Repeated interactions allow lenders to uncover private information about their clients, decreasing the informational asymmetry between a borrower and his lender but introducing one between the lender and competing financiers. This paper constructs a credit-based model of production to analyze how learning through lending relationships affects monetary transmission. I examine how monetary policy changes the incentives of borrowers and lenders to engage in relationship lending and how these changes then shape the response of aggregate output. The results demonstrate that relationship lending prevails in equilibrium, smoothes the steady state output profile, and induces less volatile responses to certain monetary shocks.
An important issue in both welfare and development economics is the interaction between institutions and economic outcomes. While welfarists are typically concerned with how these variables contribute to overall wellbeing, empirical assessments of their joint contribution are limited. Development economists, on the other hand, have focused extensively on whether institutions cause or are caused by growth yet the relevant literature is still rife with debate. In this article, we use a notion of distributional dominance to tackle both the measurement of multivariate welfare and the evaluation of inter-temporal dependence without hindrance from the mix of discrete (political) and continuous (economic) variables in our data set. On the causality front, our results support the view that institutions promote growth more than growth promotes institutions. On the welfare front, we find that economic growth had a positive impact from 1960 to 2000 but declines in institutional quality over the earlier part of this period were sufficient to produce a decline in overall wellbeing until the mid-1970s. Subsequent improvements in institutions then reversed the trend and, ultimately, wellbeing in 2000 was higher than that in 1960.
Is risk-taking ever a privately optimal response to agency problems within banks? In a model where borrower types only matter for safe projects, I show that the answer to this question depends on the nature of the agency problem - that is, whether loan officers are hired to screen or whether they are hired to both screen and monitor. Incentivizing screening favors no risk-taking but involves a non-monotone relationship between performance and compensation. This non-monotonicity undermines incentives to monitor so, when both screening and monitoring are important, the bank instead prefers a strategy which pushes low types into risky projects. That selected risk-taking emerges under impediments to non-monotone compensation is also illustrated in an environment with rank-order tournaments and no monitoring.
I analyze whether banks are efficient at allocating resources across intermediation activities. Competition between lenders means that resources are needed to draw borrowers into credit matches. At the same time, imperfect information between lenders and borrowers means that resources are also needed for screening. I show that the privately optimal allocation of resources is constrained inefficient. In particular, too many resources are spent on getting rather than vetting borrowers but, once properly vetted, not enough matches are retained. Uninformed lending is thus inefficiently high, informed lending is inefficiently low, and a tax on matching activities helps remedy the situation.
We investigate the effectiveness of central bank communication when firms have heterogeneous inflation expectations that are updated through social dynamics. The bank's credibility evolves with these dynamics and determines how well its announcements anchor expectations. We find that trying to eliminate high inflation by abruptly introducing low inflation targets generates short-term overshooting. Gradual targets, in contrast, achieve a smoother disinflation. We present empirical evidence to support these predictions. Gradualism is not equally effective in other situations though: our model predicts aggressive announcements are more powerful when combating deflation.
The Rise of China's Shadow Banking System (with Zheng Michael Song), January 2015
New -- Preliminary version available upon request
Research in Progress
Too Small to Fail: The Hidden Cost of Small Bank Failures in the 1930s (with Jon Cohen)
Bail-in bonds have gained a lot of attention among bank regulators. In principle, these bonds raise the hurdle for a government bailout by converting into loss-absorbing capital once the issuing bank runs into trouble. We show that banks can short-circuit bail-in requirements by offering investors implicit guarantees against conversion. The bond itself appears as a bail-in bond on the issuer's balance sheet while the guarantee is booked off balance sheet until the bond converts. If the regulator is expected to penalize guarantees, then the issuer's incentive to offer and honor them can be eliminated. However, guarantees are not discovered by the regulator until the issuing bank runs into trouble so penalties may not be credible. Our model shows that the penalty needed to eliminate guarantees is indeed credible if a bank's probability of running into trouble only depends on an aggregate shock. However, if banks are privately heterogeneous in their probabilities, then there are several equilibria where the penalty is not credible. More precisely, heterogeneity introduces a strong signaling motive to provide guarantees which the regulator must counteract with a substantial penalty. The size of this penalty is not credible since it would involve bankrupting a non-trivial fraction of the banking system. Bail-in requirements are thus much less effective under heterogeneity.