This article is part of the online forum: Taxation as a development tool

There is no dearth of reform proposals to raise sub-national taxation. The problem is how to remove the veto power of the localities that are potential losers

The literature on fiscal decentralization has traditionally emphasized the benefits of assigning significant own-revenue sources¹  to sub-national governments (SNGs), on both macro-economic and efficiency grounds. Potential macro-economic advantages include: creation of sustainable fiscal space for the provision of local public goods and services; promotion of sub-national fiscal responsibility; and reduction of volatility and uncertainty for sub-national budgets. Efficiency gains include: closer reflection of local preferences in the level and structure of taxation, and increased political accountability of local budget authorities to their electorates.

At the same time, the literature has also recognized the significant economic, institutional and political economy obstacles to revenue decentralization. The main economic obstacles include the greater mobility of goods and factors across sub-national jurisdictions than across national ones, and often large differences in revenue-raising potential among regions and localities. Institutional obstacles relate mainly to differences in administrative capacity, and increased compliance costs for taxpayer operating in multiple sub-national jurisdictions. The main political economy obstacles are the frequent resistance of central governments (CGs) to devolving taxation powers, and reluctance of SNGs to absorb the political costs of raising own revenues.

As a result of the interplay of these factors, whose relevance and intensity varies significantly both across countries and over time, the degree of revenue decentralization also tends to vary widely. International comparisons indicate that, on average, revenue decentralization in LAC is significantly less advanced than in the OECD, and also than in comparable emerging economies in other regions. Since, at the same time, decentralization of spending responsibilities has proceeded relatively fast in the region in recent decades, significant vertical imbalances have emerged (Figure 1). These imbalances have been filled through a combination of intergovernmental transfers and sub-national borrowing, with reliance on the latter being better controlled in recent years, in the wake of sub-national debt crises in some countries during the 1990s.

However, these average trends mask substantial differences within the region:

• Across countries, with Brazil presenting a relatively high share of sub-national own revenues in total revenues; Mexico, Ecuador, Peru and Central American relatively low ones; and other countries in between;
• Across levels of government, with local governments typically, but not always, enjoying greater revenue autonomy than regional ones; and
• Across sub-national units within each level of government, reflecting disparities in both revenue-raising capacities and efforts.

Background case studies²  for the recently released IDB flagship report “More than Revenue: Taxation as a Tool for Development” analyze the main factors affecting sub-national revenue potential and efforts in selected LAC countries. They find that the revenue potential of individual jurisdictions is mostly explained by their level of development, but other factors, such as the composition of output and of the labor force, the poverty level, and the extent of urbanization and of informality, also play important roles, the last two factors being especially relevant in explaining the revenue potential of local property taxes.

The degree of sub-national tax effort is found to be influenced by both institutional factors (such as the administrative capacity of SNGs, and the levels of transparency and corruption) and political economy factors (such as the political alignment of the sub-national jurisdiction with the national ruling party). The design of inter-governmental transfers is also found to play an important role: discretionary transfers from the CG (as well as a history of repeated sub-national bailouts by the latter) tend to dis-incentivize sub-national own-revenue efforts, as they create moral hazard. In contrast, appropriately designed, formula-based equalization transfers are found to promote own-revenue mobilization in some countries (e.g., Colombia and Peru).

The composition of sub-national own revenues in LAC also varies significantly across countries and levels of government. Typically, revenue assignments to local governments are in line with good international practice, as they focus on the taxation of property (land, buildings and vehicles). In contrast, at the regional level most countries provide limited revenue authority to their SNGs. This is the case not only for unitary countries, where intermediate levels of government tend to be the result of de-concentration of national administrations rather than of real decentralization, but also for some federal ones (Venezuela and Mexico). Only in Brazil and Argentina have the states and provinces, respectively, traditionally enjoyed substantial revenue autonomy. Moreover, the main regional taxes tend to be fraught with important distortions, such as cascading; extensive tax expenditures and predatory competition; and other weaknesses in design and enforcement.

These considerations highlight the need for substantial reforms in the sub-national revenue systems of LAC. These reforms can be grouped under two broad headings:

1. Assigning appropriate new tax handles to sub-national (especially regional) governments

Based on the extensive existing literature on sub-national revenue assignments³, the following main options should be considered in the Latin American context:

• A sub-national PIT (or surcharge on national PIT)
  This option would work best in the context of desirable reforms of the national PIT, to increase its revenue and redistributive potential , as discussed in the above mentioned IDB book.
• A regional or local retail sales tax
  The main obstacles to the adoption of such a tax are its regressivity, and likely enforcement difficulties in retail sectors characterized by strong fragmentation and high levels of informality.
• A regional VAT
  The main advantage of such a tax is its significant revenue potential, but, as the example of the Brazilian state VAT (ICMS) shows, it also raises substantial difficulties especially in the taxation of interstate trade; and in its relatively high administration and compliance costs.
• Regional or local excises (or surcharges on national excises)
  These taxes have a relatively good revenue-raising potential, but for some of them (e.g. on tourism), the bases may be heavily concentrated in certain regions; others (e.g. on electricity consumption) affect production costs and might undermine competitiveness.

