A Note on Monetary Policy
John Whittaker
In the UK and other countries with developed financial markets, central banks set the
short-term nominal interest rate for the currency that they issue: the short-term interest
rate is the monetary instrument. In recent years, monetary policy has been supplemented
in the major economies by quantitative easing: the purchase by the central bank of
government debt and other assets.
The first part of this note describes how the central bank interacts with financial markets
and how its policy choices are transmitted to the wider economy. The second part
describes how the central bank makes its policy choices.
1. What is monetary policy?
1.1 The functions of commercial banks
1
1.2 The Bank of England’s operations
2
1.3 Interest rates in the money-market, and banks’ retail rates
3
1.4 Effects of the financial crisis and the recession
5
1.5 Bank regulation
6
2. How does the central bank choose the interest rate?
2.1 The effects of changes in interest rate: the transmission
mechanism of monetary policy
9
2.2 The role of expectations
10
2.3 The central bank’s choice of interest rate
11
2.4 Quantitative easing
13
Appendix: Monetary base control
15
Economics Department
Lancaster University
January 2017
http://www.lancaster.ac.uk/staff/whittaj1/
1. What is monetary policy?
1.1 The functions of commercial banks
account holder makes a payment (see the
simplified balance sheet below). But loans are
the largest component of banks’ assets and they
are illiquid: they may, for instance, be loans to
firms which are used to purchase fixed assets or
mortgage finance to householders, and cannot
easily be sold if the bank needs to repay its
depositors.
The traditional business of commercial banks is
to take deposits and to lend, and they make
profits from the interest margin between the rate
they charge for loans and the rate they pay to
depositors. Many large banks also perform
other functions such as managing investment
funds and dealing in securities, although
regulation is now attempting to enforce
separation of these functions from retail
banking, in response to the financial crisis.
To cope with net withdrawals of deposits, banks
need liquid assets, the most liquid being
holdings of currency (banknotes) and reserves.
Reserves are the banks’ own deposits at the
central bank (hence they appear as liabilities on
the Bank of England’s balance sheet) and they
are liquid because they may be immediately
withdrawn in the form of currency.
The other important function of commercial
banks is to operate the payments system. For
this purpose, about half of private sector
deposits in UK banks are sight deposits that
may be withdrawn without notice when the
__________________________________________________________________________________
UK banks and building societies: consolidated account. Sterling balances, £ billions
assets
£ currency (vault cash)
reserve deposits at BoE
govt and private securities
loans to UK private sector
Feb.06
5.7
0.8
126.5
1776.7
1909.7
Aug.16
10.9
333.2
434.2
2335.0
3113.3
liabilities
private sector deposits
public sector deposits
repo loans from BoE
capital and net other
Feb.06
1668.7
38.2
26.7
176.1
1909.7
Aug.16
2628.7
41.6
24.2
418.8
3113.3
Feb.06
36.9
0.8
0.8
3.8
Aug.16
72.7
333.2
4.1
83.4
42.3
493.4
Bank of England (BoE), £ billions
assets
government securities
repo lending to banks
Feb.06
15.6
26.7
Aug.16
469.2
24.2
42.3
493.4
liabilities
£ currency issued
bank reserve deposits
government deposits
net other
Simplified balance sheets for UK banks (£ sterling balances) and the Bank of England. The
earlier date (Feb. 2006) is included to show the difference before interest was paid on reserve
balances and before quantitative easing: see section 2.4.
Source: Monetary and Financial Statistics, Bank of England (‘Bankstats’), tables B1.1.2, B2.1, B2.2, B2.2.1
1
As currency earns no interest, the banks keep
stocks of ‘vault cash’ that are only a small
fraction of their deposit liabilities (see balance
sheet). They prefer to hold their liquidity in the
form of reserves 1 which (since May 2006) earn
interest at the Bank of England’s official Bank
Rate (0.5% since 2009, reduced to 0.25% in
August 2016) or interest-earning securities such
as treasury bills and government bonds. These
assets are liquid as they may easily be sold in
financial markets or used as collateral security
for loans from other institutions. The banks’
holdings of reserves have increased markedly
since 2009 as a result of quantitative easing
(see Section 2.4 below).
changes on the consolidated balance sheet of
the banks presented above and are not relevant
for monetary policy. The transactions that are of
interest are those that involve the central bank.
1.2 The Bank of England’s operations
While monetary systems in all the major
economies operate in similar ways, the
following is a simplified description of
transactions that involve the Bank of England.
Suppose that individuals decide to hold more
currency, as they regularly do during holiday
seasons, and they draw on their bank deposits
for this purpose. Since commercial banks only
hold small stocks of currency (‘vault cash’),
they must obtain the extra from the Bank of
England. In the above balance sheets, deposits
at banks fall while the currency liability of the
Bank of England rises.
When a retail payment is made, for instance by
a cheque drawn on the payer’s bank account
which the payee deposits in a different bank, or
using a debit card, the paying bank owes the
amount of the payment to the receiving bank.
The wholesale debts between banks thus
created by retail payments may be settled in a
variety of ways: for example by transfers of
bank reserve deposits at the Bank of England,
by sales of assets, by loans that are secured on
assets (usually in the form of ‘repos’: sale and
repurchase agreements 2) or by unsecured
interbank loans. Settlement by transfers of
reserves is now more common, given the banks’
large holdings of reserves.
The banks pay the Bank of England for this
currency by drawing on their reserve deposits:
the relevant bookkeeping entry is a reduction in
bank reserves equal to the amount of the extra
currency.
As another example, suppose the government
receives a payment of tax which the payer
draws from his deposit at a commercial bank. If
the government pays the tax into its deposit at
the Bank of England, this again causes a debt of
the commercial bank to the Bank of England
which is settled by an increase in the bank’s
reserves. The opposite transactions take place
when the government spends out of its Bank of
England deposits.
Note that these transactions are between the
banks – one bank’s loss of a deposit is another
bank’s gain – hence they will not cause any
1
Some central banks oblige their banks to hold required
reserves as a minimum ratio (e.g. 2%) of their own shortterm deposit liabilities. However, the ability to vary this
ratio is not an additional tool of monetary policy, given
that the central bank must always supply sufficient
reserves to cover any such requirement (see section 1.2
and Appendix). The Bank of England does not apply
reserve requirements.
