minimum salaries for all categories of contribution in year 2007 as some currently levies with lower taxable base being gradually reduced. 84 Annex 3 A3.68 There is also a mandatory employee profit sharing plan, which accounts for 10% of the taxable profits excluding the inflation effect. However, losses of prior years are not deductible in computing profit to be shared. Furthermore, the portion of profits shared by employees is not deductible for the income tax purpose. However, new enterprises are exempt from profit sharing for the first year of operation and those engaged in manufacturing a new product are exempt for the first 2 years of operation. Tax incentives A3.69 The main features of the Mexican tax incentive regime are its preferential tax treatments towards mostly primary industries, smaller taxpayers and taxpayers outside the three largest metropolitan districts. 6 More specifically, there are four types of tax incentives as described below. A3.70 Cash-flow-based reRime. This regime allows firms engaged in agriculture, livestock, forestry, fishery and land transportation activities to calculate their taxable income on a cash- flow basis, where only resources taken out of the entity are subject to tax. In other words, firms are able to defer their tax liability until recover all their capital expenditure and operating expenses. A3.71 Special rate reRime. Under this regime, a lower CIT. rate of 17% is applied to firms engaged in agriculture, livestock, forestry, fishery, silviculture and publishing. The applicable CIT rate will be higher (i.e., 25.5%) if the taxpayers within these industries except publishing commercialize or industrialize their products. A3.72 Special reRime for small taxpayers. Taxpayers with an annual gross income below 2.2 million pesos (or roughly below $350,000) fall into this regime under which taxpayers subject to simplified tax of 0.25 - 2.5% of gross income. A3.73 An immediate deduction on depreciable assets. Under this regime, qualified taxpayers may choose, instead of taking annual depreciation allowance under ordinary rules, an immediate depreciation deduction for certain assets. This deduction is a percentage of original cost, which equals the present value of the annual depreciation allowances using a real discount rate of 3 percent. For example, the percentage is 74% for buildings, 74-95.7% for machinery and equipment, and 94% for computers and peripherals. A3.74 Qualified taxpayers include those outside the three largest metropolitan districts -- Mexico City, Monterey, and Guadalajara -- and taxpayers regardless of their location with gross income and assets not exceeding 7 and 14 million pesos (roughly $ 1.1 millions and $2.2 millions) may enjoy an immediate deduction for their capital investment. The rate of deduction equals the present value of the annual depreciation allowance using a real discount rate of 3%. 36. There is also a tax-incentive-package related to maquiladoras. It includes a rather generous safe-harbor rule, which set the minimum taxable income as 5% of the total value of assets used in the operation. However, there is presently a clear trend towards treating maquiladoras for tax purposes in the same way as any other Mexican corporation. 85 Annex 3 I The United States A3.75 US firms are subject to federal taxes on their worldwide income, including income of foreign branches (whether or not the profits are repatriated). In general, a US firm is not taxed by the United States on the earnings of a foreign subsidiary until the subsidiary distributes dividends or is sold or liquidated. Numerous exceptions to this deferral concept may apply, resulting in current US taxation of some or all of the foreign subsidiary's earnings. The capital tax provisions A3.76 The corporate income tax rate. A progressive CIT scheme is applied to the taxable income. Firms with taxable income between $335,000 and $1 million are effectively taxed at 34% on all taxable income. Corporations with taxable income of less than $335,000 receive partial benefit from graduated rates of 15% and 25% that apply to the first $75,000 of taxable income. A firm's taxable income exceeding $15 million but not exceeding $18,333,333 is subject to an additional tax of 3%. Firms with taxable income in excess of $18,333,333 are effectively subject to tax at a rate of 35% on all taxable income. These rates apply both to US corporations and to the income of foreign corporations that is effectively connected with an US trade or business. A3.77 In addition, most states and some local governments levy an income tax up to 13%. (An average rate of 6.5% is used for our effective tax simulation). This type of sub-national income tax is deductible for the federal income tax purpose. By using an average state rate of 6.5% and the highest CIT rate at the federal level, the combined CIT rate is about 39%. A3.78 The taxc depreciation rule. Tangible depreciable assets placed in service after 1986 is generally depreciated under a modified accelerated basis (MACRS). Under the MACRS system, assets are grouped into eight different classes and each class is assigned a recovery period and a depreciation method. For example, an asset with a useful life of 10 to 17 years is classified as a seven-year property, A seven-year property is recovered using the 200% declining-balance method with a half-year rule for the first year and a switch to the straight- line method in the sixth and seventh year, using the depreciation rate of the fifth year; and then a residual is written off in the eighth year. Based on the MACRS and the capital structure by industry, we estimated the equivalent tax depreciation rates based on the declining balance, which varies by industry and is above 5% for buildings and well above 30% for machinery and equipment. A3.79 Inventora accounting method. Both FIFO and LIFO are allowed as inventory accounting method for tax purposes. However, the method chosen must be applied consistently. In practice, about 75% firns in the US using the LIFO accounting method. A3.80 Loss carrv-overs. Business losses, or net operating loss, may be carried back 3 years and forward 15 years, or until the loss is used up. 86 Annex 3 A3.81 The withholding tax rate on dividends. Dividends paid by an US company to a non- resident shareholder (e.g., a foreign multinational firm) are subject to a withholding tax. According to the treaties, the withholding tax rate on dividends paid to both Canadian and Mexican firms is 5% assuming that, among other conditions, the recipients is a corporation owning a specified percentage of the voting power of the distributing corporation. A3.82 The provertv tax. The property tax is levied at the municipal level. As a result, the tax rate varies by location, and no sensible estimate is available for our effective tax rate calculation. The transaction taxes A3.83 The main transaction tax that affects the capital investment in the US is the state sales taxes applied to some capital goods. According to the Mintz Report, the effective sales tax rate on capital goods is about 4.2% across the sectors. 87 Annex 3 Table Al Business Tax Provisions Applicable to Manufacturing and Service Industries Canada Mexico The United States' 1. The CaDital Taxes Corporate income 36/43 incL the provincial Crr 35 39.5 incl. the state Assets-based tax 0.35% 1.8 None Thin capitalization rule Yes None Yes Tax depreciation rate37 Buildings 5.0 DB 5.0 SL Equivalent to 5.0+ Machinery 30.0+ DB 10 and up SL Equivalent to 31.0+ Inventory accounting FI FO Equivalent to LIFO optional Loss Carry-over38 3-yrs (B) and 7-yrs 10-year (F) 3-yrs (B) and 15-yrs WH tax on dividends To Canada 10.0 5.0 To Mexico 10.0 5.0 To theU. S. 5.0 5.0 Urban Propertv taxes Vary by location FD. 0.131-0.647 Vary by location Property transfer tax 3.3 Sector-oriented Yes Yes None 2. The Indirect Tax on Capital goods Effective sales tax Around 3.0 none 4.2 Import duty 0 11 (average) 0 37. As the classification of depreciable assets varies by country, please refer to the text for details. Also note that DB = declining-balance method, and SL = straight-line method. 38. Following the number of years for loss carry over, the letters in brackets indicate the following: F = forward, B = backward, and R = certain restriction in the value of loss to be written off. Please refer to the text for details. 