The transmission of unconventional monetary policy in UK government debt markets.
Meaning, Jack ; Warren, James
Through its quantitative easing programme the Bank of England has
looked to manage the supply of nominal UK government securities in order
to lower interest rates. In doing so it has removed more than 25 per
cent of the overall supply of those securities from the publicly
accessible market. The benchmark New Keynesian model suggests this
should only have an I impact on interest rates insofar as it affects
expectations of future policy rates, whilst alternative theoretical
frameworks imply a direct effect of changes in supply onto yields. Our
aim is to test for the existence of these potential transmission
mechanisms. We find empirical evidence to support the existence of both
channels. Our analysis suggests the Bank's quantitative easing
programme reduced yields by around 25 basis points through the supply
channel alone. Importantly, we find that such supply effects have
remained significant in recent years, suggesting that as quantitative
easing is unwound the increase in publicly available supply will put
upward pressure on interest rates. Lastly we highlight the monetary-
fiscal interaction inherent in our result and discuss some of the issues
it raises for policymakers.
Keywords: quantitative easing; unconventional monetary policy;
asset purchases; monetary-fiscal
JEL Classifications: E520; E580; E630
Introduction
Since March 2009, the Bank of England's quantitative easing
programme has seen the UK's central bank purchase 375bn [pounds
sterling] worth of assets from the private sector, funded by the
creation of reserves. These purchases have predominantly been of UK
government securities. In fact, the Bank of England, via the Asset
Purchase Facility (APF), has bought and now holds over 25 per cent of
the total outstanding supply of nominal gilts. This has expanded the
Bank's balance sheet dramatically, but also represents a
substantial reduction in the quantity of those securities available to
the private sector.
One important aim of the policy has been to lower medium and
longer-term interest rates and thus stimulate economic activity through
the traditional Keynesian transmission mechanism. (1) However, exactly
how such policies can achieve this aim has been a question of ongoing
academic debate.
According to the expectations hypothesis, which lies at the centre
of the benchmark finance model, interest rates on term bonds are
determined by the expected sum of short-term interest rates over the
life of the bond (equation 1). If they weren't, this would open up
an arbitrage opportunity which perfectly functioning and complete
financial markets would instantly exploit until it no longer existed. In
its slightly weaker form a wedge is often introduced between this
expectational term and the prevailing interest rate to represent a
premium but this wedge is assumed to be constant.
[R.sup.n.sub.t] = [n.summation over (t=1)] [i.sub.t] + [rho] (1)
where [R.sup.n.sub.t] is the interest rate on a bond with maturity
n, [i.sub.t] is the short-term nominal interest rate and p is the
premium associated with the bond.
This assumption is at the heart of the canonical New Keynesian
model and, along with assumptions that the representative household is
infinitely lived and rational, allows the constellation of interest
rates that exist in practice to be collapsed down to a single,
one-period interest rate. As shown famously by Eggertson and Woodford
(2003), under the assumptions of the baseline New Keynesian model
central bank asset purchases are only effective insofar as they act as a
signal of the future path of short-term policy rates. It is posited that
they may do this by strengthening the incentive for the central bank to
hold rates lower for longer, as increasing rates will impose a cost to
the central bank's balance sheet, or by clarifying the central
bank's commitment to accommodative policy.
Eggertson and Woodford's result is dependent on the
no-arbitrage element of the New Keynesian model with which they work,
and a second fundamental concept, that of Wallace neutrality. Laid out
by Wallace (1981), this is a form of Ricardian equivalence for central
bank asset purchases. It states that a transaction which moves assets
between the private and public sectors should be inconsequential for
prices. This is because the asset the private sector agent has given up
is now held by the public sector, of which the private sector agent is
the ultimate owner. Therefore the private sector agent is still exposed
to the same payoffs as before in all states of the world, albeit that
some will now be realised through taxation or reduced government
spending as opposed to the direct impact of holding the asset. Knowing
this, in order to counteract the increased exposure through the state,
the private sector agent will reduce her demand for the asset sold,
hedging and exactly offsetting the change in supply and leaving the
price unchanged. (2)
It has long been widely accepted that Wallace neutrality fails with
respect to money, at least away from the zero lower bound. This is
because money provides a non-pecuniary benefit to the holder, for
instance, utility emanating from its ability to pay taxes, or easily
effect transactions. Private sector agents therefore do not perceive
money held on their own balance sheet as equivalent to that held for
them by the state as they do not receive the non-pecuniary benefit. This
utility is itself subject to decreasing marginal returns and thus
changes in the supply of money held by the public, for a given level of
money demand, will affect the market clearing price, the short-term
interest rate. This premise has been at the heart of the open market
operations central banks have used to conduct policy for decades.
