Mexico: Reforming Intergovernmental Transfers to More Effectively Advance Their Policy Objectives I. Motivation and Objectives 1. Asymmetries between revenue and expenditure decentralization are common in federal systems. Due to the sensitivity of major tax bases, personal and corporate income taxes and consumption taxes tend to be more efficiently administered at the federal level. However, subnational governments (SNGs) are often better equipped to provide public goods and services that meet the needs of local populations. The resulting combination of centralized tax collection and decentralized service provision— known as “asymmetric decentralization�—creates vertical fiscal gaps in federal republics worldwide. 2. The extent of asymmetric decentralization varies from country to country. Because emerging economies often have highly centralized revenue systems, efforts to decentralize service provision tend to generate large vertical fiscal gaps. In many emerging economies, the expenditure responsibilities of SNGs far exceed the revenue potential of their assigned tax bases. The resulting disparities between subnational expenditure needs and subnational revenue capacity are bridged by transfers from the federal government to SNGs, the terms of which vary according to each country’s fiscal federalism framework. 3. The importance of intergovernmental transfers is determined both by the size of the vertical fiscal gap and by the size of the gap between expenditure needs and fiscal capacities of individual SNGs, the so called horizontal fiscal gap. In addition to the vertical fiscal gap, regional differences in socioeconomic conditions create horizontal fiscal gaps, as SNGs vary in terms of their expenditure needs and ability to generate own-source revenue. Countries that face large horizontal fiscal gaps often adjust intergovernmental transfers to reduce disparities in the provision of public goods and services across SNGs. Intergovernmental transfers can also be used to advance national social and economic objectives or to boost spending on public goods and services that generate positive spillovers across subnational jurisdictions. 4. Because Mexico faces large vertical and horizontal fiscal gaps, intergovernmental transfers play a major role in the country’s fiscal federalism framework. Mexico has one of the largest vertical fiscal gaps among comparable countries worldwide. Due to the centralization of revenue collection at the federal level, combined with the decentralization of expenditures at the state level, Mexico’s SNGs are now responsible for around 50 percent of total spending, yet they collect only 10 percent of tax revenue —the largest vertical fiscal gap among OECD countries. Meanwhile, large regional disparities create especially acute expenditure needs in Mexico’s less-developed regions, where the own-source revenue capacity of SNGs is particularly limited. In some of these regions, intergovernmental transfers finance more than 95 percent of subnational spending. To advance national development objectives, the federal government earmarks a large share of federal transfers for key social sectors, such as education, health, public security, and basic social infrastructure. Finally, limited access to credit markets among certain SNGs, especially local governments, requires that capital transfers finance a significant share of public infrastructure investment. 1 5. Intergovernmental transfers are designed to attenuate vertical and horizontal fiscal gaps, broadly equalize public service provision across regions, and advance national development goals, but the heavy dependence of many Mexican SNGs on federal transfers creates serious risks. While the optimal vertical fiscal gap for each federal system is difficult to determine, the empirical analysis presented in this note indicates that large fiscal gaps and low levels of subnational own-source revenue weaken incentives for expenditure efficiency. Large intergovernmental transfers can also undermine incentives for SNGs to strengthen own-source revenue collection, creating a cycle of dependence. 