Changes in banks’ reserves at the Bank of
England are the balancing entry whenever any
of the central bank’s other liabilities or assets
change. The above transactions are
‘autonomous factors’ that change banks’
reserve balances, (i.e. they are not under the
control of the Bank of England). If they cause
banks, in aggregate, to run short of reserves, the
Bank of England itself makes further reserves
available. Normally, it does this by routinely
2
Under a ‘repo’ or ‘sale and repurchase agreement’, the
borrowing bank sells the lending bank a security (such as
a government bond or ‘gilt’) with an agreement to
repurchase it later. A repo thus amounts to a loan backed
by the collateral security. Interest is effected as the
difference between the sale and repurchase price of the
bond.
2
granting them short-term ‘repo’ loans 3. The
interest rate charged for repo lending is close to
Bank Rate that it pays on reserve deposits.
Since 2009, however, the banks have had large
excess reserve balances because of quantitative
easing (QE, Section 2.4). There has been no
need for the Bank of England to supply them
with more reserves and it has suspended its
routine short-term repos. In this situation, Bank
Rate, as the rate paid on reserve balances, is the
opportunity cost of short-term funding for the
banks. This rate is transmitted to the wider
economy via the sterling interbank market, as is
explained in the following section.
The important point is that, by its own dealing
with the banks, the Bank of England supplies
the banks with reserves (provides liquidity) as
needed to ensure that, collectively, they always
have positive reserve balances. It cannot do
otherwise. This is fundamentally because all
financial liabilities denominated in sterling are
ultimately claims on sterling currency, the most
obvious example being ‘sight’ deposits in
commercial banks which the depositor can
withdraw as currency on demand, and because
the Bank of England is, by law, the only issuer
of sterling currency. This is the defining
difference between the central bank and other
financial institutions.
1.3 Interest rates in the interbank market
and banks’ retail rates
With the Bank of England lending to banks as
required at its chosen Bank Rate and paying
interest to banks on their reserve balances, also
at Bank Rate, short-term wholesale lending
between banks also normally takes place at
rates that are close to Bank Rate. At present,
with abundant reserve balances and little doubt
about bank solvency, competition between
banks ensures that no bank has to pay more than
Bank Rate for a short-term loan from another
bank, and no bank would lend at less than Bank
Rate when it is paid at Bank Rate on its reserve
deposits. The routine practice of the central
bank, of lending to banks on demand and
paying interest on the banks’ reserve balances,
means that the wholesale rates for interbank
lending remain close to Bank Rate (Figure 1). 5
If the Bank of England failed to provide
reserves as needed, or even if there were a
suspicion that some bank would be unable to
satisfy demands for currency withdrawal, the
likely consequence would be a bank run, as
happened to Northern Rock Bank in 2007.
Since the Bank of England cannot refuse such
lending, it has to choose the interest rate at
which it lends. This Bank Rate is the marginal
cost of funds to the banks: it is the rate at which
banks know they can always obtain short-term
funds (provided they have sufficient collateral
that is acceptable by the Bank of England for its
repos 4).
argued that central bank lending against illiquid assets
amounts to solvency support; see section 1.4.
5
Interbank rates became detached from Bank Rate during
the financial crisis, reflecting lack of trust between the
banks. For instance, three month libor reached as far as
1½% above Bank Rate in late 2008 (Figure 1). During
2011, libor rates rose again; this may have reflected the
UK banks’ exposure to some eurozone banks.
In the eurozone, there is the added complication that the
rate paid by the ECB for excess reserves, currently minus
0.4% (since March 2016; the rate on its ‘deposit
facility’), is less than its lending rate, currently 0% (its
‘main refinancing rate’). Given that eurozone banks have
large excess reserves, it is the rate paid on those reserves
that is relevant and the interbank rate (EONIA) is
currently minus 0.35%, close to this rate.
A further complication has been added by new
regulations (the Liquidity Coverage Ratio, LCR) which
oblige banks to hold liquid assets, mainly reserves and
3
See footnote 2. On balance sheets, Bank of England
repo lending is shown as an asset on the Bank of
England’s balance sheet and a liability of the borrowing
bank. Under standard accounting conventions, the
borrowing bank’s collateral security stays on its balance
sheet even though it is held by the Bank of England
during the tenure of the repo.
4
This raises the question of what happens when banks
run low on liquid assets that the central bank deems to be
eligible collateral for its repo loans. At the onset of the
financial crisis when many banks became illiquid, major
central banks reduced the quality of assets against which
they will lend. The quality of assets that the European
Central Bank is prepared to accept for its repo lending
remains much lower than before the crisis. It can be
3
their lending by borrowing in the interbank
market at rates close to the central bank’s
dealing rate, so they do not wait for deposits
before lending, as would be implied by the
textbook ‘money multiplier’ theory (see
appendix). When a bank has an opportunity to
lend, its decisions over whether to lend, how
much to lend and at what interest margin, can
be taken mainly on the basis of the perceived
default risk. The banks are said to practice
‘liability management’: lend and then find the
necessary funds (liabilities).
Figure 1. Bank Rate and an interbank rate.
libor = London Interbank Offer rate, for unsecured
lending
source: Bank of England
What interest rate will banks offer for deposits?
Deposits and interbank borrowing are
alternative sources of funds so, as competitive
profit maximisers, they set the rate offered on
deposits at some margin below Bank Rate to
cover administration and transactions costs
(Figure 2 displays this pattern up to 2008). For
short-term deposits used for transactions, the
margin must also cover liquidity risk: the risk
that the deposit may be withdrawn without
notice.
Figure 2. Retail deposit and lending rates
source: Bank of England
Similarly, the rate charged by banks for their
lending to individual and corporate borrowers
exceeds Bank Rate by a margin to cover
transactions costs, also default risk and the cost
of holding capital to support the loan. In other
words, the banks set their retail lending and
deposit rates at appropriate margins above and
below the official Bank Rate of the central
bank. When Bank Rate changes, the banks’
retail rates normally change approximately in
parallel.