88 Annex 3 APPENDIX 2 Impact of Non-tax Parameters on the Estimate of Effective Tax Rates Expected Inflation Rate The expected inflation rate affects the effective tax rate on capital through its impact on the nominal interest rate. For a given real interest rate, the higher the inflation rate, the higher the nominal interest rate will be. When there is no regulation for adjusting the inflation impact, the nominal interest rate interacts with taxes mainly through the following three channels. Firstly, interest cost is deductible for income tax purposes at the nominal rate. As a result, the higher the nominal interest rate in relation to a fixed real interest rate, the lower the real after-tax financing cost, and hence the lower the METR. This effect is particularly favorable for leveraged land financing. Secondly, the accumulated present value of a given annual tax depreciation allowance decreases as the nominal interest rate rises. Since higher inflation lowers the present value of tax depreciation allowance, it increases METR on depreciable assets. And finally, if the first-in-first-out method is used for the inventory accounting, it may results in inflated taxable income and, hence, a higher METR on inventory when prices rise. Since inflation thus affects METR on different assets in different directions, its net impact on capital will depend upon the capital structure related to a given industry. (See the end section of this appendix for further explanation of the capital structure by industry). Expected Real Interest Rate The impact of the real interest rate on the effective tax rate is in. part similar to the impact of inflation. For example, as the real rate rises, so will the nominal rate, thus increasing the effective tax rate on depreciable assets. For a given debt-asset ratio, however, unless inflation is high, there is unlikely to be much of a distortion in effective tax rate arising from the deductibility of interest. We use the US real interest rate for our study assuming a full mobility of investment fund within NAFIA and the American's dominant role in the North America financing market. As shown in Table A2, the real interest rate in the US is 6.1% corresponding to the nominal interest rate of 8.4% and the inflation rate of 2.3%. Debt-asset Ratio The ratio of debt to assets is sometimes referred to as the financing structure. As already noted, the impact of this ratio on the effective tax rate is related to the expected inflation rate and (real) interest rate. For a given inflation rate and real interest rate, the higher the debt-asset ratio, the more the potential benefit from the tax deductibility for debt financing cost, or interest expenses. A higher debt-asset ratio may thus reduce effective tax rate through lowering the real after-tax cost of financing. For simplicity, we apply a debt to assets ratio of 40% across sector and across border in our study. 89 Annex 3 Economic Depreciation The economic depreciation rate interacts with the tax depreciation allowance to affect the effective tax rate. Suppose, for example, under our assumption of fully mobile capital and technology that a given type of machinery is depreciated at the same economic rate everywhere around the world. Countries with faster tax depreciation allowances for this type of machinery will then encourage this type of capital investment through a lower effective tax rate. Capital Structure A real capital investment generally involves two categories of capital: depreciable and non-depreciable assets. These two categories can be further divided into four types: buildings and machinery (both depreciable) and inventory and land (non-depreciable). Capital investments in different industries are as a rule structured differently. Moreover, under the same statutory tax rate(s), different types of assets may incur different effective tax rate due to the various interactions between tax provision and non-tax parameters discussed above. In the absence of other information, we use the same capital structure by industry, based on the Canadian data, to aggregate these differentiated effective tax rate on various type of capital for a given industry across countries. 90 Annex 3 Table A2. Non-Tax Parameters (in percent) Canada Mexico U. S. Expected inflation rate 1.7 21.7 2.3 Expected real interest rate39 6.1 6.1 6.1 Debt to assets ratio Debt raised abroad to home capital 40.0 40.0 40.0 Debt to assets ratio in home country 40.0 40.0 40.0 Economic depreciation rate Manufacturing Buildings 3.8 3.8 3.8 Machinery 16.4 16.4 16.4 Services Buildings 3.5 3.5 3.5 Machinery 24.4 24.4 24.4 Capital structure by asset type Manufacturing Buildings 24.0 24.0 24.0 Machinery 38.1 38.1 38.1 Inventory 35.9 35.9 35.9 Land 2.0 2.0 2.0 Services Buildings 60.6 60.6 60.6 Machinery 11.7 11.7 11.7 Inventory 9.5 9.5 9.5 Land 18.2 18.2 18.2 39. The expected real interest rate of 6.1 % is derived from the US inflation rate and bank lending rate based on the IMF, International Financial Statistics, March 2000. 91 Annex 3 References Bird, Richard (1994), "Colloquium on NAFTA and Tradition: Commentary: A View for the North," Tax Law Review, 49 (Summer). Blomstrom, Magnus and Ari Kokko (1997)"Regional Integration and Foreign Direct Investment: A conceptual Framework and Three Cases," Policy Research Working Paper 1750, The World Bank, Washington D.C. (April) Boadway, Robin and Neil Bruce (1992), "Pressures for the Harmonization of Income Taxation between Canada and the United States," in John B. Shoven and John Whalley, (editors) (1992), Canada-US Tax Comparisons. Chicago: University of Chicago Press. Brown, Drusilla K., Alan V. Deardoff, and Robert M. Stern (1992), "A North American Free Trade Agreement: Analytical Issues and a Computational Assessment," The World Economy, January, pp 15 29. Bulmer-Thomas, Victor, Nikki Craske and Monica Serrano (1994) Mexico and the North American Free Trade Agreement: Who Will Benefit?" St. Martin's Press. New York Cockfiled, Arthur J.(1998), International Tax Policy Under NAFTA: The Impact of National Tax Differences on Capital Flows Under Regional Trade and Investment Integration, a dissertation submitted to the Department of Law, Stanford University. Davila Flores, Alejandro (2000), " Impactos Econ6micos del TLCAN en la Frontera Norte de Mexico," in Leycegui, Beatriz and Rafael Fernandez de Castro (editors), (2000), Socios Naturales? Cinco Afios del Tratado de Libre Comrcio de America del Norte . Instituto Tecnol6gico Aut6nomo de Mexico, Mexico D.F. Fernandez, Raquel (1997), "Returns to Regionalism: An Evaluation of Non-Traditional Gains form RTAS," NBER Working Paper No. 5970 (March). Fernandez de Castro, Rafael (2000), " Las Instituciones del TLCAN: Una Evaluaci6n a los Cinco Anios," in Leycegui, Beatriz and Rafael Fernandez de Castro (editors), (2000), Socios Naturales? Cinco Afios del Tratado de Libre Comrcio de America del Norte Instituto Tecnol6gico Aut6nomo de Mexico, Mexico D.F. Gordon, Roger and Eduardo Ley (1994), "Implications of Existing Tax Policy for Cross Border Activity Between the United States and Mexico after NAFTA" National Tax Journal, 47, pp 35-41 Gravelle, Jane G. (1986) "International Tax Competition: Does it Make a Difference for Tax Policy? 39 National Tax Journal, pp 375-384 Hinojosa Ojeda, Raul, DaVid Runsten, Fernado De Paolis, and Nabil Kamel (2000), 'The US Employment Impacts of North American Integration After NAFAT: A Partial Equilibrium 92 Annex 3 Approach," North American Integration and Development Center , School of Public Policy and Social research, UCLA (January) Kats, Isaac (2000),"EI Impacto Regional del. TLCAN: Un Analisis de la Industria Manufacturera," in Leycegui, Beatriz and Rafael Fernandez de Castro (editors), (2000), Socios Naturales? Cinco Afios del Tratado de Libre Comrcio de America del Norte Instituto Tecnol6gico Aut6nomo de Mexico, Mexico D.F. Krugman, Paul (1993), "It's foreign policy, stupid," Foreign Affairs, Nov/Dec Krueger, Anne 0. (1999), "Trade Creation and Trade Diversion Under NAFTA," NBER Working Paper No. 7429 (December). Leycegui, Beatriz and Rafael Fernandez de Castro (editors), (2000), Socios Naturales? Cinco Afios del Tratado de Libre Comrcio de America del Norte . Instituto Tecnol6gico Aut6nomo de Mexico, Mexico D.F. McDaniel, Paul R. (1994), " Colloquium on NAFTA and Tradition: Formualry Taxation in North America Free Trade Zone," Tax Law Review, 49 (Summer). McLure, Charles (1986), "Tax Competition: Is What 's Good for the Private Goose Also Good for the Public Gander?" 39 National Tax Journal, pp. 341-48. Perez de Acha, Luis Mauel (1994), "An Examination of the Tax Treatment of Dividends in Mexico," Tax Notes International, July 11. Rubin, Seymour J. and Dean C. Alexander (1995), NAFTA and Investment Kluwer Law International, The Hague, The Netherlands. Shoven, John and John Whalley, (editors) (1992), Canada-US Tax Comparisons . Chicago: University of Chicago Press. Studer, Isabel (2000), " El Sector Automotor," in Leycegui, Beatriz and Rafael Fernandez de Castro (editors), (2000), Socios Naturales? Cinco Afios del Tratado de Libre Comrcio de America del Norte . Instituto Tecnol6gico Aut6nomo de Mexico, Mexico D.F. Tornell, Aaron and Gerardo Esquivel (1995), "The Political Economy of Mexico's Entry to NAFTA," National Bureau of Economic Research, Working Paper No. 5322, Cambridge, MA Trigueros, Ignacio (2000), " El TLCAN y la Situaci6n Macroecon6mica de M6xico," in Leycegui, Beatriz and Rafael Fernandez de Castro (editors), (2000), Socios Naturales? Cinco Afios del Tratado de Libre Comrcio de America del Norte . Instituto Tecnol6gico Aut6nomo de Mexico, Mexico D.F. 93 Annex 3 Von Wobeser, Claus (2000), " EL R6gimen Legal de la Inversi6n Extranjera en el TLCANy sus Efectos en los Flujos de Capital hacia Mexico," in Leycegui, Beatriz and Rafael Fernandez de Castro (editors), (2000), Socios Naturales? Cinco Afios del Tratado de Libre Comrcio de America del Norte . Instituto Tecnol6gico Aut6nomo de Mexico, Mexico D.F. Weiner, Joan (2000), "Institute for Fiscal Studies Issues Report on EU Corporate Tax Harmonization" Tax Notes International, 22 May pp 2267-73. World Bank (2000), "Export Dynamics and Productivity: Analysis of Mexican Manufacturing in the 1990s," Draft Report, Mexico Country Department, Latin America and the Caribbean region. 