The Federal Reserve Bank of New York has released a staff report by Morten L. Bech and Elizabeth Klee titled The Mechanics of a Graceful Exit: Interest on Reserves and Segmentation in the Federal Funds Market:
To combat the financial crisis that intensified in the fall of 2008, the Federal Reserve injected a substantial amount of liquidity into the banking system. The resulting increase in reserve balances exerted downward price pressure in the federal funds market, and the effective federal funds rate began to deviate from the target rate set by the Federal Open Market Committee. In response, the Federal Reserve revised its operational framework for implementing monetary policy and began to pay interest on reserve balances in an attempt to provide a floor for the federal funds rate. Nevertheless, following the policy change, the effective federal funds rate remained below not only the target but also the rate paid on reserve balances. We develop a model to explain this phenomenon and use data from the federal funds market to evaluate it empirically. In turn, we show how successful the Federal Reserve may be in raising the federal funds rate even in an environment with substantial reserve balances.
This issue has been discussed on PrefBlog before, two posts being Effective Fed Funds Rate Continues to Confuse and Effective Fed Funds Rate: A Technical Explanation?.
The authors state the problem succinctly:
Between the October and December 2008 FOMC meetings the average effective federal funds rate was 32 basis points, while the target was 1 percent; the interest rate paid on reserves was 65 basis points or higher for the entire period. This deviation from the theoretical prediction was surprising for even the most astute observers of the federal funds market (Hamilton, 2008, for example). Why did interest on reserves provide an imperfect floor for the federal funds rate, even after the policy was changed so that the interest rate paid on reserves was set equal to the target rate?2 Why would any financial institution lend out funds below the rate paid by the central bank? And even if that were the case, arbitrageurs would surely relish the opportunity of making a pure profit by borrowing cheaply in the market and placing the proceeds with the central bank, and by doing so, move the market rate toward the floor.
… and propose:
The explanation for the puzzling outcome is at least threefold. First, not all participants in the federal funds market are eligible to receive interest on their reserve balances. Government-sponsored enterprises (GSEs) in particular, which are significant sellers of funds on a daily basis, are not legally eligible to receive interest on balances held with Reserve Banks. This heterogeneity across participants created a segmented market with different rate dynamics. Second, banks’ apparent general unwillingness or inability to engage in arbitrage has produced a market structure in which those banks that are willing and able to buy funds from the GSEs have been able to exercise bargaining power and pay the GSEs rates below the interest rate paid on reserves. The lack of arbitrage possibly has been driven in part by banks seeking to control the size of their balance sheet more closely in part to avoid stressing regulatory capital and leverage ratios, as mentioned in Bernanke (2009a). Third, a combination of financial consolidation, credit losses, and changes to risk management practices has led at least some GSEs to limit their number of counterparties in the money market and to tighten credit lines. This trimming of potential trading partners has likely decreased bargaining power with the remaining counterparties. In addition, the GSEs have become a larger share of the federal funds market in recent history and hence have pulled down the weighted average federal funds rate.
The paper is organized as:
First, we briefly discuss the institutional details of the federal funds market. Second, we turn to the history and implementation of the interest-on-reserves regime. Third, we present our model of a bifurcated federal funds market with banks and GSEs. We also sketch out how the model can be extended to describe a trifurcated market in which some banks are slow to adopt the new policy regime. Fourth, we calibrate our model to federal funds market data and back out the relative bargaining power of the different market participants. Fifth, we explore the factors that affect our computed bargaining parameters. We show that the level and distribution of reserve balances, rates in other overnight funding markets, and the riskiness of the buyers all have significant predictive power in explaining the bargaining power of the different types of sellers. With this information, in our sixth section, we forecast the effective federal funds rate under a variety of exit scenarios from the current accommodative stance of monetary policy. The seventh section concludes and offers directions for further research.