However, for some banks that had become
over-reliant on short-term interbank market
funding, this method of working broke down
dramatically after the onset of the financial
crisis in 2007, when the quality of their assets
fell under suspicion and other banks were
unwilling to roll over their interbank loans.
At the end of 2008, the relationship between
Bank Rate and the commercial banks’ retail
rates also changed as Bank Rate was rapidly
reduced from 5% to 0.5% while retail rates did
not fall so far (Figure 2). Lending margins have
remained higher than before the crisis and
banks are still competing to attract deposits by
offering rates for some short-term deposits that
are greater than Bank Rate. This is indicative of
a change in the banks’ attitude to risk together
with tighter Basel III regulations (section 1.5).
After aggressive expansion of lending up to
2007, the banks are still rebuilding their balance
sheets to increase the proportion of lending that
Another consequence of the central bank’s
routine provision of liquidity (and the payment
of interest on excess reserves) is that the
commercial banks’ retail lending decisions are
not dependent on their having sufficient
deposits. They know they can normally fund
government bonds, equal to 100% of their short-term
interbank borrowings. This has markedly reduced the
volume of lending in the interbank market.
4
be divided into tranches with a prescribed
priority order of pay-out when some of the
underlying loans default on payments. In a
simple structure containing a senior, a
mezzanine and an equity tranche, the first losses
are borne by the equity tranche, and the
mezzanine tranche only becomes impaired
when the losses exceed the entire amount of the
equity tranche. Similarly, the senior tranche
only suffers if losses exceed the amounts of the
mezzanine and the equity tranches, making it
much more secure than the original pool.
is funded by deposits rather than interbank
borrowing, to increase their holdings of liquid
assets to meet new statutory requirements, and
to raise their capital ratios (both risk-weighted
capital and leverage ratios).
The irony is that, whilst the government is
raising banks’ costs by regulation, it is also
pressing the banks to increase their lending. The
government cannot have it both ways.
1.4 Effects of the financial crisis and the
recession
The senior tranche is thus credit-enhanced and,
before the problems of 2007, it invariably
gained the highest rating from one of the credit
rating agencies, thus becoming an attractive
investment to other institutions such as pension
funds. This process enabled banks to turn risky
illiquid loans into assets that could be sold or
used as collateral for repo funding in the
financial markets, freeing up cash for further
lending. Under existing bank regulations, it also
enabled institutions that owned senior tranches
to hold smaller amounts of regulatory capital
than if they held the original assets, thereby
increasing leverage.
In late 2007, it became clear that some banks
had been lending too freely and too cheaply,
particularly for home loans, driving up the
prices of houses and other assets to
unsustainable heights (Figure 3). In both the
United States and Britain, mortgage loans were
granted in amounts that were six times the
borrower’s annual income and up to 125% of
the value of the secured property. When house
prices started falling and significant numbers of
loans began to default (initially in the
‘subprime’ market in the US) there was a
widespread fall in confidence in the banks and
several had to be bailed out or recapitalised by
governments.
In the years prior to 2007, the issue of assetbacked securities grew at an increasing rate,
reaching a global value of around $10 trillion in
mid-2007 6, with banks in most developed
countries becoming involved as originators of
these securities, or buyers, or both. And as the
market expanded, more complex financial
instruments were introduced such as
collateralized debt obligations which are, for
instance, securitised assets created by
combining several mezzanine tranches of
previous securitisations. These new assets were
hard to value and they had the effect of
concealing risk. When the market in these
assets broke down, many were written down to
a fraction of their face values, earning the label
‘toxic assets’.
Figure 3. UK house prices
One of the factors which drove this expansion
in lending and also played a major part in the
banking collapse was securitisation, whereby a
bank pools together several hundred loans to
create an asset-backed security. This may then
6
Estimate from Bank of England Stability Report,
October 2007.
5
During 2008, this deterioration in the value of
banks’ assets quickly led to justifiable doubts
about the solvency of many banks including
several of the world’s largest. As a result, banks
became reluctant to lend both to their retail
customers and to other banks. The sudden
reduction in the availability of credit,
accompanied in Britain and the US by sharp
falls in property prices, was a major contributor
to recession 7 in all major economies.
measures, in particular quantitative easing (see
section 2.4)
1.5 Bank regulation
The other response to the financial crisis has
been a large increase in regulation under the
international Basel Accords, with the intention
of preventing future financial crises and their
consequences. The core of the Basel regulations
has always been to stipulate minimum levels of
bank capital.
Faced with a potential disruption of payments
systems, central banks gave support to
vulnerable banks, by providing liquidity for
longer periods and against lower quality
collateral. And governments provided banks
with new capital, in some cases amounting to
partial or complete nationalisation.
As with any firm, profits add to the firm’s
capital stock (or ‘net worth’ or ‘equity’: the
shareholders’ claim on the firm) and losses
subtract from it, and the firm is solvent (in
accounting terms) if the value of its capital is
positive. It follows that the larger the value of
the firm’s capital, the greater its ability to
sustain losses and the more likely it is to remain
solvent.
Governments also responded with expansionary
(Keynesian) fiscal policy, but the scope for
extra government spending or for tax reduction
as limited because of high government debt
levels. Government debts had risen sharply
because of the recession-induced reduction in
tax revenue and increased welfare spending (the
‘automatic stabilisers’) and because of the
support given to the banks. The high levels of
government debt in many countries continue to
impede economic recovery.
The insolvency of any firm implies losses for
creditors, job losses for the firm’s employees,
and interruption of production. In the case of a
bank, the creditors are the bank’s depositors
who would lose their uninsured savings, and an
important bank ‘product’ is the operation of the
payments system (Section 1.1). If one bank is
unable to process a payment, this can lead to a
chain of failed payments and to the weakening
of other banks. Widespread breakdown of the
payments system would be catastrophic for the
economy. Thus, while the failure of nonfinancial firms and small banks might be
tolerated, there is a strong case for government
and/or central bank intervention to prevent the
failure of large banks; large banks may be
considered as ‘too large to fail’ (TLTF). Indeed,
governments of all major economies provided
financial support (‘bail-out’) to a number of
banks which suffered heavy losses after the
onset of the financial crisis in 2007-08.
