Electoral reforms can cut down on corruption, fraud, and vote-buying, and make elections fairer. New research on an electoral reform in Brazil suggests that they might do more than that — they might even save lives.
“Voter-Buying, Electoral Reform and Health Outcomes in Brazil” is a new working paper by Ridwan Karim, an assistant professor in the Lazaridis School of Business and Economics at Wilfrid Laurier University (though the study was written while he was at the University of Toronto). It traces the effects of an electoral reform enacted in 2007 that made vote-buying in Brazil harder, cracking down on a common practice that frequently altered election outcomes.
Karim finds that in the wake of the reform — which required immediate re-registration by voters in some municipalities — incumbents running for local office lost more often. Even more important, in the three years following the reform, politicians implemented new policies that spent more money on public health-care services. And the improved health services made a difference: “We observe a drop in infant mortality, an increase in prenatal visits by expecting mothers, and a rise in birth weight,” Karim told me.
How do electoral reforms produce better health care outcomes? The paper suggests that reforms with an eye toward making elections freer and fairer can have the effect of making politicians more responsive to their constituents. That, in turn, can lead to policies that are important to the broader public — including, in this instance, health-care policies that led to life-saving outcomes.
Brazil has been in the news recently for its national election of Jair Bolsonaro, a right-winger who has courted the idea of a military dictatorship and disparaged indigenous communities, and who rose to power partially through Brazil’s disillusionment with corruption in politics. This research looks at a reform from a decade ago, so it doesn’t speak to the recent election. It’s focused on the effects, on a small scale, of making fraud more expensive.
Those small-scale effects can be a big deal, though. Tackling corruption in fragile democracies is a concern worldwide. If the results in Brazil can be replicated elsewhere, then that’d be an encouraging sign we just need to make corruption more expensive.
Designing a study to test the effects of electoral reform
By 2007 — the period the study looks at in Brazilian politics — it had become common for voters to be registered in jurisdictions they didn’t live in. Politicians would pay these non-resident voters to fraudulently participate in elections. Nearly 200 mayors had been charged for engaging in fraudulent relocation of voters from 2004 to 2008, according to Brazil’s Federal Prosecutor’s Office.
That year, Brazil’s Federal Electoral Court cracked down, requiring a full re-registration of voters in districts that had a suspiciously high registration rate. That meant these fraudulent voters were unregistered, and candidates faced an electorate that was harder to bribe. (Voters in the jurisdiction tend to be more expensive to bribe and more likely to vote their conscience, Karim told me.)
It’s hard to measure the effects of fairer, freer elections. Even if infant mortality drops and health outcomes improve after an electoral reform, it’s tough to be confident that the electoral reform caused it. After all, maybe the general social forces that led to electoral reform also led to improvements in health outcomes. Maybe health outcomes were improving as a result of long-standing, totally unrelated trends.
The design of Karim’s study, called a regression discontinuity design, allows us to tell how much of the difference is really due to the 2007 electoral reform, which forced re-registration in order to end the common practice of paying people who lived elsewhere to vote in a municipality.
Fraudulent voter registrations (usually of real people who lived elsewhere) were common across Brazil, but Brazil’s courts didn’t crack down everywhere. They ordered a complete re-registration of voters if a municipality met a specific requirement: “a ratio of registered voters to total estimated population in 2006 greater than 80%.” Voter rolls in those suspicious counties got thrown out, obliging everyone to re-register — and that made it harder for fraudulent registrations to happen.
That’s the perfect setup for a regression discontinuity design, which works by comparing regions that just barely qualified for the intervention — having a number of registered voters just over the threshold — to a region that just barely failed to qualify by having just slightly too few registered voters. The changes in outcomes between these regions after the electoral reform is likely to be a result of the electoral reform.
Karim found sizable differences. In regions where the reform went into effect, elections were more competitive. More incumbents lost. There wasn’t a statistically significant increase in overall spending, but there was a statistically significant increase in spending on public health.