The assignment of new tax handles may be a necessary condition, but is unlikely to be sufficient to ensure increased sub-national revenues. A reform of the intergovernmental transfer system, to make it less discretionary, a tightening of borrowing controls, and avoidance of expectations of bailouts may also be necessary to induce SNGs to adequately exploit any new tax handles provided to them.

2. Improving existing sub-national taxes

Reforms of existing sub-national taxes in LAC should focus on both increasing their revenue potential and reducing some of their distortive features.

Possible steps to increase revenues include:

• Base-broadening through the elimination of exemptions and special treatments (which the case studies found to be quite widespread)
• Rate increases, in the frequent case that they cluster at the bottom of an allowed range
• Adapting the design of real estate taxes to cope with pervasive informality, e.g., by modifying the real estate tax to link it to improvements in the urban environment; and by charging the tax to occupants, even if not owners
• Modernizing property cadasters and updating them more frequently; and
• Other improvements in local tax administration, including systematic exchange of information with the regional and national ones.

The main obstacle to the adoption of efficiency-enhancing reforms is likely to be the fact that they, even if revenue neutral for the sub-national government level in question as a whole, could entail losses and gains for individual SNGs, necessitating compensating measures and/or spreading their effects over an extended period of time. Two significant examples in this respect are reforms of the turnover tax (ingresos brutos) in Argentina and of the state VAT (ICMS) in Brazil.

One of the background studies mentioned above⁴  analyzes the option of replacing the Argentine turnover tax with a provincial surcharge on the national VAT, or with a retail sales tax. It finds that, although preferable on some grounds, the first option would face two main obstacles, namely the problems created by the taxation of interstate trade, and the fact that, if the tax were levied at a uniform rate calibrated to ensure overall revenue neutrality, it would entail significant gains and losses for individual provinces. Given the difficulties of administering effectively a retail sales tax, the study argues for the maintenance of the turnover tax, but excluding the sales of primary and intermediate industrial goods from its base and adjusting rates to offset the resulting revenue loss.

The Brazilian ICMS, albeit very productive in terms of revenues, raising over 7 percent of GDP, is also fraught with well-known substantial distortions⁵. Since indirect taxes levied by the federal government are also fraught with significant flaws (including a degree of cascading), ideally reform of the ICMS should be accompanied by a revamping of those taxes, and by a consolidation with the ISS (the municipal tax on services), to create a dual (federal and state) VAT, with the state portion on interstate trade collected on an origin basis, but redistributed on a destination basis.

This is indeed the thrust of a number of reform proposals put forward in recent years, including by the federal government. However, progress in this respect has been stymied so far by the failure to form a consensus on how to deal with losses and gains that the shift to a destination basis would entail for individual states.

1. Own revenues refer here to those for which SNGs enjoy substantive autonomy in setting at least the rate structure. Accordingly, revenues shared with higher levels of government on a derivation or other basis are not considered own revenues, but transfers.

2. Background studies were prepared for Argentina, Bolivia, Brazil, Colombia, Mexico, Peru and Venezuela. Some have been already issued as IDB working papers, and all are expected to be published in a separate volume on sub-national revenue mobilization and reforms in LAC. Several of the studies utilize stochastic frontier techniques (see Fenochietto and Pessino, 2010) to estimate sub-national revenue potential and effort.

3. This literature emphasizes that in principle sub-national taxes should exhibit a number of desirable characteristics: a relatively low mobility of the tax base; minimization of distortions, and of risks of tax exporting, or predatory tax competition; a relatively even distribution of the tax base across the national territory; a significant revenue- raising potential; low sensitivity to cyclical fluctuations and other exogenous shocks; relative ease of administration; low compliance costs for taxpayers; and political viability. All potential sub-national tax handles fall short to varying degrees of one or more of the desirable criteria, and trade-offs need to be made among them, taking into account the specific country’s economic, institutional and political context.

4. Artana, Cristini and Moskovits et al. (2012)

5.  See Ter-Minassian (2012) for an overview of literature on ICMS distortions and on reform proposals.

Watch the video for this article