The Fed's Mispricing of Liquidity: Nothing New Under the Sun

bank subsidies, federal reserve, liquidity, interest on reserves, interest on excess reserves Website/Blog/Electronic Publication - Non Commercial: Use at one URL for an indefinite period, where the web site does not take paying subscribers or advertisers and is not used to promote a commercial entity. A blog is defined as a regularly updated weblog presented in a "diary" format sorted by date. It may be hosted on the user's own web server, or through a third party blogging service such as CartoonStock grants the image user a non-transferable, non-exclusive right to copy the CartoonStock image(s) stated above onto their computer and to use the CartoonStock image(s) and any derivatives or copies (collectively, the "image(s)") on their personal or company computer(s). This license is valid only with respect to the image(s), provided the applicable license fee stated above is paid within the time period set out above. This license grants the image user only use of the image(s) for the purpose stated above. The image user may not use the image for any other purpose without prior written agreement from CartoonStock. The image user may not publish a modified version of the image(s) or any part of the image(s) without prior written agreement from CartoonStock. He/she may however modify the image(s) caption font to his/her satisfaction without prior agreement from CartoonStock. The specified image(s) may not be sold or sublicensed alone or as part of any product other than the product/publication to which this license refers. Use of the image(s) on or in a product does not grant or purport to grant the right to duplicate or distribute the image(s) to any third party or end user.

federal reserve, interest on reserves, interest on excess reserves, liquidityIn his recent post exploring the welfare implications of the Fed paying interest on reserves, George Selgin observed that if banks failed to maintain optimal reserves, the problem was either incorrect remuneration of reserves themselves, or the Fed’s mispricing of short-term (intraday or overnight) credit. Regarding the latter, he claimed mispricing was once serious, but has since been partly rectified.

I agree that the problem of mispriced Fed intraday credit was once serious, and that it has been “partly rectified.”  The Fed made several attempts to improve its overdraft policies, and more recently, the vast increase in banks’ excess reserve holdings has made daylight overdrafts unnecessary.

Yet important questions remain concerning the extent to which the Fed continues to misprice its services, and to thereby violate the spirit and letter of the 1980 Depository Institutions Deregulation and Monetary Control Act (DIDMCA, or MCA for short). These questions, along with Selgin’s concerns about the legality of the Fed’s above-market interest rate on bank reserves, suggest a need for heightened scrutiny of the Fed by Congress.

Here I survey the Fed’s post-MCA payments-pricing practices. I hope to show that, even though the mispricing of some of these services may have been “partly rectified,” there are still good reasons to doubt that the Fed has complied with the MCA’s provisions calling for it to fully recover “all direct and indirect costs” of the payment services it provides to banks.

Some History on the Fed-Subsidized Services

The Fed has published financial statements for its priced payment services ever since the MCA was passed. The statements report the Fed’s revenue, expenses, and cost recovery for those services under accounting standards set not by an independent standard setter, but by the Federal Reserve itself. Unsurprisingly, the Fed’s standards have allowed it to report that its revenue has fully recovered its costs over time, even when conventional standards would not have.

The Fed has thus been able, using its own financial statements, to defend itself when outside parties (and some rival, private suppliers of payment services) have accused it of subsidizing its payment services. For example, in a 1997 Congressional hearing, Alice Rivlin, then-vice chair of the Federal Reserve Board of Governors, testified:

The MCA requires the Federal Reserve Banks to charge fees for their payment services, which must, over the long run, be set to recover all direct and indirect costs of providing the services. In addition, the MCA requires the Federal Reserve Banks to recover imputed costs, such as taxes and the cost of capital that would have been paid and imputed profits that would have been earned if the services were provided by a private firm.

Over the last ten years, the Federal Reserve has fully recovered the total costs of its priced services, including imputed costs as required by the Monetary Control Act. … Shortly after the MCA was enacted, the Board of Governors adopted pricing principles that are more stringent than the requirements of the MCA and that require the Federal Reserve Banks to recover priced service costs, not just in the aggregate, but for each major service category. Our check service, for example, has fully recovered its costs over the last ten years.[1]

In turn, in an appendix to the body of this testimony, Rivlin reiterated that

Taxpayers do not subsidize the cost of the Federal Reserve's check transportation. It is understandable that whenever a public entity competes with the private sector in providing services, the issue of subsidies arises. The Monetary Control Act of 1980 (MCA) addressed that issue by requiring the Federal Reserve, over the long run, to set fees for its priced payment services to recover all direct and indirect costs of providing those services. [Emphasis original.]

