Vote-buying—a particularly corrosive form of political corruption—is present in many jurisdictions, especially in the Global South. And not only is vote-buying itself a form of corruption, but the practice exacerbates other forms of corruption, because politicians need to raise enough money to buy enough votes to beat their opponents (who are also engaged in vote-buying), and in order to raise enough money, politicians often enter into deals with private parties who (illegally) “lend” the politician the money he or she needs to buy enough votes to win the election, and then, once in office, the politician pays back the private parties—either directly, with embezzled funds, or with inflated government contracts (see, for example, here and here).

But sometimes a candidate might worry that she won’t be able to buy enough votes from voters who actually live in the district where the candidate is running. This is especially true when competing candidates are trying to buy votes. In a competitive district, a relatively small number of votes can swing the election, so politicians have an incentive to scrounge for extra votes. This, understandably, also drives up the “price” for votes in the district. In Colombia, one way that politicians have developed to increase the pool of voters they can “buy,” and hence keep the price down, is to pay voters to illegally register in a district other than the district where they actually live. (In Colombia, as in many countries, adult citizens may only register to vote in the district where they actually reside.) So, the politician pays these voters twice—first to illegally register in another district, and then to vote in that district for the politician—and on both registration day and election day the politician will arrange for the transportation of these non-resident voters to the district where the politician is running. This practice, known as electoral transhumance, is illegal, yet there have long been concerns that it is pervasive in many parts of the country.

In October 2015, Colombia introduced a new tool to fight this sort of electoral fraud. Using so-called “big data analytics,” the authorities were able to cross-reference the National Electoral Registry databases with the System for the Identification of Potential Beneficiaries of Social Programs (known as SISBEN), and as a result of these checks, nearly 1.6 million voter registrations to vote were declared void—a large number in a country with 33 million registered voters. That seems like a big win, and a nice example of how new technologies can help crack down on pervasive corruption (here, electoral corruption). But a closer look reveals that the picture is not as rosy as it first appears: Despite its good intentions, and some positive results, this purge of the voter rolls ended up disproportionately disenfranchising low-income voters.

The main problem is that the cross-referencing of names in the National Election Registry and the SISBEN is not perfect, and was likely to incorrectly flag some number of individuals as registered to vote in a district other than the their district of residence. And because the SISBEN database focuses on potential beneficiaries of social welfare programs, the voters whose registrations were checked were disproportionately low-income voters. It’s true that voters whose registrations were voided by the cross-checks were entitled to challenge the decision, but that safeguard was insufficient, for a couple of reasons. First, the National Electoral Authority announced its decision to void invalid registries just two weeks before the election. Many voters who might have wanted to challenge the voiding of their registries might have been unable to file their claims within the tight constraints within the timeframe, and if a substantial number of them had in fact done so, it’s unlikely that the National Electoral Authority would have had the time to review and resolve those claims before the date of the election.

But even if the screening had taken place with ample time before the election, there’s a deeper problem: Because the vast majority of those whose voter registrations were declared void were those of poor people, and because poor people generally don’t have much education, it’s unlikely that the affected voters knew that they had the right to challenge the decision, knew how to file such a challenge, or had the means to challenge the decision. 

Still, the underlying idea behind this program was a good one, even if the execution in 2015 raises serious concerns about the disproportionate disenfranchisement of the poor. In the future, should Colombia (or other, similarly situated countries) decide to implement similar voter registration checks, there are a few relatively simple modifications that could ensure that efforts to combat illegal voter registration are more fair:

• First, and most straightforwardly, any decision to remove voters from the rolls due to alleged registration fraud needs to be far enough in advance before an election that voters have a realistic opportunity to challenge an inaccurate or otherwise unlawful decision. As a rule of thumb, such decisions ought to be made at least six months before the election.
• Second, and relatedly, reviews of voter registration lists should be accompanied by public information campaigns that educate citizens about the process of the review of electoral registries, as well as their right to challenge decisions to void registrations deemed unlawful.
• Third, though cross-registering voter registration lists with other databases to assess whether the district of registration matches the district of residence is a clever idea, it’s problematic when this cross-checking is only done using databases, like SISBEN, that include (potential) social welfare recipients. Instead, the cross-checks should be done with a wider variety of databases that also include residence information, such as those held by public and private pension funds, utility companies, cellphone operators, health insurance companies, and banks, among others. 
• Fourth, rather than relying only on voters to challenge decisions voiding registrations, the databases themselves, and the cross-checks, should be reviewed by independent auditors to ensure that they are not being manipulated for political gain.  

Making these changes to the process of reviewing electoral registries can help to ensure electoral integrity, while simultaneously protecting the right to vote and the right to equal treatment for all citizens, regardless of their income.