The unchartered territory of Quantitative Tightening
The major central banks are normalising their monetary policy, both in terms of policy rates and non-conventional policies. The Fed has already started to reduce its balance sheet again (Quantitative Tightening), and the ECB is expected to follow suit from 2023 on. The rationale for starting QT along with raising policy rates includes freeing up capacity for future QE, escaping the threat of fiscal dominance and reducing the risk of capital losses in an environment of rising interest rates. Although the effect is difficult to assess, early estimations by Fed staff suggests that QT has a small but sustained tightening impact on financial conditions. However, the impact of QT is probably asymmetric to the effect of QE. One reason is the smaller policy signalling role of QT compared to QE, when policy rates are typically near their effective lower bound. Moreover, the impact of QT is probably non-linear, increasing as the remaining amount of reserves shrinks towards the critical threshold that markets need to function properly. This all implies that central banks are entering unchartered territory with their QT programmes.
In 2021-22, all major central banks carried out a turnaround towards policy normalisation, and in some instances even towards a restrictive policy stance. The Fed and the ECB have ended net purchases under their respective Quantitative Easing (QE) programmes in 2022, and started a policy rate hiking cycle soon thereafter. ‘Forward rate guidance’ as a policy instrument in its own right has effectively come to an end. The next logical step in the sequencing of policy tools is the unwinding of the large stock of assets (Quantitative Tightening (QT)) that central banks have accumulated during the period of QE programmes.
The Fed started its QT programme in June 2022. In September 2022, it was fully phased in. It works by not reinvesting maturing Treasury bonds and Mortgage Backed Securities (MBS), with a target pace of balance sheet deflation of 95 bn USD per month. Meanwhile, the ECB has communicated that it intends to reinvest maturing assets of its Pandemic Emergency Purchase Programme (PEPP) until at least the end of 2024. The ECB’s total PEPP portfolio currently amounts to about 1.7 trillion EUR. However, the ECB is expected to soon announce details about the (partial) unwinding of its general Asset Purchase Programme (APP) portfolio (of currently about 3.3 trillion EUR). This is likely to start in 2023.
In contrast to the Fed, the ECB not only conducted QE through outright asset purchases, but also through large-scale longer-term refinancing operations (refis) to the financial sector. Since the European sovereign debt crisis, these liquidity provisions have gradually extended the traditional toolkit of the MROs by instruments with longer duration, such as the LTRO’s, VLTRO’s, PELTRO’s and most recently the various vintages of TLTRO’s .
In October 2022, the ECB started a first move towards QT by making the terms of the remaining TLTROs less attractive (effective from 23 November on). The objective is to stimulate banks to repay them prematurely, thereby withdrawing liquidity from the financial system (i.e. QT). The ECB targets the phasing out of the TLTROs before adapting its APP portfolio. Quantitative monetary policy via longer-term refis is relatively unreliable. Unlike QE via the outright purchase of financial assets, the effectiveness of policy based on refis depends on the stability of the financial sector’s demand for the liquidity offered. Instability may interfere with the ECB’s policy target, as happened in 2013-14, when demand for liquidity fell and the ECB balance sheet shrank as a result, leading the ECB to start the outright APP.
Why QT ?
It is sometimes argued that balance sheet reduction via QT may not be necessary, because the policy stance can be tightened by as much as needed by raising policy rates since there is no effective upper bound on rates. However, there are some consensus arguments in favour of balance sheet normalisation.
Since quantitative monetary policy has by now been accepted as part of the conventional monetary toolkit, QT would free up capacity for future rounds of QE when the policy rate hits the effective lower bound again. Without QT, if the effective lower bound rate is hit often, central banks’ balance sheets would be on an ever-increasing and unsustainable trajectory throughout the cycles. This would affect the proper function of financial markets. For example, there would be the risk of creating a structural shortage of ‘low-risk’ collateral assets.
Moreover, there is a case for the Fed to get rid of its holdings of MBS, since these holdings affect the sectoral allocation of capital, by having an impact on the relative cost of capital between sectors. Such a sectoral policy would rather be a task for the elected budgetary authorities than for a central bank.
There are also the fiscal implications of quantitative monetary policy. For central banks, QT is a way of reasserting their independence and reducing the risk of fiscal dominance on monetary policy. As long as government bonds remain on the central banks’ balance sheets, they ultimately amount to monetary financing of fiscal deficits and public debt. This occurred visibly with the PEPP during the pandemic.
Keeping these large bond portfolios on their balance sheets would also expose central banks to a significant interest rate risk as they start a rate tightening cycle. There would be a capital loss on the bond portfolio held by central banks, potentially eroding the central banks’ own capital. Balance sheet reduction is a way for central banks to reduce this risk (Cavallo et al. 2019).
Asymmetry
Some economists (e.g. Cúrdia et al. (2010), Bullard) argued that, theoretically, quantitative monetary policy should have no impact, since central banks essentially swap one government liability (base money) for another (bonds). Similarly, former Fed president Bernanke remarked that “[...]the problem with QE is it works in practice, but it doesn’t work in theory”. In practice, however, this neutrality proposition seems implausible, since these government liabilities are not perfect substitutes and differ, for example, in terms of duration and access constraints.
The precise impact of QT on the economy is, however, difficult to assess. The only empirical experience is from the Fed in the period 2017-19. The current QT is similar, albeit at a faster pace (Figure 1). Fed staff (Crawley et al. 2022) estimate, with considerable uncertainty, that a balance sheet reduction by about 2.5 trillion USD over the next few years would be equivalent to a rise of the policy rate of about 50 basis points on a sustained basis.
Research suggests that the economic impact of QT may be more limited and not symmetric to the impact of QE (Sablik (2022), Bullard (2019)). The main asymmetry in impact lies in the relative strength of the signalling effect. QE signals to the market that the central bank intends to keep its policy rate close to its effective lower bound for an extended period of time. This signal is reinforced by the implicit precommitment of the central bank to first reduce its balance sheet again before it can raise policy rates, in order to avoid capital losses on its accumulated portfolio (see also above). In times of QT, these signalling effects about policy rates are much weaker or even non-existent.
Finally, the impact of QT may also be non-linear and kick in fully only once the critical level of reserves for the market to function properly has been reached. As reserves shrink, each additional amount drained may have a larger effect on interest rates. Properly calibrating QT is indeed unchartered territory and a huge challenge for central banks.