6. The following note examines Mexico’s intergovernmental transfer system, assesses its fiscal- equalization effects, and offers recommendations to improve its efficiency and effectiveness. The note is organized into four sections. Following this introduction, Section 2 describes the intergovernmental transfer system and reviews its evolution since 2002. Section 3 assesses its fiscal equalization effects and identifies constraints on its efficiency and capacity to redistribute resources to less-developed regions. Section 4 concludes the analysis and recommends reforms to improve the effectiveness of the intergovernmental transfer system. II. Mexico’s Intergovernmental Transfer System 7. Intergovernmental transfers can be classified into four types according to their policy objectives. These include: (i) revenue-sharing transfers used to close the vertical fiscal gap; (ii) equalization or compensatory transfers used to reduce the horizontal fiscal gap; (iii) conditional or earmarked transfers used to finance expenditures that advance national development objectives or generate positive spillovers across jurisdictions; and (iv) capital transfers used to overcome subnational credit constraints that would otherwise inhibit public investment. Some countries have also established regional development funds designed to promote economic growth in underdeveloped areas. A hybrid of types (ii) and (iii), these funds typically provide equalization transfers that are earmarked for specific expenditure categories. 8. Mexico’s fiscal federalism framework includes all four types of intergovernmental transfer (Figure 1). Revenue-sharing transfers: The largest source of revenue-sharing transfers (participaciones) from the federal government to state governments is the General Fund for Revenue-Sharing Transfers (Fondo General de Participaciones, FGP). The FGP is financed by 20 percent of all federal tax revenue and federal income from extractive industries. States, in turn, are required to transfer at least 20 percent of FGP resources to municipalities, and each state establishes its own allocation criteria. In addition, the Revenue Collection and Auditing Fund (Fondo de Fiscalizacion y Recaudacion) receives 1.25 percent of projected federal sharable revenue (recaudación federal participable); the Fund for the Promotion of Local Governments (Fondo de Fomento Municipal) receives 1 percent; and the Fund for Municipalities Located along the Border or in Coastal Areas (Fondo para Municipios Colindantes con la Frontera o los Litorales) receives 0.136 percent, which it allocates according to the value of international trade processed at seaports and border crossings in each municipality. Additional revenue-sharing mechanisms also apply to specific federal taxes and other revenue streams, including excise taxes, taxes on gasoline, taxes on the oil industry, customs duties, income taxes paid by civil servants, etc. Equalization transfers: In addition to covering the vertical fiscal gap, the revenue-sharing transfers described above are also intended to address disparities in expenditures across jurisdictions. The federal 2 revenue shared through the FGP is primarily collected in Mexico’s most-developed regions, and less- developed regions tend to receive more than they contribute. However, the FGP is not designed to achieve full fiscal equalization, and many of its sub-transfers have other policy goals. Conditional transfers: Mexico’s fiscal federalism framework also includes conditional transfers (aportaciones) that finance sector-specific spending by SNGs. These transfers are distributed through eight funds under the Ramo 33 budget line, the largest of which finance education, health, and public security services provided by state governments. Conditional transfers also include discretionary transfers under Ramo 23, which subsidize state programs that contribute to national objectives. Finally, federal resources are channeled to state governments through individual agreements (convenios) negotiated on a case-by-case basis. Capital transfers: Capital transfers finance investments in basic social infrastructure by state and local governments. These transfers are conditional and distributed according to equalization criteria. The subsidies provided under Ramo 23 and the resources transferred through individual agreements with state governments can also be used to finance capital spending. Figure 1: Taxonomy of Mexico’s Intergovernmental Transfer System Revenue sharing Equalizing Conditional Capital Aportaciones Basic Social (Education (FONE), Fondo General de Fondo General de infrastructure for state Health (FASSA), Public Participaciones Participaciones (FISE) and municiplaities Security (FASP), others (FAIS) under Ramo 33. Other revenue-sharing schemes (Fomento Municipal, oil revenues, Subsidies under Ramo tax collection of 23 individual taxes, etc) Agreements Source: Fiscal Coordination Law, SHCP 9. The total value of Mexico’s intergovernmental transfers rose from 6.6 percent of GDP in 2002 to 8.1 percent in 2017—equal to about half of all federal revenue or 90 percent of SNG