In microfinance circles, people tend to be fond of asking the question, “Does microfinance work?” Over the last decade, countless studies have attempted to answer this question by studying the net impact of microcredit on the lives of borrowers. Yet, these impact studies don’t necessarily tell us much about the nuances of how organization-level factors might influence the final impact of microcredit. NYU Economist Hunt Alcott and FAI Affiliate Sendhil Mullainathan have a recent paper that notes that the MFIs that participate in rigorous impact evaluation aren’t like MFIs in general. But there is a very important deeper level of analysis that is important. Little attention has been paid to how individual groups and actors shape the nature of microfinance services – that is, how the behaviors of funders, bank executives, and front-line loan officers might fundamentally alter the delivery and outcome of microlending.
Here at FAI, we’re not just interested in financial products but in how systems and people interact to make the right (or wrong) products available (or unavailable). For example, why do loan officers behave as they do? What incentives affect a loan officer’s job performance and how? How does the relationship between the loan officer and the client influence the borrowing and repayment process?
Two new working papers presented at the 2012 Innovations for Poverty Action (IPA) Conference, shed light on these important questions. In “Rewarding Calculated Risk-Taking: Evidence from a Series of Experiments with Commercial Bank Loan Officers,” Shawn Cole, Martin Kanz and Leora Klapper provide experimental evidence on the relationship between compensation practices and loan officer behavior. They investigate how loan officers make lending decisions for small retail loans in India. In particular, they focus on officers’ decisions regarding uncollateralized loans made to entrepreneurs with very limited credit histories – in other words, the types of loans that require loan officers to rely on their judgment and expertise to make a decision. The authors study these lending decisions against the backdrop of four common incentive schemes: a flat wage; a bonus for originating loans; a bonus for originating only successful loans; and a more complicated scheme that penalizes officers for loans that become delinquent, but rewards them for successful loans.
The researchers find that compensation structures have a significant impact on the performance of loan officers. Loan officers who are rewarded for the amount, rather than the quality, of loans they originate tend to approve a much higher share of the loans they evaluate. On the other hand, loan officers who face penalties for poor-performing loans tend to be more conservative in their approval rates. The nature of compensation also affects loan officers’ subjective risk assessment, with loan officers who receive bonuses for approving performing loans being more likely to have inflated risk assessments. Finally, the authors look at how deferred compensation practices affect the behavior of loan officers. (Any incentive scheme that rewards loan officers on the performance of loans must inherently involve deferred compensation because the outcome of the loan is not initially known.) The authors find that incentive schemes that delay payment by three months reduce the effort expended by loan officers by fifty percent!
The authors’ findings provide institutions that are seeking to align the behavior of front-line staff with the mission and goals of the organization with some interesting things to think about. The authors hope to use these initial findings as a base from which to explore other dimensions of the loan underwriting process, including how the role of individual characteristics of loan officers – such as age, gender and work experience – might interact with incentive structures. The authors propose that the “one-size fits all approach” compensation practices might not be optimal. Instead, they suggest that banks might consider allowing officers to select their incentive schemes from a menu of options.
In an interesting corollary study, Antoinette Schoar explores how the relationship between loan officers and their clients affects the behavior of borrowers. In “The Personal Side of Relationship Banking,” Schoar conducts a randomized experiment in India to test whether the communication strategies of loan officersaffect loan repayment rates. She finds that borrowers who are regularly called by a loan manager have higher repayment rates and display greater customer satisfaction than borrowers who do not receive follow up calls or only receive calls when they are delinquent. Schoar finds that these results are both statistically and economically significant—meaning that the cost to the bank of having loan officers make frequent phone calls is more than made up for the cost
Taken together, these two studies have important implications for the field of microfinance. Our thinking and discussion needs to evolve beyond simply asking “whether microfinance works” and looking at products and contracts. We also need to be paying attention to all the actors in the chain—from investors to management to loan officers to customers to figure out how to make sure optimal products are consistently delivered in optimal ways. That’s the “last mile” of making microfinance work.