Principal investigator: Pedro Jorge Holanda Alves
Article title: THE POLITICAL RESOURCE CURSE
Article authors: Fernanda Brollo; Tommaso Nannicini; Roberto Perotti; Guido Tabellini;
Location of the intervention: Brazil
Sample size: 1202 municipalities
SectorEconomic Policy and Governance
Main variable of interestCorruption
Evaluation method: Experimental Evaluation (RCT)
The Problem of Politics
Imagine a situation where a country discovers a new source of wealth or more funds are transferred from a general government to a local region. These resources would generate unexpected growth in government revenue, but would they be beneficial to society? According to the resource curse theory, the answer would be no, since, paradoxically, countries with abundant natural resources tend to have lower economic growth due to an inability to manage these resources and market mechanisms that appreciate the real exchange rate.
According to extensive literature, one of the reasons for economic stagnation is the malfunctioning of governmental institutions. This malfunction is linked to aspects related to corruption, which are defined in the literature as a complementary effect: higher unexpected revenues induce greater corruption and attract lower-quality individuals to politics.
Brollo et al. (2013) analyze how an unexpected increase in government revenue in Brazil affects politicians' decisions to become corrupt or not. In light of these (theoretical) results, the authors used a near-ideal scenario to assess the empirical implications for Brazilian municipalities. To do this, they used the Municipal Participation Fund (FPM), Brazil's main constitutional transfer fund, as a way to find the real causes of how additional revenue affects the quality of politicians. But how would this be possible?
The Implementation and Evaluation Context
With the determination of the Brazilian Federal Constitution of 1988, it is possible to use municipalities near the Municipal Participation Fund (FPM) as an evaluation experiment to discover the effects of an increase in government revenue on corruption policies and the quality of politicians. In the case of municipalities, the main instrument for transferring funds is the FPM (representing on average 40% of municipal revenue and 75% of all federal transfers).
The information on revenues from the Municipal Participation Fund (FPM) was obtained by the authors from the National Treasury Secretariat (STN), and the population estimate from the Brazilian Institute of Geography and Statistics (IBGE) for two full terms of mayoral administrations (2001–2004 and 2005–2008). To define the scope of the experiment, the authors defined their sample as comprising all municipalities near the first 7 income brackets, represented in Table 1. The rationale is based on the fact that the larger the population size of the municipality, the smaller the change in behavior of individuals near that bracket.
Table 1 – FPM Coefficients
| Population Range | FPM coefficients |
| under 10.189 | 0,6 |
| 10.189 - 13.584 | 0,8 |
| 13.585 - 16.980 | 1 |
| 16.981 - 23.772 | 1,2 |
| 23.773 - 30.564 | 1,4 |
| 30.565 - 37.356 | 1,6 |
| 37.357 - 44.148 | 1,8 |
| 44.149 - 50.940 | 2 |
| Above 50.940 | 2,2 to 4 |
Since the objective is to evaluate how these municipalities behave regarding corruption, the authors used corruption data obtained from the General Comptroller's Office of the Union (CGU) through the Brazilian anti-corruption program. The program was launched in 2003 and consists of randomly selecting municipalities through a lottery to be audited. The database represented approximately 1.202 observations, with 802 municipalities audited between 2001 and 2004 and 400 municipalities between 2005 and 2008. Other electoral information was collected from the Supreme Electoral Court.
Policy/Program Details
Under the Brazilian anti-corruption program, auditors gather information about the municipalities inspected and, a few months after the audit, publish the reports to all levels of government, making them available on the website of the Comptroller General of the Union (CGU). With information from 2001 to the present, the report contains a detailed list of all irregularities found by the auditors, such as fraud, non-competitive bidding, excessive billing, embezzlement, etc.
Thus, the report makes it possible to detect the types of irregularities that the local public government (municipal public administration) is committing, becoming a valuable basis for assessing corruption. To define the variables to be studied, the authors state that corruption can be defined as broad and narrow. Broad corruption includes irregularities that could also be interpreted as mismanagement, while narrow corruption refers to serious irregularities that are more likely to be visible to voters.
Assessment Method
Table 1 shows that these bands are discrete within a population interval, meaning that transfers to municipalities change discontinuously between the population band intervals. In other words, due to the constitutional provision, municipalities near the FPM band receive different amounts of funds in varying proportions, reflecting the differences in population size simply because they are on opposite sides of the band.
Let's use a case as an example: municipalities with 10.150 inhabitants and 10.200 are very similar in terms of population and therefore should not have different levels of tax revenue. However, due to the income brackets, municipalities below 10.189 receive less than those above. Thus, it is possible to see that municipalities close to the FPM (Municipal Participation Fund) brackets have discontinuity in revenue, and this effect can be attributed to an unexpected increase in revenue. This effect can serve as a channel for other mechanisms, which is what will be discussed here.
The data were divided into two samples of municipalities: a small sample and a large sample, and the sample was restricted to municipalities with fewer than 50.940 inhabitants due to limitations in the audited local governments. Given the institutional configuration described, it is possible to assign a typical treatment mechanism based on a fuzzy (or discontinuous) regression design. fuzzyThis method assigns to the treatment the dependence on a variable that is continuous, but due to external factors, generates discontinuous results. For this exercise, that variable would be the population, which is continuous across the municipalities, but due to its constitution generates discontinuity for those municipalities close to the ranges.
The type of regression discontinuity method fuzzy This is implemented because there are cases of incorrect allocation around the population cutoff points. In other words, not every municipality near the range has significant growth in its revenues from the FPM (Municipal Participation Fund). Therefore, the RDD (Regime Disciplinar Diferenciado - Differentiated Disciplinary Regime) of this type... fuzzy This study only considers municipalities that experienced discontinuities in their behavior, meaning that its data only represent individuals who exhibited this behavior. The authors' objective is to discover how these municipalities were classified in the corruption results obtained from the audit data.
Main results
The results show that a 10% increase in federal transfers increases the incidence of a broad measure of corruption by 6 percentage points and the incidence of a more restrictive measure by 16 percentage points. At the same time, a 10% increase in transfers worsens the quality of political candidates challenging those seeking reelection, decreasing the fraction of opponents with at least a university degree by 7 percentage points. As a result, candidates receiving larger transfers increase their probability of reelection by 6 percentage points.
Overall, empirical evidence showed that a large amount of transfers increases the likelihood of corruption and reduces the average educational level of mayoral candidates. The empirical findings of the authors point to the existence of what we call the "curse of political resources," that is, a negative impact of unexpected resources on political corruption and political selection. The abuses detected by the audits suggest that Brazilian municipalities are a fragile institutional environment where problems of political agency are widespread.
Lessons in Public Policy
It may be that an inheritance of resources does not have the same deleterious effects in other contexts, such as societies with a long tradition of good governance and abundant social capital. However, additional resources are often allocated precisely to regions or countries with weak institutions. Therefore, it is necessary that these transfer policies be approached with caution, because even if they aim to reduce regional inequalities or generate beneficial incentives for the local population, these results, as found by Brollo et al. (2013), can produce the opposite outcome.
References:
BROLLO, Fernanda et al. The political resource curse. American Economic Review, v. 103, no. 5, p. 1759-96, 2013.
