Bank Mergers and Crime?

March 17, 2006
Officer Krupke, the Merger Made Me Do It
By FLOYD NORRIS
DO big bank mergers lead to more crime? And is that a reason to toughen antitrust enforcement?
The idea sounds unlikely, and admittedly the effect is a small one. But a persuasive new study published in the April issue of The Journal of Finance finds a relationship.
"This does not explain crime patterns in the United States, but it does indicate that finance matters for things you might not have expected," said Mark J. Garmaise, an assistant professor of finance at U.C.L.A.
Mr. Garmaise, who conducted the study with Tobias J. Moskowitz, a professor of finance at the University of Chicago, notes that studies around the world have shown that finance matters when it comes to economic growth, and that prosperity affects the level of property crime, such as car thefts and burglaries.
What they did was get extensive data on commercial lending — for such things as apartment buildings and stores — in 11 states. They had the exact location of each piece of property, and information on the mortgage lender, and thus could measure how competitive the lending market was by looking not at countywide data but at lending within 15 miles of each project.
Then they looked at large bank mergers, where each bank had assets of more than $1 billion, and at areas where the two banks had previously been competitors.
Here is what they found: If bank concentration increases by 10 percent, you can expect higher interest rates on loans, and smaller loans. And perhaps, although the relationship is not statistically significant, fewer loans will be made.
Property crime rates increase 1.8 percent over the rate in areas where there is no increase in banking concentration.
There is, as would be expected, no change in violent crimes. Murder and rape are not economic crimes.
The effects seem to fade away in three years, presumably because competition returns as other banks take note of the higher margins, but three years is a long time.
That can be particularly true in areas that face the risk of decline. With less lending, areas can seem older and less attractive.
Through statistical measures, the authors show that the causation does not run the other way, that declining areas do not lead to bank mergers. By excluding smaller mergers, they also tried to be certain that mergers were not caused by signs of decline in a bank's lending area.
"We're not saying that all mergers are bad," Mr. Garmaise said in a telephone interview. "But it is an implication of the paper that bank mergers have effects on construction, income levels and crime."
To make it worse, he said, "the crime effects are much larger in poor neighborhoods." The message, he added, "is that regulators should consider variables like that in deciding whether to approve a merger."
A 10 percent increase in bank concentration, measured the way antitrust regulators measure it, can occur from many mergers. Imagine a very competitive market with six lenders, four with 20 percent market shares and two with 10 percent each. If the bottom two merged, concentration would rise 11 percent. It would increase 22 percent if a larger bank bought a smaller one.
Another way to look for evidence that the pattern works is to see what happened to crime in states that allowed statewide branch banking after barring it earlier, thus making competition easier. That has led to an average 1.5 percent decline in burglaries.
That does not mean that criminals should be allowed to plead that mergers caused them to violate the law. But it does show that a lack of competition in banking can hurt even those who never seek a loan.