Jack Worlidge
Purchases of government bonds have been a prominent tool that has helped central banks meet inflation objectives when short-term interest rates have been constrained by their effective lower bounds. But how does QE work? There are a range of channels through which QE can/might operate, though there remains uncertainty over the relative size and importance of these channels. This post presents new evidence from granular transaction data consistent with a portfolio rebalancing channel. Specifically, during the Bank’s latest QE programme (known as QE5) investors were found to have bought less new gilt issuance and bought more risky assets like corporate bonds.
The latest round of such Quantitative Easing (QE) in the United Kingdom, launched in response to the effects of Covid in 2020, took the total size of gilt purchases to £875 billion. That is almost half the size of the valuation of the entire FTSE 100.
The portfolio rebalancing channel is one way that QE is thought to work in practice…
The ‘portfolio rebalance’ channel is one of the most prominent channels through which QE is thought to operate. It works if some investors prefer to hold government bonds for reasons not related to their price. For example some investors may have a preference for the long dated maturities that gilts can offer, and have a preference for certain gilts, known as a ‘preferred habitat’. Such ‘preferred habitat’ investors for the UK gilt market have been identified as being foreign central banks as well as insurance companies and pension funds (known as ICPFs).
Because of their preferences, these investors require compensation for selling their gilt holdings, and so the price of government bonds has to increase when central banks buy government bonds (and correspondingly the yield falls). Several authors have demonstrated how QE works to lower government bond interest rates. In turn, investors are unlikely to prefer holding the cash that they obtain from parting with gilts, and so the result is that they are likely to rebalance their portfolios away from gilts and towards other – riskier – assets like corporate bonds. That might be because they would still like assets with long dated cash flows, or it might be because QE encourages them to engage in a broader ‘search for yield’, looking for assets that generate a higher return. And in the end, this should equate to easier financing conditions in the real economy.
Lots of academic work has looked at prices to assess whether a portfolio rebalancing channel operates in practice. But there is a much smaller body of literature that tries to look at evidence from portfolios – or quantities. How different sectors respond to QE is a question that remains open. Do certain types of financial companies switch into riskier assets?
New data allows for novel and unconventional ways of assessing unconventional monetary policy…
New granular transaction level data has opened up new opportunities to explore portfolio choices. With the MiFID II financial regulation implemented in 2018, regulators have gained a new avenue for research based on insight into gilt and corporate bond transactions of individual financial firms. This work follows several papers making use of transaction data to examine government bond markets. Following work to examine the effect of QE in Sweden, I examine MiFID II data from a UK perspective for a sample period of 2018–20, including the first £260 billion of QE5.
I find evidence that Insurance companies and pension funds (ICPFs) reduced their purchases of newly issued gilts during QE5 in the United Kingdom
The transaction data suggests that ICPFs and foreign central banks were not major active net sellers of gilts in the open market during the QE programme in 2020. Chart 1 shows the total estimated net purchases of gilts by different sectors during the QE5 period (March–December 2020), expressed as a share of the QE purchases made over that period by the Bank of England. It shows that ICPFs, asset managers, foreign central banks and banks were all net buyers of gilts during the QE5 period. That is contrary to what might be expected from the typical description of portfolio rebalancing, which describes the portfolio channel as operating by inducing sales of gilts by preferred habitat investors.
Chart 1: ICPFs were not a major active seller of gilts
Net purchases of gilts by sector March 2020–December 2020
Sources: MiFID II, DMO and author calculations.
Chart 2: But ICPFs and asset managers bought fewer gilts issued by the DMO than in the past Net purchases/sales by sector, of gilts issued by the DMO (+-1 week)
Sources: MiFID II, DMO and author calculations.
But ICPFs did rebalance their holdings away from gilts. What’s interesting is that, during QE5, some sectors rebalanced by reducing their net purchases of newly issued gilts relative to their past behaviour over the preceding two years. Chart 2 shows net purchases/sales by sector in bonds that the DMO issued in the QE period in 2020 (pink) and in the previous two years (blue). During QE5 in 2020, ICPF and asset managers acquired a much smaller share of gross issuance.
Rolling substitutions – what is the impact for the corporate bond market?
To see whether this reduction in the rate of net acquisition of newly issued gilts has any implication for portfolio rebalancing into other assets, I look at firm activity in the sterling investment grade corporate bond market. This is featured in the same transaction data and is a likely candidate for observing any rebalancing that may be occurring because it is also a fixed income asset. I estimate net demand for corporate bonds as a function of the realised net investment in newly issued government bonds. To do this, I run a (weekly) panel regression with firm and time fixed effects.
Over the sample period, there is evidence of substitution to the investment grade corporate bond market by firms that reduced their net investment in newly issued gilts (a negative estimated coefficient). Chart 3 visualises regression estimates and 95% confidence intervals of regressing weekly firm level investment of corporate bonds on firm investment in the gilt primary market, interacted with different sector types. The chart presents estimates for two models. First, a firm level fixed effects model which includes financial controls (like the S&P 500, VIX and 10 year yields) is shown in blue, and a second specification which includes time fixed effects, in yellow. Both models produce similar results. The estimates for ICPFs statistically significant and suggest a substitution effect between the two markets, although the magnitude of the estimate is very uncertain. For every £1 billion of foregone investment in gilts, ICPFs are estimated to have invested £180 million more in investment grade corporate bonds (estimate for ICPFs in blue in Chart 3). Of course, there could also have been rebalancing into other markets, which is not looked at here.
Chart 3: Estimated marginal effect on net corporate bond investment of firms’ purchases of newly issued gilts
For two different model specifications
Sources: MiFID II, DMO and author calculations
There is evidence consistent with a portfolio rebalancing channel being in effect during QE5 in 2020
The conclusion is that ICPFs reduced their net acquisitions of gilts during the QE period and increased their investment in corporate bonds, consistent with a portfolio balance channel. These results are in line with the previous UK experience for ICPFs, and point to evidence of a portfolio rebalancing channel in operation during QE5. To put the regression results into a sense of scale, the estimates suggest that for the £260 billion the QE5 programme over 2020, reduced purchases of gilts by ICPFs implies increased net investment in corporate bonds of around £4 billion, or approximately 1.6% of the size of the market.
Jack Worlidge works in the Bank’s Market Intelligence and Analysis Division.
If you want to get in touch, please email us at bankunderground@bankofengland.co.uk 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.