revenue. The continuous growth of revenue-sharing transfers reflects a steady increase in federal tax revenue over the period, while the expansion of conditional transfers has been driven by the decentralization of service delivery. Conditional transfers are not automatically linked to federal revenue and can, in principle, be adjusted by policymakers. In practice, however, a large share of earmarked transfers is effectively mandatory, as their level and allocation are highly inertial. Nevertheless, conditional transfers have declined since the federal government launched its current fiscal consolidation effort in 2015. The federal 3 authorities exercise a greater degree of discretion over the subsidies provided under Ramo 23 and the decentralization agreements, both of which contracted sharply between 2014 and 2017 (Table 1). Table 1: Intergovernmental Transfers in Mexico, 2002-2017 2002 2006 2010 2014 2017 % Fed. % Fed. % of % Fed. % of % Fed. % of % Fed. % of GDP % of GDP Revenue Revenue GDP Revenue GDP Revenue GDP Revenue TOTAL 6.6% 49.5% 7.1% 48.2% 7.7% 49.8% 8.7% 52.8% 8.1% 46.3% Revenue Sharing 2.9% 21.7% 3.1% 21.1% 3.3% 21.0% 3.3% 20.3% 3.5% 20.1% FGP 2.4% 18.3% 2.6% 17.6% 2.5% 16.0% 2.7% 16.1% 2.6% 14.6% Other 0.5% 3.4% 0.5% 3.5% 0.8% 5.0% 0.6% 3.9% 0.5% 5.5% Conditional 3.7% 27.8% 4.0% 27.0% 4.5% 28.8% 5.4% 32.5% 4.6% 26.1% Aportaciones (Ramo 3.2% 24.3% 3.2% 21.9% 3.2% 20.9% 3.2% 19.5% 3.0% 17.3% 33) Education 1.9% 14.3% 1.8% 12.5% 1.9% 12.0% 1.8% 10.6% 1.7% 9.5% Health 0.4% 2.8% 0.4% 2.6% 0.4% 2.6% 0.4% 2.6% 0.4% 2.4% Social Infrastructure 0.3% 2.2% 0.3% 1.8% 0.3% 2.0% 0.3% 2.0% 0.3% 1.8% Other Aportaciones 0.7% 5.1% 0.7% 4.9% 0.7% 4.3% 0.7% 4.2% 0.6% 3.7% Other Conditional 0.4% 2.8% 0.4% 3.0% 0.4% 2.3% 0.4% 2.2% 0.3% 1.9% Transfers Subsidies (Ramo 23) 0.0% 0.0% 0.3% 1.8% 0.5% 3.0% 0.8% 5.1% 0.6% 3.4% Agreements 0.5% 3.5% 0.5% 3.4% 0.8% 4.9% 0.9% 5.7% 0.6% 3.7% Health Insurance n.a. n.a. n.a. n.a. n.a. n.a. 0.4% 2.3% 0.3% 1.8% Source: SHCP 10. The multiple objectives of the FGP and other revenue-sharing transfers distinguish Mexico’s fiscal federalism framework from more conventional intergovernmental transfer systems. In a conventional system, vertical and horizontal fiscal gaps are addressed by two different instruments—pure revenue-sharing transfers distributed according to devolutionary or derivation criteria and pure equalization transfers designed to reduce disparities in expenditure capacity after accounting for the fiscal impact of revenue-sharing transfers—and neither mechanism is used to create incentives for SNGs or advance policy goals beyond its core purpose. In Mexico, however, revenue-sharing transfers are used to address both the vertical and horizontal fiscal gaps, to stimulate regional economic growth, and to incentivize own-source revenue collection by SNGs.1 In addition, the numerous small transfers financed by federal sharable revenue are designed to achieve a wide range of objectives, which further dilute the fundamental revenue-sharing and fiscal-equalization goals of Mexico’s transfer system. 11. The excessive complexity and inconsistent purposes of Mexico’s revenue-sharing transfers prevent them from contributing to a single, unified policy objective; as a result, they do not effectively address vertical or horizontal gaps, accelerate regional economic growth, or incentivize subnational tax collection. Revenue-sharing transfers were introduced in 1978, when the central government replaced many indirect state and local taxes with a national value-added tax. The initial purpose of the transfers 1 This is also a feature of fiscal transfers in other federal republics in Latin America, including Argentina and Brazil. 4 was to compensate SNGs for the loss of own-source revenue, and the original formula distributed resources based on the amount of federal revenue collected in each state.2 Revisions to this formula in the early 1990s shifted its focus from devolution toward relative population size and own-source tax collection, effectively introducing a fiscal-equalization component into the allocation criteria. However, the new formula had only a weak equalizing effect, as it did not directly incorporate fiscal capacity or expenditure needs. In 2007, the distribution formula was revised again, and resource allocation is now based on the growth rate of gross state product (a devolutionary component), relative population size (an equalization component), and the growth rate of the state’s own-source revenues (an incentive to improve tax effort). In addition, a “hold harmless� clause was included to ensure that no SNG may receive fewer resources than it received before 2007. Due to this clause, the revised formula retains elements of both the original formula’s devolutionary component and the population-size criterion from the 1990 reform. The confused structure and inconsistent objectives of Mexico’s revenue-sharing transfers have produced a system that is neither devolutionary nor equalizing and that creates only weak incentives for SNGs to support local economic growth or strengthen own-source revenue collection. 