An argument against the effectiveness of recent quantitative easing
put forward by Eggertson and Woodford is that, by definition, at the
zero lower bound the non-pecuniary benefits of money have been exhausted
and so the mechanism by which money violates Wallace neutrality is no
longer operative. However, there is a growing body of empirical evidence
that suggests changes in the supply of assets other than money can also
have implications for their prices and yields. D'Amico and King
(2010) and Meaning and Zhu (2011) both use individual bond level data to
find significant changes in bond prices as a result of changes in the
supply of government securities in recent central bank asset purchase
programmes in the US and UK.
Joyce et al. (2010) find similar evidence for the UK, as do Breedon
et al. (2012) and Banerjee et al. (2012). The challenge has been
developing a rigorous theoretical framework to explain such results,
with ex-Federal Reserve Board Chairman Ben Bernanke describing asset
purchases as policies which "work in practice, but not in
theory" (Bernanke, 2014).
The theoretical arguments for such an effect have their roots in
the work of Tobin (1963,1969) and Culbertson (1957), who showed,
relatively intuitively, that if there was imperfect substitutability
between assets then changes in supply would induce movements in rates of
interest. Modigliani and Sutch (1966) then began to develop a more
rigorous theoretical explanation as to why imperfect substitutability
may exist based on the idea that investors had a preference for certain
assets or areas of the term structure. For instance, they may wish to
hold assets with the same maturity profile as their liabilities or, in
an extreme case, be mandated by law to hold assets with certain
characteristics. (3) In the context of our earlier discussion about
money, this imperfect substitutability between assets can be thought of
as deriving from a non-pecuniary benefit which that asset provides to
the holder and which other assets cannot perfectly replicate. More
recently, the idea that investors have preferred-habitats, and that this
gives rise to imperfect substitutability between assets, has been
formalised in more complex economic models by authors such as Andres et
al. (2004) and Vayanos and Vila (2009) and has taken a prominent role in
how the Bank of England has framed its quantitative easing policy
(Benford et al., 2009).
One way of modelling this is by modifying equation 1 such that p
becomes a function of the supply both of the asset and its substitutes.
[R.sup.n.sub.t] = [n.summation over (t=1)] [i.sub.t] +
[[rho].sub.t] where [[rho].sub.t] = f[[s.sub.t], [sub.sub.r]] (2)
The aim of our econometric investigation is to test for the
existence of these two channels.
Uncovering potential transmission mechanisms
Using equation (2), the exercise then aims to identify each of
these channels and their impact on interest rates. To do this we begin
by building a data set of UK government securities at the individual
bond level.
First we take data on the prices and total supply outstanding for
each individual bond in the UK government's debt portfolio. This
raw data is sourced from the British Government Securities Database,
constructed and maintained by David Wilkie and Andrew Cairns of
Heriot-Watt University. We restrict our analysis to nominal securities
only, and also to those with a remaining maturity of over three months.
Our full sample period runs from January 2000 to July 2015 at a monthly
frequency, although due to limitations in other data series we focus
primarily on the period from January 2009 onwards. This leaves us with a
sample of 103 securities.
