The electoral bias: the political economy of subnational transfers in Brazil, Mexico, Chile and Colombia

Xabier Gainza, as part of his Regional Studies Association Membership Research Grant, debates the electoral bias in the distribution of investment in Brazil, Mexico, Chile and Colombia, and its implications for territorial cohesion. The aim is to determine if electoral concerns drive the allocation of investment from the central or state government to the local level, rather than efficiency or spatial equity goals.


Introduction
Since the late nineties, most Latin American countries have developed subnational transfer programs to improve the provision of social infrastructure and reduce arbitrariness in the distribution of public goods. Earlier, political parties relied on local level spending to cement their electoral strategy, assigning funds where they should obtain the highest returns in the ballot. However, as decentralization progressed, many countries implemented subnational transfers to encourage the autonomy of local governments and limit parochialism. Particularly important in this regards was the introduction of distribution formulas as they were supposed to disseminate funds based on observable measures of area characteristics.
Although transfers were a major institutional reform against government discretion, in the recent past a burgeoning research on distributive politics has documented how politicians target resources to obtain electoral advantage (Golden & Min, 2013). This literature considers that the desire to retain office motivates politicians, and consequently, they will use intergovernmental fiscal transfers to win, exercise, and retain power (Bonvecchi & Lodola, 2010, p. 179). So, if transfer still leaves room for discretion, two main questions come to mind: what is the best electoral tactic, diverting resources to strongholds or favouring disputed areas? Does the tactic change depending on countryspecific conditions?
This short article introduces some of the issues addressed in research aimed at determining whether transfers to local governments in Mexico, Brazil, Chile and Colombia remain mediated by electoral concerns. In particular, we explored if national governments systematically favoured aligned mayors in the distribution of funds, and what political factors explained such dysfunctional institutional performance. Our research considered panel data on the main investment transfer and we focused on these four countries because they hold different settings in terms of local level autonomy and the party system, so we could test whether tactics differ depending on institutional conditions. In what follows, we summarize the main findings.

The political economy of subnational transfers
A large body of research has documented how incumbent politicians deviate resources to support their fellow candidates. It is however, contested about what strategy yields the greater advantage: targeting the municipalities where aligned mayors dominate, or where co-partisans face stronger competition. In other words, if political machines should concentrate subnational transfers in core or swing constituencies. Golden and Picci (2008) argue that the answer to this question depends on the electoral system and the strength of political parties. In proportional electoral systems like the ones in Europe and Latin America, every vote is worth the same, thus organizations have incentives to shower core areas because they are more secure than swing districts. By contrast, municipal elections hold a majoritarian system logic in the sense that what it matters is to win mayoral offices, not total ballot. Consequently, electoral machines place higher value in disputed municipalities because they make the difference for spreading partisan power throughout the territory.
These predictions depend, nonetheless, on political organizations' capacity to impose their strategy over candidates. Local party leaders often try to attract resources for their bailiwicks. When it comes to municipal elections, it may well happen that strong politicians are able to attract resources for their constituents, in spite of being in the interest of political machines to favour swing areas. Table  1 summarizes the expected outcomes depending on the electoral system and the power of parties to impose their interests over free-whiling candidates.
National elections (proportional system logic) Local elections (majoritarian system logic) Strong party party strongholds disputed municipalities Weak party party leader bailiwicks core municipalities There are few issues that make one procedure prevail over the others. The first is which government tier can claim the credit for transfers, i.e. will constituents reward the national or the local government when they see a new infrastructure investment in the municipality? If the central government holds credibility, then it can rely on transfers to pursue its own outcomes in national elections and the electoral strategies would recreate a proportional system logic (the left column in Table 1). However, we believe there are reasons to assume that intergovernmental transfers provide credit spillovers to mayors because funds are administered by municipalities and used to finance basic infrastructure that can easily be claimed by local politicians. Municipal elections would hence reproduce the expectations found in a majoritarian system (the right column in Table 1).
Another concern refers to the determinants of party strength. In principle, the local electoral geography is much more disputed and fragmented than national politics: the local arena is the breeding ground for small parties, personalist candidacies and empowered contenders exploiting their personal appeal, in spite of flying the national party banner. One may expect, thus, lower levels of party control in municipal policy-making than in national politics. There are, however, institutional constraints that help disciplining nominees. For instance, in some countries electoral rules limit the number of times elected officials can head the government. Under these conditions, politicians have incentives to follow the party guidelines rather than to satisfy their constituents because their career depends on their loyalty towards national political machines.

