The idea that checks and balances in the government—such as legislative oversight of the executive branch—can reduce corruption is intuitive, but quantitative empirical evidence for or against this hypothesis is relatively scant. Moreover, the effect of a separation of powers on the extent of corruption may depend on whether the same political party or faction controls both branches of government, or whether different factions control the legislature and the executive. Indeed, some legal scholars have argued that the true separation of powers is not between branches of government, but rather the political parties in the government, and that the traditional view of the separation of powers—ambition counteracting ambition—only works if different branches are controlled by different political parties. But the likely effect of such partisan separation on corruption is not entirely clear: If the legislature is controlled by a party or coalition opposed to the party that controls the executive branch, this could mean increased legislative oversight and lower corruption, but alternatively, increased opposition may simply drive the executive to bribe the opposition to go along with his or her agenda, leading to more corruption.
Carlos Varjão and I investigate this the question empirically in our recent working paper, “Political Opposition, Legislative Oversight, and the Performance of the Executive Branch.” We focus on municipal governments in Brazil, which are particularly suitable for this sort of study for a number of reasons: there are many municipalities with a similar overall government structure, there’s a wealth of data on various forms of corruption (mainly embezzlement, procurement fraud, and over-invoicing) from Brazil’s public audit reports, and there’s considerable variation in both the level of corruption and the political control of the branches of the municipal governments. Our findings are striking and unambiguous: increased representation of the political opposition in the local legislature is associated with more legislative oversight of the executive, less executive branch corruption, and better public service delivery.
First, some background context: Brazilian municipal governments are led by an elected mayor (the head of the executive branch) and a separately-elected city council (the legislative branch). The city council usually has nine members, and is led by a city council president chosen by a majority of the council. The mayor manages public services (such as health clinics, schools, and public infrastructure), while the city council is responsible for reviewing and auditing municipal spending. In addition, if the city council concludes that something has gone awry, the city council president can authorize a special team of city councilors—an investigative commission—which has the judicial investigative power to subpoena documents and interrogate witnesses. Although an investigative commission can’t directly prosecute a mayor for criminal offences, the commission can forward its findings to prosecutors and public media, and can even recommend to the rest of the city council that the mayor be impeached.
So in principle the city council has powerful tools at its disposal to investigate and deter corruption by the executive branch. But if the city council is controlled by politicians allied with the mayor, the council might not be interested in investigating the mayor properly, or forwarding any findings to the relevant authorities. After all, city council members who are part of the mayor’s coalition have little to gain from being tough on the mayor, especially given that if the mayor is revealed to be corrupt, the mayor’s allies will likely suffer at the polls in the next election. By contrast, city counselors who are politically opposed to the mayor do have an incentive to keep the mayor honest, and denounce him if he isn’t.
Indeed, one of our main findings is that when the opposition increases its representation by one seat—and in particular when this shift of one seat is enough to flip control of the city council from the mayor’s coalition to the opposition coalition—legislative oversight goes up, corruption goes down, and public services improve. More specifically:
- Increasing opposition representation on the city council—primarily when this increased representation is enough to shift in control of the council from the mayor’s coalition to the opposition—increases the likelihood that the city council opens an investigative commission by a whopping 55 percentage points. So just in terms of oversight activity, we see clear evidence that political opposition matters a lot.
- Does this increased oversight translate into lower corruption? The answer is a resounding yes. According to the data from Brazil’s random municipal audits, the effect of flipping one city council seat from mayoral to opposition control—especially when doing so changed majority control of the council—is a dramatic decrease in corruption(a 0.85 standard deviation decrease, for those who are statistically minded). Interestingly, of the three forms of corruption we looked at, the one that was affected by opposition control of the legislature was embezzlement (situations where the mayor or his subordinates transfer public funds to private bank accounts, without receipts or other documents proving that funds were spent in an appropriate way). In contrast, we didn’t find that increases in political opposition were clearly associated with lower levels of procurement fraud or over-invoicing. Perhaps the explanation is that it is much easier for the city council to ensure that all payments are supported by proper documentation than to identify corrupt procurement practices.
- The reduction in corruption associated with opposition control of the city council has a direct and tangible effect on the welfare of the people who live in these municipalities. We find that much of the decrease in embezzlement associated with opposition control of the city council was in the healthcare sector, and we further find that this reduction in embezzlement of public healthcare funds translated into large improvements in the performance of municipal clinics, as measured by waiting times, the likelihood that doctors and dentists will be present to see patients, and, most importantly, measurable health outcomes. The reduction in healthcare sector embezzlement associated with opposition control of the city council caused a large reduction in infant mortality rate (3.4 fewer deaths per 1000 births) among the uneducated poor—the group most likely to benefit from government healthcare services.
These results may seem surprising, especially to those familiar with local governance in Brazil and Latin America more widely, because these local city councils are generally not viewed as having much power, and policy differences at the local level of government are small. Yet even in this setting, the adversarial nature of the partisan system does its work well.
What broader lessons might we learn from all of this? These are a few:
- First, and most fundamentally, democratic institutions¾ which decentralize power and provide checks and balances on individual behavior¾ do work to decrease corruption, but only if a formal separation of institutional power is coupled with partisan competition.
- Second, because the absence of strong opposition in the legislative branch means that corruption is less likely to be uncovered at the local level, national-level anticorruption agencies and prosecutors should focus their investigative efforts on areas where the political opposition does not have a strong presence. In places where there is strong opposition, the legislative branch should be capable of doing the needed investigations and forwarding findings to relevant authorities.
- Third, the legislative branch needs tools that will empower it to hold the executive branch accountable. Legislative oversight capacity varies greatly across nations, with one World Bank study finding that in a sample of 39 countries (25 of them OECD countries), only 25% had a specialized budget research organization associated with the legislature. Our analysis suggests that strengthening legislative oversight, by ensuring that legislative branches have both the authority and capacity to investigate the executive branch, could be an effective way to reduce corruption.