12. The current distribution formula distributes the increase in FGP resources since 2007 at close to an equal per capita basis, which heightens its equalization effect but undermines incentives to boost economic growth or increase tax effort. Although the formula is designed to reward increases in state economic activity (proxied by federal tax revenue from each state) and encourage own-source revenue collection at the state level, improvements in economic activity or tax effort above the national averages generate only a small increase in revenue-sharing transfers, as population size determines the bulk of the allocation.3 The distribution of the post-2007 increase in FGP resources on a simple per capita basis further dilutes the fund’s incentive structure. When the rising trajectory of FGP transfers renders the hold- harmless clause obsolete, zeroing out the weight of the pre-2007 distribution, the FGP will come close to being an equal per capita transfer for most states. 13. Conditional transfers were established in 1998 to finance the decentralization of basic education services, healthcare, and social infrastructure. Unlike traditional conditional grants, which are typically used to encourage SNGs to allocate more resources to certain public goods or services than they otherwise would, Mexico’s conditional transfers were created to finance the decentralization of staff and facilities from the federal government to SNGs. Consequently, the allocation criteria for transfers under Ramo 33 reflect supply-side considerations rather than demand, and focusing on inputs rather than outcomes weakens the efficiency of subnational service provision. Moreover, because the criteria for conditional transfers were defined when services were decentralized, they have a strong inertial component, and cost considerations are often absent from their allocation formulas. 14. Mexico’s most important capital transfer is distributed through the Transfer Fund for Social Infrastructure (Fondo de Aportaciones para la Infraestructura Social, FAIS). The FAIS represents 2.5 percent of federal shareable revenue and is by far the largest financing source for capital investment at the municipal level.4 The FAIS could also be described as an equalization transfer, since its resources are allocated to states and municipalities based on poverty indicators. However, the FAIS formula combines 2 A compensatory fund provided additional resources to states that received less than the average per capita revenue-sharing transfer. 3 Herrera, 2012. 4 FAIS transfers are disbursed through two funds: (i) the Fund for State Social Infrastructure Transfers ( Fondo de Aportaciones para la Infraestructura Estatal, FISE), and (ii) the Fund for Municipal Social Infrastructure Transfers (Fondo de Aportaciones para la Infraestructura Social Municipal, FISM). FISM accounts for about 88 percent of FAIS transfers. 5 state and municipal indicators, which attenuates its equalizing effects, as a poorer municipality in a relatively wealthy state would receive fewer resources than a similar municipality in a poorer state. 15. Ramo 23 is a more flexible budget line consisting of discretionary transfers, the size and beneficiaries of which are determined by the federal government and approved by Congress. Resources from Ramo 23 are distributed to subnational governments through specific funds, including regional- integration projects, the metropolitan fund, the culture fund, and a border-areas fund, among others. Decentralization agreements between federal ministries and state governments finance the implementation of federal programs at the state level. III. The Equalization Impact of Intergovernmental Transfers in Mexico 16. Mexico exhibits substantial regional disparities in socioeconomic development. Differences are especially large between the industrialized north and center-north (where annual per capita GDP averages MXN 181,100 or US$9,800) and the less-developed south (where annual per capita GDP averages MXN 67,800 or US$3,300). In 2016, the country’s wealthiest federative entity, Mexico City, had a GDP per capita of MXN 