94 Annex 3 Table A3.1 Mexico's Trade (Millions of USD) 1991 1992 1993 1994 1995 1996 1997 1998 Total Trade 112,061.90 129,115.20 138,589.40 163,267.81 211,333.00 219,675.00 218,760.00 253,624.00 Exports of goods and services 51,459.50 55,406.00 61,391.00 71,396.40 110,505.00 113,568.00 109,285.00 122,956.00 Imports of goods and services 60,602.40 73,709.20 77,198.40 91,871.40 100,828.00 106,107.00 109,475.00 130,668.00 Trade Balance -9,142.90 -18,303.20 -15,807.40 -20,475.00 9,677.00 7,461.00 -190.00 -7,712.01 (Annual IG growth) 1991 1992 1993 1994 1995 1996 1997 1998 Exports of goods and services 5.07% 4.98% 8.09% 17.80% 30.19% 18.23% 10.81% 9.72% Imports of goods and services 15.18% 19.62% 1.86% 21.25% -15.04% 22.88% 22.80% 14.20% (% of GDP) 1991 1992 1993 1994 1995 1996 1997 1998 Total Trade 35.64% 35.51% 34.35% 38.48% 58.17% 62.26% 60.79% 64.45% Exports of goods and services 16.36% 15.24% 15.22% 16.83% 30.42% 32.18% 30.37% 31.25% Imports of goods and services 19.27% 20.27% 19.13% 21.65% 27.75% 30.07% 30.42% 33.21% Source: World Bank LDB. Table A3.2 Mexico's EXPORTS (Millions of USD) Total Oil Non-Oil Oil Non-Oil 1991 42,687.7 8,166.4 34,521.0 19.13% 80.87% 1992 46,195.5 8,306.6 37,889.0 17.98% 82.02% 1993 51,886.0 7,418.4 44,467.4 14.30% 85.70% 1994 60,882.2 7,445.1 53,437.3 12.23% 87.77% 1995 79,541.6 8,422.4 71,119.0 10.59% 89.41% 1996 95,999.7 11,653.7 84,346.1 12.14% 87.86% 1997 110,431.3 11,323.0 99,108.2 10.25% 89.75% 1998 117,459.4 7,134.3 110,325.2 6.07% 93.93% 1999 136,703.2 9,920.2 126,783.0 7.26% 92.74% Source: INEGI. 95 Annex 3 Table A3.3. Mexico's Merchandise Trade by Type of Industry. Exports (Millions of USD) 1991 1992 103 1994 1905 1996 1997 190 TOTAL EXPORTS 42,688 46,196 51,886 60,882 79,542 96,000 110,431 117,460 Maquiladoras 15,833 18,680 21,853 26,269 31,103 36,920 45,166 53,083 Non-maquiladoras 26,855 27,516 30,033 34,613 48,438 59,079 65,266 64,376 Agriculture and forestry 2,373 2,112 2,505 2,678 4,016 3,592 3,828 3,797 -Agrculture 1,877 1,679 1,961 2,221 3,324 3,197 3,408 3,436 - Livestock 414 373 488 395 579 188 247 254 -Fisheries 82 60 55 62 114 207 173 107 Manufacturing Industries 32,307 36,168 42,500 51,075 67,383 81,014 95,665 106,550 - Food, Beverages and Tobacco 1,421 1,365 1,590 1,896 2,529 2,930 3,325 3,508 - Textile, Apparel and Leather Ind. 2,014 2,317 2,770 3,256 4,899 6,339 8,815 9,844 - Lumber and derivatives 443 499 574 586 619 861 1,047 1,057 - Paper, Printing and Publishing 622 685 662 562 872 895 1,063 1,164 - Oil derivatives 643 624 719 544 653 664 683 561 - Petrochemicals 259 263 214 263 340 247 278 174 - Chemicals 2,120 2,298 2,344 2,756 3,972 4,011 4,403 4,610 -Plastic and Rubber Products 697 794 1,005 1,064 1,218 1,416 1,707 1,801 - Other non-Metallic Mineral Products 836 919 1,125 1,215 1,405 1,718 2,025 2,290 -Iron and Steel 1,261 1,145 1,399 1,S35 3,088 3,085 3,655 3,282 - Mining-Metallurgy 827 929 1,024 1,085 1,801 1,705 1,703 1,657 - Metallic Prods. Mach. And Equipment 20,463 23,711 28,352 35,324 44,681 55,736 65,166 74,783 - Other Manufacturlng Industries 701 649 722 989 1,306 1,406 1,696 1,821 Extractive Industries 7,812 7,776 6,764 6,994 7,975 11,192 10,840 6,865 -Oil and natural gas 7,265 7,419 6,485 6,638 7,430 10,743 10,362 6,399 Extracton of Metallic Minerals 251 158 135 184 311 249 278 280 - Extraction of other minerals 294 198 144 173 234 200 200 186 - Other extractive industries 1 0 0 0 0 0 0 0 Non classifled products 196 139 118 134 168 202 198 247 Source: Banco de M4xdco. Figures may not add-up due to rounrDng off 96 Annex 3 Table A3.4. Mexico's Merchandise Trade by Type of Industry. Exports (Annual Growth) .1-0 l92-0 100-4 1994-3 195.996 196-97 19967-0 1994-98 TOTAL EXPORTS 8.22% 12.32% 17.34% 30.65% 20.69% 15.03% 6.37% 92.93% Maquiladoras 17.98% 16.99% 20.21% 18.40%/6 18.70% 22.33% 17.53% 102.07% Non-maqulladoras 2.46% 9.15% 15.25% 39.94% 21.97% 10.47% -1.36% 85.99% Agriculture and forestry -11.00% 18.61% 6.91% 49.96% -10.66% 6.57% -0.81% 41.78% - Agriculture -10.55% 16.800/a 13.26%/ 49.66%A .3.82% 6.60%/o 0.82°A 54.71% - Livestock -9.90% 30.839% -19.06%A 46.58%A -67.530/a 31.38%/o 2.83% -35.70% - Fisheries -26.83%c -8.33%c 12.73% 83.87% 81.58%e -16.43% -38.15% 72.58% Manufacturing Industries 11.95% 17.51% 20.18% 31.93% 20.23% 17.96% 11.49% 108.61% - Food, Beverages and Tobacco -3.94%h 16.48%/c 19.25% 33.390% 15.S6% 13.48% 5.50% 85.02% - Textile, Apparel and Leather Ind. 15.04%/ 19.55% 17.55% 50.460 29.390/6 39.06%o 11.67% 202.33% - Lumber and derivatives 12.64%/o 15.03%e 2.09%' 5.63Y 39.10Y 21.60% 0.96Y 80.38% - Paper, Ptlntlng and Publishing 5.31% 1.07% -15.11% 55.16% 2.64% 18.77%o 9.50% 107.12% - Oil derivatives -2.95% 15.22%c -24.34% 20.04% 1.68% 2.86% -17.86%c 3.13% - Petochemicals 1.54%/o -18.63%/o 22.900/c 29.28% -27.35% 12.55% -37.41% -33.84% - Chemicals 8.400/c 2.000/c 17.580% 44.12°% 0.98% 9.77% 4.70% 67.27% - Plastic and Rubber Products 13.92Ye 26.57% 5.87% 14.47/o 16.26% 20.55% 5.51% 69.27%Y - Other non-Metallic Mineral Products 9.93% 22.42Yc 8.00°/c 15.64% 22.280%o 17.87/o 13.09°Y 88.48%Y - lrn and Steel -9.200/% 22.18°A 9.72%o 101.17%/ -0.10% 18.48% -10.21% 113.81% - Mining-Metallurgy 12.33% 10.23%h 5.96% 65.99% -5.33%o -0.12%o -2.70Ye 52.72% - Metallic Prods. Mach. And Equipment 15.870/ 19.57%/o 24.59Y 26.49%/ 24.74% 16.92% 14.76% 111.71% - Other Manufacturing Industries -7.42°h 11.25% 36.98°h 32.05% 7.66%o 20.63% 7.37/o 84.13% Extractive Industries -0.46% -13.01% 3.40% 14.03% 40.34% -3.15% -36.67% -1.84% -Oil and natural gas 2.12% -12.59%/o 2.36%/ 11.93'# 44.590% -3.55% -38.25% -3.60°h - Extraction of Metallic Minerals -37.05% -14.56% 36.30% 69.02%h -19.94% 11.65% 0.72Yo 52.17% - Extraction of other minerals -32.65% -27.27%Y 20.14% 35.26%h -14.530/ 0.00% -7.00% 7.51% - Other extractive industries -100.00°h 0.00° 0.00%o 0.00%Y 0.00% 0.00Y 0.00% 0.00°h Non classified products -29.08% -15.11% 13.56% 25.37% 20.24% -1.98% 24.75% 84.33% Source, Banco de Mexico. Figures may not adk-up due to rounding off 97 Annex 3 Table 3.5. Mexico's Merchandise Trade by Type of Industry. Imports (Millions of USD) 1991 1992 1993 1994 1995 1996 1W7 1998 TOTAL IMPORTS 49,967 62,129 65,367 79,346 72,453 89,469 109,808 125,373 Maquiladoras 11,782 13,937 16,443 20,466 26,179 30,505 36,332 42,557 Non-maquiladoras 38,184 48,192 48,924 58,880 46,274 58,964 73,476 82,816 Agriculture and forestry 2,130 2,858 2,633 3,371 2,644 4,671 4,173 4,773 - Agriculture 1,687 2,402 2,324 2,993 2,479 4,346 3,660 4,281 - Livestock 434 443 293 352 140 308 486 455 - Fisheries 9 13 16 26 17 17 27 38 Manufacturing Industries 46,967 58,237 61,568 74,426 67,500 81,138 101,587 116,431 - Food, Beverages and Tobacco 2,635 3,336 3,356 3,989 2,616 3,115 3,587 3,931 - Textile, Apparel and Leather Ind. 2,237 3,023 3,525 4,167 3,618 4,603 6,146 7,441 -Lumber and derivatives 428 551 571 695 350 390 461 544 - Paper, Printing and Publishing 1,812 2,189 2,366 3,039 2,899 2,887 3,280 3,536 - Oil derivatives 1,335 1,458 1,368 1,275 1,243 1,626 2,515 2,319 - Petrochemicals 479 513 600 759 920 942 1,217 1,188 - Chemicals 3,695 4,413 4,855 5,818 5,521 6,884 8,226 9,157 - Plastic and Rubber Products 2,534 3,153 3,404 3,972 4,157 5,275 6,470 7,070 - Other non-Metallic Mineral Products 568 717 820 1,010 910 1,264 1,462 1,538 Iron and Steel 2,994 3,461 3,312 3,931 3,693 4,542 5,469 6,235 - Mining-Metallurgy 792 1,048 968 1,195 1,203 1,407 1,813 2,282 - Metallic Prods. Mach. And Equipment 26,903 33,731 35,673 43,490 39,709 47,462 59,792 69,689 - Other Manufacturing Industfies 555 644 750 1,086 662 741 1,149 1,501 Extractive IndustrIes 386 520 390 438 600 649 854 916 -Oil and natural gas 31 180 90 73 106 59 106 120 - Extraction of Metallic Minerals 73 104 76 84 122 127 204 245 - Extraction of other minerals 251 181 161 214 260 322 350 359 - Other extractlve industfies 31 55 62 67 112 141 195 190 Non classified products 483 514 776 1,112 1,709 3,011 3,194 3,253 Source: Banco de MWxico Figures may not add-up due to rounCing off 98 Aninex 3 Table A3.6 Export, Assembly Plants (Maquiladoras) Production Account * [ Employees Average Annual Payment ** Production Intermediate Cons. Gross Value Added Total Factory Other Total Factory Other 1990 50,163,134 40,276,610 9,886,524 451,169 418,035 33,134 11,432 9,685 33,472 1991 52,804,962 43,489,508 9,315,454 434,109 401,086 33,023 13,807 11,730 39,033 1992 60,732,377 49,718,230 11,014,147 503,689 465,112 38,577 16,168 13,643 46,618 1993 68,158,225 56,628,991 11,529,234 526,351 487,298 39,053 17,715 14,886 53,016 1994 87,375,493 74,607,081 12,768,412 562,334 522,345 39,989 19,661 16,706 58,256 1995 107,344,659 93,171;078 14,173,581 621,930 578,286 43,644 25,032 20,809 80,990 1996 132,810,723 115,845,784 16,964,939 748,262 694,296 53,966 31,952 26,388 103,538 1997 157,072,932 137,704,846 19,368,086 899,167 834,968 64,199 38,820 32,412 122,172 * Thousand pesos at 1993 constant prces Current Pesos per worker *** 1993 = 100 Source: INEGI Table A3.7- Exports Assembly Plants (Maguiladoras) Production Account * Employees [ Average Annual Payment ** Production Intermediate Cons. Gross Value Added Total Factory Other Total Factory Other 90-91 5.27% 7.98% -5.78% -3.78% -4.05% -0.34% 20.78% 21.12% 16.61% 91-92 15.01% 14.32% 18.24% 16.03% 15.96%. 