Shrinking the Fed’s Balance Sheet Is Easier Said Than Done

At his Senate confirmation hearing in April, Federal Reserve Chairman Kevin Warsh reiterated his desire to shrink the size of the Fed’s holdings. “Slowly and deliberatively, I believe we need a smaller central bank balance sheet,” he said.

Yet slimming the Fed’s $6.7 trillion balance sheet, even slowly and deliberately, is easier said than done. In a recent Brookings paper, Darrell Duffie, an emeritus professor of finance at Stanford Graduate School of Business, details several ways the Fed could proceed. “Even though I don’t say the Fed should reduce its balance sheet, I do say that it should have the tools that would allow it to reduce its balance sheet if it were to need to,” says Duffie, who is also a senior fellow at the Stanford Institute for Economic Policy Research and senior fellow (by courtesy) at the Hoover Institution.

Two decades ago, the Fed held about $900 billion in assets. Then came the 2008 financial crisis. In rounds of quantitative easing, the Fed bought up bonds to stabilize the economy. Its assets continued to grow, peaking in 2022 at $8.9 trillion. It paid for these assets by creating more reserve balances, which are deposits held at the Fed by commercial banks. Warsh, a former lecturer at Stanford GSB, has criticized this expanded balance sheet as “a proxy for the Fed’s growing imprimatur on the economy.”

Duffie won’t say just how much of its $3 trillion in reserve balances the Fed could realistically unload, emphasizing the need for further research. “The U.S. banking system is very complicated,” he says. “I’m reticent about putting a number on these [proposals]. The Fed hasn’t yet done the research necessary to put a solid number on that.”

Duffie spoke with Insights about ways to make a dent in the Fed’s balance sheet and the economic and political considerations behind them.

Let’s start with a question that you ask in your paper: “What has changed so much since 2007… that has caused the financial system to now need trillions of dollars of reserve balances?”

Darrell Duffie: Most importantly, the Fed changed the way that it fights inflation. Before the crisis, it used to restrict the supply of money. Now it pays interest on reserves, which means that when a bank is thinking about making a loan, they compare the loan interest to the interest they get from the Fed. When the Fed raises the rates that it offers to banks, then banks optimize by raising rates on loans to firms and households. That slows down inflation.

Once the Fed started to pay interest to banks on their reserves, the banks have had no reason to skimp on the reserves that they want. That’s terrific for the banks, but it means the Fed now supplies a lot more reserves than would be necessary if the Fed were to make some of the changes I analyze in my paper.

The other big change is that the Fed introduced a bunch of liquidity regulations. Some of these regulations require the banks to prove that they can be self-sufficient for liquidity. So the banks hold even more reserves to make sure they don’t suddenly go to the Fed asking for more.

How much of this is driven by the banks becoming attached to a larger supply of reserve balances?

Darrell Duffie: I think it’s mostly that the banks don’t mind having a lot of reserves. Once banks get more reserves, they don’t want to give them up. In a well-known Jackson Hole paper, Viral Acharya and Raghu Rajan called this a “ratchet effect.” Reserves are like the Swiss Army Knife of finance: They’re good for everything and they pay interest. So what’s not to like?

If the Fed were to reduce the supply of reserves significantly from where it is today, banks would not want to lend them to financial market participants except at a very high interest rate, and that messes up the Fed’s monetary policy. So, the Fed has become conservative. They let the banks have plenty of reserves. That’s actually OK. Nothing explodes if there are more than enough reserves, while too little can lead to serious problems with hoarding and spiking interest rates.

What is the argument, as you understand it, for shrinking the Fed’s balance sheet?

Darrell Duffie: Those who think it’s a problem have a vague reason — although I think it is an important reason — which is politics. For people who don’t understand how the Fed works, a $6.7 trillion balance sheet looks like the Fed is doing more than just setting interest rates. There’s this subtext of concern that maybe the Fed is helping the government finance itself by buying a lot of government bonds. That’s not what the Fed is supposed to do. It’s not supposed to support the fiscal ability of the government. It’s just supposed to set interest rates and keep inflation under control. In actuality, the Fed does not have a large balance sheet because it wants to provide fiscal support. The large balance sheet is caused by the very high demand for reserves that I described. But even incorrect perceptions that the Fed has “gone fiscal” do matter.

