Maren Froemel, Mike Joyce and Iryna Kaminska
During 2020 the MPC announced a further £450 billion of QE purchases, slightly more than the total amount of assets purchased over the preceding ten years, taking the target QE stock to £875 billion of gilt holdings and £20 billion of sterling investment-grade corporate bonds. We study the high-frequency reaction of gilt markets to these QE announcements in light of the surprises to market expectations of the future QE path. We find the yield reactions to be broadly consistent with news about the expected medium-term stock of QE. This is in line with recent commentary, which has focused on the ‘pace of purchases’, as a faster/slower pace translated into a larger/lower stock of expected purchases, and could capture the effects of the local supply channel. The reaction to news about purchase pace could also be potentially consistent with an impact on expected liquidity premia or expected policy rates.
The large expansion of QE during 2020 was announced in three separate instalments, starting with the special MPC meeting on 19 March – where the Market Notice published a few hours after the main announcement provides us with an additional QE event – and followed by the two scheduled MPC policy meetings on 18 June and 5 November. Using an event study approach and the information from market intelligence about the expectations of the future QE path, we try to draw some lessons from the intraday reaction of gilt yields following each of these announcements.
The high-frequency reactions of markets to monetary policy decisions enable us to examine the incremental impact of QE, isolating it from other potential drivers. It is important to get the appropriate-sized window, although there is always a difficult trade-off between excluding other unrelated events and capturing the full market reaction. Here, we focus on narrow windows using high-frequency data to minimise the possibility of the results being contaminated by other similar international policy moves, which have been important over this period. In particular, we consider narrow 30-minute windows centred on each of the four announcements, as the initial intraday analysis of gilt prices around the announcements suggests that the peak of the reaction tended to be within the first 15 minutes.
Surprises to the expected QE path
The three announcements for QE in 2020 varied in size, but also in the pace and length of each announced programme. The sequence of communications was:
- the initial unscheduled announcement on 19 March of a £200 billion expansion of QE, which was a surprise increase of around £100 billion in the expected stock of QE, according to market intelligence;
- a subsequent Market Notice on the same day indicating an unusually fast pace of purchases (stating explicitly that ‘the Bank intends – at least initially – to make purchases at a materially higher pace than in the recent past’) and a change in the definitions of ‘medium’ and ‘long’ maturity purchase buckets (resulting in higher expected purchases of longer-dated gilts);
- an announcement on 18 June of a further £100 billion expansion of purchases, which was in line with market expectations but at a slower-than-expected pace, implying a smaller overall programme (resulting in a surprise tightening of £50 billion, measured by end-year expectations); and
- the larger-than-expected £150 billion of purchases announced on 5 November, which nevertheless did not alter the expected path of purchases for the remainder of 2020 and matched market expectations for the QE stock in 2021 (suggesting there was no material surprise in the broad stance of policy).
By examining surprises to QE expectations and gilt yield reactions jointly, we can make inferences about the stock effects of the QE programmes. The initial analysis, described in the rest of the post, suggests that what broadly seems to matter is the expected stock of purchases over the medium term, although there is some evidence consistent with a distinct impact from the expected purchase pace in March.
The intraday yield reaction
Table A shows immediate yield reactions and the corresponding surprises to the expected stock of purchases at various horizons for each of the QE announcements (based on information from market intelligence), demonstrating the relationship between surprises in yields and QE expectations more explicitly. ‘Headline’ measures the surprise to expectations of the Bank’s target stock of QE, whereas ‘Medium term’ and ‘Longer term’ denote the surprise 3–6 and 12 months out, respectively; all based on market intelligence. While the immediate yield curve reaction to nearly half a trillion of asset purchases might seem relatively modest, it is consistent with the corresponding size of the surprises to the expected stock over the medium term.
Table A: Changes in zero coupon 10-year gilt yield and surprises to expected stock at various horizons
|Gilt move (10-year)||Headline||Medium term||Longer term|
|March||-16||£100 billion||£100 billion||£100 billion|
|June||+7||£0 billion||-£50 billion||£0 billion|
|November||-3||£50 billion||£0 billion||£0 billion|
The yield curve reactions themselves, at least in a small window around the announcements, appear consistent with the operation of the local supply channel (ie the effect of asset purchases on gilt yields arising from the increased scarcity of specific gilts). As Chart 1 shows, the size of the yield reactions was proportional to the market surprises in the expected size of QE over the medium term: the yield reaction to the unexpected March QE easing announcement was roughly twice as large in absolute terms as the yield curve reaction to the surprise slowing of QE purchases implied by the June QE announcement and had a similar pattern across maturities; by contrast, the overall QE stance announced in November was not a surprise to the market and as a result the yield curve barely moved in reaction.
One implication is that communication about the pace and length of QE purchase programme is an important tool for policymakers because markets use it as an indicator to infer longer-term changes in the target stock of QE.
The Market Notice published on 19 March, a couple of hours after the initial announcement of the size of the purchases, provides an interesting additional case study. This event had a substantial effect on the curve, depressing the yields by further 8–15 basis points (Chart 1). The large immediate reaction of long-term yields is consistent with a local supply channel, as the announced changes to maturity buckets implied larger purchases at longer maturities. However, shorter-maturity yields fell as well by around 9 basis points. In principle, this fall could also be in line with the local supply channel, as the faster pace of the purchases inferred from the Notice could have encouraged market participants to upgrade their expectations of the target QE stock by end-2020. But the fact that yields decreased substantially also for very short maturity gilts, ineligible for purchase, indicates that other channels were in operation. For example, implied larger and quicker purchases could have also signalled a lower path for future policy rates (signalling channel). Or, the quicker pace of purchases scheduled to start imminently during a period of severe market stress could have lessened the liquidity pressures expected by market participants, resulting in lower expected liquidity premia across the whole curve (liquidity channel).
Chart 1: Change in zero coupon curves around the short window for all announcements
Sources: Bank of England and Bloomberg.
While it seems plausible that the effect of QE announcements on gilt yields will be stronger during periods of high market dysfunction, such as during the global central bank responses to the Covid-19 pandemic in March 2020, our event study analysis does not allow us to form a definitive conclusion on this. On the one hand, as Chart 1 shows, the immediate yield response was roughly proportional to the size of the implied surprise (if not more pronounced for longer yields). On the other hand, while market liquidity indicators (eg bid-ask spreads) showed a high level of market stress prior to 19 March, we do not see a significant improvement in these indicators during our event study windows.
Combining information on market expectations about the path of QE and high-frequency gilt yield data around the 2020 QE announcements, we find that the market reaction was affected by news about the pace of purchases and the length of each announced programme, as well as by the headline quantity of purchases. In other words, news about the whole path of future QE purchases was potentially relevant. Gilt yield reactions to last year’s QE announcements were broadly in line with the effects of the local supply channel, but the evidence suggests other transmission channels have also played a role, in particular during times of market stress. The evidence from the market reaction to the March Market Notice is potentially consistent with ‘pace’ having a distinct impact on expected liquidity premia in conditions of severe market stress, but we also cannot rule out some role for the signalling channel either, as larger and quicker purchases could have also signalled a lower path for expected future policy rates.
Maren Froemel, Mike Joyce and Iryna Kaminska work in the Bank’s Monetary and Financial Conditions Division.
If you want to get in touch, please email us at email@example.com or leave a comment below.
Comments will only appear once approved by a moderator, and are only published where a full name is supplied. Bank Underground is a blog for Bank of England staff to share views that challenge – or support – prevailing policy orthodoxies. The views expressed here are those of the authors, and are not necessarily those of the Bank of England, or its policy committees.