16.82% 17.10% 16.31% 19.43% 92-93 12.23% 13.90% 4.68% 4.50% 4.77% 1.23% 9.57% 9.11% 13.72% 93-94 28.20% 31.75% 10.75% 6.84% 7.19% 2.40% 10.99% 12.23% 9.88% 94-95 22.85% 24.88% 11.01% 10.60% 10.71% 9.14% 2732% 24.56% 39.02% 95-96 23.72% 24.34% 19.69% 20.31% 20.06% 23.65% 27.64% 26.81% 2Z84% 96-97 18.27% 18.87%o 14.1 7% 20.17% 20.26% 18.96% 21.49% 22.83% 18.00% 94-97 79.77% 84.57% 51.69% 59.90% 59.85% 60.54% 97.45% 94.01% 109.72% * Thousand pesos at 1993 constant prices ** Current Pesos per worker *** 1993 = 100 Source: INEGI. 99 Annex 3 Table A3.8. Mexico's Trading Partners. Exports (Millions of USD) I t 990[ 1991 1992 1993 1994 1995 1996 1997 1998 1999 TOTAL 35,171 40,711 42,688 46,196 51,886 60,882 79,542 96,000 110,431 117,460 136,703 ArIca 30,209 34,683 37,171 41,160 47,667 S6,209 73,295 89,067 103,281 110,665 NA. North AmerIca 28,398 32,748 34,956 38,420 44,609 53,t77 68,260 82,746 96,458 104,612 122,920 - United States 28,121 32.290 33,930 37,420 43,068 51,680 66,273 80,574 94,302 103,093 120,609 -Canada 277 458 1,025 1,00 1,541 1,497 1,987 2,172 2,157 1,519 2,311 Centrai America 50 483 617 612 645 684 95S 1,160 1,494 1,673 1,597 -Costa Rica 82 70 80 107 99 95 142 188 221 282 250 -El Salvador 91 111 116 121 112 127 148 i58 214 218 244 -Guatemala 106 114 225 153 204 218 310 360 498 590 544 - Nlcaragua N.A N.A 18 18 21 21 31 53 64 57 65 -Panamna 100 78 99 109 145 124 224 281 334 351 303 -Othter 181 90 79 104 64 99 96 140 163 174 191 South America 736 908 991 1,370 1,598 1,631 2,904 3,499 3,813 3,024 2,214 *Argentina 113 120 186 180 278 248 313 520 498 384 256 -BolMa 4 4 13 9 17 13 24 30 32 35 32 - Brazil 194 168 187 408 291 376 800 878 703 536 400 -Cofombia 110 110 156 219 236 306 453 438 513 449 368 - Chile 83 96 127 152 194 204 490 689 842 625 366 -Peru 56 66 78 63 94 110 179 211 238 196 178 -Venezuela 62 137 127 199 227 174 380 424 675 54E 436 -Other 114 206 118 140 261 200 265 309 312 253 178 Antilles StS 56! 607 758 81t 717 1,180 1,642 1,516 1,356 NA. Europe 2,81 3,772 3,515 3,5156 2,819 2,989 4,005 3,995 4,462 4,305 NLA. Gennany 361 453 530 491 427 395 515 641 719 1,152 2,073 Austria 36 21 25 70 40 10 13 10 16 11 11 BegIpum-Luwenbourg 137 219 321 283 282 271 487 409 373 230 247 Spain 1,134 1,457 1,150 1,235 874 864 797 907 939 714 944 Franoe 481 552 690 567 429 518 483 426 430 401 289 HoUand 152 336 183 163 123 174 177 192 262 339 487 136 211 172 146 76 86 197 134 273 181 171 United Kingdom 182 . 187 219 233 215 267 481 532 664 639 747 Sweden 15 13 t 22 26 17 24 30 20 53 46 24 Swltzertand 69 206 121 130 141 158 608 360 344 258 445 Former USSR 51 24 17 7 12 5 17 152 14 6 N.A Other s5 93 156 205 163 217 200 212 376 329 N.A. Asia 1,982 2,128 1,856 1,381 1,307 1,648 2,078 2,757 2.420 2,221 NA. Korea 71 113 63 41 26 41 91 198 66 73 154 Taiwan 90 69 76 43 21 23 44 42 43 50 91 Hong Kong 66 43 87 62 62 174 504 434 283 217 178 Israel 196 215 164 187 103 3 11 10 30 18 38 Japan 1,394 1,506 1,241 793 700 1,001 979 1,393 1,156 851 777 Singapore 11 33 37 104 131 67 173 235 387 449 480 China 0 9 63 20 45 42 37 38 46 106 126 Other 155 139 126 131 219 197 239 407 407 456 N.A Africa 73 61 70 42 14 16 47 81 120 94 N.A. Oceanla 53 5 76 57 56 69 7b 75 8s 123 NA Australia 38 37 51 49 48 54 63 58 76 109 N.A Other 16 0 25 a 8 15 12 17 12 14 N.A. Rest ot thseWo,td 38 10 0 0 22 52 42 25 60 52 NA. *Source: State of the Nation Report from 1989-98, Seoofl for 1999, both with Banco de M6xico data. *Exports Includes transportation and Insuranoe expenses Figures may not add-up due to rounding off 100 Annex 3 Table A3.9. Mexico's Trading Partners. Imports (Millions of USD) 1989 I 1990 1 1991 1 1992 1 199 1994 1 1995 i 1998 I 1997 I 1998 J 1999 TOTAL 34,766 41,5931 49,967 62,1291 65,367 79,346 72,453 89,469 109,808 125,373 142,064 America 28,859 32,887 39,405 47,683 50,176 59,391 57,082 71,481 86,770 98,626 N.A. North Anerica 27,36t 31,268 37,484 45,268 47,630 56,382 55,276 69,280 83,969 95,549 108,305 - United States 26,948 30,810 36,814 44,216 46,467 54,762 53,902 67,536 82,001 93,258 105.357 - Canada 421 458 670 1,052 1,163 1,621 1.374 1.744 1,968 2.290 2,949 Centrel America 181 189 246, 192 18o 175 97 179 221 238 342 -Costa Rira 5 38 21 15 22 28 1t 58 77 87 191 -El Salvador 4 3 19 12 14 19 8 19 24 25 18 -Guatemala 42 41 87 77 61 82 51 77 80 81 83 -Nicaragua NA. N.A. 14 18 1 1 1 1 8 12 1 1 14 15 -Panama 122 83 93 58 61 24 9 7 19 16 26 -Other 15 24 12 12 11 11 5 6 10 14 9 South America 711 1,283 1,538 2,038 2,158 2,588 1,416 1,734 2,273 2,561 2,835 -Argentina 137 401 385 241 251 333 191 300 236 264 212 -Bolivia 5 5 10 17 16 19 ! a 10 7 8 -Brazil 361 482 803 1,109 1,193 1,226 565 690 86. 1,038 1.129 -Colombia 22 34 50 72 83 121 97 97 124 151 220 -Chile 46 61 50 98 130 230 154 171 372 552 684 -Peru 26 76 102 190 170 210 99 117 142 143 180 -Venezuela 57 171 140 207 227 297 214 234 421 303 297 -Other 57 52 1i 106 88 152 91 117 98 103 105 Antilles 90 14e 137 185 208 245 293 289 307 279 N.A. Europe 4,080 5,72S 6,746 8,290 8,358 9,741 7,237 8,335 10,732 12,589 N.A. Gemiany 1,368 1,840 2,328 2,477 2,832 3,101 2,687 3,174 3,902 4,543 5,032 Austria 25 45 71 113 103 121 8e 113 139 192 170 Belgium-Luxembourg 157 246 328 308 269 337 210 239 327 355 305 Spain 329 520 573 875 1,152 1,338 694 629 978 1,257 1,321 Frante 564 712 967 1,305 1,077 1.527 979 1,019 1,182 1,430 1,394 Holland 130 225 215 240 241 240 218 225 262 328 326 itaty 365 455 623 984 818 1,021 771 999 1,326 1.581 1,649 UnitedKi tdom 327 491 499 619 590 707 532 679 - 915 1,056 1,133 Sweden 222 316 356 333 261 277 201 229 354 339 700 Switzeriand 314 333 379 497 497 490 389 457 559 589 720 Fomter USSR 7 17 16 49 75 141 64 59 180 246 N.A Other 271 522 391 492 443 442 404 513 607 672 N.A Asia 2,097 2,618 3,584 5,798 6,419 9,645 7,778 9,061 11,526 13,123 N A Korea 247 265 434 617 662 734 974 1.178 1,831 1,951 2,964 Taiwan 195 312 429 543 658 1,029 716 891 1,137 1.527 1,557 Hong Kong 184 229 309 403 62 250 159 129 189 218 253 Israel 10 17 24 43 45 85 47 79 112 137 173 Japan 1,309 1.470 1,596 3,041 3,369 4.780 3,952 4.132 4.334 4,537 5.083 Singapore 49 46 86 104 158 213 289 383 426 493 540 China 0 30 142 425 353 428 521 760 1.247 1,617 1,921 Other 104 247 564 622 1,112 2.126 1,117 1,509 2,250 2,645 N.A Africa 69 97 80 98 131 149 129 221 271 368 NA. Oceanta 11i 258 151 285 268 317 178 261 318 401 N.A. Australia 38 65 80 105 113 167 99 128 -166 244 N.A. Other 81 191 72 153 155 150 79 133 151 1S6 N.A Rest of the World 42 13 1 2 14 103 109 191 267 NA. *Sourmc State of the Nation Report from 1989-98, Secotil tor 1999, both with Banco de Mexdco data. *Imports are in custom prtces I I I I I I I I Figures may not add-up due to rounding oft I I I I I 101 Annex 3 Table A3.10 Foreign Direct Investment By Country of origin (Millions of dollars) Year Total USA U. Kingdom Germany Japan Switzerland France Spain Sweden Canada 1980 1,622.6 1,078.6 48.6 170.8 123.1 111.4 19.5 80.0 10.9 17.5 1981 1,701.1 1,072.1 40.9 146.3 212.1 74.9 10.3 101.8 15.3 5.2 1982 626.5 426.1 7.4 39.9 65.4 23.1 6.8 40.4 -2.0 8.1 1983 683.7 266.6 49.2 110.0 3.8 16.2 110.0 12.7 29.1 22.1 1984 1,429.8 912.0 44.3 152.5 35.6 59.8 8.7 11.7 61.1 32.5 1985 1,729.0 1,326.8 56.3 55.5 79.3 141.2 10.7 14.0 5.5 34.9 1986 2,424.2 1,206.4 104.3 218.5 142.2 34.1 316.9 93.7 24.6 40.6 1987 3,877.2 2,669.6 430.9 46.9 132.8 95.2 31.2 125.8 36.7 19.3 1988 3,157.1 1,241.6 767.6 136.7 148.8 86.3 152.4 34.1 32.5 33.9 1989 2,499.7 1,813.8 44.7 84.7 15.7 194.4 16.5 44.0 6.9 37.5 1990 3,722.4 2,308.0 114.4 288.2 120.8 148.0 181.0 10.4 13.3 56.0 1991 3,565.0 2,386.5 74.2 84.7 73.5 68.5 500.5 43.8 13.9 74.2 1992 3,599.6 1,651.7 426.8 84.9 86.9 315.3 69.0 37.2 2.0 88.5 1993 4,900.7 3,503.6 189.2 111.4 73.6 101.7 76.9 63.5 2.4 74.2 1994 10,493.1 4,825.1 593.4 305.0 630.9 53.9 90.5 145.1 9.3 740.4 1995 8,077.1 5,265.4 213.5 548.5 155.7 200.2 119.5 41.6 61.1 168.7 1996 7,396.4 4,966.5 74.4 193.9 139.3 76.1 118.9 59.8 96.6 482.0 1997 10,795.6 6,460.6 1,841.3 467.6 342.3 28.7 59.0 263.5 7.2 202.5 1998 4,470.6 3,153.4 109.5 130.2 84.6 10.1 47.6 113.5 9.6 123.2 Data from 1980-93 and 1 994-98 are not strctly comparable do to a change in the methodology Source: INEGI. 102 Annex 3 Table A3.11 Foreign Direct Investment By Country of origin (CondMon) Year USA UJUngdom German Japn Swtzzwland Fmnoe Spain Sweden Caniada Others 1980 66.47%/ 3.00% 10.53% 7.59% 6.8P%o 1.20% 4.93%o 0.67%o 1.08% -2.33% 1981 63.02% 2.400/% 8.60% 12.47P% 4.40%/o 0.61% 5.98% 0.90%/o 0.31% 1.31% 1982 68.01% 1.18% 6.370/ 10.44% 3.69% 1.090/0 6.45% -0.32% 1.290/o 1.80%/o 1983 38.990/a 7.20% 16.090/o 0.560/o 2.37% 16.090/o 1.860/a 4.26%/o 3.23% 9.36% 1984 63.790/o 3.10% 10.67%0/ 2.49% 4.18% 0.61% 0.82% 4.270/ 2.27% 7.81% 1985 76.74% 3.26% 3.21% 4.59°h 8.1P/o 0.620/o 0.81% 0.320/o 2.020% 0.28% 1986 49.760/o 4.30%/o 9.01% 5.870/o 1.41% 13.070/o 3.87/o 1.01% 1.670/o 10.020/o 1987 68.85% 11.11% 1.21% 3.43% 2.460/% 0.80% 3.24% 0.95% 0.50%/a 7.45% 1988 39.33%/o 24.31% 4.330/o 4.71% 2.730/a 4.83%/ 1.08% 1.03% 1.07% 16.57/o 1989 72.56%/ 1.790/o 3.390/o 0.630/o 7.78% 0.660/a 1.76% 0.28% 1.500/% 9.66% 1990 62.000/a 3.07/o 7.74% 3.25% 3.98% 4.860/o 0.28% 0.36% 1.500/o 12.960/o 1991 66.94% 2.080/a 2.38% 2.06% 1.920%/ 14.04% 1.230/a 0.390/o 2.08% 6.88/o 1992 45.890/o 11.86% 2.360/o 2.41% 8.76% 1.92% 1.03% 0.06% 2.46% 23.260/o 1993 71.490/o 3.860/a 2.27°h 1.50% 2.08% 1.57% 1.30% 0.05% 1.51% 14.370/o 1994 45.98% 5.660/o 2.91% 6.01% 0.51% 0.860/a 1.38% 0.090/o 7.06% 29.54% 1995 65.190/a 2.64% 6.790/o 1.93/ 2.480/a 1.480/o 0.52o/s 0.76% 2.090/a 16.13%/o 1996 67.15% 1.01% 2.62% 1.880/a 1.030/a 1.61% 0.81% 1.31% 6.52% 16.070/a 1997 59.84% 17.06% 4.33% 3.170/a 0.270/a 0.55% 2.44% 0.07% 1.880/a 10.4Y0% 1998 70.54% 2.45% 2.91% 1.890/a 0.23% 1.06% 2.54% 0.21% 2.760/a 15.41% Data ftom 1980.93 and 199498 are not strcty cowMarabe dD to a chanwe in the methocbo1gy Source: INEGI. 