At first, the two Ph.D. students couldn't believe what the data told them: when companies in countries with high levels of theft added extra security guards, theft actually increased.
David Perez-Reyna and Enoch Hill had worked together two full years on an econometric model that was now yielding counterintuitive, even absurd-sounding, results.
Perez-Reyna is from Colombia, and Hill worked in Guatemala before graduate school. Having witnessed theft problems firsthand in those countries, where armed guards patrol nearly every business and ride along on most delivery trucks, they persisted in their work.
"We found that our initial perception was correct: if there's a lot of theft, then companies are driven to hire more security," says Hill.
"But the amount of theft kept growing even when companies were spending more on private security," Perez-Reyna explains. "So we realized there had to be another factor."
Deeper into their research, the model indicated yet another apparent anomaly: even when a country spends more money on public law enforcement, it doesn't necessarily help the overall economy. At least, not immediately.
"If a country moves a little in spending more on law enforcement, it's not a good investment," Hill says. "Poorer countries are actually worse off when they spend just a little more. The overall costs to the economy aren't recovered because they have to hire more police and build more prisons, and there will be fewer workers when more people are put in prison. Plus, the chances of getting caught remain relatively low, so not much deterrence occurs for potential thieves."
They found that it's not until law enforcement spending reaches a medium level that the graph starts turning positive. As enforcement increases, there is greater deterrence, and consequently less incarceration, thus putting less indirect pressure on the economy.
That discovery helped Perez-Reyna and Hill explain the huge variations in theft among
countries that invest in law enforcement at different levels, such as Somalia versus Guatemala versus the United States. Spending a lot more really does matter, they're finding.
Their finding flew in the face of an economic rule of thumb: effects of an economic event are greatest at the outset, and diminish gradually after that.
"Usually in [economic changes], the greatest gains come in the early stages, but not here," Hill says.
Implications for public policy
Top scholars both -- Hill* had the best academic record of any student in his first-year grad school class, and Perez-Reyna* was tied for second -- they have been fast friends for three years. Sharing an office in the economics department at Hanson Hall, they work together on research studies and play soccer on the weekends.
The model, which Hill and Perez-Reyna have updated with new data at least 10 times already, includes a look at the relationship between a firm's size and its security costs. Not surprisingly, bigger firms experienced more theft and therefore hired more guards.
Their professor, Timothy Kehoe, a prominent economist and adviser to the Federal Reserve Bank of Minneapolis, praised the project. While the two students still have work to do, he says, their project has the potential to provide valuable guidance on what levels of law enforcement are cost-effective in various business and civic scenarios.
Was it unusual for economics students to work on what, on the surface, seems to be a
"Economics, at its basic level," Hill explains, "is about how to make choices under scarcity." That means the principles of economics can be applied to any type of supply problem, even the cost of security to deter theft at manufacturing plants around the world. "Everything we've done is related to optimizing resources," he says.
Furthermore, their work has focused on macro-forces in the marketplace -- an economics issue -- rather than on specific firms or countries experiencing security problems.
Individual cases are difficult to model, Hill says. "So much is seemingly random and there's no way to capture all the factors. But across thousands of firms, you can see a pattern and write equations about what a bunch of people are likely to do."
Those patterns have helped Hill and Perez- Reyna reach some macro-conclusions on the
problem in less developed areas.
For example: "The biggest cost of theft isn't actually the theft; you have to waste so much on employees standing around with a gun and not working," says Hill. "And if people steal a lot, it lowers the wages for everyone and there's even more incentive to steal."
"But in an economy with less theft, the wages tend to be higher and there's actually less
incentive to steal," adds Perez-Reyna.
Their hope is to fine-tune the model such that it will be useful to planners. Says Hill, "Someone working on public policy might look at our paper and consider implementing it in their country."
Joe Kimball, a former columnist and reporter for the Star Tribune, now writes for MinnPost.com. He is the author of the bestselling Secrets of the Congdon Mansion.
This article includes a number of corrections from the print version. Corrections are indicated with an asterisk. (*) We regret the errors.