A glance at the electoral geography of Brazil, Mexico, Colombia and Chile
Following the discussion above, we deem explanatory a brief description of the party systems and the local electoral geographies in Brazil, Mexico, Colombia and Chile. Although the four countries feature competitive elections in a multiparty setting, they differ in terms of the party system and organizations' strength to control local policymaking. In Brazil and Colombia, vote fragmentation and volatility compel parties to rely on post-electoral coalitions. At the local level, elections are rather contested and organizations find hard to discipline politicians. In Chile, competition for national ballot is much more stable because parties join into electoral coalitions. However, the panorama splits at the municipal level, as parties that gather into coalitions for presidential elections do present their own alternatives. As a result, it is difficult for national organizations to command and control municipal disputes.
In sharp contrast to Colombia, Brazil and Chile, in Mexico three parties (PRI, PAN, PRD) dominate national and local politics. They account for 90% of the total vote and national organizations have been successful in retaining control over resources and nominees. Important in these regards is an electoral rule that bans consecutive re-elections. This constraint weakens the incumbentconstituency connection and helps disciplining local politicians, who find incentives for following the organizational guidelines if they want to make progress within the party ranks (Greene & Sánchez-Talanquer, 2018).

The electoral bias of transfer distribution
Do national governments over-finance aligned mayors to favour their re-election strategy? What is the preferred tactic, benefiting core or swing constituencies? In spite of distribution formulas, our estimations show that assignments in the four countries still respond to an electoral logic: in Brazil and Colombia, aligned mayors receive 19% extra capital transfers per capita, whereas in Chile they obtain, on average, 10% supplementary investment. Mexico is the only country where partisan links between mayors and the federal government have not reported benefits to aligned municipalities.
Our findings confirm that funds are distributed across partisan lines, but there are certain particularities in the way the electoral tactic is displayed. In Mexico, political parties ahead of the federal government support their local strongholds facing stronger competition. In the theory sketched above, we argue that targeting swing municipalities is the preferred strategy to extend partisan power across the national territory, but it is contingent on organizations' capacity to discipline local leaders. In Mexico, strong political machines are capable to control local candidates, largely thanks to an electoral rule that bans consecutive re-election. Consistent with this predictions, our estimations confirm that national political machines are able to command a swing strategy and favour their fellow candidates that win by a narrow margin.
On the contrary, in Colombia, Brazil and Chile, transfers are concentrated on safer municipalities where aligned mayors win by a wide margin. This result is explained by the difficulties facing coalition governments to command a swing strategy. In coalition governments, parties do not control local actors and incumbent politicians have to respond to a plethora of diverse, sometimes contradictory, interests. Under this scenario, organizations face an agency problem that complicate the aim of distributing resources for the benefit of aligned mayors confronting stronger competition. Political machines are not capable to enforce their interests and get trapped in the second-best option of diverting transfers towards safer areas.

Conclusion
The implementation of subnational transfers in many Latin American countries was a major institutional reform to encourage the autonomy of local governments, improve the social infrastructure and reduce arbitrariness in the provision of public goods (Diaz-Cayeros, Estévez, & Magaloni, 2016). Nevertheless, our work shows that electoral motivations do yet play a role in the distribution of collective resources. We have found evidence of aligned mayors being systematically over-financed, in spite of distribution rules that should reduce parochialism. What strategy to follow depends on political parties' capacity to command their strategy. In countries with empowered organizations, parties target aligned mayors facing stronger competition because this way they can extend territorial power. But when there are several parties in coalition, such a strategy is not possible to accomplish. There are plenty of bottom-up demands and particularistic interests; hence, the central government has no choice but to feed strongholds.