From this, for each security we create a series of publicly
available supply by subtracting the amount held on the Bank of
England's balance sheet from the total. Figure 1 shows the
aggregated series for publicly available supply as we have defined it
plotted alongside total supply and supply held by the Bank of England on
the APF. Since 2009, the publicly available supply of gilts has been
considerably lower than the total gilt stock would imply, as the Bank of
England has removed gilts through quantitative easing.
We then use these securities to construct series of publicly
available substitutes, which we define as all securities in our sample
with a remaining maturity within two years, or +/- 50 per cent of the
remaining maturity of the security in question, whichever is greater. In
other words, if security (i) has ten years left to maturity, all
securities with a remaining maturity between five and fifteen years are
considered substitutes and as such their supply is summed to generate
the sub series for security (i). (4)
[FIGURE 1 OMITTED]
To capture the expectations of future interest rate policy,
consistent with the term [n.summation over (t=1]) [i.sub.t], in equation
(2), we use overnight index swap (OIS) rates. As discussed in Joyce et
al. (2010) OIS rates serve as a reasonable proxy for markets'
expectations of the path of Bank Rate as they are formed from geometric
averages of overnight interest rates over the specified term. For each
month we match the remaining maturity of each bond to the OIS rate with
corresponding term, as measured by the Bank of England's fitted OIS
curve. In this way we get a bespoke measure of the expectation of policy
rates over the life of that bond in that month and thus should have an
accurate representation of the expectations component of term rates. (5)
It should be noted here that expectations of the short-term nominal
interest rate will move in response to many things which are themselves
considered determinants of bond prices, such as expectations of
inflation or future output growth. The key for our analysis is that by
using OIS rates we can isolate the extent to which these factors change
agents' perceptions about the future path of monetary policy and
identify the impact associated with the expectations hypothesis-style
arguments outlined earlier.
Monthly data on OIS rates in the UK is only available back to
January 2009, which makes this the limiting variable in our main sample.
We then use this panel of data to estimate the following variant of
equation (2)
[p.sub.t] = [alpha] + [beta][s.sub.t], + [[beta].sub.2][sub.sub.t]
+ [[beta].sub.3]OI[S.sub.t] + [[delta].sub.it] (3)
where [p.sub.t] is the percentage change in the price of a security
from the previous period, [s.sub.t] is the publicly available supply of
the security and [sub.sub.t] is the supply of securities which are
considered to be substitutes for that security. [OIS.sub.t] is the
overnight index swap rate which corresponds to the remaining maturity of
the security and [[delta].sub.it] is a vector of control variables. We
apply a fixed effects estimator to control for idiosyncratic features of
any particular bond.
A potential concern is endogeneity between the dependent and
explanatory variables, particularly between the supply variables and
price changes. In theory, a prudent fiscal debt manager who is looking
to issue a given quantity of debt at the lowest possible cost to the
government may find it optimal to issue more of that debt in the form of
a particular bond when its price rises and so it becomes relatively
cheaper to do so. This would imply that not only were bond prices
responding to supply, but that supply was also responding to changes in
price. For instance, Greenwood et al. (2014) show that in the United
States, the US Treasury increased the maturity of their debt portfolio
just as the Federal Reserve was removing maturity and raising the price
of longer-dated bonds relative to shorter-dated ones. They show that
this, at least partially, offset some of the efficacy of the Fed's
policy and they highlight the inherent monetary-fiscal coordination
issues such a result implies. A similar argument is made by D'Amico
and King (2010), but from the monetary policy side, suggesting that the
Federal Reserve may have been more likely to purchase bonds which were
underpriced and thus would have expected an increase in their price
anyway. To control for this they use an instrumented variable within
their regression. In our framework, both these endogenous monetary and
fiscal responses would be captured within the estimated relationship
between price and our publicly available supply series.
However, within our study we do not consider such endogeneity to be
a cause for concern for two reasons. The first is that the overall
supply of debt which the UK Debt Management Office (DMO) has to issue is
unlikely to be determined significantly by price. Rather it is
determined primarily by the government's funding requirement. It
could be argued that a government which sets its debt issuance optimally
over time would seek to issue more debt when costs are low and the
economy is weak, so financing through taxation is costly to economic
activity. For the larger part of our sample however, the UK government
has explicitly shunned this idea and sought to reduce the amount of debt
it issues, making any such endogenous link improbable. Given the
relative stability of the nominal gilt portfolio's characteristics
over time (figure 2), it is these changes in total aggregate supply
which are likely to drive changes in the total supply at the individual
bond and substitute bond level, not changes in debt issuance practices.
