The Role of Interbank Markets in Monetary Policy: A Model with Rationing Xavier Freixas Universitat Pompeu Fabra and CEPR José Jorge CEMPRE, Faculdade Economia, Universidade Porto
Motivation Starting point: Paradox in the way we model the microfoundations of monetary policy. 1. The role of banks is justified by the existence of asymmetric information. 2. Yet, when it comes to monetary policy transmission mechanisms, the Interbank Market is assumed to be a perfect market that plays a neutral role in the allocation of resources. 2
Empirical evidence on liquidity and the interbank market Micro data: Kashyap and Stein (2000) found that the impact of monetary policy is more pronounced for banks with less liquid balance sheets. Ashcraft (2006) on bank holding companies Hence 1) liquidity matters and 2) interbank markets cannot be perfect. 3
Related literature 1. Money view vs. Broad Credit Channel 2. Stein 1998 spread augmented interest rate channel 3. Interbank market models 4
Methodology 1. We introduce asymmetric information in the interbank market. 2. Derive the contracts available in the interbank market. 3. Derive the individual net demands. 4. Derive the equilibrium in the interbank market. 5
Main Results Under perfect information, the money view holds. Under asymmetric information, there is rationing in the interbank market. Our model is consistent with the main empirical results. 6
Motivation Model Perfect Asymmetric Markets Information Conclusion The Model Firms invest one unit at date 0. Firms face a liquidity shock (cost overrun) ν at date 1. Firms either borrow funds and continue or are liquidated. If firms are not liquidated, they produce Y with certainty. 7
Project 0 1 2 I = 1 ν ν, ν Y (or 0) We focus on date 1. [ ]
Motivation Model Perfect Asymmetric Markets Information Conclusion The supply of bank retail deposits is inelastic in the short run Relationship banking is assumed. The amount of equity is fixed zero-in the short run. Simplification: each bank lends only to a single firm (portfolios are not perfectly diversified and banks suffer different shocks). 9
Motivation Model Perfect Asymmetric Markets Information Conclusion Bank s liquidity shocks Firms liquidity shocks ν (positive or negative) Deposit shocks D (positive or negative) Buffers against liquidity shocks: T-Bills Interbank loans 10
The defaulting threshold for the firm depends on the contract established between the firm and the bank (Y- Repayment of Initial Loan)/(1+r L ) The continuation of the project depends on the banks only and, in particular, on the interbank market: ν > Y/(1+r L ) 11
Motivation Model Perfect Asymmetric Markets Information Conclusion Benchmark : Perfect Information Case The liquidity shock ν is public knowledge. Banks obtain funding in the interbank market. Only projects with high cost overrun are liquidated. Equilibrium entails no rationing. 12
Asymmetric Information The information about ν is private to the relationship firm-bank. Moral Hazard / Gambling for resurrection Bankers have access to an alternative private benefits project which yields pledgeable return K and private benefits. The choice of this project is not observable. Private benefits are proportional to L => Bad borrowers want the largest loan available in the interbank market. 13
Interbank market contracts Contracts are defined by (L, r L ). Interbank market contracts: Banks obtain a loan at the risk free rate up to the size of L. L = K D 1 ( r) 1+ r L
Result: Pooling Equilibrium: bad/ insolvent prefer to borrow the maximum amount. borrowers pool at the level of the fully collateralized loan. No risk spread in the interbank market. 15
Intuition: If there was a larger risky loan, lenders could separate good from bad borrowers by offering a loan of a slightly smaller size at a slightly lower interest 0 rate. Such a contract would attract only good borrowers, as bad ones always apply for the largest loan. But then the contract with bad borrowers is not profitable. Hence there is no separating equilibrium. 16
Implications Banks are rationed in the interbank market. Banks face limited available liquidity that depends on monetary policy! 1. Bank deposits 2. Balance sheet channel. Projects are liquidated if 1. Excessive cost overrun (as in perfect information case) 2. Bank itself is rationed 17
Comparative statics The relevance of imperfections in the interbank market for monetary policy depends on: 1.The dependence of firms on bank finance. 2.The extent of relationship banking. 3.Heterogeneity on banks liquidity positions. 18
Consistency Two empirical puzzles (Bernanke and Gertler 1995): Magnitude effect: The impact of monetary policy is larger than we would expect from firms individual decisions. Composition effect: Monetary policy has most direct effects over short term interest rates, but it affects both short and long term investment. 19
The model s Results Magnitude effect: There are positive NPV projects that are not undertaken because of rationing. Interest rates differ from opportunity cost of capital. 20
Results Proposition 4 (Magnitude Effect) The aggregate effect of an interest rate shock is larger than the sum of the individual effects ects of an increase in the user cost of capital. Proof: First, aggregate output equals ( B B)( ν ν ) Second, sum of individual effects: 1 B F dy 1 Υ = YdνdB and ε r B ν dr Y dy dy 1 ε uc( rf ) dν = dν = 0 dr y = ν 1 ˆF ν dr ν F y y is the output from an individual firm r F is the interest rate on bank loans There is a number of firms for which an increase in r F reduces output, BUT it has measure zero. F ) () r = 0 0 >
Motivation Model Perfect Asymmetric Markets Information Conclusion Composition Puzzle Banks ration their clients regardless of the maturity of their investments. 22
The model s Results (2) Kashyap and Stein Liquidity Puzzle: 1. The effect of interest rates occurs (mostly) through the rationing channel. 2. Banks with less liquidity, are more likely to be rationed. 3. Hence monetary policy has a larger effect on banks with less liquidity. 23
What empirical predictions? Cumulative Deviation of Output from Trend 3 2,5 2 1,5 1 0,5 0 0 1 2 3 4 5 6 7 Predicted Efectiveness of Monetary Policy Data for the United Kingdom, Netherlands, France, United States, Italy, Austria, Germany and Japan. 24
To conclude Asymmetric information may imply rationing in the interbank market. Results are consistent with the findings by Kashyap and Stein. The model allows to understand better the magnitude and the composition puzzles. The model predicts that monetary policy would be more effective when interbank market imperfections are stronger. 25