At the same time, monetary policy was used
aggressively in the attempt to mitigate the
recession and central banks rapidly reduced
their interest rates towards zero. Recently, some
central banks have even reduced their rates
below zero: the European Central Bank
currently pays negative 0.4% on excess reserve
deposits and the Bank of Japan negative 0.1%.
However, central bank deposit rates cannot fall
far below zero, otherwise banks would gain by
holding (zero-earning) currency instead of
(negative-earning) reserve deposits.
In view of this lower limit on official rates,
central banks have turned to other stimulatory
If the bank’s shareholders were to absorb all
losses, they would be inclined to ensure that the
bank carries prudent levels of equity relative to
their risky assets. However, the owners of large
banks know that they can expect bailout in the
7
A common but arbitrary definition of recession is two
successive quarters of negative real economic growth.
6
event of failure. They therefore have the moral
hazard incentive to operate with lower levels of
equity in order to maximise their return on
equity, knowing that society bears the risk. This
is the justification for regulation that forces
banks to have minimum levels of capital.
However, a bank prefers to hold illiquid assets
as they yield higher interest. This is the
justification for the LCR which requires the
bank to hold ‘high-quality liquid assets’
(HQLA) (mainly reserves and government
bonds) as a minimum percentage of its liquid
liabilities. For instance, this percentage (the
‘run-off’ rate) for deposits may be 3%, 5% or
10% depending on the likelihood of withdrawal,
it is 40% for repo borrowing from the central
bank for less than 30 days, and 100% for
unsecured interbank borrowing.
The first international regulation, Basel I 8 was
agreed in 1988 and widely implemented by
1993. Banks’ assets were subdivided into
classes, with each class assigned a weight
according to its perceived risk (e.g. lending to
business 100%, mortgage debt 50%,
government bonds 0%). Each bank was obliged
to hold a minimum value of capital as a ratio
(8%) of the total risk-weighted assets, where the
main components of ‘capital’ are shareholders’
equity (Tier 1) and subordinated debt (Tier 2).
After much revision and amendment, Basel I
was supplanted by Basel II in 2004-2008, the
main change being in the methods of assessing
risk weights and the definitions of permissible
capital.
Note that illiquidity may imply insolvency and
vice-versa. If a bank is short of liquid assets, it
may still try to satisfy deposit withdrawals by
selling its illiquid assets such as its loans. But
such sales, especially if undertaken quickly,
may raise less than the book value of the assets,
reducing the value of the bank’s equity and
moving it towards insolvency. Conversely, if
there are doubts about a bank’s solvency, its
depositors would tend to withdraw their
deposits for fear that they would be lost: there
would be a ‘run’ on the bank. Thus either
illiquidity or insolvency may be the root cause
of a bank’s failure. They tend to occur together.
However, these rules clearly failed to prevent
the crisis, and the latest round, Basel III, has
raised the capital ratios and made them
dependent on bank size and the state of the
business cycle: banks are required to have more
capital during periods of high economic growth
which is then available to absorb losses in the
downturn. Basel III has also added further
regulatory ratios, notably the Liquidity
Coverage Ratio (LCR).
Another new ratio is the Net Stable Funding
Ratio (NSFR) which says a bank must hold
long term liabilities (e.g. capital and time
deposits) as a minimum percentage (100%) of
its illiquid assets. Finally, the Leverage Ratio
sets a minimum ratio of capital to total assets
(Basel says 3%; national regulators may set a
higher percentage).
Besides being solvent, a bank needs to be liquid
(see Section 1.1) in order that it can make
payments from its customers’ deposits as and
when ordered to do so. It must have sufficient
reserves (deposits at the central bank) and other
liquid assets that are immediately saleable or
acceptable as collateral for (repo) borrowing
from other financial institutions or the central
bank, such as government bonds and bills.
All these ratios involve large numbers of
arbitrary parameters such as the capital riskweights, the LCR run-off rates and the
allowable proportions of the different sorts of
capital. Moreover, there is clearly overlap
between the different ratios. A bank that
satisfies the NSFR should necessarily also
satisfy the LCR; if it does not, this is because
these two ratios use different classifications of
liquid and illiquid liabilities and assets. And the
capital ratio and leverage ratio both stipulate
minimum capital levels. The reason why the
leverage ratio was introduced was that the
8
The Basel Accords are the work of the Basel Committee
on Banking Supervision (BCBS), a subdivision of the
Bank for international Settlements (BIS).
7
capital to risk-weighted asset ratio might be
treating some assets too leniently.
recommended that banks should ‘ring-fence’
their retail banking activity from other activities
such as investment banking.
As a final point of criticism, note that these new
Basel III ratios only become 100% binding in
year 2019, while several new rules and
amendments of existing rules are already under
discussion. Moreover, Basel III also contains
many other measures governing regulation and
supervision. In addition, there are national
regulations such as the wide-ranging DoddFrank Act in the US which prohibits banks from
trading on their own account, amongst other
things. In Britain, the Vickers report
Has this comprehensive regulatory onslaught
made banks safer? Probably, but the complexity
of the regulation and its continuous change
makes banks less efficient and hinders their
normal business of accepting deposits and
lending. The predictable result is
disintermediation: the shifting of financial
intermediation into the less regulated ‘shadow
banking’ sector.
8
2. How does the central bank choose the interest rate?
more. There are also other influences on
demand which are not easily measurable or
predictable such as ‘habits’ and ‘confidence’. 9
In recessions, when individuals and firms are
not confident about their future incomes they
are naturally reluctant to incur further debt,
and so lower rates may have little effect on
spending. Japan has been suffering from this
malaise for a number of years during which
interest rates have hardly risen above zero.
2.1 The effects of changes in interest rate:
the transmission mechanism of monetary
policy
The next task is to consider how the central
bank should choose the value of its control
instrument: the nominal short-term interest
rate (Bank Rate in the UK). For this purpose,
the objective of monetary policy needs to be
defined and and it is generally a target for
inflation. The target may be explicit, such as
the UK’s inflation target of 2% (consumer
price index) with a margin of error of ± 1 %,
or it may be stated more loosely as in the
directive to the European Central Bank to aim
for “low and steady inflation” which it
interprets as below but close to 2%.