Rivlin’s remarks had been anticipated by then-Fed chairman Alan Greenspan, who in his 1996 testimony to the Senate Banking Committee insisted that the Fed’s

priced services are subject to the inherent discipline of the marketplace as the Federal Reserve must control costs in order to meet the statutory directives for cost recovery in the Monetary Control Act. The risk-management decisions that we make concerning the way we provide payment services to depository institutions are tested directly in the marketplace. These services comprise more than one-third of the Federal Reserve banks' total budget and the Monetary Control Act requires that, over the long run, we price these services to recover their costs, as well as costs that would be borne by private businesses, such as taxes and a return on equity. If we provide these services inefficiently, we price ourselves out of the market.

Over the past decade, our track record has been good. The Reserve Banks have recovered 101 percent of their total cost of providing priced services, including the targeted return on equity. I should also note that, by recovering not only our actual costs but also the imputed costs that a private firm would incur, the Federal Reserve´s priced services have consistently contributed to the amount we have transferred to the Treasury. During the past decade, priced services revenue has exceeded operating costs by almost $1 billion.[2]

Yet even with the help of their special accounting practices, Fed officials weren’t able to consistently maintain that its payments services weren’t subsidized. Just two months before Rivlin assured Congress that “Taxpayers do not subsidize the cost of the Federal Reserve's check transportation,” Greenspan, in making a case for the Federal Reserve Board to serve as lead regulator for new financial holding companies, had been claiming just the opposite:

a number of observers have argued that there is no subsidy associated with the federal safety net for depository institutions — deposit insurance, and direct access to the Federal Reserve's discount window and payment system guarantees. The Board strongly rejects this view.

While some benefits of the safety net are always available to banks, it is critical to understand that the value of the subsidy is smallest for very healthy banks during good economic times, and greatest at weak banks during a financial crisis. … What was it worth in the late 1980s and early 1990s for a bank with a troubled loan portfolio to have deposit liabilities guaranteed by the FDIC, to be assured that it could turn illiquid assets to liquid assets at once through the Federal Reserve discount window, and to tell its customers that payment transfers would be settled on a riskless Federal Reserve bank? [Emphasis added.][3]

Although Greenspan refers generally to subsidies offered by the Fed’s safety net, that net includes access to Fedwire as well as to the discount window.

In related 1997 testimony, Greenspan noted explicitly how the safety net (including Fedwire access) shifted risk to the government, and how, because that risk constituted a cost, the practice amounted to a subsidy provided to banks. "The use of sovereign credit in banking," Greenspan said,

even its potential use — creates a moral hazard that distorts the incentives for banks: the banks determine the level of risk-taking and receive the gains therefrom, but do not bear the full costs of that risk. The remainder of the risk is transferred to the government. This then creates the necessity for the government to limit the degree of risk it absorbs by writing rules under which banks operate, and imposing on these entities supervision by its agents — the banking regulators — to assure adherence to these rules.[4]

Fedwire — the Fed’s wholesale funds transfer system — is one of the largest payment services the Fed provides. Prior to the recent crisis, banks paid one another more than $1 trillion every day using this service. The funds being transferred consisted of bank reserve account balances — balances on which the Federal Reserve now pays banks billions of dollars a year in interest.

Before the present "abundant reserves" era began, banks settled huge Fedwire payments volume on a relatively thin aggregate cushion of excess reserves, with the extension of intraday overdraft credit from the Fed a key enabler. Only a thin cushion sufficed because banks that ran short of reserves with which to meet either their settlement needs or their minimum legal requirements could, for a price, borrow both either from the Fed or, for overnight loans, from other banks, to avoid shortfalls. The Fed provided billions of dollars of intraday credit daily, the result of the fact that the Fed guaranteed Fedwire payments to receiving banks, notwithstanding that the sending bank might be short funds at the end of day — thereby exposing the Fed itself to credit risk.