335,200 (US$18,000), six times that of the poorest state, Chiapas, at MXN 54,000 (or US$3,000) (Figure 2). Per capita GDP in Nuevo León, one of Mexico’s richest states, was close to the level of Poland, while Chiapas had a per capita GDP similar to that of Honduras or East Timor. Poverty and other socioeconomic indicators follow a similar pattern: 68 percent of people in extreme poverty live in just six of Mexico’s thirty-two states: Chiapas (16 percent), Veracruz (14.2 percent), Oaxaca (11.6 percent), State of Mexico (11.3 percent), Guerrero (8.8 percent), and Puebla (6 percent). Figure 2: Per Capita GDP by State, 2016 (MXN thousands) Source: INEGI 6 17. Socioeconomic disparities are also reflected in the own-source revenues of SNGs. Because wealthier states tend to have much larger tax bases and more efficient tax administrations, their own- source revenue capacity often far outstrips that of their poorer counterparts. SNG in the north and center- north collect an average of 4.5 times as much revenue per capita as SNGs in the south. Mexico City collects 15 times more state and municipal revenue per capita than Chiapas, Oaxaca, and Tlaxcala and over 10 times more than Guerrero and Zacatecas, the two states with the lowest levels of own-source revenue (Figure 4). Figure 3: Per Capita Own-Source Revenue by State, 2016 (MXN) Source: INEGI 18. On balance, the distribution of revenue-sharing transfers is devolutionary rather than equalizing and tends to magnify disparities in state revenue. The amount of federal revenue-sharing transfers received by each state broadly correlates with its level of economic activity. The six states that contribute the most to national GDP—Mexico City, State of Mexico, Nuevo León, Jalisco, Veracruz, and Guanajuato—are also the states that receive the largest shares of revenue-sharing transfers (Figure 4). However, the correlation between economic activity and revenue-sharing transfers is inconsistent. The economic contributions of Mexico City, Nuevo León, Coahuila, and Campeche are much greater than the shares of revenue-sharing transfers they receive. Conversely, the share of transfers received by State of Mexico, Veracruz, Puebla, Chiapas, Oaxaca, and Guerrero far outweigh their contributions to national GDP. 7 Figure 4: Shares of GDP and Revenue-Sharing Transfers by State, 2017 18.00 16.00 14.00 GDP Participaciones 12.00 % of Total 10.00 8.00 6.00 4.00 2.00 0.00 Nayarit Michoacán Estado de México Campeche San Luis Potosí Hidalgo Oaxaca Yucatán Aguascalientes Guerrero Zacatecas Colima Tlaxcala Coahuila Sinaloa CDMX Tabasco Puebla Sonora Querétaro Chiapas Morelos Nuevo León Baja California Sur Veracruz Guanajuato Chihuahua Baja California Durango Jalisco Quintana Roo Tamaulipas Source: SHCP, INEGI. World Bank staff calculations 19. Due to their mostly devolutionary nature, the distribution of revenue-sharing transfers is regressive. SNGs with higher levels of per capita GDP receive more in revenue-sharing transfers than do their poorer counterparts (Figure 5). While conditional transfers and other transfers (i.e., Ramo 23 subsidies and individual agreements) are all regressive, revenue-sharing transfers and other transfers are substantially more regressive than conditional transfers (Figure 5, Figure 6, and Figure 7). As all of its components are regressive, the overall intergovernmental transfer system is also regressive (Figure 8). Figure 5: Per Capita GDP and per Capita Revenue- Figure 6: Per Capita GDP and per Capita Sharing Transfers, 2000-2016 Conditional Transfers, 2000-2016 9.5 9.5 9 9 Log per capita Participaciones Log per capita Aportaciones 8.5 8.5 8 8 7.5 7.5 y = 0.5425x + 1.887 y = 0.2245x + 5.6934 R² = 0.5815 R² = 0.1453 7 7 10 11 12 13 14 10 11 12 13 14 Log per capita GDP Log per capita GDP 8 Figure 7: Per Capita GDP and per Capita Other Figure 8: Per Capita GDP and per Capita Total Transfers, 2000-2016 Transfers, 2000-2016 9.5 10.5 9 8.5 10 Log per capita Other Transfers Log per capita Total Transfers 8 9.5 7.5 7 9 6.5 6 8.5 5.5 y = 0.8536x - 2.8196 y = 0.4263x + 4.172 R² = 0.2062 R² = 0.3556 5 8 10 11 12 13 14 10 11 12 13 14 Log per capita GDP Log per capita GDP Source: World Bank staff calculations 20. Due to their regressive nature, intergovernmental transfers do little to reduce the horizontal fiscal gap and enable SNGs to attenuate regional socioeconomic disparities. Per capita spending levels among state and local governments in the northern states are more than 25 percent higher than those of their southern counterparts. Public spending in Mexico City is 35 percent higher than public spending in Michoacán, Guerrero, and State of Mexico (Figure 9), and disparities in per capita spending are reflected in access to basic services. Between 25 and 40 percent of the population of southern states lacks access to health services, compared to less than 15 percent in the northern states (Figure 10). The share of houses without access to sanitation services in northern states is less than 5 percent, while in southern states this share reaches 15 percent (Figure 11). And 25 to 30 percent of children in the southern states experience educational-attainment gaps, versus just 10 to 15 percent of children in the northern states (Figure 12). 