Why public health? It’s a government service that poor households care about a lot, Karim observes, so politicians find it an effective way to win over voters. Finally, he finds that the changes in public health spending got real results for the people in the affected regions: “a 6.6 percent increase in prenatal visits [by] expecting mothers, a 0.5 percentage point (15 percent) decrease in the incidence of low birthweight births, and a 0.1 percentage points (5.3 percent) reduction in infant mortality rates.”
Karim ran some other checks to verify that these results were a consequence of the electoral reform. For example, vote-buying through fraudulent registration works better in smaller municipalities, since you don’t need to buy as many votes to swing the election, so we’d expect that the effects of the reforms would be more pronounced in smaller municipalities. Karim finds exactly that.
We might find that effects would be larger in municipalities that had previously had been particularly corrupt; Karim observes that pattern, too.
How does cracking down on fraud cause more spending on health care?
There are several plausible ways that cracking down on electoral fraud would produce politicians who support more public spending on health care. For example, maybe cracking down on fraud means that different candidates are elected — candidates who care more about health care.
Or maybe resetting the voter rolls, as the Brazilian courts required, means that politicians have a different pool of constituents and pursue the policies that their new constituent base wants. Alternately, maybe cracking down on fraud makes vote-buying no longer as effective a path to remaining in office, so politicians devote their energy to health care instead of voter-buying.
Karim examines the evidence and is more confident in the last effect: “clientelistic practices became relatively more expensive for incumbent politicians, leading them to shift expenditures towards public health services to maximize their re-election probabilities.”
In other words, politicians are trying to get reelected through whatever means work best. Before reform, voter buying did the trick; after it, providing services to one’s constituents seemed the more effective route. “This reform is ratcheting up the cost of voter buying”, Karim told me.
That might be very cynical indeed, but in a way it’s actually encouraging. It means we don’t need a new crop of virtuous politicians and we don’t need to make fraud impossible. We just need to make fraud more of a hassle than honest service provision, and politicians will take the path of least resistance — in this case, providing health care.
What can we learn from this for democracies elsewhere?
This is only one study, it’s still going through the peer-review process, and it looks only at one specific electoral reform — re-registration in Brazil in response to widespread fraud there. The mechanism for fraud in these Brazilian municipalities — re-registering out-of-jurisdiction voters and then busing them in to vote — only works in small municipalities in countries with a lot of corruption, so it’s not clear whether other electoral reforms would have similar effects. We’ll want to see similar work looking at other electoral reforms elsewhere before we try to generalize too broadly.
And there are other reasons to be skeptical of this study’s story. Other social science research has found that voters struggle to attribute policy outcomes to the political actors who were actually responsible. And requiring voters to re-register disenfranchises some legitimate voters — Karim checked that the benefits aren’t a result of reduced enfranchisement, but perhaps it has some other costs elsewhere, hard to detect. The effect also relies on local politicians having a lot of influence over health care access. Simply put, we can’t necessarily expect most electoral reforms to produce effects like these.
But if Karim is identifying the mechanism correctly, different kinds of electoral reform might have similar results elsewhere. Combatting gerrymandering, for example, could make elections more competitive, which might similarly motivate politicians to provide more of the services their constituents want.
There’s a lot of interest in improving political institutions to combat corruption and make poor countries richer. Often, it’s hard to identify interventions that really work. Brazil’s voter reform, though, seems to have been reasonably successful at making for more democratic and competitive elections.
Combating fraud — at least somewhat — turned out to be as simple as making it easier to win elections by providing services than by buying votes. That’s fascinating, and might be suggestive of where our priorities should lie in other democratic reforms.
Do you ever struggle to figure out where to donate that will make the biggest impact? Or which kind of charities to support? Over five days, in five emails, we’ll walk you through research and frameworks that will help you decide how much and where to give, and other ways to do good. Sign up for Future Perfect’s new pop-up newsletter.