Consider the consequences of a reduction in
the Bank of England’s official Bank Rate.
Most bank loans in the UK are at a variable
interest rate, such as variable-rate mortgage
loans for house purchases and overdraft loans
which are the main type of finance for smaller
firms. As the cost of this finance falls, this
tends to cause an increase in borrowing, both
for consumption and investment. The
reduction in interest rates also reduces the
reward to saving, which also encourages
spending. 10
The central bank may also be instructed to aim
for high economic growth and employment, as
in the US, but a single instrument (the interest
rate) cannot be used to target more than one
objective at the same time. Raising growth
would call for lower interest rates, while
reducing inflation would require higher
interest rates. In practice, even though central
banks are always conscious of the effects of
their interest rate choices on economic growth,
the main focus is on inflation.
Another important influence on demand is the
wealth effect. When interest rates fall, the
prices of both physical and financial assets
rise. Financial assets such as bills and bonds
are claims to specified future nominal amounts
of cash; hence their nominal values rise to
bring the return on these assets into line with
the lower rate of interest. For physical assets
such as property, prices rise from raised
demand due to the lower cost of borrowing.
To fulfil its mandate to control inflation, the
central bank needs to know how inflation is
affected by its interest rate choices. The main
impact of interest rate changes is on aggregate
demand: a reduction in interest rate tends to
raise demand, and vice versa. Depending on
supply capacity, higher demand leads to higher
real output or higher inflation or both, where
the change in real output is usually thought to
be temporary as described by Phillips Curve
theory.
9
A more complete description of the effects of interest
rates is in “The Transmission Mechanism of Monetary
Policy”, Bank of England Quarterly Bulletin, 39.2 (May
1999), p.161-70.
10
While this is believed to be the dominant effect of a
reduction in interest rates on savings, it could be the
opposite. Lower interest rates may cause greater saving
rather than less, if savers want to maintain a given
income stream from their interest payments. The
outcome depends on the relative magnitudes of the
‘income and substitution’ effects in standard
microeconomic theory of intertemporal choice.
However, there is considerable uncertainty in
the magnitude and timing of the response to
interest rate changes: the peak of the response
of inflation to a change in short-term interest
rate is thought to be lagged by 18 months or
9
The higher value of these assets makes
individuals and firms more confident about
spending and also provides increased collateral
security for increasing their borrowing.
by the central bank (Bank Rate is an overnight
rate), at a level that is designed to cause the
inflation rate to converge towards its target;
therefore, an important influence on long-term
rates is the expectation of future inflation. If
the expectation of future inflation is revised
upwards, this raises the expectation of a higher
Bank Rate in the future and this, in turn, will
raise current long-term rates.
A further influence on demand works through
the foreign exchange rate: a lower interest rate
is usually associated with a lower foreign
exchange value of the currency which
stimulates demand for exports. However, this
linkage is unreliable because there are other
stronger influences on the return to investment
in foreign currencies.
It also follows that, when Bank Rate changes,
this can cause changes in long-term rates in
either direction, depending on how the change
in Bank Rate causes expectations of future
Bank Rates to be revised. If a change in Bank
Rate is fully anticipated and therefore does not
affect expectations, longer-term rates are
unaffected. But if a Bank Rate change is a
surprise, or smaller or larger than expected,
this represents new information and longer
rates also change. An unexpected fall in Bank
Rate, for instance, would cause longer rates
also to fall; a fall in Bank Rate that is smaller
than expected would cause longer rates to rise.
These are the main channels whereby interest
rates are understood to influence aggregate
demand. Despite the uncertainties, there is
agreement that lower interest rates stimulate
demand and vice versa, as would be indicated
by any standard model of the ‘IS’ curve.
Another matter that needs attention is that
Bank Rate is a short-term rate while much
expenditure, particularly for investment, is
more dependent on longer-term rates. We must
consider how the yield curve (the relationship
between interest rates of different maturities)
is influenced by expectations of future shortterm rates.
Clearly, all information that causes a revision
of expectations of the future path of Bank Rate
is important. If, for instance, the governor of
the central bank makes a statement implying
that Bank Rate will be lower in the future and
this statement is believed, then this also
reduces long-term rates. For a time, this
connection was exploited deliberately in the
US and the UK as ‘forward guidance’: the
central banks attempted to hold down longterm rates by making explicit promises that,
subject to conditions, the policy interest rate
would not rise during some specified future
period.
2.2 The role of expectations
The pure expectations hypothesis of the term
structure of interest rates states that the interest
rate for a (risk-free, zero-coupon) loan of
maturity T years is an average of current and
expected future short-term rates over the Tyear period. 11 The very short-term rate is set
11
While the expectations hypothesis is valuable for
explaining the term structure of interest rates, there are
other influences. The rates for government bonds, for
instance, are influenced by the supply of bonds by the
government and the demand for them by investors.
Yields also include a premium for default risk, which
can be large for ‘junk’ corporate bonds and even some
governments. The yields on government debts of several
‘peripheral’ eurozone countries such as Greece and Italy
remain elevated relative to Germany, reflecting the
ongoing risk of default or leaving the euro. Finally,
yields on long maturity bonds are presumed to include a
The general pattern is that long-term rates
follow short-term rates (Figure 4) as would be
predicted by the expectations hypothesis. Over
time, a lower Bank Rate is associated with
lower interest rates across the term structure,
and lower interest rates tend to stimulate
demand and inflation, albeit with a high degree
‘term premium’ to compensate for the higher sensitivity
of their price to interest rate changes.
10
of uncertainty, as analysed above (section 2.1).
It is against this background that the Bank of
England must choose the Bank Rate that is
most appropriate for achieving its inflation
objective.
forecasts can be used as a guide, the choice of
central bank policy rates has to rely on
judgement. In the UK this judgement is the
work of the members of the Bank of England’s
monetary policy committee.
The purpose of central bank independence is
to insulate monetary policy choices from
interference from the government, which may
desire lower interest rates than would be
consistent with the inflation target. It is often
assumed that governments might be tempted to
hold down interest rates to stimulate spending
ahead of an election, disregarding the longerterm potential consequence of inflation.
