Center for Financial Innovation and Stability

Notes from the Vault

Larry D. Wall
January 2015

The Federal Open Market Committee (FOMC) currently expects to raise the target range for short-term funds in 2015.¹ However, recent and proposed changes in bank regulations will likely raise the cost and potentially impair the effectiveness of the Federal Reserve's primary tool for raising rates.

The Federal Reserve's tools for raising rates are discussed in the FOMC document "Policy Normalization Principles and Plans." That document states that the Federal Reserve "intends to move the federal funds rate into the target range set by the FOMC primarily by adjusting the interest rate it pays on excess reserve balances."²

The theory behind the FOMC's plan is that changes in the interest on excess reserves (IOER) will affect other short-term interest rates through its effect on the rate banks are willing to pay for additional funds. For example, an increase in IOER increases the profitability to a bank of buying short-term funds and holding the proceeds as excess reserves (hereafter, called IOER arbitrage). Banks would respond to the increased profitability by increasing their demand for short-term funds, which would put upward pressure on short-term rates. The incentive to demand more funds would continue until bank funding costs increased to the point where IOER arbitrage is no longer profitable. As a result, if IOER arbitrage were costless, banks would bid up short-term rates until they almost equaled IOER.

However, IOER arbitrage is not costless due in part to regulatory policies, and these regulatory costs will increase if a pending regulation is adopted in its current form.³ The Federal Deposit Insurance Corporation (FDIC) insurance premiums paid on banks' reserves impose one such cost. The pending regulation is the Federal Reserve proposal to incorporate wholesale funding into the capital surcharge required of global systemically important banks (GSIBs). When binding, the wholesale funding surcharge would impose higher capital requirements to the extent that banks use wholesale funding in IOER arbitrage.

This post analyzes the merits of exempting IOER arbitrage from FDIC assessments and capital requirements. IOER arbitrage could be exempted from these costs by excluding all excess reserves and an equal amount of overnight wholesale funding from the assessments and capital requirements. I begin with a brief discussion of the mechanics of the Federal Reserve's plans for raising rates. The next section explains the regulatory costs in more detail. The following two sections discuss the potential benefits and costs of imposing regulatory costs on IOER arbitrage. The post ends with some concluding remarks.

Monetary policy normalization
The Federal Reserve's primary tool for conducting monetary policy prior to the financial crisis was to raise or lower the federal funds rate via open market operations that changed the amount of reserves outstanding. However, this mechanism became inadequate during the Great Recession, as both employment and inflation remained below targets even though the FOMC had reduced the federal funds rate to near zero (0.25 percent). To stimulate the economy further, the FOMC sought to lower longer-term interest rates by engaging in large-scale asset purchases, buying quantities of U.S. Treasury and Agency obligations.⁴ As a result, the Federal Reserve's portfolio grew from $775 billion at the start of 2007 to $4.2 trillion at the end of 2014, with reserve balances over that period increasing from $44 billion to $2.7 trillion.

Given the large increase in total reserves, the Federal Reserve's precrisis approach of controlling the funds rate with small changes in reserves would not be effective. To clarify its plans for controlling short-term interest rates, the FOMC issued its statement on normalization principles in September 2014. That statement not only indicated that rates would be controlled primarily via changes in IOER but also specified that the FOMC is prepared to use supplementary tools such as overnight reverse repurchase agreements (ON RRPs—discussed in greater detail below), to control the federal funds rate as needed.

When the Federal Reserve Board first announced the payment of interest on reserves in October 2008, the expectation was that doing so "should help to establish a lower bound on the federal funds rate." In practice, that was not the case with the federal funds recently trading around 12 basis points whereas the Federal Reserve pays 25 basis points for excess reserves. The discrepancy is due in part to the inability of government-sponsored enterprises to earn IOER but also due to the presence of regulatory costs that lower the net benefit of holding reserves.⁵

