Relationship Lending in the Interbank Market and the Price of Liquidity 
The previous literature has argued that private information about counterparty risk plays a crucial role in enhancing efficiency of the liquidity allocation in the interbank lending market. Despite the interbank market’s importance for financial stability and monetary policy, due to a lack of transaction-level data, there is little empirical evidence on the precise role that private information plays in the allocation and pricing of credit. With a unique dataset composed of a daily panel of unsecured overnight loans between 1,079 distinct pairs of German banks, the authors use different sample selection models for the availability and pricing of credit from March 2006 through mid-November 2007, a period which in a smaller way mimicked the larger constrained credit conditions present in 2008 with the onset of the global financial crisis. In particular, the authors determine if bilateral lending relationships help mitigate search frictions and overcome informational asymmetries related to uncertainty about counterparty credit risk.
Key Findings
- Search frictions and uncertainty about credit risk play a role in interbank markets and repeated interactions among bank pairs mitigate these frictions. In particular, established lending relationships and reciprocal relationships are important factors for improving a bank’s access to liquidity. Bank pair relationships in the interbank market are persistent and the formation of these relationships is not random but dependent on past bilateral lending frequency.
- Conditional on a granted loan, borrowers pay a significantly lower spread to lenders with whom they have had a previous trading relationship. A bank pair that interacted on any given day in the past 30 days will agree on an interest rate that is up to 12.9 basis points lower than the spread agreed on by a bank pair that did not trade during the past month. The inclusion of various control variables that capture different facets of search frictions and other control variables does not change the significant impact of lending relationships on the matching probability.
- Relationship lending has a particularly strong negative effect on the bilateral interest rate when the loan is granted in a situation that suffers from high credit risk uncertainty, such as lending to opaque borrowers or lending during the liquidity crisis in summer 2007. This result holds after controlling for search frictions and suggests that frequent borrowers benefit from a lower lemons premium related to asymmetric information problems about counterparty risk when borrowing from their relationship lenders.
Implications
There are three main implications. First, having a decentralized unsecured interbank market structure confers potential benefits, but these need to be weighed against the larger systemic risks this market structure imposes. Second, if a bank that serves as a relationship lender fails, it may affect its borrowers as well as its lenders—contagion effects may work through the bank’s asset side as well as its liability side. Third, since the network of interbank overnight relationships is very persistent, this means that counterfactual contagion analyses and the identification of SIFIs can be based on interbank data with a relatively small time dimension.