9 Figure 9: Total SNG Spending per Capita Figure 10: Share of the Population Lacking (thousands of 2013 MXN) access to Healthcare Services (%) Figure 11: Share of the Population Lacking Figure 12: Share of Children with Educational- Access to Sanitation Services (%) Attainment Gaps (%) Source: INEGI, SHCP, World Bank staff calculations 21. As noted above, a high degree of dependence on intergovernmental transfers weakens incentives for SNGs to increase their own-source revenue collection, and Mexican SNGs collect only an estimated 30-50 percent of their potential tax revenue. The international evidence shows that SNG tax effort is negatively correlated with the share of intergovernmental transfers in their total revenue.5 A 5 See Blanco (1996) for Brazil, World Bank (2016) for Peru, and World Bank (2018) for Argentina. 10 stochastic frontier analysis using panel data for all 32 federative entities from 2000-2016 confirms that the negative correlation between SNG tax effort and the share of intergovernmental transfers in total revenues holds true for Mexico (Table 2 and Figure 13).6 Table 2: Intergovernmental Transfers and Fiscal Effort, 2000-2016 (1) (2) (3) (4) Fiscal Fiscal Fiscal Fiscal effort effort effort effort Total Transfers (% total revenues) -0.03 *** -0.03 *** 0.01 0.01 Participaciones (% total revenues) 0.03 0.03 0.02 0.02 Aportaciones (% total revenues) -0.10 *** -0.10 *** 0.02 0.02 Others (% total revenues) 0.00 0.00 0.02 0.02 Debt (% GDP) 0.05 0.00 0.05 0.05 Debt ( % IDL) 0.00 0.00 ** 0.00 0.00 Living in urban areas (% population) 0.00 0.00 *** 0.00 ** 0.00 ** 0.00 0.00 0.00 0.00 Intercept 0.39 *** 0.38 *** 0.54 *** 0.54 *** 0.04 0.03 0.05 0.05 Observations 448 448 448 448 F test 6.52 6.23 9.16 9.18 F test all ui =0 415.56 440.78 314.44 340.41 Overall R-sq 0.09 0.11 0.26 0.27 *p<0.10, **p<0.05, ***p<0.01 Source: World Bank staff calculations Figure 13: Own-Source Revenue, Fiscal Capacity, and Tax Effort among SNGs (2003-2016 average) $12,000 70% $10,000 60% Own Revenues Pesos Per Capita 50% $8,000 % Own Revenues Effort 40% $6,000 30% $4,000 20% $2,000 10% $0 0% CDMX México Colima Nayarit Durango Michoacán Veracruz Zacatecas Chiapas Chihuahua Tabasco Yucatán Oaxaca Campeche Puebla Tlaxcala Guerrero Nuevo León Querétaro Quintana Roo Baja California Sur Guanajuato Jalisco Morelos Sonora Tamaulipas Aguascalientes Coahuila San Luis Potosí Sinaloa Baja California Hidalgo Own Revenues Fiscal Capacity Effort (right) Source: World Bank staff estimates 6 Estimating tax capacity and tax effort is a three-step process. First, stochastic frontier analysis is used to estimate each state’s tax capacity. The tax effort of each state is then calculated as the ratio between observed and potential tax collection. Finally, the tax effort is regressed to account for the share of transfers in total revenue. 11 22. In summary, Mexico’s intergovernmental transfers system does not effectively devolve fiscal resources to the regions where they were collected, mitigate regional disparities in the provision of public goods and services, promote policies that accelerate local economic growth, or encourage greater tax effort among SNGs. The devolutionary component of the transfer system outweighs its equalizing component, and thus transfers do little to offset horizontal fiscal imbalances or improve the overall equity of public spending. However, the equalizing component is strong enough to undermine the relationship between local economic growth and the amount of revenue-sharing transfers received by each SNG, preventing transfers from effectively encouraging pro-growth policies. Meanwhile, the large share of transfers in state revenue weakens incentives for SNGs to boost own-source revenue collection. IV. Options for Reform 23. Transfers will continue to play a central role in Mexico’s fiscal federalism framework, but significant reforms will be necessary to improve their equity and enhance their efficiency. While Mexico’s enormous vertical and horizontal fiscal gaps will continue to make large-scale intergovernmental transfers necessary for the foreseeable future, policymakers have considerable scope to reform the intergovernmental transfer system and strengthen its contribution to Mexico’s development priorities. 