Aspiring bank robbers, take heed. New statistical analyses of confidential bank data suggest that mountains of riches aren't in your future, but that a jail cell is.

"The return on an average bank robbery is, frankly, rubbish," wrote the authors of a new article about the economics of British bank robberies in Significance, a quarterly statistics journal published by the American Statistical Association and Royal Statistical Society. "[I]t is so low that it is not worth the banks' while to spend as little as £4,500 [$7,000] per cashier at every branch on rising screens to deter [bank robberies]."

These and other conclusions emerge from a confidential data set obtained by economists Neil Rickman and Robert Witt at the University of Surrey, and Barry Reilly at the University of Sussex. The trio spent months negotiating with the British Banker's Association, or BBA, to obtain confidential records detailing 364 bank heists that occurred over three years in the U.K.

The data reveal how many robbers pulled off with each heist, if they were armed, what security measures were present, how much they escaped with, whether or not they were caught, bank location (and thus proximity to police stations), and more. Such detailed bank robbery data is not available in the U.S. where, by contrast, if U.S. banks even record most of these details, they are lost in anonymity in the FBI 's quarterly reports on bank robberies.

"There hasn't been a lot of work done on this problem because most data sets only tell you if a robbery occurred. The size of a robber's takings is sensitive to banks, so they keep it confidential,"  Rickman says. "We're very lucky to get the data we have."

Game of numbers
About 80,000 robberies occur in the U.K. each year, and of these less than 0.5 percent were bank robberies. Averaged across 10,500 bank branches in the region, the odds of a bank robbery at any particular branch in recent years work out to about one in 100. "One element to being a successful bank robber is not only walking away uncaptured, but the amount of money you walk away with," Rickman says. "Walking away with £10 [$15], for example, isn't something I'd consider successful."

By averaging the work of the best criminals with the worst—and every crook in between—from 2005 through 2008, the data paint a bleak picture for criminals eyeing a future career in larceny.

The data reveal that the average U.K. bank robbery consists of 1.6 robbers and nets £20,330.50 ($31,600) per heist, with a standard deviation of £53,510.20 ($83,000). Assuming an equal share, the average take was £12,706.60 ($19,700) per robber, per heist—roughly equivalent to a coffee shop barista's annual salary.

Wielding a gun increased the average haul by £10,300.50 ($16,000), as did adding more robbers to a raid. But going alone netted the average robber more money, because the amount gained by increasing a heist party did not outpace the hit taken by splitting an individual robber's spoils.

Dangerous career choice
The amounts are not chump change, Rickman noted, but bank robbing is risky business. Roughly 33 percent of U.K. bank heists end with no robber earning anything, and about 20 percent of raids end in capture. The odds of arrest get worse as a robber attempts more heists. On a robber's fourth heist, for example, the odds of capture compound to 59 percent. "Somehow, I expected most bank robbers to be doing much better than what the data actually show," Rickman says. "Would I pick up a firearm to share 20,000 quid with another person? Probably not."

What did the average U.K. bank robber take home annually? Criminals themselves may not know because Rickman says there's "no data on the criminals" associated with the heists—not even how many exist or the average number of heist attempts each year.

Yet economic models assume that criminals act rationally if they perceive the rewards outweigh the risks of their actions. To that end the authors did a back-of-the-envelope calculation to roughly quantify what amount might convince a robber to give up his or her life of crime. Factoring in the per-heist, per-robber haul amount, failure rate and capture rate, that number, called the expected net benefit, ended up being £33,545.40 ($52,100)—roughly 125 percent of the average U.K. journalist's annual salary.

"Bank robbers [are] not living in mansions popping champagne—or most of them are not, anyway," Rickman says.

Robbed of data
But not all bank robbers are created equal. A select few methodically plan their work whereas the majority give little thought to minimizing their risks and upping their hauls, says economist Giovanni Mastrobuoni at the University of Turin's College of Carlo Alberto in Italy.

Mastrobuoni specifically took issue with the article's lack of attention to professional bank robbers, who presumably raked in most of the £7.4 million ($11.5 million) robbed from U.K. banks between 2005 and 2008. "The article suggests, for example, that fast-rising security screens mean less money for robbers," Mastrobuoni says, referring to one of the authors' conclusions that the presence of fast-rising security screens—expensive bulletproof screens that rapidly rise up between a lobby and a teller window at the push of a button—reduced a robbery's success by one third. "But I'd argue that the dumb robbers target banks with fast screens while professional ones scope out the place thoroughly before robbing," Mastrobuoni says. "Since they are better in all dimensions, they get more money."

But the researchers' access to the data, although unprecedented, didn't come without catches. "One condition of access to the data was that we did not analyze different banks themselves," Rickman says. Such limitation bars more meaningful analyses, yet smart bank robbers could use them, for example, to find prime targets for their next heist.

Another problem in examining the professionals is a lack of information about them. "That requires two sets of pretty sensitive data to be collected and matched: banks and criminals," Rickman says. With the BBA having a lock on the bank data and the police a hold on criminal data, he notes, his "hunch is that there is no easy way to match these in the U.K."

Economist Hedayeh Samavati of Indiana University–Purdue University Fort Wayne, who studies U.S. bank robbery economics, says the new work underscores the need for more and better data on the matter. She notes that the enterprise is worth the risks because, ultimately, taxpayers and bank customers "end up paying the bills" for bank robbers' actions in public services, bank fees and more.

"This is a sensitive issue for banks, so there isn't much usable data or data sufficient for people to investigate this issue. It's a major problem," Samavati says. "But I think it's worth the costs, both social and otherwise."