CHICAGO – Hardly a day goes by without a settlement between a bank and a US government agency or regulator. The latest one is between Bank of America and the government-backed mortgage underwriter Fannie Mae: $3.55 billion in cash; $6.75 billion in repurchases of bad mortgages that Countrywide, later acquired by BoA, sold to Fannie; and an additional $1.3 billion in cash to resolve Fannie’s servicing issues with BoA.And yet, as in most settlements, BoA did not admit any fault. Was BoA tripped up by bad luck or sloppiness? Or did it and other financial institutions behave with intent to defraud investors?As a result of the settlement, we will never learn the truth (at least the version that is established beyond a reasonable doubt in open court). In each settlement, the investigations are stopped, and whatever has been found is sealed.Fortunately, academic research may be able to shed some light on this important matter. In a recent paper, the economists Tomasz Piskorski, Amit Seru, and James Witkin compare the characteristics of securitized mortgages as they were disclosed to investors at the time of sale with these loans’ characteristics as they were recorded in the banks’ proprietary databases. The data cover non-agency residential mortgages, which are not the same as those targeted by the settlement, and the data come from a pool of banks, so we do not know which banks were present. But their data are similar enough to enable us to learn a lot about what was going on.If the massive wave of defaults on securitized mortgages was purely the result of bad luck, we would expect that the characteristics reported to investors would not differ from those recorded in the banks’ databases. This is especially true of characteristics that are relevant to default risk, such as whether the borrower was an owner or an investor, or whether there was a second lien on the property. In fact, the authors find that more than 6% of mortgage loans misreport the borrower’s occupancy status, while 7% do not disclose second liens.Of course, any database contains errors. Are these errors large enough to be worrisome? And how can we be sure that the banks knowingly misrepresented this information, rather than that they merely were sloppy in reporting it?The authors provide some interesting evidence in this context. They show, for example, that the misrepresentation is correlated with higher defaults down the line: delinquent payments on misreported loans are more than 60% higher than on loans that are otherwise similar. Thus, the errors do not seem to be random, but purposeful.What the authors do not find is also interesting. The degree of misrepresentation seems to be unrelated to the incentives provided to the top management and to the quality of risk-management practices inside these firms. In fact, all reputable intermediaries in their sample exhibit a significant degree of misrepresentation. Thus, the problem does not seem to be limited to a few bad apples, but is pervasive.This makes a solution more difficult to achieve. After all, if the financial industry’s leaders are misleading investors, it is the culture of the entire industry that needs to be changed.Unfortunately, these giant settlements are unlikely to make a difference. They are paid by shareholders and taxpayers (most of these settlements are tax-deductible), leaving the people who were ultimately responsible not only to walk away, but to walk away very rich.For example, Angelo Mozilo received almost $470 million during his tenure as CEO of Countrywide Financial. As far as I know, he paid nothing of BoA’s $11.6 billion settlement cost. Mozilo did settle with the SEC, but on different charges (insider trading). And he paid only $47.5 million of the $67.5 million settlement, owing to a $20 million indemnification agreement that was part of his employment contract.Why have Mozilo and others been allowed to pay a relative pittance and avoid criminal proceedings? A few prominent criminal convictions have sent a powerful signal in the fight against insider trading. If the target is an industry-wide culture of deception, shouldn’t the criminal justice system be mobilized against banks’ mortgage fraud as well?Luigi Zingales is the Robert C. McCormack Professor of Entrepreneurship and Finance at the University of Chicago Booth School of Businesswww.project-syndicate.org