By Alessandro Rebucci and Sinem Yagmur Toraman
Johns Hopkins Carey Business School, CEPR and NBER
Johns Hopkins University, Economics Department
On Wednesday September 28, 2022 the Bank of England announced a short-term Gilt purchase program to bail out the UK pension system. The 30-year yield tanked, UK pension funds were rescued, and, quite surprisingly, the British pound appreciated sharply against the US dollar. This note argues that financial stability QE is different than monetary policy QE and can appreciate the exchange rate if it squeezes a buildup of levered intermediaries’ short positions against a currency.
On Wednesday September 28, 2022 at 11 am local time, the Bank of England announced it was intervening in the Gilt market, committing to do whatever it takes, for a limited period of time, to restore financial stability amidst severe stress in the UK pension system. A classic lender of last resort intervention. Two minutes after the announcement, the 30-year Gilt yield dropped about 30bps, as intended. At the same time, contrary to the predictions of standard models and unlike previous QE episodes during the Global Financial Crisis, the US Dollar–British pound exchange rate (GBP/USD rate) appreciated rather than depreciating (Figure 1).
During the subsequent 60/120 minutes, both the 30-year yield and the GBP/USD rate at least twice reversed their course but then resumed moving in the initial direction (Figure 2). At about 12:30 pm UK time, the GBP/USD rate started to rally, continuing to appreciate for the rest of the day, to close at 1.0887, or 1.4 percent stronger than the previous close (Figure 3). Moreover, the GBP/USD rate continued to appreciate on Thursday, 29, and was stable on Friday, 30, through the end of the trading week on global foreign exchange markets.
This QE episode is a good example of how different time windows can lead to different conclusions on the asset price impact of a policy intervention, as there is no “right” answer to what the right window should be. Intraday studies often consider 5 minutes before and after the event, but it could be anything from 15 minutes to 2 hours afterward. In this QE episode, the 5-minute change shows a little appreciation using tick data (Figure 1). The daily change points to a significant and sustained appreciation over a 1-day, 2-day, or 3-day period. At the same time, windows of varying lengths between 30 minutes and two hours would not provide conclusive evidence or point to a depreciation. Nonetheless, a consensus has emerged in the markets and the policy community that this QE intervention significantly appreciated the GBP/USD exchange rate.
Financial stability QE is different than monetary QE
From a theoretical perspective, QE appreciating the exchange rate is puzzling, as standard models predict a depreciation (e.g., Greenwood, Stein, and Sunderam, 2020; Gourinchas, Ray, and Vayanos, 2020). Available evidence on GFC-era QE also shows that monetary policy QE geared toward lowering long rates or providing liquidity to banks leads to reflate the economy also depreciated the adopting currency (e.g., Rogers, Scotti, and Wright, 2018). The latest BoE’s QE intervention is particularly puzzling as it materialized in the context of a previously announced large fiscal expansion, a higher-inflation environment at home than abroad, and a period of extreme US Dollar strength—three factors that pushed the pound to historic lows against the dollar in the days before the dramatic intervention and arguably fueled the buildup of speculative positions against the pound.
Like the BoE, several other advanced economy (AE) and emerging market (EM) central banks adopted QE to ease domestic financial conditions at the onset of the COVID-19 pandemic in March 2020. Unlike during the GFC, when EMs attempted to slow depreciations with foreign exchange intervention, many of these countries saw their exchange rates vis-à-vis the US Dollar depreciating less than in AEs, and in some cases, even appreciating (Benigno, Hartley, García-Herrero, Rebucci, and Ribakova 2020 and Rebucci, Hartley, and Jiménez 2022). This is a finding corroborated by several studies, including for example Arslan, Drehmann, and Hofmann (2020), Hordahl and Shim (2020), Sever, Goel, Drakopoulos, and Papageorgiou (2020), and Fratto, Vannier, Mircheva, de Padua, and Ward (2021).