103 Annex 3 Table A3.12 In % 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 GDP at market prices 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 Net indirect taxes 7.98 9.69 8.13 8.30 8.50 8.53 8.56 8.24 8.02 8.61 9.06 9.48 8.54 GDP at factor cost 92.02 90.31 91.87 91.70 91.50 91.47 91.44 91.76 91.98 91.39 90.94 90.52 91.46 Agriculture, value added 9.48 8.74 7.26 7.11 7.18 6.88 6.11 5.78 5.28 5.00 5.53 5.01 4.82 Industry, value added 32.12 34.31 29.50 26.93 26.00 25.64 25.69 24.62 24.70 25.53 25.84 25.87 26.03 Construction, value added 4.26 4.06 3.67 3.46 3.59 3.76 4.12 4.40 4.87 3.72 3.78 4.03 4.29 Gas, electricity and water, value added 1.45 1.40 1.20 1.22 1.24 1.36 1.46 1.46 1.35 1.16 1.06 1.08 1.07 Mining and quarrying, value added 3.61 5.03 2.71 2.17 2.14 1.69 1.60 1.29 1.23 1.58 1.42 1.38 1.22 Manufacturing, value added 22.80 23.82 21.92 20.08 19.03 18.83 18.52 17.47 17.25 19.06 19.58 19.38 19.45 Services, value added 50.42 47.26 55.11 57.65 58.31 58.95 59.63 61.37 62.00 60.86 59.56 59.63 60.61 Transport, storage and communication, value added 7.39 7.24 8.69 8.37 8.32 9.12 8.71 8.54 8.79 9.15 9.26 9.59 9.90 Trade, value added 20.51 19.12 23.29 22.88 22.63 21.18 20.86 19.99 19.41 19.15 19.57 19.34 18.30 Banking, value added 7.68 7.01 8.91 11.12 12.13 12.53 13.24 14.55 14.89 16.79 13.67 12.10 12.61 Public administration and defense, value added 16.13 15.19 15.63 15.88 16.29 17.30 18.92 20.97 21.90 20.69 19.31 19.93 20.80 Other services, value added -1.30 -1.31 -1.41 -0.60 -1.05 -1.18 -2.10 -2.68 -2.99 -4.91 -2.25 -1.33 -1.00 Source: World Bank LDB. 104 Table A3.13 Withholding Tax Rates (as of 1997) Tax Treaties Canada- Canada- U.S.-Mexico Mexico U.S. Parent/Subsidiary Dividends 10% 5% 5% Portfolio Dividends 15% 10% 15% Interest 15% 10% 15%-10% Royalties 15% 10% 10% Capital Gains 0 0 0 Source: Cockfield (1998). 105 Table A3.14 Effective Corporate Tax Rate (%) on the Foreign Capital Investment IA. Mexico as the host non-exporters Manufacturing Services U. S. Canada U. S. Canada Buildings 9.5 11.8 8.7 10.3 Machinery 31.2 32.3 40.9 41.7 Inventory 26.9 28.5 26.9 28.1 Land 22.9 24.5 22.9 24.2 Aggregate 25.2 26.7 18.9 20.3 IB. Mexico as the host, for exporters (L.e., with import duty exemption) Manufacturinj Services (for ilustration only) U. S. Canada U. S. Canada Buildings 9.5 11.8 8.7 10.3 Machinery 17.1 18.9 25.8 27.0 Inventory 22.9 24.5 22.9 24.2 Land 22.9 24.5 22.9 24.2 Aggregate 17.9 19.8 15.6 17.0 2. Canada as the host Manufacturine Services Mexico U. S. Mexico U. S. Buildings 31.4 23.0 29.5 20.9 Machinery 25.4 16.3 46.1 39.4 Inventory 41.0 33.7 45.4 38.7 Land 32.0 23.7 35.1 27.3 Aggregate 33.5 25.3 35.0 27.1 3. The U. S. as the host Manufacturine Services Mexico Canada Mexico Canada Buildings 23.8 21.7 22.4 19.8 Machinery 25.4 22.2 35.5 32.1 Inventory 21.1 19.8 21.1 19.2 Land 21.1 19.8 21.1 19.2 Aggregate 23.4 21.2 23.8 21.3 106 Annex 3 Figure A3.1 North American GDP 100.00% 90.00% 80.00% 70.00% 60.00% - 50.00% 40.00% 30.00% 20.00% 10.00% 0.00% 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 I-Canada - Mexico -United States 107 Annex 3 Figure A3.2 Composition of Exports by Country 14,000,000 12,000,000 - 10,000,000 - 8,000,000 _ 6,000,000 4,000,000 ' 2,000,000 0 0 0 - - - CO tO CO - *D DO - a 8 Q @ X -Total FOB Exports - FOB Exports to Canada -FOB Exports to US FOB Exports to Rest of the World 108 Annex 3 Figure A3.3. MEXICO in the Global Market. 1986 1992 1993 1994 1995 1998 2000 GATTis A rPEC C fOCE Costa Nicaragua I Israel r N AFTA, European Bolivia | union Colombia Venezuela Trade Agreements Currently Being Negotiated by Mexico Guatemala, Honduras and El Salvador Belize Ecuador Panama Peru Trinidad and Tobago 109 Annex 3 Figure A3.4. New Maquiladoras per State 1994-1998. 616~~ 7a8 Chihuahua - V t ~~~~~~~~~~Coahuila Baja California Nuevo Leon Tamaulipas B.C. Sur } \, (_ < t \ 213 W ) >1 nunana o ~<>Sinaloa') 85 ,axaca 1q3uerreroL|SQueretaro45 Nayarit t ? 1 3< Hidalgo 1Yuan Aguascaliente! 5GD Taxcala 22 89 Jalk 102 ueblala146 J t ~~4 f .Taba c~~~~~~~~ 110 I r 8 vVw \ Quintana Roo Distrito Federal mpeche \ |Source: SECOFI | 44 Edo. Mexico /4\ / Cmeh 59 morelo Oxc, 8uerrero 110 Annex 4 A4 ANNEX 4 Efficiency Effects from Reforms to the Mexican Corporate Income Tax A4.1 Economic neutrality, that is, the minimization of economic distortions in the allocation of economic resources, is generally accepted as a fundamental criterion in the evaluation of tax systems. In the case of the corporate income tax, excess tax burdens arise because not all types of investments are treated equally. Often, the tax treatment of investments differs by type of economic sector, by type of asset, or by the source of financing funds. These differential tax treatments lead enterprises to invest too much on those economic sectors or types of assets that receive privileged treatment and too little elsewhere. The overall result is an inefficient allocation of resources in the economy, which leads to lower national product and, over time, to lower rates of economic growth. A4.2 A different, but also important, consideration is whether the corporate income tax offers a competitive tax environment to foreign direct investment. Foreign investors generally have several choices of countries where to invest. Although there are many determinants of foreign direct investment, other things equal, the tax treatment of foreign capital has been consistently found to be a significant factor in the investment decisions of multinational corporations. Foreign direct investment, of course, is important because it contributes to capital accumulation, the import of know-how and new technologies, which lead to increase in productivity and economic growth. A4.3 The reform of the corporate income tax and the taxation of capital in general should strive to: (a) increase the efficiency of the tax system by minimizing distortions and providing for a more level playing field, and (b) to maintain a favorable environment for foreign direct investment. The goal of this annex is to research the extent to which the package of reforms to the corporate income tax enhances the economic neutrality of Mexico's tax system and keeps Mexico competitive vis-a-vis foreign investors. A4.4 A salient feature of Mexico's previous corporate income tax was the important role played by the special treatment of particular sectors and sources of incomes, such as the special treatment received by the agriculture and transport sectors. The case for the elimination of these special treatments is based not only on the distortions in the allocation of resources or excess tax burdens created, but also on increased administration 111 Annex 4 complexity and opportunities for tax evasion thus created,' and on the costs imposed on taxpayers as a whole. The revenues lost due to these special treatments need to be made up with higher taxes in other activities or sectors in the economy. These additional taxes in turn can be the source of additional distortions in the allocation of economic resources and add to horizontal and vertical inequities in the distribution of tax burdens. A4.5 In order to compare the efficiency losses induced by the corporate income tax and the tax treatment of financial incomes obtained by investors (domestic and foreign), this annex estimates marginal effective tax rates (METRs) under the previous tax regime and the new one. The METR approach looks at the impact of taxes on marginal or incremental decisions by businesses. The METRs estimate is the level of tax arising for a firm when it decides to invest one more unit of capital. The difference between the gross (before taxes) rate of return and the net (after taxes) rate of return is the tax wedge. The METR is expressed in percent terms as the ratio of the tax wedge to the gross (before taxes) rate of return. When the METRs are positive, they reflect that investment activities are discouraged, and the more so the higher the rate. Negative METRs are also possible. In this case, the econormic activity in question is being encouraged through a subsidy. Taxes on capital income, more importantly the CIT, can affect not only the level of investment but also its composition. Differences in METRs, therefore, lead to distortions in the allocation of resources, and as pointed out above, they reduce potential output and ultimately slow down economic growth. A4.6 More concretely, the marginal effective tax rate (METR) analysis provided in this annex has two focuses. First, we ask the question of whether the changes in the corporate income tax and the tax treatment of capital reduce the cross-sector tax distortions that existed under the previous tax system. For this purpose, a METR comparison between the previous tax structure and the new one is presented. Second, we examine Mexico's tax competitiveness vis-a-vis four other Latin American countries in attracting foreign investment, mainly those from the US.2 The four other Latin American countries include Argentina, Brazil, Chile and Peru. A4.7 For the analysis focusing on the domestic cross-sector tax distortion, METR simulations are conducted for a broad range of industries including agriculture, manufacturing, construction, transportation and storage, communication, public utility, wholesale trade, retail trade and the other services sector. The simulations also reveal the impact of elimination of preferential treatments for primary sectors and changes in taxing small- to medium-sized owner-operated firms.3 1 These arise form the opportunities to misuse the 50 percent tax reductions for the agricultural, livestock, fisheries and forestry sectors. Similarly, the cash flow system for agriculture and trucking is much more generous than the general regime, creating opportunities and incentives for evasion, especially in the trucking industry. 