Figure 4. Two points on the £-sterling
yield curve: Bank Rate and the yield on
10-year UK government debt.
source: Bank of England
2.3 The central bank’s choice of interest
rate
Many central banks are now independent (the
Bank of England has been independent since
1997). This means that the government
chooses the objective of monetary policy,
usually specified as an inflation target as
described above, while the task of the central
bank is to choose a time path for Bank Rate
that is most likely to achieve that objective.
Figure 5. Bank of England forecasts of
CPI inflation. Note the wide uncertainty
(the ‘fan chart’ includes 90% of the
probability distribution).
Source: Bank of England inflation report,
November 2016
As a result of the lags in the response to
interest rate changes, the practice of the central
bank may be regarded as inflation forecast
targeting. The Bank of England, for instance,
uses its macroeconomic model to forecast
inflation up to 2 years in the future (Figure 5)
and, if the mean 2-year forecast differs from
the target, Bank Rate should be adjusted
accordingly. 12 In practice, while inflation
r − π = r * + 0.5( y − y ) + 0.5(π − π *) where r is
the central bank’s policy rate (Bank Rate), π is the
inflation rate, r * is the long-run equilibrium real rate of
interest, y is real output, y is the natural rate of real
output and π * is the inflation target. This says r should
be set so that that the real rate of interest r − π is
above (below) its long-run level when output is above
(below) its long-run level and/or inflation is above
(below) target. While there is evidence that the Taylor
rule may have been a reasonable description of past
central bank interest rate policy, it is not used by central
banks as a prescription to guide their choices, one
problem being measurement of y and r * .
12
Much research has been directed towards devising
‘monetary policy rules’ which might direct central
banks in their choices of optimal interest rates to reach a
given target. One such rule is the Taylor rule:
11
European Central Bank, whose governing
council contains representatives from all the
19 national central banks of the eurozone.
Although the Maastricht Treaty specifically
states that the ECB must be strictly
independent of national interests, the political
dimension is always present and there have
been tensions.
Another reason why governments might be
prepared to tolerate some inflation is that it can
help the government budget. At a higher level
of prices, the demand to hold currency is
higher, so inflation provides governments with
extra seigniorage (the income from issuing
currency). Inflation also reduces the real value
of (home-currency, non-indexed) government
debt. But for this to be of benefit to the
government, inflation expectations must be
low when the government borrows (by selling
fixed-interest bonds); if high inflation were
expected, bond rates would also be higher to
reflect this.
The ECB’s choice of the euro interest rate is
supposedly designed to achieve the euro
inflation target (below but close to 2%).
However, as inflation rates differ across the
eurozone countries, the single official euro
interest rate can never be appropriate for all. It
is the average inflation rate across countries
that the ECB tries to steer towards the target,
and since Germany is the largest eurozone
economy, the ECB’s choices of the euro
interest rate have suited Germany better than
other countries. For example, the low ECB
policy rate (2%) during 2003-05 may have
been appropriate for Germany which was
suffering from slow economic growth but it
exacerbated inflation in the southern eurozone
countries (Italy, Spain, Greece, Portugal).
Hence, a government may try to induce the
belief that there will be low inflation so that
long-term rates are low and it can borrow
cheaply, then to cheat by allowing inflation to
write down the real value of its debts. But this
strategy cannot be repeated indefinitely
because expectations catch up, causing new
borrowing to be more expensive and negating
the advantage.
This is the main justification for central bank
independence: central bank independence is
supposed to give credibility to monetary
policy. If the central bank is insulated from
incentives that tempt the government, people
are more likely to believe that it will genuinely
try to meet its inflation target.
Whether or not central banks can be or should
be wholly independent, independence does
seem to have been helpful in reducing inflation
in developed Western economies from the
high levels that prevailed in the 1970s and
1980s (chart, page 8).
This is important because the expectation of
inflation exerts a strong influence on actual
inflation, as recognised in Phillips Curve
theory. Sellers of goods tend to build the
expected rate of inflation into their price
increases, and inflation expectations are also
built into wage agreements. This leads to
persistence in the observed time path of
inflation. Successful control of inflation is
therefore enhanced by the central bank’s
reputation for being able to hold inflation on
target.
The greater present danger, since the financial
crisis reached its height in 2008-10, may be
the opposite condition of deflation, in which
the general price level of goods and services
falls over time. Deflation is a problem because
it raises the real value of debts which reduces
demand (through the wealth effect, section 2.1
above) which, in turn, leads to greater
deflation. Moreover, with interest rates in the
UK, US and eurozone close to zero and high
government debt levels, there is limited scope
for any further attempts at stimulus, either
from monetary or fiscal policy.
Is full independence for the central bank
actually possible? It is hard to ensure that the
individuals who make up the monetary policy
committee are wholly insulated from political
influence. This is especially relevant to the
Inflation in the UK and the US is currently
(November 2016) near the target level of 2%.
However, eurozone inflation (measured as an
12
of Bank Rate in the UK) has been at 0.25%
since late 2008 (it was raised to 0.5% in
January 2016). To maintain its stimulus, the
Fed used Large Scale Asset Purchases to
acquire around $4 trillion of assets between
2008 and 2014. These purchases were split
roughly 60:40 between government debt and
mortgage-backed securities.
average over euro member states) is barely
above zero; there are fears that the eurozone
may go the way of Japan where interest rates
and inflation rates have been close to zero for
two decades.
2.4 Quantitative easing
The Bank of England is currently in its third
programme of quantitative easing (QE),
purchasing assets in the secondary market
which are mainly medium and long-term UK
government bonds (gilts). It pays for these
purchases with its own reserve liabilities: the
seller of the bonds receives a deposit in its
commercial bank and the bank, in turn,
receives a claim on the Bank of England in the
form of reserve deposits.
