In 1899, Thorstein Veblen described a type of good that is more lusted after the more expensive it is (think Ferraris). And in 1968, the economist Gary S. Becker theorized that criminals perform cost-benefit analyses just like everyone else: What are the odds of getting caught, and what’s the potential payoff? These two frameworks have lived out vibrant lives in academic journals, high-school textbooks, and college lecture halls, but, as they’re ostensibly unrelated, they’ve rarely been put in conversation with one another.

A study put out this month in Oxford Economic Papers does just that, in an effort to come up with a more nuanced understanding of the relationship between inequality and violence. There’s a good amount of research from all over the world that suggests that places with pronounced income inequality are more likely to have high rates of violent crime, a finding that makes intuitive sense: the wider the socioeconomic gap, per Becker's 1968 model, the more gains potential criminals perceive. (Not to mention, the more frustrated poorer criminals will be with society.)

But this new study takes into account the fact that it’s hard to gauge a stranger’s income—the information that potential criminals are acting on comes in the form not of pay stubs, but of proxies like expensive cars and fancy clothes. “A neighbor’s income and bank account balance,” the authors write, “are by no means perfectly observable for academic researchers or individuals considering committing crimes.” Luckily for academic researchers, though, there exist reams of data about crime and consumer spending, two topics the federal government cares deeply about.

The authors, the University of Oklahoma’s Daniel Hicks and UC Berkeley’s Joan Hamory Hicks, compared data from the Bureau of Labor Statistics’s Consumer Expenditure Survey to the FBI’s data on crime rates. (These two data sets aren’t perfectly representative of the U.S. as a whole, but they’re the most thorough available.) In order to zero in on the effects of conspicuous consumption, the researchers had to rank the consumer-spending survey’s line items according to a visibility index. (This led them to track the usual suspects—clothing, jewelry, cars—but also some less-obvious expenditures such as alcohol, furniture, and eating out at restaurants.) Having sliced the data in this way, Hicks and Hicks (whose relation is unclear) found that the link between conspicuous consumption and high crime rates is much stronger than the link between income inequality and crime.

That said, when spending is visible, the rates of only certain types of crimes tend to spike. Theft and vandalism, interestingly, aren’t significantly more present, but murder and assault are. These findings actually take a bit away from Gary Becker’s hypothesis, seeing as a visibly luxurious car apparently isn't likely to inspire theft. Instead, this study adds to what’s called “strain theory,” which is another way of making sense of criminal behavior. Strain theory suggests that when poorer people perceive inequality, they feel less of a commitment to social norms and in turn come to view crime as more acceptable. The key insight the Hicks’s study provides is that when potential criminals are giving up on social expectations, they’re doing so based on information that’s visible, not information that’s password-protected.

A 2008 study found that in São Paulo, where crime rates are relatively high, people tend to put more money into savings (and thus spend less on luxury goods, which can be stolen) than in other cities. So, on some subconscious level, it seems that people have known about this finding all along.