The Fed’s fee schedules for Fedwire transfers have evolved over time, but they have consistently been based on a fee-per-transfer basis, with "volume" discounts available not for the dollar size of the transfer(s), but the number of transfers. In the scarce-reserves era, fees were generally a fraction of $1 per transfer, even for large dollar transfers utilizing large amounts of overdraft credit. For many years after 1980, the Fed didn’t even charge fees for intraday credit utilization, only beginning the practice in 1994 — with effective interest rates a small fraction of market interest rates. In turn, the Fed didn’t include daylight overdraft fees as revenue in its cost recovery calculation for pried services — despite acknowledging that the new payment system risk policy (in 1985) and overdraft fees (in 1994) were developed “[t]o reduce the risks that depository institutions present to the Federal Reserve through their use of daylight credit and to address the risks that payment systems, in general, present to the banking system and other sectors of the economy …”[5]

With the progress in the Fed’s payment system risk policies in the late 1980s and the development of overdraft fees in the mid-1990s, some Fed leaders claimed to have addressed the subsidy issue. However, they didn’t appear to have convinced the Fed Chairman, given that he was identifying Fedwire access as a subsidy in the late 1990s.

In light of the low fees being charged, even allowing for some improvements here and there, Fedwire sure looked like a subsidy. But for many years, the Fed was also asserting that Fedwire was not a subsidy, given that the Fed priced the services to fully recover "all direct and indirect costs."

In 1999, Congress passed the Financial Services Modernization Act. This law established the Federal Reserve Board of Governors as the lead regulator and supervisor for financial holding companies. In making their case to Congress for the lead position, Fed leaders acknowledged, like Greenspan did in that 1997 testimony, that the Fed’s assumption of risk on Fedwire constituted a subsidy. The Fed was very concerned about the risk it was exposed to, the story went, and so the Fed had to be in the lead position as a regulator (and supervisor).

The Fed’s case for a lead supervisory role was (and is) ironic in another sense. Greenspan asserted that the shifting of risk to the government from safety net features like Fedwire drove the need for costly government supervision. So it could have been argued that the costs of this supervision were another element of the "direct and indirect costs" to be recovered by Fedwire pricing. But they were not.

Yes, Fedwire was subsidized. The Fed knew it, but didn’t account for it. And Congress stood idly by, as the Fed’s accountants and lawyers resorted to window-dressing to feign compliance with the law.

How About the New “Reserve-Abundant” Era? 

In 2007 and early 2008, reserves at the Fed were running about $40-45 billion, with Fedwire accomplishing more than a $1 trillion a day in funds transfers. In those "old" days, some Fed officials actually argued that high Fedwire volume on a thin reserve base was evidence for how "efficient" the system was.

With the massive easing of monetary policy amidst the blooming financial crisis and arrival of IOR, reserves at the Fed mushroomed from $46 billion in August 2008 to more than $800 billion at the end of the year. And from there, reserves reached new heights, rising to nearly $3 trillion in 2015. Not surprisingly, as excess reserves ballooned, daylight overdrafts dropped to record low levels.

Let’s take a closer look at what was going on as the 2007-2009 financial crisis intensified. In the weeks leading up to mid-September 2008, peak daylight funds overdrafts were running at about $130 billion. From September 10 to September 24, however, peak overdrafts jumped to $200 billion, and higher still to $245 billion in the reserve period that ended October 8. The September 10 – October 8 period coincided with the onset of the burgeoning reserve balances — and rising fears (and failures) in the financial markets.

In other words, in the first weeks of sharply higher reserves, daylight overdrafts were actually still growing dramatically, with large financial firm failures exploding around the payment system minefield. It was only later, after the acceleration in QE, that daylight credit began falling.

In making their recent case for the efficiency of paying interest on reserves, Fed economists have praised the reduction in risk to the Fed, and the public, from the reduced utilization of intraday credit. For example, in a 2012 FRBNY article “How The High Level of Reserves Benefits the Payment System,”[6] Fed economists praised the new system: “we can document an important and overlooked benefit of the high level of reserves: a significantly earlier settlement of payments on Fedwire." They continue, "[a]t the same time, the Fed has been extending less intraday credit, which reduces the public’s risk exposure.”

Can the Fed have it both ways? Was the Fed truly recovering "all direct and indirect costs" from 1980 until the financial crisis? Was Fedwire efficient back then, and even more efficient today?

