Forward guidance: The other side of QE and QT

From: Bill Nelson <[email protected]>

Date: Wednesday, March 12, 2025 at 6:59?AM

In the canonical QE operation, the central bank purchases a longer-term security, shifting interest-rate risk from the public to the central bank, reducing term premiums.? The lower term premiums and therefore lower longer-term interest rates stimulate the economy through the normal interest rate channel.

There is, of course, another side to the operation.? The central bank has to finance the security purchase somehow, either by increasing a liability – normally reserve balances (deposits of commercial banks at the central bank) – or reducing another asset.? Under the Board of Governors’ economist orthodox view, this other side doesn’t matter much.? For example, the Maturity Extension Program (a/k/a “operation twist”), in which the Fed allowed shorter-term securities to roll off as it purchased longer-term securities, leaving reserve balances unchanged, was judged to have had basically the same effect on the economy as the similarly-sized QE2 in which the securities purchased were funded by an increase in reserve balances (see staff memos to FOMC on the MEP here and here).

However, there is a range of views among central bankers globally about whether this other side matters and, in particular, whether the quantity of reserve balances matters.? Some central bankers judge that the added reserves provide additional stimulus by boosting bank lending and others judge that the reserves slow the economy by reducing bank lending.? If the overnight rate is pinned at zero and the economy is teetering on the brink of a deflationary spiral, and buying long-term securities adds stimulus, then quibbles about the economic cost or benefit of reserve balances are understandably cast aside.? But when determining how to implement monetary policy in normal times – with a bloated balance sheet or a lean one – the cost of reserve balances can be dispositive.

Reserve balances are beneficial

While this is not the orthodox Board economist’s view (and not my view), some economists and policymakers judge that the increase in reserve balances is a beneficial side effect of QE (or even the whole point).? In a recent speech, Piero Cipollone, member of the executive board of the ECB, stated that

Research documents the strong relationship between loan supply and structural sources of liquidity, such as reserves obtained through credit easing programmes or those injected through quantitative easing interventions…More specifically, a €1 change in non-borrowed reserves or credit easing reserves is associated with a corresponding change in credit of approximately 15 cents or 10 cents respectively.

Cipollone cites? Altavilla, Rostagno, and Schumacher (2024) as the source of these estimates (link). In a separate paper, the authors argue that the ECB should stick with a floor system in which it oversupplies reserves rather than shifting to a ceiling system in which bank demand determines the quantity of reserves because the added reserves are beneficial. (link)? (The ECB decided to switch to a ceiling system.)

In yet another paper, the economists describe how the added reserves stimulate the economy using terms lifted from the old “money multiplier” story.?

For a bank, creating a loan means expanding deposits as well, as the borrower's loan principal is booked into his sight account with the lender. But the new deposit has a high probability of being monetised in short order, so it requires the backing of more reserves. (link)

In this telling, liquidity requirements serve the role of reserve requirements in the money multiplier to link reserves to deposits and lending.? Indeed, the authors estimate that a 1 percentage point increase in nonborrowed excess reserves results in a 1? percentage point increase in loans to firms, both expressed as a percentage of bank assets.? They define “nonborrowed excess reserves” as reserves minus required reserves minus any outstanding loan from the ECB.? If the same relationship held in the United States, the current $3.3 trillion in nonborrowed excess reserve balances would be creating $5 trillion in business lending, more than a third of all bank loans to financial and nonfinancial firms.?

I confess that I do not understand this view, but that may be because I have never understood how money growth is actually supposed to stimulate the economy – the last mile, as it were.? Moreover, the relationship between reserves and deposits through liquidity requirements is much looser than through reserve requirements.? Lastly, the authors find that when “nonborrowed reserves” go up/down at an individual bank, loans at that bank subsequently increase/decrease.? There are many reasons why this could be true at the bank level that do not imply that a central bank purchase of a security will cause aggregate bank lending to increase.?