2 Theinternational comparison builds on Bird and Chen 2000. 3 Under the Mexican tax regime, firms are taxed differently according to the nature of business, the size and the location of the firm. 112 Annex 4 A4.8 For the analysis focusing on the cross-border tax competitiveness, the METR simulations cover only the manufacturing and service sectors, because these are the main destination of foreign direct investment. Assuming that multinational firms in these sectors are generally large and located in major industrial areas, these simulations do not include any special tax treatment for small to medium taxpayers or any tax provisions for underdeveloped regions. A4.9 Because debt-financed investments and equity-financed investments are treated differently in the tax laws, generally the METRs results are sensitive to the assumptions made about the financial structure of investments. The financing structure for all simulations, unless otherwise noted, will be assumed to be a combination of 40 percent debt and 60 percent equity. I. A comparison of marginal effective tax rates on domestic capital investment under the previous and the new tax regimes A4.10 The changes in the taxation of capital income focus on improving both the efficiency and simplicity of Mexico's tax system. The most important reforms, which are incorporated into our METRs calculations, include the following: * The 5-percentage point deferral (for profit distribution) in corporate income tax collection is eliminated. * The withholding tax on dividends and the gross-up taxing scheme are eliminated. * The preferential cash-flow tax regime for primary sectors (mainly agriculture and ground transportation) are eliminated. * The mandatory sharing of total profits with employees at a minimum of 10 percent, which is not deductible under the previous system, will be deductible for CIT purposes (This reform was proposed in 2001 and enacted, but would take effect in 2002 only if GDP growth would exceed 3 percent. This provision affects more the overall level of METRs, rather than their variability across sectors). * The progressive scheme of presumptive tax applied to small owner-operated taxpayers will be simplified to a single rate of 1%. In the meantime, the maximum revenue for defining small taxpayers will be reduced to 1 million pesos of annual gross income. * A new category of corporate taxpayers would be identified as medium owner- operated taxpayer, with up to 4 million pesos of annual gross income. The cash- flow regime previously applied to the primary sector would be applied to this category of taxpayers. * The states will be given new authority in taxing small to medium owner-operated taxpayers. More specifically, the state governments will be allowed to levy a surtax of up to 5 percent on the CIT payable by the medium owner-operated taxpayers. * Finally, there are two major changes in personal income taxes (PIT) on financial investment that would affect our estimate of integrated METR on capital investment. First, the 24 percent withholding tax on interest income, which acts as a final tax under the previous system, will become a true withholding part of a 113 Annex 4 global PIT, where interest income is added to all other types of income to compute the taxpayer's final liability. Thus, under the new regime, interest income will be taxed as regular income and subject to the progressive PIT rate structure (3% to 35%). Second, under the previous system, gains from selling publicly traded stocks are exempted. In the new scheme, this exemption will be eliminated. For our METR simulation, a 35 percent PIT rate will be used for capital gains and for the interest income under the new tax system. This estimate is based on the assumption that most but not all of the financial investments are made by high-income people. A4.12 Our METR simulations for domestic capital investment start with the general corporate METR. The goal is to illustrate the impact of changes at the firm's level (Tables A4.1, A4.2A and A4.2B). Then we present a simulation for integrated METR that combines taxes on capital at both firm level and personal level (Tables A4.3A and A4.3B). 1. A general comparison by industry: the previous vs. the new tax structures A4.13 The computed general corporate METRs on domestic capital investment are presented in Table A4. 1. These computations exclude the small to medium owner- operated firms. The simulations for the previous tax structure take into account the cash- flow tax regime for the primary sectors, mainly agriculture, and for the ground transportation sector. In contrast, the simulations for the new tax regime eliminate the cash-flow provision for these sectors and the 50 percent reduction in the CIT rate for publishing. Furthermore, as mentioned above, with the elimination of the CIT deferral for dividends distribution, corporate taxpayers have to pay the standard CIT rate of 35 percent regardless of their decision on the distribution of profits. In contrast, under the previous system, the actual CIT rate might be well below 35 percent as long as there was no full distribution.4 However, due to the disallowance for the mandatory employee profit sharing (which we assumed to be at the minimum 10 percent), the effective CIT rate on taxable income net of profit shared with employees could have been higher than 35 percent in the previous regime. Therefore, the CIT rate used for simulating the previous system is slightly higher (35.6 percent)5 than that used for the new system (35 percent). A4.14 Three important conclusions are derived form the results in Table A4.1. * First, the METRs are generally lower under the new tax system than under the previous system. The only exceptions are for the agricultural and transportation industries. The general result is mainly due to the new deduction allowance for 4Assuming that the average rate for profit distribution is 40%, a 32% (=30%+5% x 40%) CIT rate is used in our simulation for the previous system. Of course, the actual CIT rate may differ across taxpayers depending on the distribution portion. 5 Refer to footnote 4. By applying 32% CIT on the taxable income net of the 10% profit shared with the employees, which is not deductible for tax purpose under the previous regime, the effective CIT rate is 35.6% (=32%/(I-10%)). 114 Annex 4 the mandatory employee profit sharing. But notice that the reduction in the METRs is not significant for most industries. This is because the elimination of the CIT deferral for dividends distribution is partly offset by the impact of allowance for profit shared by employees. * Under the new tax regime, the METR is somewhat higher for agriculture and significantly higher for the transportation industry, compared to the previous regime. This is mainly a result of elimination of cash-flow regime for both industries. It is also noteworthy that METR for the transportation industry under the previous system is negative, which implies a tax subsidy for this industry. This is mainly because depreciable assets such as buildings, machinery and equipment (e.g. motor vehicles) account for almost 90 percent of the total capital invested in the transportation industry. Under the cash-flow regime, depreciable assets can be fully written off before calculating any income tax, which, combined with other tax allowance (e.g. interest deduction), may entail a total write off for the transportation industry well above 100 percent of income generated from capital investment for a long period. * Third, the most important result is that the inter-sector tax distortion, as measured by the inter-sector METR dispersion, will be significantly reduced under the new tax structure. This is the result of the elimination of preferential treatment for particulars sectors ( the cash-flow regime favoring on lye the primary sectors, such as agriculture, and the ground-transport sector). A4.15 In summary, the METRs are generally lower under the new tax system than under the previous system, and the dispersion of METRs is also smaller under the new system. Therefore, the changes in the tax laws seem to be successful in reducing distortions and therefore in increasing efficiency in the allocation of resources and the prospects for growth in the Mexican economy. 2. A more detailed comparison across industries by type of assets A4.16 The computations for the METRs by economic sector can hide significant differences in the tax treatment of different types of assets within each sector. Tables A4.2A and A4.2B provides corporate METR simulation by industry and by asset type. These two tables are simply the extension of Table A4. 