Loan Officer Incentives, Internal Rating Models, and Default Rates
Tobias Berg, Manju Puri, Jörg Rocholl
Review of Finance, Volume 24, Issue 3, May 2020, Pages 529–578, https://doi.org/10.1093/rof/rfz018
Manipulation of information has been at the center of a wave of recent investigations into fraudulent bank behavior. Examples include misrepresentation of debt-to-income ratios in securitized mortgages, rigging of LIBOR and FX rates, or the recent wave of investigations into money laundering activities. These cases have two things in common: first, manipulation was conducted by employees down the line such as loan officers, traders, or bank advisors. Second, all these cases involved manipulation of hard information, i.e., information that is in principle objective, quantitative and verifiable.
In this paper, we analyze how incentive systems within banks drive manipulation behavior by loan officers. Using almost a quarter million retail loan applications, we show that loan officers who face volume-based incentives significantly manipulate ratings even in settings where ratings are computed using hard information only. Manipulation is widespread across loan officers, with low-performing loan officers manipulating more toward the end of the year. These incentives have a first-order effect on bank profitability, reducing return on equity by 1.5 percentage points.
We conclude that reliance on hard information does not overcome loan officer agency problems. Even with hard information, banks and regulators must either ensure that incentives are not skewed, or have robust processes and procedures in place to detect misrepresentation of hard information by employees. We also discuss the role of risk management, compliance, and top management in alleviating fraudulent behavio