Study shows banks that have good working relationships with their customers reduce loan defaults.

CAMBRIDGE, Mass. — You’ve probably seen advertising campaigns in which banks describe how much their customer relationships matter to them. While such messaging might have been cooked up at an ad agency, it turns out there is some truth underlying these slogans.

As a newly published study co-authored by an MIT professor shows, strong working relationships between bankers and clients reduce the likelihood of loan delinquencies and defaults, at least in the context of an emerging economy.

Using propriety data from a large bank in Chile, the study finds that when loan officers go on leave, their clients in good standing — often small businesses — increase their probability of becoming delinquent on loans by almost 22 percent. For already-delinquent clients, the probability of default on loans increases by 18 percent.

At the same time, there is a 5 percent reduction in the approval rate for loans among those clients. That means banks are both issuing fewer loans and suffering more defaults when relationships between individual bankers and clients are interrupted.

Intriguingly, the study indicates that this aggregate phenomenon is due to actions by both the bank and its customers. The banks in question are probably not retaining or using the “soft information,” or informal knowledge, that absent bankers have collected about their clients. And the clients themselves seem more willing to take their business elsewhere, even if it means defaulting, when their bankers disappear.

“It is a two-way channel,” says Antoinette Schoar, the Michael M. Koerner Professor of Entrepreneurship and a professor of finance at the MIT Sloan School of Management, and a co-author of the paper reporting the findings. She attributes the results to “a type of loyalty building up between the client and the loan officer” that “could not only affect how good the loan information is … but also the willingness of the client to default.”

The findings also indicate that we should not think of credit ratings as simple, immutable predictors of client types that are unaffected by the ways that banking relationships are conducted.

“This interplay between how the bank treats the customer and how the customer reacts to the bank therefore has an impact on the bank’s repayment history,” Schoar explains. “It actually has an impact on the observed credit quality of the customer.”

She adds: “A financial system that is more effective in how it treats customers endogenously creates better creditors.”

Why loan officers leave

The paper, “Do Relationships Matter? Evidence from Loan Officer Turnover,” has been published in the journal Management Science. The authors are Schoar and Alejandro Drexler of the Federal Reserve Bank of Chicago.

To conduct the study, the scholars obtained detailed transaction data from BancoEstado, Chile’s largest lender to small businesses; that data involved 187,000 borrowers from 2006 to 2008. The study took advantage of the fact that loan officers leave banks for different reasons, some of which can be anticipated by the bank more easily. For instance, when a loan officer takes leave due to a pregnancy, the bank is better able to anticipate and replace the interrupted professional relationship, compared with instances when the loan officer suddenly takes sick leave.

When the loan officer takes a maternity leave, the paper notes, that banker’s clients “show no propensity to go to a bank outside of the current relationship.” However, clients of officers who take sick leave are 2 percent more likely to get a loan from another bank, which is 13 percent higher than the probability that the average client in the study will do so. And clients of bankers on sick leave are about 20 percent less likely than the average client to get a new loan from the same bank.

“When the loan officer has time to prepare the client for the leave, all these negative effects are much mitigated,” Schoar observes. “Ultimately it seems like this relationship and information is transferrable, given enough time.”

About 53 percent of the loan officers in the study were women, and they each had an average of 339 active clients.

More testing, better policies

As Schoar acknowledges, the study’s findings stem at least in part from the particular circumstances of its location and time, since detailed information about clients might be easier to obtain in other countries. On the other hand, the research fits with other types of studies across a variety of countries, including the U.S., pointing to the idea that person-to-person relationships influence outcomes in finance.

In Schoar’s view, the policy applications of the findings are relatively clear: Banks can do business more effectively and efficiently by instituting better practices to smooth over interruptions in banker/client relationships. To be sure, that also means banks would probably have to invest a bit more in those new practices. But as Schoar notes, that is a question that can at least be subjected to direct cost-benefit studies, which ultimately could produce more beneficial relationships between bankers and customers.

“The financial industry has to do more testing of the optimal mix between the use of technology and human interaction,” Schoar says.