[FIGURE 2 OMITTED]
What is more, in contrast to the US, the UK DMO was explicitly
tasked with not setting its issuance to take advantage of the price
changes brought about by the Bank of England's supply management
over this period. The extent to which this commitment is credible is
arguable, especially after such a long period of low rates when a
counterfactual is hard to derive. However, as can be seen in figure 2,
the average maturity of the nominal portfolio has been broadly constant
over our sample period, indicating that the DMO has not issued
particularly more longer-dated debt relative to short, despite the
relative fall in longer-term yields.
Results
The results of our estimations are presented in table 1. Our
baseline case is the estimation from January 2009 to July 2015,
including all bonds with a remaining maturity of greater than three
years, whose results are reported in the first column of table 1. (6)
The first point of note is that OIS rates have a significant impact
on the price of bonds. An expectation that short-term interest rates
will be higher over the life of the bond acts to reduce the price of
that bond and thus increase the rate of interest it pays. Importantly,
this result is estimated over a period when the short-term nominal
interest rate itself was constrained by the lower bound and, bar the
interest rate cut in March 2009, Bank Rate had been unwavering at 0.5
per cent per annum in the UK. Thus it supports the idea that it is not
just current policy which determines longer rates, but rather the
expectation thereof, implying that shaping these expectations in one
manner or another represents a genuine channel by which monetary policy
can gain traction, even when constrained by the lower bound on nominal
rates.
Second, there is a significant influence from both changes in the
supply of a security, and changes in the supply of its substitutes.
Changes in both are correlated negatively with prices, meaning that a
policy such as quantitative easing, which reduced the publicly available
supply of government securities, pushes up on prices and lowers interest
rates. Given that our work is carried out in price space, the
coefficients require some careful interpretation, but using the modified
duration of the total nominal gilt portfolio over our sample period, our
estimation implies that the removal of approximately 30 per cent of the
publicly available supply of gilts, as has occurred through the Bank of
England's quantitative easing, would lead to an average reduction
in yields of between 20 and 25 basis points. This falls towards the
lower end of the range of estimates within the literature, but it should
be noted that it does not take account of any impact QE purchases have
had through changing expectations of the future path of policy rates.
(7,8) What is more, there is likely to be significant variation across
bonds, with some segments of the yield curve where purchases were
heavier and bonds more sensitive experiencing greater reductions in
yields, while others moved very little. The same estimation over a
sub-sample which includes only longer-dated bonds with a remaining
maturity in excess of ten years shows they were more sensitive to
changes in their supply, and also to changes in supply of substitutes.
This would seem to make sense as the degree of market segmentation is
likely to be higher further along the yield curve when compared to the
shorter end where, for instance, the liquidity benefits of holding a
3-month bond are not dissimilar to holding a 6-month bond.
Quantitative easing in 'normal' times and exit
It could plausibly be argued that the period of recent financial
crisis may be a special case. Heightened financial market stress may
have elevated the degree of imperfection in substitutability between
assets as investors came to value more the non-pecuniary benefits
provided by specific asset types, such as safety or liquidity. In times
of less financial stress then, it is possible that the degree to which
assets are imperfect as substitutes for one another falls as markets
become closer to the idealised no-arbitrage markets of theoretical
models. If that were the case, then this would render quantitative
easing largely irrelevant, other than as a tool in specific periods of
extreme financial stress. As posited by Miles and Schanz (2014), it
would also have implications for how a central bank should exit from
asset purchase programmes. If the assets are purchased at a time of
financial turmoil, then they would be expected to reduce rates. However,
those assets could then be returned to the market when normal market
functioning resumed with little to no impact on rates, other than their
perceived informational content regarding the path of Bank Rate. This
would simplify the exit process greatly, as the Bank of England could
reduce its balance sheet without much impact on the monetary stance.