24. The transfer system’s most pressing challenges stem from its excessive complexity and inconsistent objectives. The proliferation of small revenue-sharing and conditional earmarked transfers has needlessly complicated Mexico’s fiscal federalism framework, while the multiple objectives and conflicting allocation criteria used by individual transfer mechanisms weakens their impact. Different elements of the system were designed to achieve the incompatible goals of devolution and equalization, and while on balance federal revenue raised in wealthier regions is redistributed to poorer ones, the distribution formula for the largest revenue-sharing transfer falls far short of achieving interregional equity. Moreover, the allocation criteria for most conditional transfers are based on supply-side factors rather than demand, which negatively affect expenditure efficiency, and the heavy dependence of SNGs on federal transfers creates perverse incentives that nullify the impact of mechanisms designed to encourage own-source revenue generation. In this context, the overarching objective of the reform process should be to align Mexico’s intergovernmental transfer system most closely with international best practices by simplifying and streamlining transfer mechanisms to advance clear and consistent objectives. 25. The authorities could balance the conflicting objectives of revenue devolution and fiscal equalization by reorganizing the current array of revenue-sharing transfers into a single, pure revenue- sharing transfer and a single, pure equalization transfer. Creating two consolidated transfer mechanisms could help ensure the transparent devolution of fiscal resources while enabling policymakers to more precisely calibrate the degree of horizontal equalization. 26. The formula for the pure revenue-sharing component could be designed to ensure that transfers reflect the amount of tax revenue collected within the jurisdiction of each SNG (i.e., a derivation or origin-basis criterion). Most of the federal revenue that finances revenue-sharing transfers comes from tax bases that are under the exclusive purview of the federal government (e.g., value-added tax and corporate income tax), but each state’s contribution to GDP can be used to approximate its contribution to federal revenue. Basing revenue-sharing transfers on state-level economic activity encourages state governments to adopt policies that boost local economic growth, but the convoluted nature of the current formula obscures this incentive. States may also be more likely to collaborate with 12 the federal government to improve tax administration if a share of the additional revenue will be transferred directly to SNGs. 27. The distribution criteria for a pure equalization transfer should capture differences in expenditure needs and own-source revenue capacity across SNGs, the two fundamental dimensions of the horizontal fiscal gap. Expenditure needs (not actual expenditures) are the amount each SNG would need to spend to provide a standard level of public services based on its population size, local socioeconomic conditions, and local public administration costs. Revenue capacity (not actual revenue collection) is the ability of each SNG to raise own-source revenues based on an average level of administrative effort, adjusted for the size of the government’s assigned tax bases and the funds received through revenue-sharing mechanisms. 28. The design of an equalization transfers should be informed by a clear equity standard, which may or may not target full fiscal equality. Equalization transfers need not eliminate subnational differences in expenditure needs and fiscal capacity altogether. Instead, a more limited equalization transfer may seek to reduce such differences to a socially tolerable level. For example, the distribution formula could be designed to achieve a minimum level of service delivery across all subnational jurisdictions. The more ambitious the minimum level, the larger the equalization transfer. 