MONEY SUPPLY AND DEMAND

1
MONEY SUPPLY AND DEMAND
Summary Money supply. We consider, and
criticise, the standard textbook story of the
money supply process, according to which control
of the (broad) money supply, M4, is achievable
via control of the money base M0, as a result of
a supposedly constant money multiplier, mM4/M0.
If that story is wrong, what does drive the
supply of M4 (i.e. cause M4 to be what it
is)? Money demand. The central plank of
monetarism is the Classical money demand
function, according to which
If this equation is correct, and if Y is
exogenous (i.e. independent of monetary policy,
and if v is constant, and if M is controllable,
then M causes P. So inflation control then
translates into a simple money supply rule, such
as Milton Friedmans 4 growth proposal
(presidential address to the American Economic
Association, AER 1968). But do the assumptions
hold?
2
Money supply The money multiplier can be derived
as follows (Mankiw 5th ed, ch 18.1)
The first two identities are definitions and the
remaining three identities follow. The money
multiplier, m, therefore depends on the reserve
ratio (rr) and the currency ratio (cr). If cr0,
then m is just the inverse of rr, so if rr0.1,
then m10. As cr rises, m falls, so for instance
if rr0.1 and cr0.4, then m2.8. This algebra
suggests a simple story whereby an open-market
purchase of bonds by the Central Bank, raising H,
will raise M by m times as much. It is worth
thinking through the process involved.
3
So the elementary view of the money supply
process is that H is determined exogenously and
this, with rr and cr assumed fixed, determines M
via the money multiplier. So there are three
assumptions to be examined (1) rr is exogenous
and known (2) cr is exogenous and known (3) H is
the monetary policy instrument Note that in (1)
and (2) fixed has been weakened to exogenous
and known. Clearly we can control M with H even
if rr or cr are predictably time-varying rather
than fixed constants. Are (1) and (2) valid?
The answer is no. The essential point is that rr
and cr are, in advanced countries, best viewed as
behavioural variables. In the UK, there is a
statutory rr of 0.15 (of eligible liabilities).
This is nothing to do with monetary control and
exists in order to provide revenue to the Bank of
England. So banks must decide what quantity of
reserves to hold. That is a complex calculation
which depends on the opportunity cost of reserve
holding vis-à-vis the expected benefit (linked to
the costs incurred should a bank run out of
reserves).
4
Similarly with cr. This is determined by a
complex interaction between the banks and their
depositors. Confidence in bank solvency, the
interest rate paid on bank deposits, the ease of
accessing non-cash means of payment all affect
cr, and the banks exert a good deal of influence.
The implication is that pushing on a string -
trying to expand M via H in a recession - is not
very effective and, also, restraining M via H in
a boom may also be difficult given the
countermeasures available to banks. rr, cr and
therefore m are endogenous, so that even if H
were exogenous, M would be endogenous. The table
below, taken from Mankiws text, illustrates this
well. Why did rr and cr rise so much between the
two dates shown?
5
What about assumption (3), that H is the monetary
policy instrument? Is it true, and could it be
true? The answer to the first question is no.
Ever since the widespread bank failures in the
Great Depression in the US, Central Banks have
operated a lender of last resort facility,
whereby they are always willing to provide cash
to the banks. The Central Banks principal policy
instrument is the short rate of interest. What
about the second question? The answer is yes,
under floating exchange rates, within limits.
However the best-known attempt to control H in
modern times, in the US between 1979 and 1982,
was a failure. It led to unacceptable
fluctuations in short interest rates and the
exchange rate. So the money multiplier
identities do not provide a very useful account
of the money supply process. What does, then,
determine money supply? An alternative theory
begins from what are called the credit
counterparts of money. To understand this, we
need to explore financial system balance sheets.
6
Central Bank Assets Liabilities
Consolidated Banking System Assets
Liabilities
M4
Commercial banks
7
The preceding slide, based on some charts in
Burda/Wyplosz, shows how the financial system
balance sheets fit together. The red and green
boxes show items that cancel out when the
Commercial Bank and Central Bank balance sheets
are consolidated. The end result is (net)
foreign assets M4 lending lending to
government M4 net worth Or, simplifying a
little, and writing D for domestic credit -
defined as total lending by the banking system to
domestic non-banks - we can write R D M N
(R is forex reserves, N is total banking
system net worth) Can we infer any causation
directly from this identity? No. But it may be
that M is largely driven by D, that is lending to
government plus M4 lending.
8
Consider first the government component of D. If
governments are constrained in the amounts they
can borrow from abroad and from the non-bank
private sector, then only the domestic banking
system is left. Extra government borrowing is
therefore monetized, unless there are offsetting
falls in reserves or M4 lending. So in developing
countries monetary growth may be essentially
driven by the fiscal deficit. In advanced
countries, where M4 lending is much greater than
bank lending to government, and where the short
rate of interest is directed largely to inflation
targetting, then monetary growth may be largely
driven by M4 lending. However this leaves open
the question of how much of the borrowing by
households and firms is intermediated through the
banking system. Some agents do not have a choice
in the matter, and can borrow only from the
banks. We will return to this later in the course.
9
Money demand There will only be time for a brief
discussion in the lecture. We want to understand
the demand for both M4 and M0 (H). The classical
model plus the money multiplier identities offer
a very simple story demand for M4 proportional
to PY and demand for M0 proportional to the
demand for M4. Unfortunately this story does not
remotely fit the facts, as we can see for the
UK. The next slide shows M4/PY, ie the inverse of
M4 velocity, for the UK post-1970. M4 divided by
quarterly disposable income has risen from about
3 in 1980 to about 5.5 in 2000. Why? For the
assessment of monetarism, ie the policy of
targetting the growth rate of M4 at some constant
rate, the important point is that the velocity of
M4 has proved highly unpredictable. In such
circumstances, targetting of monetary growth will
produce unintended fluctuations in the growth
rate of nominal income (PY).
10
BoE Inflation report, August 2001.
11
The final slide, next, shows the behaviour of MO
velocity in the UK. While M4 velocity has fallen,
M0 velocity has risen. Why? That question is left
for the seminar. Note the association on the
chart between the short rate of interest and M0
velocity.
12
(No Transcript)