To test this hypothesis, we estimate the framework outlined above,
but for the pre-crisis period of January 2000 to December 2008.
Unfortunately data limitations mean that we cannot include OIS rates as
a control variable, as they are unavailable at maturities greater than
two years over this sample period. Instead, we proxy their effect with
the inclusion of Bank Rate. What our results show is that imperfect
substitutability did exist over this pre-crisis sample, but that the
coefficients on the two supply variables are reduced when compared to
the post-2008 sample. This could be indicative of a change in the degree
of imperfect substitutability between the two periods. However, we
caution that Bank Rate amounts to an imperfect substitute for OIS rates
as it has no forward looking component and cannot be matched to the
maturity of the bond in question. Therefore it does not capture the
theoretical principle of the expectations hypothesis in the same way. To
see how much of an influence this fact has on the results, we run the
same regression, which includes Bank Rate in place of OIS rates, for the
full sample from January 2000 to July 2015, so encompassing the crisis
period. The coefficients on the supply variables were broadly similar
when compared to the pre-crisis sample, suggesting that at least some of
the weakness in the pre-crisis estimation was due to the change in
control variable.
Regardless, even prior to the financial turmoil of the recent
crisis and the adoption of quantitative easing as an explicit policy
tool, it appears there was enough market segmentation for changes in
supply of assets to have significant effects on prices and yields. This
result echoes that of D'Amico et al. (2012), who use a similar
methodology on a pre-crisis sample for the US and find significant
relationships between supplies and yields on US Treasury securities.
Addressing the point of Miles and Schanz (2014) on exit from QE, we
estimate our original model over a sub-sample which runs from January
2013 to July 2015. In this period we should note that UK financial
markets were operating without any particular stresses and largely as
they may be expected to function when the Bank of England comes to
unwind its asset purchase programme. We find from this sample that there
is still a significant degree of imperfect substitutability between
assets and, therefore, the return of bonds to the market would be
expected to place downward pressure on prices and upward pressure on
yields, equating to a monetary tightening. What is more, the implied
magnitude of the effect is similar to that over the whole post-2008
sample, though the weighting has shifted somewhat between changes in the
supply of a security itself and those of substitute securities.
Therefore, as the quantitative easing programme is unwound, we
should expect to see upward pressure on UK sovereign bond yields in
excess of those emanating solely from an increasing Bank Rate and
expectations thereof.
An aside on expectation management
An interesting point of note is that in our later subsample, the
coefficient on changes in the OIS rate is considerably larger than in
the 2009- sample. This may be unsurprising, as over recent years the
adoption of forward guidance as an explicit policy tool has meant that
increased focus has been put onto markets' expectations of policy
rates, and the Bank of England's attempts to shape them. What is
more, with the prospect of rate tightening seemingly on the horizon, all
financial markets are likely to be acutely aware of the possibility of
an upcoming turning point in the monetary policy cycle. This may have
led to an increased sensitivity of market behaviour, and thus prices, to
changes in perceptions about the future path of policy. From the
Monetary Policy Committee's point of view this could pose a
potential problem, as it means that any volatility in market
expectations of Bank Rate will transmit more heavily into volatility in
bond prices and interest rates, something the MPC might wish to avoid.
Clear and effective communication might be particularly valuable if the
costs of uncertainty are particularly high.
All in all, our results suggest that changes in the supply of UK
government securities have a significant impact on the prices and
interest rates associated with those securities and securities which are
close substitutes for them. This supports the existence of a channel by
which the monetary authority can directly influence interest rates via
quantitative easing in excess of any signalling of the future policy
rate. Whilst the strength of this channel may be amplified by market
turmoil, it remains significant in periods of relatively
well-functioning markets. As we will discuss below, our results also
raise a number of questions about how these monetary policy moves
interact with fiscal policy.