Nothing New Under the Sun

A strong case can be made, and Selgin has made it, that in paying above-market interest on banks’ reserves, the Fed is not just subsidizing banks, but subsidizing them illegally.

But far from being an exceptional, new development, this practice is only the latest instance of the Fed’s longstanding practice of offering subsidized payments services, which it mainly did in the past by helping banks fund trillions of dollars in Fedwire payments with risky overdraft credit, without pricing that risk to fully recover the cost of providing Fedwire as required by law.

Daylight overdrafts have declined to next to nothing since 2008, while the problem of mispriced overdrafts, though seemingly in abeyance, has not actually been addressed. The Fed’s pricing of Fedwire and funds overdrafts continues to be at odds with the spirit, if not the letter, of the MCA’s cost recovery provisions. While daylight overdraft risk no longer poses the direct problem it used to, in "eliminating" it the Fed has created an even more serious problem (and indirect cost) of balance sheet risk for itself, and ultimately for taxpayers. In short, the Fed is now paying banks billions of dollars a year for the "privilege" of "reducing" the risk they pose to the Fed, and shifting it to the taxpaying public!

New Congressional hearings should explore these developments, with the aim of helping the Fed comply with the law, at long last.


[1] Alice M. Rivlin, "Role of the Federal Reserve in the Payment System," testimony before the Subcommittee on Domestic and International Monetary Policy of the Committee on Banking and Financial Services, U.S. House of Representatives. September 16, 1997.

[2] Alan Greenspan, "Recent reports on Federal Reserve operations," testimony before the Committee on Banking, Housing, and Urban Affairs, U.S. Senate. July 26, 1996.

[3] Alan Greenspan, "The Financial Services Competition Act of 1997," testimony before the Subcommittee on Finance and Hazardous Materials of the Committee on Commerce, U.S. House of Representatives. July 17, 1997.

[4] Alan Greenspan, "Modernization of the Financial System," testimony before the Subcommittee on Financial Institutions and Consumer Credit of the Committee on Banking and Financial Services, U.S. House of Representatives. February 13, 1997.

[5] Stacy Panigay Coleman, “The Evolution of the Federal Reserve’s Intraday Credit Policies,” Federal Reserve Board of Governors Bulletin (February 2002)

[6] Morton Bech, Antoine Martin, and Jamie McAndrews, "How the High Level of Reserves Benefits the Payment System," Liberty Street Economics, Federal Reserve Bank of New York. February 12, 2012.

  • Mattyoung

    Currency risk accumulates in Congress in the form of inflation insurance, price fixing, and social insurance. Once per generation we default.

    • Ray Lopez

      That's true. And it raises the question of 'so what'? If businesspeople are greedy and cannot compute money neutrality, that's their problem. For the rest of us, we diversify into stocks with cost-plus pricing power, real estate, gold, silver, etc. Life goes on.

      Bonus trivia: technically even laypeople can expand the money supply, George Selgin 'crowd sourcing' style, if they adopt the habit of using endorsed checks with no special instructions (i.e., just sign the back of the check and don't write 'FOR DEPOSIT ONLY') and then using these endorsed checks as 'cash'. Apparently a form of this was done back in the 19th century with US 'free banks' and their paper. Which proves that a country's money supply is very elastic. Even in hyperinflation Zimbabwe the people adopted, though even I agree there were real negative effects with Zimflation (i.e., money is not neutral during hyperinflation nevertheless it's not the end of the world).

  • Hu McCulloch

    Interesting post, Bill. How was the fee on daylight overdrafts set before 2008? Why was this inadequate?

    Am I correct that any end-of-day overdrafts are simply rolled over to the next day at the discount rate (currently 50 bp, I believe, above the FF target)? Daylight overdrafts make me nervous, but I haven't studied them closely.

  • JP Koning

    Good post on the dangers of Fed-provided intraday credit. I agree with much of it.

    In Canada, we have an interesting payments system whereby the private bank participants provide each-other with intraday lines of credit up to limits that each lender is comfortable with. If a borrower defaults, the lenders are all on the hook for that amount. So the member banks have a huge incentive to monitor each other. I think that makes a lot of sense since it offloads a lot of the costs of regulation onto those with the best resources.