If a big balance sheet doesn’t look right to critics, then that could eventually impair the Fed’s independence. Some in the government might say, “I’m not sure why the Fed needs to be so big; maybe we need to control it more.” That might impair the Fed’s independence, and that would not be good.

Beyond the political considerations, is there an economic argument for shirking the Fed’s balance sheet?

Darrell Duffie: In terms of the economics, there’s nothing really wrong with a large balance sheet, with one or two exceptions. One of these is called “crowding out.” A bank must meet its capital requirements. If the Fed supplies a lot more reserves to banks, banks might choose to reduce the loans they provide to firms and households. That’s not a big issue for now — but remember the ratchet effect. Suppose another crisis comes along, and the Fed needs to buy a whole lot more assets, maybe $2 or $3 trillion more. Whatever cost there is of having a large balance sheet just got bigger. It’s hard to get the Fed’s balance sheet back down if there is truly a ratchet effect.

Suppose the Fed does decide to shrink its balance sheet. What are some of the constraints that prevent it from doing this quickly or on a large scale?

Darrell Duffie: When people think about the balance sheet size, they’re usually looking at the assets. The naive reaction is, “Well, just sell them.” But they’re forgetting that assets equal liabilities. Which liabilities are you going to reduce? The second biggest item on the Fed’s liability side is paper currency. Would the Fed announce: “Would everyone please turn in some of your $100 bills?” That’s not going to happen.

The other big item is reserve balances — the deposits that commercial banks have at the Fed. That’s the only place where the Fed could realistically get a lot of reduction. But banks have about the amount that they want, and that amount is growing with the size of the economy. So how would the Fed reduce the quantity without causing havoc with interest rates? I list a few options in my paper, but none of them are that easy.

When the Fed started to pay interest on reserves to control inflation, the response of banks was, “Terrific, now we can hold a lot of reserve balances, and we won’t lose any interest. We’ll get about the same interest rate from the Fed that we would get from anybody else.”

The Fed could change that policy to tiered remuneration. The new policy would be: Once a bank has the quantity of reserves that it needs to do its payments and enough to run its other business operations, including meeting its liquidity requirements, the Fed could say to a bank, “Sorry, we’re only going to give you full interest rate on this needed first $10 billion. Anything after that, we’ll pay you a lower rate, say 0.5% less.” The Fed can do that because it administers this interest rate.

Banks would then conclude, “Well, we’re not going to hold all these extra balances—no way— because why would we hold money at the Fed paying 3% when we can go out into the market and get 3.5%? So we’ll ditch any unnecessary reserves.”

In new research with my doctoral student collaborators Francesco Spizzuoco and Thanawat Sornwanee, we showed that the Fed could reduce the quantity of balances that banks hold by a lot without changing monetary policy implementation, and without stressing the banks for liquidity. In theory, the Fed could make that decision overnight and it would probably work pretty much overnight. But there are regulations involved — the Fed would probably need to provide advance notice of this and perhaps do a cost-benefit analysis. And then there could be a counter lobbying effort by banks, because this could reduce the profitability of the largest banks.

Are your proposals being considered inside the Fed?

Darrell Duffie: I’ve presented this work inside the Federal Reserve Board and lots of other places, including the Brookings Institution, which sponsored this research. Let’s see if it eventually has any impact. That will take at least until the Fed’s new task force on its balance sheet provides a report, probably around the end of this year. That task force might explain that the current size of the balance sheet is fine. And it might suggest some need for additional research and planning. Maybe the task force will also focus on the asset side of the Fed’s balance sheet and discuss how the Fed could change the composition of its assets so as to lower the volatility of its net income. That’s a story for another day!