Unconventional monetary policy and fiscal policy
We now turn to the monetary-fiscal interaction inherent in
movements in the publicly available supply of government securities. Our
discussion so far has been concerned with how the monetary authority can
exploit the relationship between supply and prices through quantitative
easing. However, manipulating the supply and composition of the publicly
available government debt portfolio amounts to debt management, which
has traditionally been the preserve of the fiscal authority. Referring
back to figure 1, the same profile of publicly available debt could also
have been achieved by the UK Treasury simply issuing less debt.
So, does it matter which arm of policy--monetary or fiscal--acts to
reduce the publicly available supply of government securities? The
answer is that it clearly does. To achieve the reduction in supply from
the fiscal side, the fiscal authority must reduce its debt issuance and,
all else equal, tighten the stance of fiscal policy, either through
increased taxation or less expenditure. This would be contractionary and
would contribute to offsetting any stimulus which might come from lower
interest rates. (9) This problem is likely to be even more pronounced
when monetary policy is constrained by the zero lower bound, as the
central bank cannot use its traditional instrument to loosen policy and
offset the fiscal contraction.
Alternatively, if the changes in supply are effected by the
monetary authority, the fiscal authority is free to provide a stimulus
which works to reinforce the expansionary monetary policy. In essence,
coordinating policy in this way both allows policymakers to pursue their
own objectives independently of each other; the fiscal authority can set
total debt in accordance with its funding requirements, conditioned on a
set of macroeconomic and policy goals as well as the automatic
stabilisers, whilst the monetary authority can manipulate the amount of
that debt which is publicly available to be consistent with its own
target for interest rates and thus inflation. Such a policy obviously
raises questions about the appropriate degree of coordination and the
safeguarding or otherwise of central bank independence, but it does not
follow that there is no scope for welfare improving coordination between
two independent policy arms, just perhaps an understanding that monetary
policy decisions may have fiscal implications, and fiscal decisions may
have consequences for the monetary authority.
Our result also serves to highlight an often overlooked, but
important channel through which unconventional monetary policy may
operate. By allowing the fiscal authority to issue an increased level of
debt without the associated rise in interest rate burden, central bank
purchases of government debt may help to create fiscal space which, if
utilised, could stimulate activity and thus inflationary pressure in
line with the monetary objective. This phenomenon is by no means new, or
confined to unconventional monetary policy. In fact it occurs every time
the central bank cuts the traditional policy rate. It is also far from a
free lunch for the fiscal authority though, as the same mechanism will
apply in reverse when the central bank tightens policy and unwinds its
balance sheet. As discussed in Kirby and Meaning (2015), the fiscal
benefits of central bank asset purchases therefore amount to an
intertemporal transfer, the optimality and net benefit of which require
deeper consideration.
Conclusions
Our analysis shows that both the expectations of short-term nominal
interest rates and changes in supply of assets matter in determining the
prices of UK government securities. This allows for the possibility of
multiple transmission mechanisms for monetary policy at the zero-lower
bound with the Bank of England seemingly able to shape expectations
through forward guidance and signalling, whilst also able to change the
publicly available supply of assets through quantitative easing.
Our estimates suggest that the roughly 30 per cent reduction in the
publicly available supply of nominal gilts that has occurred as a result
of the Bank of England's current quantitative easing programme
would lead to an average of a 20-25 basis point fall in yields on those
securities. This estimate is seen as a lower bound to the efficacy of
the programme. The uncovered relationships are robust to various
specifications and controls. Importantly, they are also relatively
robust across time periods, suggesting there were significant supply
effects prior to 2009, and that a significant relationship between the
publicly available supply and the prices of gilts has persisted in
recent years. Should this continue then as the Bank of England unwinds
its expanded balance sheet, we can expect this to reinforce the monetary
tightening associated with tighter expectations on the Bank Rate.
Our results also serve to demonstrate the interactions between
monetary and fiscal policy when the former operates in sovereign debt
markets. Managing supply, as was done through quantitative easing, in
many ways amounts to debt management, leaving open the prospect that the
same changes in the publicly available supply could have been effected
by the fiscal authority. There may be important differences between the
two approaches though and the correct balance between the two requires
careful consideration.