In the eurozone, the European Central Bank
only began Q in 2015, although it had earlier
conducted several rounds of 3-4-year longterm repo lending to the banks, which amount
to QE with a contract that the purchases be
reversed. Then in March 2015, the ECB began
purchasing €60bn per month of eurozone
government debt, later increased to €80bn (in
reality, the purchases are made by the National
Central Banks of the eurozone countries and
the assets purchased are mainly the countries’
own government debt). The total purchase to
date (November 2016) is €1.4 trillion and the
scheme is due to continue at least until March
2017.
In the first QE programme in 2009-10, the
Bank bought £200bn of gilts. This was raised
to £375bn during 2011-12, then held at this
level by issuing new debt to replace existing
holdings as they matured. In August 2016, the
Bank began a third programme of purchases to
raise its stock of gilts to £435bn, in order to
provide stimulus in the wake of the ‘Brexit’
vote. The value of government debt held by
the Bank of England is currently (November
2016) about 26% of the total debt outstanding
(see balance sheet, page 1).
The particular difficulty with QE in the
eurozone is that the eurozone contains 19
countries. How should the €80bn purchases be
divided amongst them? In the event, the ECB
determined that the purchases should be
proportional to the size of each country’s
economy as measured by its share in the
capital of the ECB. But this means that the
main stimulus is felt in Germany and the other
‘core’ countries that do not need it, with less
stimulus in struggling ‘peripheral’ countries
such as Spain and Portugal. Another difficulty
is a shortage of assets for purchase; the ECB
has addressed this by declaring certain
corporate bonds to be eligible for purchase.
The Bank has used QE because Bank Rate has
been close to its zero lower bound since 2009
(it is currently 0.25%, November 2016).
Therefore there has been little scope to use its
usual method of stimulating economic activity,
which is reducing Bank Rate. QE may thus be
considered as an extension of monetary policy;
it is effectively the same as textbook ‘open
market purchases’. However, it might also be
viewed as fiscal policy as it provides
‘monetary finance’ for the government budget,
casting doubt on the central bank’s
independence.
The country that has conducted the largest
amount of QE, relative to country size, is
Japan. Starting in 2013, the Japanese central
bank now holds some ¥400 trillion ($3.6
trillion) of Japanese government bonds.
QE has also been used elsewhere, in similar
circumstances, and brief descriptions follow.
In the US, the interest rate paid by the Federal
Reserve on excess reserves (the US equivalent
To the extent that QE has had useful effects,
the main channel by which it has worked is
thought to have been via a reduction in
medium and longer-term rates of interest. QE
13
reserves means an increase in the monetary
base, M0 (defined as banknotes and coin plus
bank reserves) and the quantity theory of
money says that the price level rises roughly in
proportion to the money supply, measured as
M0 or some broader measure such as M4
which includes deposits in banks (see
appendix on monetary base control).
causes greater demand for government debt,
raising its price and reducing its yield (see
footnote 10). As an example, interest rates on
long-dated UK government bonds fell by
about 1% during 2009 and corporate yields
also fell as investors sought substitutes for the
bonds bought by the Bank of England (the
‘portfolio balance’ effect). While monetary
policy in ‘normal’ times is the choice of Bank
Rate which is an overnight interest rate, QE
may provide an additional handle over longerterm interest rates.
Lower long-term rates are helpful to
businesses that can finance themselves by
issuing debt. A more important channel may
be the ‘wealth effect’ (Section 2.1), whereby
the raised asset prices stimulate spending.
Another possible channel is that lower yields
tend to weaken the foreign exchange value of
the currency.
Figure 6. UK money supply growth and
inflation
It was also hoped that QE would encourage
banks to increase their lending, given that it
causes the banks to have new reserves, created
to pay for the central bank’s asset purchases
(balance sheet, page 1). However, UK bank
lending remains weak. Still conscious of losses
since the crisis, banks are looking more
carefully at risk, and retail lending margins
remain high (Section 1.3). Higher regulatory
capital requirements are also adding to the
banks’ cost of funds.
source: Bank of England and Office for National
Statistics
The problem with this reasoning is that it
ignores the mechanism by which inflation is
caused. Inflation is rising prices of goods and
services, and it is caused by excess demand for
goods and services, of which there is little
evidence at present. Consistent with this,
despite large increases in M0, average M4
growth has been low or negative for much of
the last 10 years (Figure 6) because bank
lending has been subdued, even though
monetary policy has never before been so
stimulatory.
While the early programmes of QE may have
succeeded in stimulating economic activity,
yields on government and corporate debt have
now reached very low levels in all developed
economies and particularly in the eurozone,
and there are therefore doubts about the
usefulness of any further QE.
No central bank has yet begun to reverse QE.
However, if and when inflation threatens to
return, the Bank of England says it will
gradually do this – i.e. sell government debt
back into the market. Even without a reversal
of QE purchases, the Bank still has its interest
rate instrument: it can cause market rates to
rise by raising Bank Rate, given that Bank
Rate is paid on reserve balances.
Indeed, suspicions have also been voiced that
further QE could cause a return of inflation.
This inflation worry arises from the claim that
QE constitutes ‘monetisation’ of government
debt – the Bank of England is paying for
government spending with new ‘money’ (bank
reserve deposits). The increase in bank
14
Appendix: monetary base control
To reproduce a typical textbook presentation,
suppose that for every £100 of their money,
individuals choose to hold £5 as currency and £95
as a bank deposit. Of this £95, suppose banks’
required reserves and desired excess reserves
together come to £5. Then £100 of money is
associated with £10 of monetary base: the ‘moneymultiplier’ is 10. When the central bank wants to
raise the money supply, it raises the monetary base
by buying government bonds from the private
sector in exchange for new reserves (an ‘open
market purchase’). 14 The money supply is then
supposed to rise through the ‘deposit expansion’
process, as follows.
The usual textbook treatment of monetary policy
claims that the central bank chooses the amount of
the monetary base M0 (alias ‘reserve base’ or ‘high
powered money’; defined as currency issued by the
central bank, plus the banks’ reserve balances at
the central bank). M0 is the central bank’s
instrument rather than the interest rate, and the
money supply M (defined as currency plus deposits
in banks 13) is assumed to be caused by the
monetary base via a predictable ‘money multiplier’
relationship. In turn, according to the quantity
theory of money, an increase in the money supply
is supposed to cause a proportional increase in
nominal output. Formally, this theory is a
statement that the velocity of money V in the
equation MV=PY stays constant where M is the
money supply, P is the price level and Y is real
output.
