Bank lending and interest on excess reserves: An empirical investigation

Highlights

• After 2008, US banks decreased lending and increased holdings of excess reserves.

• Banks hold excess reserves if the rate on reserves is above other short-term rates.

• Panel data show a switch to a high excess reserve regime after the financial crisis.

• Interest on reserves accounts for the majority of decline in lending after 2008.

Abstract

This paper econometrically tests for effects on bank lending of the Federal Reserve’s policy of paying interest on excess reserves (IOER). Following the 2008 financial crisis, US banks decreased their loan allocations and increased holdings of excess reserves. A model of bank asset allocation shows that when the rate of IOER is higher than other short-term rates, banks will switch from zero excess reserves to a regime with higher excess reserves and lower lending. Using a sample of panel data on US banks from 2000 through 2018, we find evidence of a switch to a positive excess reserve regime in the post-crisis period. Controlling for market interest rates, loan demand, and economic activity, we find that IOER accounts for the majority of the decline in bank lending after the financial crisis.

Introduction

During the financial crisis of 2008, the Federal Reserve initiated a variety of new and unprecedented policies. One such change transformed the Fed’s monetary policy framework by allowing it to pay interest on the excess reserves (IOER) that banks hold at the Fed. Though garnering fewer headlines than its quantitative easing (QE) program and its ad hoc last-resort lending facilities,¹ the Fed’s IOER policy and its effect on bank reserves is “one of the most notable and important policy issues in U.S. banking” (Dutkowsky and VanHoose, 2017, p.1). There is, however, little research on the empirical effects of IOER policy and in particular “little analysis of how reserves affect bank lending when interest is paid on reserves” (Martin et al., 2016, pp. 216–217).

Building on a model of bank asset allocation from Dutkowsky and VanHoose (hereafter “DV” Dutkowsky and VanHoose, 2017, Dutkowsky and VanHoose, 2018a, Dutkowsky and VanHoose, 2018b, Dutkowsky and VanHoose, 2020), we find that low rates of IOER increase lending but that banks reduce their loan allocations when the rate of IOER is greater than rates on other short-term assets. A Fed balance sheet expansion that increases total reserves will increase lending if the rate of IOER is below short-term interest rates, as in a corridor system of monetary policy. When the rate of IOER is above other short-term interest rates, as it was following the financial crisis, an increase in total reserves leads to a decrease banks’ loan allocations since excess reserves become a profitable alternative to loans.

This study provides an empirical investigation of the effects of IOER on bank lending. In Q4 of 2008, the Fed began paying a rate of IOER that was higher than other short-term rates, which our model indicates may have caused banks to increase their excess reserve holding and decrease their loan allocations. Using regression analysis, we test for the effects of IOER on bank loan allocations in a sample of panel data on US banks from Q1 of 2000 through Q4 of 2018. We find evidence of a change to a regime of positive excess reserves following the 2008 financial crisis. In the post-crisis regime, IOER has large negative effects on bank lending. Controlling for economic factors such as GDP growth, unemployment, uncertainty, and loan demand, we find that the rate of IOER relative to other short-term interest rates accounts for the majority of the post-crisis decline in lending as a percentage of bank assets.