Reserve balances are costly

At the other end of the spectrum, other economists judge that the added reserve balances created by QE are costly.? In a New York Fed Staff Study, Afonso, Cipriani, and La Spada (2024) evaluate the cost of reserve balances by examining what happened when the Covid-era exclusion of reserves (and U.S. Treasury securities) from the denominator of the supplementary leverage ratio expired at the end of March 2021 (link).? The supplementary leverage ratio is calculated as the ratio of equity to assets with no weighting for risk – so it requires a bank to fund its reserve balance at the Fed with the same amount of equity as the regulation requires for, say, a subprime mortgage or a construction and land development loan.? By excluding reserve balances from the denominator of the leverage ratio when Covid hit, banks were not required to fund reserves with equity; when the exclusion expired, the cost to banks of lending to the Fed increased.?

When the renewed capital requirement for reserve balances took effect, the invisible hand performed its magic, and the financial system adjusted to settle at the new low-cost way to finance the Fed.? Over $2 trillion shifted from loans from banks to the Fed (in the form of reserve balances) to loans from money funds to the Fed (in the form of reverse repurchase agreements).? The shift occurred even though even though the ON RRP rate was 10 basis points lower than the interest on reserve balances (IORB) rate, indicating that it was at least 10 basis points more costly for banks to finance the Fed.?

Because of the balance sheet costs, high levels of reserve balances may crowd out lending and therefore reduce economic activity.? In a 2023 paper Diamond, Jiang, and Ma find that “…the increase in reserve supply from 2008 to 2017 reduces firm loans extended by an average of $133.9 billion, which amounts to 8.1 cents in bank lending crowded out per dollar of reserves injected.” [emphasis added] (link).

Reserve balances are costless

In between these two views, in a recent speech, Lorie Logan, president of the Dallas Fed, stated that reserves are costless to produce and so should be produced until their marginal benefit is zero (that is, in vast amounts):

Allocative efficiency requires setting the price of resources equal to their social marginal cost. Let’s apply that idea to central bank liabilities, specifically, reserves. From the perspective of a commercial bank considering how many reserves to hold, the price of reserves is the spread between money market interest rates and whatever rate the central bank pays on reserves. That is, when money market rates exceed interest on reserves, the commercial bank pays an opportunity cost in foregone interest to hold reserves…The central bank’s marginal cost of supplying reserves is typically small. Reserves are an electronic entry on the central bank’s books.

If reserves were just a widget that the Fed manufactured to provide banks a unit of contingent liquidity, Logan’s reasoning may be correct.? But reserve balances are a forced borrowing by the central bank from the banking system with attendant costs and consequences.? While each bank chooses what level of reserve balances to maintain, the banking system as a whole has to provide the funds, and the loans to the Fed crowd out loans to businesses and households.? Reserve balances take up bank balance sheet space in part because, as noted, banks are required to fund reserves partly with equity, which is costly.? Reserves also impair banks’ stress test performance because noninterest expenses are assumed to grow with bank size, which depends in part on the amount of reserves a bank holds. Finally, reserve balances also raise the GSIB capital charge by increasing bank size.? For further discussion see Covas, Collard, and Waxman (2022).

Ways to lessen the deadweight cost of high levels of reserve balances

I have written extensively about the considerable costs associated with the Fed’s choice to conduct monetary policy with a bloated balance sheet rather than a lean one (see link and link (or link)).? In particular, an unbounded balance sheet is a risk to the Fed’s independence; an unbounded balance sheet contributes to Fed complacence about taking on interest rate risk; the more reserves the Fed creates, the more banks (and bank supervisors) demand, resulting in the Fed getting ever larger; the oversupply of reserves has crushed the federal funds market; and because banks do not need to borrow from the discount window, stigma has been increased.? The crowding out of lending to businesses and households by reserves discussed here is yet another reason to implement policy using a lean balance sheet.

To reduce the deadweight cost of requiring banks to lend the Fed $3 trillion in reserve balances in the steady state, the Fed and other banking agencies could fix the supplementary leverage ratio requirement by either simply reducing it or by excluding reserves (and potentially other safe assets) from the denominator, or both (see Covas and Rosa (2025)).? In addition, the Fed could revise liquidity regulations to recognize the value of prepositioned collateral at the discount window.? Doing so would allow banks to make loans to business and households which they would then pledge to the discount window rather than simply lending reserve balances to the Fed.

As always, please feel free to share this email with anyone.? I’m happy to add anyone to this free distribution list, just email me.? Comments and discussion always welcome.

Bill

?

Bill Nelson | Chief Economist | Bank Policy Institute | 1.703.340.4542