Revue Banque – QT is not quite QE in reverse, but it’s still tightening

Translation of an article published by Revue Banque early March 2025
By Michala MARCUSSEN, Group Chief Economist

The ECB has been shrinking excess liquidity since late 2022. In the initial phase, the decline was driven by banks repayment of targeted long-term refinancing operations (TLTRO). Since then, the phase-out of reinvesting proceeds from maturing assets, held for monetary policy purposes, has become the driving force. 2025 should see a further run-off of around €40bn per month from maturing assets, taking excess liquidity to around €2.5tn by year end from its peak of just under €4.75tn in late 2022.

While this quantitative tightening (QT) is not quantitative easing (QE) in reverse, several questions merit attention not least as the ECB shifts to a new operating model that could drive structurally higher term premia.

QE is a useful tool where interest rates are close to the lower bound

In its simplest form, QE is merely an asset swap between the commercial banking system and the central bank. The central bank buys assets, most commonly government bonds, from the banking system in return for bank reserves. While the central bank balance sheet expands, there is no corresponding expansion of the commercial banking system’s balance sheet. In this closed loop, the main channel of QE becomes the difference in renumeration of central bank reserves relative to that of the assets purchased.

This typically lower return should incentivise banks to seek replacement of the now higher share of central bank reserves held with higher yielding assets purchased from the non-bank system. This generates several channels of transmission.

First, increased demand for higher yielding assets should, all else being equal, compress term premia, lowering interest rates across the curve through a portfolio rebalancing effect. Second, the cash generated to the non-bank system creates bank deposits (the ECB does not allow non-banks access to deposit cash on its balance sheet), which in turn may be lent out to the real economy, all else being equal, increasing credit supply at lower lending rates. As QE is further credited with a signalling effect of lower for longer rates, both channels may be enhanced, making QE a useful part of the monetary toolkit when interest rates are close to the lower bound.

Two points deserve note here. First, a portfolio rebalancing and credit supply channel may also be generated by the central bank purchasing assets directly from the non-bank system. Second, although individual banks may seek to replace reserves with higher yielding assets, the banking system in aggregate cannot alter the amount of bank reserves created by QE.

QT is not quite QE in reverse, but still lifts term premia

Turn now to ECB’s present QT. The main reason why this the is not merely QE in reverse has to do with duration effects.

In allowing maturing assets, held for monetary policy purposes, to run off its balance sheet, the ECB is de facto reducing its holding of short-dated bonds, while still retaining those with longer maturities. As the respective original bond issuers (in this case primarily governments) refinance these maturing bonds, supply will increase across maturities as a function of the issuers’ decisions lifting-term premia.

Recall that under QE, central banks typically purchase longer-dated bonds, seeking to directly impact the portfolio rebalancing channels through compression of term premia. While the average weighted maturity of central bank asset holdings typically ramps up quickly under QE, the current passive QT engaged by the ECB seeks slow reduction only of the average weighted maturities, all else being equal producing a lesser impact on the term premia as it reduces its asset holdings. For QT to truly be QE in reverse, the ECB would have to sell longer dated assets.

The term premia channel of QT is nonetheless playing out and Akkaya et al (2024)¹ suggest a non-negligible impact in the pipeline. Drawing on the ECB’s Survey of Monetary Analysts, the authors find that a mechanical reduction of ECB bond holdings by €1tn would increase the term premium on the German 10-year benchmark by 35bp. This is in the lower range of previous studies suggesting that a €1tn increase of QE produces a reduction of around 35 to 65bp. Although such studies come with uncertainty, the mentioned duration effects offer a solid argument as to why the current passive QT should have a lower impact on term premia that QE.

Demand driven reserves differ from supply driven ones

Turning to the bank deposit channel, we again note two possibilities. If the central bank is merely removing the bank reserves linked to the maturing bonds from the banking system, which then buys the newly issued bonds, then we are in essence back in the case outlined above of a simple asset swap (this time in reverse). All else being equal, QT will in this case have little impact on bank deposits.

If, on the other hand, the newly issued bonds are absorbed by the non-bank sector, then this will all else being equal shrink bank deposits and reduce credit supply. At present, ECB rate cuts at least in part offset this effect as lower interest rates help support credit demand, encouraging banks to lend. Recall that bank loans create bank deposits, all else being equal.

As central bank reserves are absorbed by QT, this nonetheless has important implications for the banking system. Should the banking system desire higher level of excess reserves, due for example to the stricter prudential regulation of Basel III, then this could add both unwanted upward pressure on money market rates and add unwanted volatility.

The ECB is thus closely monitoring developments and has with its new operational framework, introduced in March 2024, already signalled that it will at some stage introduce new structural longer-term lending operations in addition to its regular shorter-term refinancing operations already at work. In further steps, we have also seen the ECB reduce the spreads between its main refinancing operation and its deposit rate from 50bp to 15bp. The ECB has also signalled that it will hold a structural portfolio of bonds. As such, the overall size of its asset portfolio will not be returning t its pre-QE levels, albeit that its future size is still unknown.

In allowing banks to access longer-dated liquidity, and this without stigma, the ECB would indeed secure the banking system access to bank reserves allowing both liquidity coverage and net stable funding ratios to be readily managed. The mechanics here are well-known from previous longer-term refinancing operations, such as the LTRO. The challenge will be in the calibration with the overall desired stance of monetary policy.

It is worth note, however, that the duration of the ECB’s past longer-term refinancing operations has typically been much lower than duration of its refinancing operations. At its peak, the weighted asset maturity on its portfolio of securities for asset purchases was close to 7 years, while the longest maturity offered for longer-term refinancing operations was 3 years. While asset purchases and longer-term refinancing operations operate through different channels, both create bank reserves, have opportunity to create banks deposits and can compress term premia.

If the ECB in the future structurally favours longer-term refinancing operations over longer-dated asset purchases, as its announcements on the future operation model suggests, then there is a question of duration to be addressed. The shorter the structural duration of the assets held by the ECB (be these asset purchases or refinancing operations), the higher the structural term premia.

¹Akkaya, Yildiz, John Hutchinson, Kasper Joergensen and Emmanuel Skeppaas, Quantitative Tightening: How do shrinking Eurosystem bond holdings affect long-term interest rates? ECB Blog, 14 November 2024.