A critical difference between AEs and EMs during the COVID episode, was that EMs adopted QE with policy rates still well above the ZLB, as in the current BoE case. This illustrates that the policy objective and constraints of financial stability QE are different from monetary policy QE, opening the possibility that the impact on the exchange rate and transmission mechanism is also different.
One important caveat here is that COVID financial stability QE episodes took place in a context of simultaneous intervention by more than 30 central banks, after a very aggressive Fed intervention in a disrupted US treasury market, and Dollar liquidity provisions through the SWAP lines. In fact, during the COVID episode, all countries except Croatia and Sweden intervened after the Fed.
However, the appreciating impact of COVID-era financial stability QE survives even after controlling for a comprehensive set of Fed actions by using Swanson’s (2021) methodology modified to account for the SWAP lines. To capture the stabilizing effect of the SWAP lines, Toraman (2022) adds the CIP deviations to the information set used in Swanson (2021) to estimate the Fed policy factors. She then estimates panel regressions similar to those used in the literature cited above controlling for four Fed policy factors: the monetary policy rate, forward guidance, asset purchases, and the SWAP lines. She finds that controlling for the SWAP lines in this manner removes the significance of the coefficient of the asset purchase surprise on the exchange rate for AEs but not for EMs. Although this evidence does not control for the difference between financial stability QE and monetary QE interventions during COVID-19, it differentiates between public and private asset purchases, finding that the appreciation of the exchange rate in response to an asset purchase surprise is stronger for purchases of private assets compared to purchases of government securities—a finding that is not inconsistent with the notion that financial stability QE may have a different impact on the exchange rate.
What might explain the appreciating impact of financial stability QE in the UK?
Understanding how financial stability QE might be appreciating the exchange rate is important, as such interventions may permit liquidity provision in local currency in instances in which stabilizing the exchange rate is a policy objective. And the special circumstance of the BoE’s intervention offers the opportunity to shed some light on this issue.
It is now well understood that the BoE’s intervention short-circuited the doom loop in which the UK pension funds got caught up. By regulation, these funds must hedge assets and liability duration, and to accomplish this efficiently they use derivative strategies. When Gilt prices suddenly moved lower after the announcement of a massive fiscal expansion a few days prior, many pension funds became exposed to margin calls that forced them to sell Gilts to meet requirements. By suddenly starting to absorb demand, the BoE’s intervention stopped the price decline in the assets in the eye of the storm and interrupted the vicious cycle.
In a similar fashion, but likely unintentionally in the case of the BoE that did not have an exchange rate objective, the QE intervention may have also induced intermediaries to unwind leveraged short positions against the pound, betting on the continuation of the pre-intervention macro trends. As we saw earlier, BoE’s QE suddenly lowered Gilt yields, and hence shifted relative returns against pound short positions in the derivative markets that were betting on continued depreciation driven by rate increases. Such a reversal increases margin requirements on these positions and requires the exposed intermediaries to run for the cover by posting additional pound-denominated collateral, appreciating the currency in the spot market. Indeed, central banks’ use of SWAP lines and non-deliverable forwards, long used in EMs to defend the currency, suggests that this might be an effective way to defend the currency and possibly less costly that intervening in the spot market by selling costly and limited official reserves.
In conclusion, the BoE episode illustrates that financial-stability QE is different than monetary policy QE and it can appreciate the exchange rate. In the case of the BoE, it was enacted above the ZLB and in the context of a monetary policy tightening. It targeted a specific segment of the market, committing an unlimited amount of purchases for a limited period of time. And it appreciated the exchange rate rather than depreciating it, possibly by forcing intermediaries to cover leveraged positions against the currency that was betting on higher interest rates. In the absence of evidence on intermediaries’ positions and a macroeconomic model, nothing definitive can be said on the precise mechanism with which such interventions might be able to support a currency or the desirability of using such a tool to intervene on the exchange rate in a situation of market distress. The BoE’s episode, however, clearly illustrates the possibility to target the exchange rate with QE rather than foreign exchange rate intervention.