1 with detailed breakdown. There are also three major conclusions arising from tables A4.2A and A4.2B. * First, on average, machinery is the lowest taxed asset, followed by buildings, while land is the highest taxed asset, followed by inventory. This is mainly because the tax law entitles enterprises to a general tax depreciation allowance for machinery that surpasses its economic depreciation rate. In contrast, land, as a non-depreciable asset, incurs financing cost indefinitely and is subject to both property transfer tax and property tax. It should be also noted that, the tax adjustment for inflation may benefit investment in depreciable assets more than investment on non-depreciable ones. There are two offsetting effects. First, due 115 Annex 4 to the inflation adjustment, only real interest cost rather than nominal interest cost is deductible for the income tax purpose. This makes the interest deduction less generous, which may affect non-depreciable assets more than depreciable assets. Second, the ability to recognize the inflationary impact on the value of depreciable assets implies a more generous tax depreciation allowance than that without the inflation adjustment. Overall, the investment in depreciable assets may gain more tax benefits than investment in non-depreciable assets. * Second, while METRs for non-depreciable assets (inventory and land) are the same across industries, METRs for depreciable assets (buildings and machinery) vary significantly across industries. This is because the only non-tax cost for non- depreciable assets is the financing cost, and the tax wedge between the before- and the after-tax financing cost of capital is the same across industries. For depreciable assets, there is an additional non-tax cost, i.e. economic depreciation cost. This economic depreciation cost (and the tax depreciation allowance) varies by industry. As a result, the tax wedge between the before- and the after-tax cost of capital for depreciable assets varies by industry. * Third, METRs for all types of assets would be generally lower in the new system compared to the previous one. The exceptions are the agricultural and transportation industries. This observation is consistent with the conclusions reached from Table A4. 1, where entire industries, without a breakdown by type of asset, are compared. A4.17 In summary, although the differential treatment of different types of assets within industries will remain under the new tax system, the METRs for all types of assets would be generally lower in the new system compared to the previous one. This, combined with the elimination of the cash-flow regime favoring the primary sectors, should also help reduce distortions in investment decisions and improve efficiency in the allocation of resources. 3. The impact offinancing structure A4.18 Debt-financed and equity financed investments are not treated the same by the tax laws. The impacts of varying financing structure on METRs are illustrated in Tables A4.3A and A4.3B. The METRs in these two tables integrate taxes on capital payable at both the corporate and the personal level. As mentioned above, the tax deductibility of borrowing costs may benefit firms with higher levels of borrowing. However, this imbalance in the tax treatment of debt and equity financing may be alleviated by properly taxing interest income at the personal level. As Table A4.3A shows, under the previous tax system, the gap in METR between pure debt financing (at 100 percent debt) and pure equity financing (at 100 percent equity) is about 12 percentage points (19 percent versus 31 percent). In contrast, as shown in Table A4.3B, the existing imbalance in taxing debt financing and equity financing will be largely eliminated under the new tax system. This is mainly because interest income under the previous tax system is taxed at 24 percent, which is much lower than the PIT rate (up to 40 percent) applicable to high-income 116 Annex 4 taxpayers. Under the new tax system, the interest income will be taxed as regular income and subject to a much higher PIT rate (up to 40%) for high-income taxpayers. A4.19 In summary, the distortions in investment decisions under the previous tax regime associated with the different tax treatment of interest and equity income are largely eliminated under the new tax regime. This again should contribute significantly to improving efficiency in the allocation of resources and investment decisions by investors, leading to better prospects for growth in the Mexican economy. II. Marginal effective tax rates on foreign capital investment. Will Mexico remain internationally competitive? A4.20 An important question on the tax reform is how it would affect Mexico's competitiveness to attract foreign direct investment, in particular from the USA. To examine this question we simulate METRs for foreign capital investment in Mexico and four other Latin American countries, including Argentina, Brazil, Chile and Peru. Each of these countries, of course, has its different tax system. Table A4.4 provides a summarized comparison in statutory tax provisions in the five countries (including Mexico) that may affect foreign capital investment. It should be noticed that the information provided in the table has incorporated the tax changes in Mexico as well as planned changes in Chile and Peru. For example, Chile started lowering its import duty since 1999 by 1 percentage point per year. As a result, its import duty will be lowered to 6% in 2003 from 11% in 1998. In the case of Peru, the CIT is 30%6 and will be lowered to 20% in 2002. A4.21 As Table A4.4 shows, Mexico's CIT rate is the same as that in Argentina and higher than the rate of the rest of the countries in the sample. However, Mexico does not impose any gross receipts tax, as is the case in Argentina and Brazil. Furthermore, Mexico is more generous in the tax allowance for building depreciation than Argentina and Brazil. Mexico also appears to have the lowest property tax rate compared to other countries. In addition, Mexico, along with Argentina, exempts new capital goods from import duty. Finally, with NAFTA, Mexico has no import duty on any production input imported from the US and Canada.7 It is interesting to note that Chile has the lowest CIT rate among the sample countries and is the only country that does not impose any tax on property transfers. However, Chile is also the only country that collects withholding tax on dividends distributed to US residents. Another noteworthy is Peru's on-going reforms. As mentioned, the CIT rate in Peru will be lowered to 20 percent in 2002 and hence become the second lowest among the five countries. The most burdensome tax system appears to be in Argentina and Brazil, which is featured by the relatively high CIT rate, the rather high gross-receipts tax, the least generous tax depreciation allowance, and the relatively high property taxes. 6A lower rate of 20% is applicable to profits reinvested in domestic productive activities. 7Actually, the liberalization of tariffs started in 1994 and will not be completed until 2009. There are some items which will never be fully liberalized. 117 Annex 4 A4.22 Due to the lack of data, we are forced to make several assumptions for our cross- country METR calculations. First, except for Chile where property tax is levied by the central government at a standard rate, the maximum property tax rate shown in Table A4.4 is used for our calculation. In the case of Brazil, there is no estimate available for maximum or average property tax rate; an arbitrary rate of 1.5% is used for our calculations. Second, import duty often varies across commodities; there is no estimate of precise rate for any countries. Therefore, a 50% of the upper end rate shown in Table A.4 is used for all countries but Brazil. The rate is applied to machinery and equipment as well as inventory assuming that only half of capital and materials (inventory) inputs for carrying out any business are imported. In the case of Brazil, again an arbitrary rate of 5.5% is used for our calculation. A4.23 Tables A4.5A and A4.5B provide our estimates of marginal effective tax rate for foreign capital investment made by US multinationals in manufacturing and services sectors, respectively. To break up the impact of different taxes and to eliminate possible errors caused by the assumptions made above, the figures presented in these two tables are divided into three panels. Panel A includes all the major taxes summarized in Table A4.4. Panel B excludes property taxes, for which there is no precise estimate except that for Chile. Panel C excludes both property tax and import duty for all countries so to focus on the impact of corporate income tax system on foreign capital investment. As a reference, the US domestic METR has been added to the table. It should be noted that there is never an estimate for the average property tax in US since it is a local tax that varies by location. Therefore, the US domestic METR is not available for Panel A. We also assume that there is no import duty on any imported inputs for production in the US. As a result, the US MiETR is the same in Panel B and C. A4.24 As Tables A4.5A and A4.5B show, when all the taxes are included (Panel A), Mexico appears to provide the most tax advantageous environment to US