Regulatory costs
The imposition of FDIC insurance premiums on reserves is a relatively recent event. Prior to the crisis, these premiums were only levied on domestic deposits, and were not applied to foreign deposits and nondeposit wholesale borrowing. Section 331 of the Dodd-Frank Wall Street Reform and Consumer Protection Act changed the assessment base so that premiums are now levied on a bank's average total assets less its tangible capital. Given that excess reserves are part of total assets, these premiums lower the net benefit of holding reserves by the FDIC assessment rate. The assessment rate varies across banks from 2.5 to 45 basis points based on each bank's size and perceived risk to the FDIC; the average assessment rate in 2013 was 7.8 basis points (Financial Section of the FDIC's 2013 Annual Report).⁶

The proposed increase in capital requirements on GSIBs' short-term wholesale funding is also a response to the crisis and is intended to help reduce these banks' vulnerability to the sudden loss of funding (runs). This cost would affect IOER arbitrage because the natural supplier of huge amounts of short-term funding to banks are large institutions, such as money market funds, that seek to invest their liquid assets in short-term wholesale markets.⁷ Although this capital charge would apply only to the eight very large U.S. banks that are categorized as global systemically important banks, the wholesale funding charge could nevertheless have a major impact on IOER arbitrage.⁸ Available data are not sufficient to indicate exactly how big a fraction of U.S. domestic banks' borrowing is accounted for by the GSIBs. However, a lower bound on their importance is seen by the fact that these banks as a group hold about 43 percent of the total banking assets held by all domestic U.S banks (as of September 30, 2014), and GSIBs are far more active in wholesale financial markets than other domestic banks.

Potential financial stability benefits
As the change in the FDIC's assessment base and the proposed GSIB capital regulations are both in response to the financial crisis, the potential for exempting IOER arbitrage raises important financial stability questions. However, whatever the overall merits of these regulations, their application to IOER arbitrage will not enhance financial stability because IOER arbitrage really is an arbitrage transaction—that is, it allows participating banks to earn positive profits with no financial risk. A bank engaged in IOER arbitrage does not face any credit risk from its investment in excess reserves, as reserves held at the Federal Reserve have no credit risk. The bank faces no interest rate risk if it borrows in the overnight market to fund increased excess reserves, as the borrowings and excess reserves are for the same period. Finally, the bank faces no liquidity risk, as the excess reserves can be used to repay the purchased funds at any time.⁹

Disadvantages of regulating IOER arbitrage
The existence of costly regulations creates a wedge between IOER, the FOMC's primary tool for controlling rates, and the level of short-term market rates. The FOMC can compensate for this wedge in two ways: pay a higher rate on IOER to compensate banks for the cost of this wedge or rely on supplemental tools such as ON RRP. Both alternatives have important disadvantages.

One disadvantage of the payment of higher rates to compensate for costly regulation is that these regulatory costs will ultimately be borne by the U.S. Treasury. The Federal Reserve pays most of its earnings to the U.S. Treasury.¹⁰ These payments will be reduced to the extent the Federal Reserve must compensate banks for the regulatory costs of engaging in IOER arbitrage.¹¹

Moreover, the beneficiaries of the higher IOER rates are not necessarily entities that Congress would like to subsidize with taxpayer funds. For example, the primary beneficiaries of the higher rates needed to compensate for FDIC insurance assessments are the FDIC insurance fund and large foreign banks. The FDIC fund benefits from the higher insurance assessments, but Congress established insurance assessments on banks so that the cost of resolving failed banks would fall on banks and not on taxpayers in general.

Another beneficiary of higher rates is large foreign banks that have U.S. branches and agencies. These branches and agencies are not insured by the FDIC and, hence, do not pay insurance premiums, but they can hold reserves that earn IOER. As a result, these foreign bank operations are among the largest holders of excess reserves, with Bank for International Settlements economists Robert McCauley and Patrick McGuire reporting that these branches held almost $1 trillion in reserves at the end of 2013. Thus, to the extent that IOER compensates domestic banks for their payment of FDIC insurance premiums, the higher IOER also provides foreign branches with additional riskless profits.