29. Several well-established methodologies can be used to estimate SNGs’ expenditure needs and fiscal capacities. The two most commonly used techniques involve: (i) estimating the amount of spending per capita necessary to finance a core suite of programs and activities in each jurisdiction, or (ii) using a weighted index of relative need to establish an average or common standard of service delivery and then dividing that amount by the number of beneficiaries. The second methodology uses data on population, geographic area, poverty rates, and other factors that determine differences in spending needs to create weighted indexes, which are then multiplied by the total population of each jurisdiction to obtain the adjusted population (for the purpose of the expenditure-need estimate). However, this methodology does not necessarily account for differences in the cost of service provision across jurisdictions. Whichever technique is used, expenditure needs must always be estimated after accounting for the impact of conditional transfers, which finance a large share of the public services delivered by Mexican SNGs. 30. Estimating fiscal capacity is often more straightforward than estimating spending needs. The concept of “need� is inherently normative, whereas fiscal capacity can be defined within objective parameters. While estimating fiscal capacity tends to be more direct, it also requires more data on tax bases, and the necessary information may not be available in all cases. The international literature provides multiple estimation methodologies, of which stochastic frontier analysis may be best suited to the Mexican context. 31. Once the expenditure needs and fiscal capacity of each jurisdiction have been estimated, the difference between the two can be used to approximate the horizontal fiscal gap. A positive fiscal gap indicates that local expenditure needs exceed local fiscal capacity, and additional resources will be needed to make up the difference. By contrast, a negative fiscal gap implies ample fiscal capacity to meet local spending needs, and thus no additional resources are required. Once the size of the gap has been established, policymakers must determine how much of the vertical gap will be covered by the transfer system and define the criteria for allocating these funds among recipient SNGs. The criteria should ensure that resources are transferred only to SNGs with positive fiscal gaps. 13 32. The criteria for distributing equalization transfers should incentivize expenditure efficiency and encourage (or at least not discourage) local revenue collection. Equalization transfers should reflect the revenue capacity of SNGs and not their actual revenue levels, which could encourage subnational authorities to neglect revenue collection. Similarly, assessments of expenditure needs should be based on clearly defined quality standards and costs rather than actual expenditure levels, which could encourage inefficient public spending. In Mexico, the numerous small revenue-sharing transfers currently used to promote subnational tax collection could be consolidated into the equalization transfer, as its allocation criteria would directly incentivize revenue effort by SNGs. 33. The allocation formulas for conditional transfers could be reformed by replacing the current supply-side criteria with demand indicators. For example, conditional transfers for education7 could be distributed on a per-student basis after adjusting for expenditure needs and the local cost of service delivery. In addition, an equalizing factor could be designed to accelerate the convergence of education indicators across states. 34. Prospective adjustments to other conditional transfers could yield additional efficiency gains, but designing effective reforms will require a more detailed analysis of each mechanism. Merging FISE with FISM8 could reduce the fragmentation of conditional transfers. FISE represents a very small share of FAIS, and municipal governments implement most social infrastructure projects, which mainly deliver local benefits with limited spatial spillovers. Consequently, merging the two transfers could yield a modest efficiency gain without weakening their overall impact. However, the authorities would need to simplify the two-stage distribution rule to prevent two municipalities with identical indicators from receiving different amounts merely because they are located in different states. 35. Finally, Ramo 23 subsidies and individual agreements could be reformed to prevent the discretionary and nontransparent distribution of federal resources. For example, Ramo 23 subsidies and individual agreements could be consolidated into FAIS to establish a comprehensive capital-transfers category as part of the fiscal federalism framework. 7 I.e., the Education Payroll and Operational Expenses Fund (Fondo de Aportaciones para la Nómina Educativa y Gasto Operativo, FONE). 8 See supra note 4. 14