5.5.2 The real costs of disinflation

The second main policy implication of the new classical approach concerns

the output–employment costs of reducing inflation. New classical economists

share the monetarist view that inflation is essentially a monetary phenomenon

propagated by excessive monetary growth. However, substantial disagreement

exists between economists over the real costs of disinflation. Here we

will compare the new classical view with that of Keynesians and monetarists.

The amount of lost output that an economy endures in order to reduce

inflation is known as the ‘sacrifice ratio’. In Keynesian models the sacrifice

ratio tends to be large, even if agents have rational expectations, owing to the

sluggish response of prices and wages to reductions in aggregate demand.

Given gradual price adjustment, a deflationary impulse will inevitably lead to

significant real losses which can be prolonged by hysteresis effects, that is,

where a recession causes the natural rate of unemployment to increase (see

Cross, 1988; Gordon, 1988; and Chapter 7). Some Keynesians have advocated

the temporary use of incomes policy as a supplementary policy measure

to accompany monetary restraint as a way of increasing the efficiency of

disinflation policies (see, for example, Lipsey, 1981). It should also be noted

that Post Keynesian economists regard incomes policy as a crucial permanent

anti-inflationary weapon. Monetary disinflation alone will tend to produce a

permanently higher level of unemployment in Post Keynesian models (see

Cornwall, 1984).

The orthodox monetarist view, discussed in Chapter 4, section 4.3.2, is that

unemployment will rise following monetary contraction, the extent and duration

of which depend on the degree of monetary contraction, the extent of

institutional adaptations and how quickly people adjust downwards their

expectations of future rates of inflation. The critical factor here is the responsiveness

of expectations to the change of monetary regime and this in turn

implies that the credibility and reputation of the monetary authority will play

a crucial role in determining the sacrifice ratio.

In contrast to both the Keynesian and monetarist models, the new classical

approach implies that announced/anticipated changes in monetary policy will

have no effect on the level of output and employment even in the short run,

provided the policy is credible. An announced monetary contraction which is

believed will cause rational agents immediately to revise downwards their

inflation expectations. The monetary authorities can in principle reduce the

rate of inflation without the associated output and employment costs predicted

by Keynesian and monetarist analysis; that is, the sacriflce ratio is

zero! As one critic has noted, ‘in a Sargent–Wallace world the Fed could

eliminate inflation simply by announcing that henceforth it would expand the

money supply at a rate compatible with price stability’ (Gordon, 1978, p. 338).

In terms of Figure 4.6, the rate of inflation could be reduced from A to D

without any increase in unemployment. In such circumstances there is no

necessity to follow a policy of gradual monetary contraction advocated by

orthodox monetarists. Given the absence of output–employment costs, new

classicists argue that the authorities might just as well announce a dramatic

reduction in the rate of monetary expansion to reduce inflation to their preferred

target rate.

With respect to the output–employment costs of reducing inflation, it is

interesting to note briefly the prima facie evidence provided by the Reagan

(USA) and Thatcher (UK) deflations in the early 1980s. Following the restrictive

monetary policy pursued in both economies during this period, both

the US economy (1981–2) and the UK economy (1980–81) experienced deep

recessions. Between 1979 and 1983, inflation fell from 11.2 per cent to 3.2

per cent in the US economy and from 13.4 per cent to 4.6 per cent in the UK

economy, while over the same period unemployment rose from 5.8 per cent

to 9.6 per cent in the USA and from 4.7 to 11.1 per cent in the UK (see Tables

1.4 and 1.5). In commenting on the UK experience, Matthews and Minford

(1987) attribute the severity of the recession in this period primarily to

adverse external and supply-side shocks. However, the monetary disinflation

initiated by the Thatcher government was also a factor. This disinflation was

unintentionally severe and as a result ‘expectations were quite unprepared for

it’. Because initially the Thatcher government had a credibility problem, the

‘accidental shock treatment’ produced painful effects on output and employment.

An important influence on credibility in new classical models is the

growth path of government debt. New classical economists insist that in order

to engineer a disinflation without experiencing a severe sacrifice ratio, a fiscal

strategy is required which is compatible with the announced monetary policy,

otherwise agents with rational expectations will expect a policy reversal (‘Uturn’)

in the future. As Matthews and Minford (1987) point out, ‘A key

feature of the Thatcher anti-inflation strategy was a parallel reduction in

government budget deficits.’ This ‘Medium Term Financial Strategy’ was

aimed at creating long-run credibility (see also Minford et al., 1980; Sargent

and Wallace, 1981; Sargent, 1993, 1999) .

In the USA a ‘monetary policy experiment’ was conducted between October

1979 and the summer of 1982. This Volcker disinflation was also associated

with a severe recession, although the influence of the second oil shock must

also have been a contributory factor. In commenting on this case, Milton

Friedman (1984) has argued that the relevant economic agents did not have

any widespread belief in the new disinflationary policy announced by the Fed

in October 1979. In a similar vein, Poole (1988) has observed that ‘a recession

may be necessary to provide the evidence that the central bank is serious’.

For a discussion of the US ‘monetarist experiment’, the reader is referred to

Brimmer (1983) and B. Friedman (1988). Useful surveys relating to the issue

of disinflation are provided by Dalziel (1991), Ball (1991, 1994) and Chadha

et al. (1992).

From the above discussion it is clear that, for painless disinflation to occur,

the public must believe that the monetary authority is prepared to carry

through its announced monetary contraction. If policy announcements lack

credibility, inflationary expectations will not fall sufficiently to prevent the

economy from experiencing output–employment costs. Initially the arguments

relating to the importance of credibility were forcefully presented by

Fellner (1976, 1979). A second line of argument, closely related to the need

for policy credibility, is that associated with the problem of dynamic time

inconsistency. This matter was first raised in the seminal paper of Kydland

and Prescott (1977) and we next examine the policy implications of this

influential theory.