A second problem the FOMC must deal with is that the wedge between the rate it controls—IOER—and the short-term market rates it seeks to affect are likely to vary over time and over the business cycle. In part, this time-varying wedge arises due to fluctuations in the cost of regulations to the banks. The FDIC premium rate will vary both with the FDIC's need to replenish losses to its fund and variations in the financial condition of the banks as measured by the FDIC's assessment formula. The cost of the wholesale funding capital requirement will vary over time as banks perceive changes in the opportunity cost of their capital.

The wedge may also vary over time with changes in the condition and regulation of foreign banks that affect their continued willingness to be among the largest participants in IOER arbitrage.

Whether these direct regulatory costs are a significant problem depends upon the magnitude of the variation. Small variations (not more than 25 basis points) are unlikely significantly to affect the relationship between monetary policy and the FOMC's dual objectives of stable prices and maximum employment. Short-term, small fluctuations in short-term rates are far less important than market participants' expectations of the path of future short-term rates. However, even if these costs have only a minor impact on monetary policy, that impact should be evaluated against the (nonexistent) benefits of imposing regulatory costs on IOER arbitrage.

A third disadvantage of increasing IOER to compensate for regulatory costs is that it increases the cost of bank loans relative to the cost of borrowing in financial markets. The rate paid on IOER is the marginal opportunity cost of making a short-term loan to a bank and as such will be reflected in the price of bank loans. But borrowers in financial markets will pay a rate based on short-term market rates that will be below IOER. As a result, those financed through short-term markets (including shadow banks) will face somewhat lower costs than those borrowing directly from banks. Although the impact of this differential is also likely to be small, the cost should be evaluated relative to the benefits of imposing regulatory costs on IOER arbitrage.

An alternative to relying exclusively on IOER arbitrage would be for the Federal Reserve also to use supplemental tools such as ON RRP. The Federal Reserve Bank of New York describes an ON RRP as a transaction in which the Federal Reserve "sells a security to an eligible RRP counterparty with an agreement to repurchase that same security at a specified price at a specific time in the future." The transaction is essentially a collateralized loan to the Federal Reserve, which temporarily takes excess reserves out of the banks.

The primary advantage of raising rates through ON RRP is that it allows the Federal Reserve to deal directly with suppliers of short-term funds and thereby avoid the regulatory costs imposed on IOER arbitrage. Whether it would be effective in raising short-term market rates up to the Federal Reserve's target in practice would depend upon the quantity of repurchase activity committed to by the Federal Reserve. In order to maximize its effectiveness, the Federal Reserve would need to be willing to "reverse repo" as many assets as market participants demand at the stated rate (that is, the Federal Reserve would have to provide "full allotment" to the providers of funds).

On the other hand, the provision of a full allotment ON RRP would bring some potential financial stability concerns. Among the concerns of most of the participants at one FOMC meeting was "that in times of financial stress, the facility's counterparties could shift investments toward the facility and away from financial and nonfinancial corporations, possibly causing disruptions in funding that could magnify the stress." Another concern of some of the participants is that "a relatively large ON RRP facility had the potential to expand the Federal Reserve's role in financial intermediation and reshape the financial industry in ways that were difficult to anticipate." Therefore, while the use of an ON RRP facility may help to support an IOER inhibited by regulatory costs, it comes with its own potential risks and is not a panacea.

Concluding thoughts
Exempting IOER arbitrage from regulatory costs could both reduce the costs borne by the U.S. Treasury and increase the effectiveness of IOER in raising market interest rates. Moreover, exempting IOER arbitrage would not increase the riskiness of banks, as this transaction is financially riskless.

The regulatory costs associated with IOER arbitrage could be avoided if the FOMC instead relied on supplemental tools such as ON RRP to control short-term rates. The use of an ON RRP facility avoids regulatory costs by allowing the Federal Reserve to deal directly with other short-term suppliers of funds without using banks as intermediaries. Still, reliance on an ON RRP facility is likely to result in a variety of unintended consequences as the financial system rearranges itself to more profitably use the new Federal Reserve facilities. Among these unintended consequences is an escalation of the risk of increased instability of bank funding during periods of stress.

