The world is filled with bad guys. It seems, therefore, like it would be pretty easy to simply put together an army, say “go get ‘em,” and watch as the bad guys are slowly but surely eliminated.

Things aren’t that simple, however. At the very least, getting the bad guys takes time and effort. At the very worst, it could cost you your life. Once a soldier’s paycheck has cleared, he no longer has an incentive to find and kill bad guys. People usually aren’t going to take on a lot of extra risk when there are high costs and low benefits.

It seems, though, like there’s another simple solution: pay for performance. You could reward soldiers and their commanders with extra pay, extra vacation time, or more career advancement if they increase their bad guy body count. In any introductory economics class, you learn the law of supply: holding everything else constant, increasing the rewards for something will elicit more of that thing. Want your soldiers to kill more bad guys? Pay more for each dead bad guy.

Once again, it’s not that simple. Incentives are fickle, and they produce unintended consequences. In some cases, those unintended consequences are great: “Paris gets fed,” exclaimed the 19th century French economist Frederic Bastiat, even though that’s not people’s explicit goal when they’re growing, harvesting, and hauling grain. In other cases, those unintended consequences are awful. Price controls are classic examples: rent control creates shortages, price gouging laws create shortages, and minimum wages make it harder for poor people to get jobs.

A recent study by a group of economists showed that when the Colombian military introduced “high-powered incentives”—pay-for-performance, where performance was defined as “dead bad guys”—they got more than they bargained for. In their paper “The Perils of High-Powered Incentives: Evidence from Colombia’s False Positives,” the economists Daron Acemoglu, Leopoldo Fergusson, James Robinson, Dario Romero, and Juan F. Vargas found that high-powered incentives meant more “false positives,” or “innocent civilians...killed and misrepresented as guerillas” (p. 1). The paper appears in the August 2020 issue of American Economic Journal: Economic Policy.

The economists rely on the Data Bank on Human Rights and Political Violence assembled and maintained by a Colombian Jesuit NGO, the Center for Research and Popular Education. They hypothesize that municipalities with weaker judicial institutions and where more brigades are commanded by colonels who have the most to gain career-wise by racking up more confirmed kills are more likely to have more false positives.

This, in fact, is what they find. In their words, “removing the colonels reduces false positive cases and casualties by about 6.5 percent...while getting rid of judicial inefficiency has a slightly larger impact” (p. 30). This effect, they estimate, is larger than the effect on “true positives” (confirmed kills of actual bad guys).

So what happened in Colombia? The authors give us a brief history beginning with the election of Alvaro Uribe in 2002. After he was elected, the Colombian military almost tripled in size (from 160,000 to 430,000 soldiers). Uribe’s Democratic Security Policy allocated funds to fight guerillas and paramilitary groups, and on November 17, 2006, Army Directive 29 created lucrative rewards for catching and killing guerillas and members of paramilitary groups. On May 5, 2006, Presidential Decree 1400—which was revoked on May 14, 2007— “rewarded army members or DAS functionaries with up to 12 times their monthly salary for participating in successful operations of ‘national importance’ against the insurgency” (p. 9).

The incentives were structured poorly. They quote the University of Chicago economist Canice Prendergast: “One of the first rules of pay for performance is that you never offer pay for performance in circumstances where a person both diagnoses and cures the problem.”

When you think about it for a minute, the problem becomes clear. Colombian military leaders had incentives to kill more bad guys, but they were also in charge of figuring out who the bad guys were in the first place. They had incentives, therefore, to mis-identify and even intentionally misrepresent civilians as enemy combatants. They might get more bad guys, but they’ll also catch and kill more people who aren’t bad guys. These are more than mistakes and “collateral damage.” The authors quote a Human Rights Watch report documenting examples of soldiers killing civilians and then planting weapons on them to make it look like they are the enemy.

As the economists argue, “the high-powered incentives without a strong accountability system can backfire” (p. 7). In addition, they analyze the effect of these incentives on the quality of judicial institutions and find that judicial institutions tended to erode in municipalities where colonels faced high-powered incentives that led to false positives. This is especially interesting because of the importance of the judiciary to long-run economic performance. In their textbook, the economists Tyler Cowen and Alex Tabarrok list “a dependable legal system” as one of the “ultimate causes” of economic progress. A legal system that treats people arbitrarily or that privileges a select few—like ambitious and unscrupulous military leaders—at the expense of the many generates uncertainty and confusion that thwarts economic progress.

During the period the economists studied, innocent people lost their lives. Why, one might ask, were the military leaders on the ground so careless or even so cruel? For an important part of the answer, we need to look no further than their incentives.

I thank Sarah Estelle for bringing this paper to my attention.