NOTES
(1) Alternative motivations, such as increasing the money supply
and promoting bank lending, have also been discussed as possible
transmission mechanisms of the policy, but the bulk of discussion in the
UK has been on the impact on interest rates.
(2) On a theoretical level, such an irrelevance proposition hinges
on the assumption of a representative agent.
(3) Clear examples here are the UK pensions market which is legally
required to hold long-term government securities, or the need to hold
liquid assets under the latest Basel Liquidity Coverage Ratio.
(4) We conducted our analysis using a range of different measures
of substitute securities including just bonds with a remaining maturity
within two years or less of a bond, those with a remaining maturity
within 50 per cent of that of a bond, and those within 25 per cent. We
found that the key results were little changed, but that when one
tightened the definition of substitutes some of the magnitude of the
subs effect waned as many longer-dated securities, where the
substitution effect is found to be strongest, were by definition without
substitutes as there were no bonds within two years of them.
(5) The Bank of England's OIS curve only reports rates up to
five years, so for all bonds in our sample with a remaining maturity
over this term we apply the 5-year OIS rate.
(6) It should be noted that we estimated a number of different
specifications, including bonds of all maturities, bonds with maturity
of between 5 and 25 years, which was the Bank of England's initial
purchase range, and controlling for inflation and the exchange rate. Our
core result was robust to all of these and so the results presented were
chosen for their intuition and parsimonious nature.
(7) Event studies such as Caglar et al. (2015) find this channel to
be weak, but are limited in their methodology to a narrow window across
which expectations might change, whereas in practice expectations may
adapt significantly in advance of policy announcements if they are
anticipated.
(8) Estimates of the impact of QE in the United Kingdom in the
existing literature vary between 20 and 150 basis points. For a summary
of the empirical literature, see Joyce et al. (2012).
(9) In fact, the impact on interest rates itself is ambiguous here,
as the additional risks that a tighter fiscal stance might imply for the
broader economy could potentially push up on rates by more than the
change in supply lowers them.
REFERENCES
Andres, J., Lopez-Salido, J.D. and Nelson, E. (2004), Tobin's
imperfect asset substitution in optimizing general equilibrium',
Journal of Money, Credit and Banking, 36(4), pp. 665-90.
Banerjee, R., Latto, D. and McLaren, N. (2012), 'Using changes
in auction maturity sectors to help identify the impact of QE on gilt
yields', Bank of England Quarterly Bulletin, Q2.
Benford, J., Berry, S., Nikolov, K. and Young, C. (2009),
'Quantitative easing', Bank of England Quarterly Bulletin, 49,
2, pp. 90-100.
Bernanke, B. (2014), 'Central banking after the Great
Recession: lessons learned and challenges ahead', speech given at
the Brookings Institute.
Breedon, F., Chadha, J.S. and Waters, A. (2012), 'The
financial market impact of UK quantitative easing', Oxford Review
of Economic Policy, Oxford University Press, 28(4), pp. 702-28, Winter.
Caglar, E., Chadha, J.S., Meaning, J., Warren, J and Waters, A.
(2015), 'Central bank balance sheet policies: three views from the
DSGE literature' in Chadha, J.S. and Holly, S. (eds), Interest
Rates, Prices and Liquidity, Cambridge, Cambridge University Press.
Culbertson, J.M. (1957), 'The term structure of interest
rates', Quarterly Journal of Economics, 71, pp. 485-517.
D'Amico, S., English, W., Lopez-Salido, D. and Nelson, E.
(2012), 'The Federal Reserve's large-scale asset purchase
programmes: rationale and effects', Economic Journal, Royal
Economic Society, 122(564), pp. F4I5-46, November.
D'Amico, S. and King, E. (2010), 'Flow and stock effects
of large-scale treasury purchases', Federal Reserve Board Finance
and Economics Discussion Series, 2010.52.