Larry D. Wall is the executive director of the Center for Financial Innovation and Stability at the Atlanta Fed. The author thanks Mark Fisher, Scott Frame, and Paula Tkac for helpful comments on the paper. The views expressed here are the author's and not necessarily those of the Federal Reserve Bank of Atlanta or the Federal Reserve System. If you wish to comment on this post, please e-mail atl.nftv.mailbox@atl.frb.org.

_______________________________________

¹ Federal Reserve Board Chair Yellen stated that "almost all" of the FOMC members "believe that it will be appropriate to begin raising the target range for the federal funds rate in 2015."

² The Federal Reserve has long conducted monetary policy by targeting the federal funds rate, the rate at which banks trade reserves with one another. The federal funds rate is not by itself an important cost of funding to banks. However, given that banks and other market participants actively seek the most favorable prices, changes in the FOMC's federal funds rate target have been associated with similar moves in other short-term markets.

³ These are not the only two regulatory costs that may affect IOER arbitrage. Another possible regulatory cost arises from regulatory requirements limiting the ratio of a bank's capital to total assets ratio. Elizabeth Klee told the Federal Reserve Board that staff analysis suggested this regulatory cost would be "limited and readily offset" (see page 17 of the April 8, 2014, Open Board Meeting). Nonetheless, an impact on rates that would have a small impact on the effect of monetary policy on the broader economy may, when combined with other regulatory costs, have a material impact on the extent to which short-term market rates diverge from IOER.

⁴ The large-scale asset purchases are more popularly, but less accurately, known as quantitative easing or QE.

⁵ An additional cost in the form of a credit risk premium charged by the suppliers of funds to banks likely also reduces the profitability of IOER arbitrage. This credit risk premium compensates investors for the risk that the bank's overall portfolio will lose value and the bank will fail with losses to those supplying short-term funds. While this premium can be large (as at times during the crisis), in general the premium should be tiny, as the risk that a seemingly solvent bank will fail overnight is small, especially if the bank is large enough to be one of the big participants in wholesale funding markets.

⁶ The FDIC's Quarterly Banking Report, Deposit Insurance Fund Trends for the Fourth Quarter of 2013 reports that banks with assets greater than $100 billion were 61.7 percent of the assessment base, suggesting that the average assessment rate is indicative of the rate paid by the largest banks.

⁷ In general, short-term wholesale funding consists of both unsecured funds and unsecured funding from nonretail customers and counterparties with a remaining maturity of less than one year. An important exclusion from this broad definition is operational deposits, such as corporate accounts used to disburse the firm's payroll.

⁸ The wholesale funding charge would not necessarily be binding on all U.S. GSIBs. The Federal Reserve's proposal requires GSIBs to calculate their required increase in capital using two formulas, only one of which includes a charge for wholesale funding. However, even if the wholesale funding requirement is not immediately binding for some GSIBs, the requirement would limit these banks' ability to increase their wholesale funding without becoming subject to the higher requirements.

⁹ Although exempting excess reserves from the FDIC assessment base would not enhance financial stability, an argument could be made that exempting them would materially worsen smaller banks' competitive ability to engage in IOER arbitrage. However, wholesale activity is currently concentrated in the largest banks as noted above and imposing higher costs on the largest banks is unlikely to result in a material increase in institutional investors' willingness to deal with a large number of small banks.

¹⁰ A primer on the Federal Reserve's balance sheet, earnings, and remittances is available from Federal Reserve Board economists Seth Carpenter, Jane Ihrig, Elizabeth Klee, Daniel Quinn, and Alexander Boote.

¹¹ This is not to argue that the payment of interest on excess reserves per se is a net cost to the Treasury. If the Federal Reserve had purchased fewer Treasury securities there would be less excess reserves but this decrease would be fully offset by an increase in the amount of outstanding Treasury debt held by the public.

¹² This is an example of an "unknown, unknown" discussed by Mark Fisher that can arise from the Federal Reserve operating with a very large balance sheet.