investors. This is mainly a combination of its relatively generous CIT regime (i.e. more generous depreciation allowance and no gross-receipts tax) compared to Argentina and Brazil, plus the lowest property tax among all the countries in the sample, plus the zero import duty within NAFTA. A4.25 By excluding property tax (Panel B), METRs on buildings and land drop more significantly in all the other four countries by comparison to Mexico. This is because the property tax in Mexico appears to be the lowest among all countries. As a result, the aggregated METR in Mexico would rise to the second from the bottom among the five countries. It is noteworthy that the impact of excluding property taxes is more significant for the service sector than for the manufacturing sector. This is because buildings and land, to which the property tax is applicable, account for a much bigger share of the capital invested in the service sector than in the manufacturing sector (79 percent versus 26 percent). A4.26 By excluding both property tax and import duty (Panel C), the competitiveness of Chile and Peru is significantly enhanced. This is mainly the result from their rather low CIT rate. It is also noteworthy that the reduction in aggregate METRs would be more 118 Annex 4 significant in the manufacturing sector than that in the service sector. This is because the import duty affects mainly investment in machinery and inventory, which account for a higher share of capital in the manufacturing than in the service sector (74 percent versus 21 percent). IHI. Conclusion A4.27 In summary, the changes in Mexican tax system will significantly reduce the inter-sector tax distortion for domestic finns. The existing imbalance between taxing debt financing and equity financing will also be largely eliminated. A4.28 For foreign investors, particularly multinational firms from the US, the new tax changes in Mexico do not seem to weaken its tax competitiveness vis-a-vis four other major Latin America countries. 119 Annex 4 Table A4.1 Effective Corporate Tax Rate on the Domestic Capital Investment For General Tax Regime, the previous vs. the new With 40% debt and 60% equityfinancing Difference Previous New (In % points) Agriculture*t 8.8 10.4 1.6 Manufacturing 20.1 18.0 -2.1 Construction 22.8 21.0 -1.8 Transportation* -8.9 13.2 22.0 Communications 16.7 15.1 -1.7 Public Utility 19.7 17.8 -1.9 Wholesale Trade 21.9 19.8 -2.1 Retail Trade 22.9 20.8 -2.1 Other Services 22.4 20.8 -1.7 Inter-sector dispersion 5.5 2.4 -3.1 * Entitled to the cash-flow tax within the previous tax structure. t Entitled to the 50% of CIT rate within both the previous and the new tax structures. 120 Annex 4 Table A4.2A METR Simulation by Industry and by Asset Type On domestic capital investment under the previous general tax regime With 40% debt and 60% equityfinancing Buildings Machinery Inventory Land Aggregate Agriculture*t 6.6 -9.0 10.4 19.6 8.8 Manufacturing 19.8 16.7 22.8 30.0 20.1 Construction 24.0 -3.4 22.8 30.0 22.8 Transportation* -2.1 -19.7 22.8 30.0 -8.9 Communications 20.0 3.6 22.8 30.0 16.7 Public Utility 17.9 21.0 22.8 30.0 19.7 Wholesale Trade 18.9 17.6 22.8 30.0 21.9 Retail Trade 19.5 23.8 22.8 30.0 22.9 Other Services 18.9 25.8 22.8 30.0 22.4 Simple Average 15.9 8.5 21.5 28.9 16.3 Table A4.2B METR Simulation by Industry and by Asset Type On domestic capital investment under the newgeneral tax regime With 40% debt and 60% equityfinancing Buildings Machinery Inventory Land Aggregate Agriculturet 16.2 -4.8 9.3 19.3 10.4 Manufacturing 18.3 14.2 20.6 19.3 18.0 Construction 22.7 -6.0 20.6 19.3 21.0 Transportation 17.2 10.4 20.6 19.3 13.2 Communications 11.2 1.0 20.6 19.3 15.1 Public Utility 16.3 18.7 20.6 19.3 17.8 Wholesale Trade 17.3 15.2 20.6 19.3 19.8 Retail Trade 18.0 21.6 20.6 19.3 20.8 Other Services 17.3 23.6 20.6 19.3 20.8 Simple Average 18.0 10.4 19.4 19.3 17.4 * Entitled to the cash-flow tax within the previous tax structure. t Entitled to the 50% of C1T rate within both the previous and the new tax structures. 121 Annex 4 - Table A4.3A Estimate of Integrated METR for Various Financing Structure On domestic capital investment under the previous general tax regime 100% debt 40% equity 60% equity 100% equity Agriculture*t 22.6 20.8 19.8 17.9 Manufacturing 21.4 27.2 29.8 34.5 Construction 24.2 29.7 32.2 36.6 Transportation* -7.0 0.8 4.3 10.6 Communications 18.3 24.2 26.8 31.6 Public Utility 21.1 26.8 29.4 34.1 Wholesale Trade 23.1 28.8 31.3 35.9 Retail Trade 24.1 29.7 32.2 36.7 Other Services 23.9 29.4 31.8 36.3 Simple average 19.1 24.2 26.4 30.5 Table A4.3B Estimate of Integrated METR for Various Financing Structure On domestic capital investment under the new general tax regime 100% debt 40% equity 60% equity 100% equity Agriculturet 40.5 34.2 30.6 21.9 Manufacturing 38.0 37.0 36.4 35.3 Construction 40.3 39.3 38.8 37.7 Transportation 34.3 33.2 32.7 31.5 Communications 35.7 34.7 34.1 33.0 Public Utility 37.8 36.8 36.3 35.2 Wholesale Trade 39.3 38.3 37.8 36.7 Retail Trade 40.1 39.1 38.6 37.5 Other Services 40.0 39.1 38.6 37.5 Simple average 38.5 36.8 36.0 34.0 * Entitled to the cash-flow tax within the previous tax structure. t Entitled to the 50% of CfT rate within both the previous and the new tax structures. 122 Annex 4 Table A4.4 Business Tax Provisions Applicable to Large Firms in Manufacturing and Service Industries Argentina Brazil Chile Mexico' Peru 1. Capital Taxes Corporate income tax 35 34 15 35 20 (2002) Gross-receipts tax 1-5 2.65 None None None Assets-based tax2 1.0 None 0.25 - 0.5 1.8 None Thin capitalization rule Yes None None None None Tax depreciation rate3 Buildinrs 2.0 SL 4.0 SL 7.5 SL 5.0 SL 3.0/10.0 SL4 Machinery 10 SL 10 SL 15.2 SL 10 SL 10.0 SL Inventory accounting5 LIFO FIFO/A FIFO immediate write-off FIFO/A Years! loss carry-over6 5 (F) Infinite(R) Infinite(R) 10 (F) 4 (F) Withholding tax on dividends Applicable to US 0.0 0.0 20.0 0.0 0.0 Property-based taxes Property tax Max. 1.5 Vary 1.425 Up to 0.647 Up to 1.0 Property transfer tax 1.5 2 None 3.3 3.0 2. Import duty On new capi goods Exempt Up to 55 6 (2003) Exempt 12 (general) On other business inputs 0 - 15 Up to 55 6 (2003) 0 (NAFTA) 12 (general) 3. Inflation Adjustment No Yes Yes Yes Yes Notes to the Table: 1. With new tax changes incorporated. See the text for details. 2. All are levied as minimum tax that can be used as tax credit for income tax payable. 3. SL = straight-line method, which means the depreciation allowance is claimed based on the acquisition cost of the capital. A different method is the declining-balance method, which means the depreciation allowance is claimed based on the balance of un-depreciated value of capital. 4. The higher depreciation allowance is applicable for buildings used for tourist business. 5. FIFO = First-in-first-out, LIFO = Last-in-first-out, and A = Average cost. It should be noticed that when there is required adjustment for inflation, there is no significant difference in tax impact between choosing one or the other method. 6. Following the number of years for loss carry over, the letter in brackets indicates the following: F = forward, B = backward, and R = certain restriction applies to the value of loss to be written off. Refer to the text for details. 123 Annex 4 Table A4.5A Effective Corporate Tax Rate on U.S Capital Investment (%) MANUFACTURING Including property tax Argentina Brazil Chile* Mexico Peru** U.S. Buildings 42.4 38.7 31.2 21.8 27.6 NA Machinery 17.8 37.3 19.4 14.2 27.6 NA Inventory 34.1 27.1 17.3 21.8 16.5 NA Land 45.0 41.5 35.5 31.0 28.3 NA Aggregate 31.6 34.5 22.3 19.3 24.0 NA B. Excluding property tax Argentina Brazil Chile* Mexico Peru** U.S. Buildings 24.8 18.3 7.1 9.7 10.4 21.8 Machinery 17.8 37.3 19.4 14.2 27.6 23.5 Inventory 34.1 27.1 17.3 21.8 16.5 22.1 Land 29.1 23.1 14.8 21.8 11.5 17.0 Aggregate 26.2 29.6 15.9 16.3 19.8 22.5 C. Excluding both property tax and import duty Argentina Brazil Chile Mexico PERU** U.S. Buildings 24.8 18.3 7.1 9.7 10.4- 21.8 Machinery 17.8 23.0 7.9 14.2 6.8 23.5 Inventory 29.1 23.1 14.8 21.8 11.5 22.1 Land 29.1 23.1 14.8 21.8 11.5 17.0 Aggregate 24.1 22.0 10.5 16.3 9.5 22.5 * For year 2003 when the import duty will be reduced from the current 8% to 6%. ** For year 2002 when the CTr rate will be reduced from the current 30% to 20%. 124 Annex 4 Table A4.5B Effective Corporate Tax Rate on Foreign Capital Investment (%) SERVICES L Including property tax Argentina Brazil Chile* Mexico Peru** U.S. Buildings 46.3 40.3 31.8 19.1 28.7 NA Machinery 38.2 54.3 28.8 35.6 41.6 NA Inventory 34.1 27.1 17.3 21.8 16.5 NA Land 45.0 41.5 35.5 31.0 28.3 NA Aggregate 44.2 41.6 31.0 24.0 29.5 NA B. Excluding property tax Argentina Brazil Chile* Mexico Peru** U.S. Buildings 31.3 21.2 8.3 6.0 12.1 20.1 Machinery 38.2 54.3 28.8 35.6 41.6 34.5 Inventory 34.1 27.1 17.3 21.8 16.5 22.1 Land 29.1 23.1 14.8 21.8 11.5 17.0 Aggregate 32.1 28.2 13.3 15.3 17.3 21.8 C. Excluding both property tax and import duty Argentina Brazil Chile Mexico Peru** U.S. Buildings 31.3 21.2 8.3 6.0 12.1 20.1 Machinery 38.2 42.6 15.2 35.6 20.5 34.5 Inventory 29.1 23.1 14.8 21.8 11.5 22.1 Land 29.1 23.1 14.8 21.8 11.5 17.0 Aggregate 31.6 25.0 11.0 15.3 13.0 21.8 * For year 2003 when the import duty will be reduced from the current 8% to 6%. ** For year 2002 when the CIT rate will be reduced from the current 30% to 20%. 125 IMAGING Report No.: 22527 ME Type: ER