When the central bank buys government bonds,
deposits in banks rise and this is reflected as an
increase in excess reserves. Banks now have more
reserves than they want, so they find willing
borrowers and they lend. But people borrow for the
purpose of paying for goods and services, which
means that the borrowed funds are paid into some
other individual’s or firm’s bank deposit, or they
may be used to repay a loan. Either way, banks
will again find they have more reserves than they
want which then causes more lending. Each time
the same initial reserves are lent and re-deposited,
a proportion is kept as currency (5%, in then above
example), and a further proportion (another 5%) is
absorbed as required reserves and desired excess
reserves, and it is therefore unavailable for lending.
In this way, deposits and the money stock M
continue to increase by decreasing amounts, as is
laboriously described in most textbooks. The
expansion process terminates when the money
stock has risen by the ‘multiplier’ (10 in the above
example) multiplied by the injection of new
reserves.
The point of this appendix is to show that this is a
poor description of reality. Further, it would not be
possible for the central bank to operate by setting
the monetary base. Essentially, this is because
bank deposits are claims to currency which only
the central bank can supply. Hence, in order to
ensure that banks can honour their obligations to
convert deposits into currency, the central bank
must stand ready to issue whatever quantity of
currency is demanded; it cannot choose the
quantity of currency it issues and must therefore
set the cost of its lending, i.e. the interest rate.
In more detail, the money multiplier theory is as
follows. With the central bank setting the value of
the monetary base, it is obviously no longer
lending (providing liquidity) to banks on demand,
and banks are therefore presumed to keep enough
excess reserves to satisfy deposit withdrawals. The
amount of this desired fractional reserve would be
based on the observed statistics of deposit
withdrawals, balanced against the loss of income
from foregone lending In this scenario, the
wholesale (interbank) interest rate becomes
market-determined at a value that equates the
given stock of reserves with banks’ demands for it.
One problem with this process is that banks do not
have a pool of approved borrowers queuing up for
loans and waiting until the banks have funds to
lend. As mentioned above, banks offer credit when
14
In this description, the tools of monetary management
are open market operations by which the central bank
changes the monetary base, the required reserve ratio
which affects the value of the multiplier, and the central
bank’s ‘discount’ rate for ‘last resort lending’ which
affects the multiplier by influencing the amount of
excess reserves that banks choose to hold.
13
There are several definitions of ‘money’: M1 is
currency held by the public + short-term deposits, M2 is
M1 + medium term deposits etc. The reported measure
in the UK is M4 – ‘broad money’ – which also includes
long-term deposits in banks and building societies. For
this discussion we can think of money as being M1.
15
a potentially profitable lending opportunity arises,
then find the funds in the money-market.
causing a loss of confidence in the banks and
breakdown of the payments system. In the absence
of access to central bank lending, the only way for
banks to be wholly confident of meeting all
possible withdrawals of currency would be for
them to hold reserves at least equal to their shortterm liabilities. This arrangement, known as 100%
reserve banking, would be a radical departure from
current practice.
Clear evidence that this multiplier process is not
operating has been provided by the Bank of
England’s recent programmes of quantitative
easing: under QE, the monetary base (M0) has
risen greatly while broad money (M4) has hardly
changed.
A more obvious problem would arise in the
opposite case of a shortage of reserves, which
might be caused by a rise in currency demand, or a
deliberate reduction in monetary base by means of
a central bank open-market sale.
It might then be argued that the central bank could
fix the amount of monetary base but undertake to
lend extra reserves only in an emergency or truly
‘last resort’ situation. But the only consistent
criterion for identifying such an emergency would
be if there is a shortage of reserves. As soon as
banks became confident of such support, at
whatever interest rate the central bank may choose,
excess reserve holdings could be reduced back to
zero, and we revert to the present system in which
the interest rate for central bank lending is the
monetary control instrument.
If an individual bank found its reserves falling
below its desired level, its reaction would be to sell
liquid assets in the money-market. If banks
collectively suffered a reduction in their reserves,
competition to raise funds by selling assets would
cause a rise in wholesale interest rates. With an
acute shortage of reserves, it might be argued that
interest rates would rise sufficiently to persuade
some individuals to deposit their currency
holdings, thus relieving the shortage. 15 A more
plausible outcome is that the banks’ behaviour
would lead depositors to doubt the banks’ ability to
pay, inducing them to withdraw more currency
rather than depositing.
The upshot is that the central bank cannot fix the
monetary base. Unless the central bank is prepared
to let banks fail because of a shortage if reserves, it
must provide reserves on demand: it must finance
money-market shortages in full against whatever
collateral the banks are able to offer. In doing so, it
cannot avoid setting its interest rate for this
finance: Bank Rate in the UK.
Even if banks generally were holding large stocks
of excess reserves, and even if the central bank
never attempts to reduce the monetary base, there
will always be a non-zero probability that net
currency withdrawals will exceed the banks’
aggregate excess reserves and vault currency. If
this happened, banks would no longer be able to
pay out currency to their depositors, immediately
Of course, this does not prevent the central bank
from targeting the monetary base. The central bank
may target any variable, meaning that it chooses
the path of its interest rate instrument over time in
attempt to achieve some desired value or range of
values of the target variable. Monetary base
targeting (and money supply targeting) has indeed
been practiced from time to time by various central
banks in the belief that this was a good method of
achieving some desired inflation rate. The
prevalent current practice is rather to target the
inflation rate itself.
15
If one follows the textbook story strictly, a shortage of
reserves would put the deposit expansion process into
reverse: banks would attempt to regenerate their
reserves by calling in loans which would reduce
deposits causing a further reduction in loans etc.
However, while banks could halt new lending, it would
be out of the question for them to call in loans unless
this is for non-performance. As stressed above, bank
loans are generally illiquid because they have been used
to purchase firms’ capital stock and individuals’
properties. When a bank liquidates a non-performing
loan, it usually recovers considerably less than the book
value of the loan, the process takes time, and it may
push the borrower into bankruptcy if the borrower is not
already bankrupt.
16