Eggertson, G. and Woodford, M. (2003), 'The zero bound on
interest rates and optimal monetary policy', Brookings Papers on
Economic Activity, I, pp. 139-233.
Greenwood, R., Hanson, S.G., Rudolph, J.S and Summers, L. (2014).
'Government debt management at the zero lower bound', Hutchins
Centre on Fiscal and Monetary Policy, Brookings. Working paper #5.
Joyce, M., Lasaosa, A., Stevens, I. and Tong M. (2010), 'The
financial market impact of quantitative easing', Bank of England
Working paper, No.393.
Joyce, M., Miles, D., Scott, A. and Vayanos, D. (2012),
'Quantitative easing and unconventional monetary policy-an
introduction', The Economic Journal, 122 (November), F271-F288.
Kirby, S. and Meaning, J. (2015), 'Impacts of the Bank of
England's asset purchases on the public finances', National
Institute Economic Review, 232, May.
Meaning, J. and Zhu, F. (201 I), 'The impact of recent central
bank asset purchase programmes', BIS Quarterly Review, December.
Miles, D. and Schanz, J. (2014), 'The relevance or otherwise
of the central bank's balance sheet', CESIFO Working paper,
No. 4615.
Modigliani, F. and Sutch, R.C. (1966), 'Innovations in
interest rate policy', American Economic Review, Papers and
Proceedings, 56(2), pp. 178-97.
Tobin, J. (1963), 'An essay on the principles of debt
management', in Fiscal and Debt Management Policies, Englewood
Cliffs, N.J.; Prentice Hall, pp. 143-218.
--(1969), 'A general equilibrium approach to monetary
theory', Journal of Money, Credit and Banking, I, February, pp.
15-29.
Vayanos, D. and Vila, J. (2009), 'A preferred-habitat model of
the term structure of interest rates', CEPR Discussion Paper 7547,
LSE.
Wallace, N. (1981), 'A Modigliani-Miller theorem for
open-market operations', American Economic Association, American
Economic Review, 71(3), pp. 267-74, June.
Jack Meaning and James Warren *
* National Institute of Economic and Social Research. E-mail:
j.meaning@niesr.ac.uk. The authors would like to thank Tony Yates, Oriol
Carreras and Rebecca Piggott for their insightful comments and Michele
Bernini for his invaluable technical assistance.
Table 1. Estimation results
Jan 2009- Jan 2009- Jan 2000-
July 2015 July 2015 Dec 2008
All bonds with Bonds with All bonds with
RM>3 years RM> 10 years RM>3 years
% change in supply -0.0068 ** -0.011 ** -0.0041
(0.041) (0.03) (0.16)
% change in supply -0.050 *** -0.057 * -0.019 *
of substitutes (0.005) (0.08) (0.01)
Corresponding -0.0065 *** -0.008 *** --
OIS rate (0.00) (0.00)
Bank Rate -- -- -0.0026 ***
(0.00)
Remaining maturity 0.00017 *** 0.00017 *** -0.00003 ***
(0.00) (0.00) (0.00)
Constant -0.013 *** -0.028 *** 0.017 ***
(0.00) (0.00) (0.00)
Adj-R2 0.017 0.017 0.011
Prob >F 0.00 0.00 0.00
Jan 2000- Jan 2013-
July 2015 July 2015
All bonds with All bonds with
RM>3 years RM>3 years
% change in supply -0.0032 -0.0065
(0.12) (0.48)
% change in supply -0.026 *** -0.072 ***
of substitutes (0.002) (0.005)
Corresponding -- -0.0095 ***
OIS rate (0.00)
Bank Rate -0.0001 --
(0.30)
Remaining maturity -0.00001 ** -0.00024 ***
(0.03) (0.00)
Constant 0.004 *** 0.075 ***
(0.00) (0.00)
Adj-R2 0.010 0.074
Prob >F 0.01 0.00
Notes: *** denotes coefficient is significant at 1% level,
** at 5% level and * at 10% level.
T-probabilities are presented in parentheses.
