The costs of inflation

By Clive Briault of the Bank’s Monetary Assessment and Strategy Division.

This article surveys the academic literature on the costs of inflation. There are many well-established
theoretical reasons why inflation—and uncertainty about future inflation—may reduce economic welfare.
Moreover, there has recently been an expansion in empirical work on the relationship between inflation
and growth, looking either at a single country or across countries. Most studies have found a significant
negative correlation between inflation and growth. And at the broadest level, the article concludes, the
available evidence supports the view that well-run economies with strong and efficient productive
structures tend to exhibit both low inflation and high growth.

The article also briefly reviews the emerging literature on the costs of reducing inflation, which suggests
that the short-term trade-off between unemployment and inflation is more pronounced in countries with
low inflation.

In his 1992 LSE Bank of England lecture on the case for work in this area,(3) but rather to offer a selective review,
price stability,(1) the then Governor listed many of the costs particularly of recent empirical contributions.
of inflation, giving particular emphasis to those arising from
unanticipated inflation. In addition to the costs that arise The first two sections describe the main theories of the costs
even if inflation is perfectly anticipated—as a result of the of anticipated and unanticipated inflation respectively, and
need to economise on real money balances and revise price some related empirical results. The third section considers
lists, and of the less than full indexation of tax systems and recent empirical work on the determinants of economic
debt contracts—there are important costs arising from growth over the longer term, in which the importance of
unanticipated inflation. These include: inflation is assessed alongside that of variables such as
investment, government spending and the presence of
● the unplanned redistribution of income and wealth; market distortions. The final section discusses the possible
trade-off between the benefits of lower inflation and the
● additional uncertainty about future prices introduced into costs of reducing inflation.
decisions about consumption, saving, borrowing and
investment; and Anticipated inflation
The simplest conceptual model in which to analyse the costs
● the higher costs of identifying changes in relative prices of perfectly anticipated inflation is one of an economy in
and allocating resources accordingly. perfect competitive equilibrium and in which there are no
distortions except for the non-payment of interest on notes
People may respond to these costs by attempting to predict and coin.(4) This can be used to illustrate the key issues,
future inflation or by searching out information on relative although it is clearly not a realistic description of the world.
prices, but the allocation of increased scarce resources to In such a model, inflation constitutes a tax on holdings of
such activities is costly for society as a whole (even if it may currency, and it imposes welfare costs as agents alter their
be privately profitable for those who undertake it). behaviour in response. At their most basic, these take the
form of ‘shoe leather’ costs: people will make more
This article expands on the Governor’s remarks, looking at frequent trips to the bank to withdraw currency (if bank
the growing theoretical and, especially, empirical literature deposits pay interest or provide depositors with other
on the costs of inflation, and in particular at the relationship services) and attempt to synchronise cash expenditures with
between inflation and growth over the longer term.(2) It is the receipt of cash income.(5) These welfare costs would
not intended to provide a comprehensive survey of academic disappear if there was deflation at a rate sufficient to drive

(1) Published in the November 1992 Quarterly Bulletin, pages 441–48.

(2) The Governor discussed the short-run trade-off between inflation and unemployment—the ‘Phillips curve’—in his lecture.
(3) There are a number of major survey articles in this area, including Driffill, Mizon and Ulph (1990), Fischer (1981a and 1994), Fischer and
Modigliani (1975), Howitt (1990), McTaggart (1992), Orphanides and Solow (1990) and Woodford (1990).
(4) It would be possible (albeit at a cost) to pay something like interest on currency, for example—as discussed by Goodhart (1993)—through a lottery
using the serial numbers on banknotes, and to remunerate banks’ reserves held at the central bank. But if this was achieved, the optimal rate of
inflation in this type of framework would become indeterminate.
(5) As described in Bailey (1956), pages 100–2.

33
Bank of England Quarterly Bulletin: February 1995

the nominal interest rate on interest-bearing and riskless non-payment of interest on currency.(6) And once additional
substitutes for cash down to zero, since people would then distortions are introduced, the optimal rate of inflation
no longer need to economise on their holdings of cash.(1) becomes more difficult to determine. For example, if
non-distortionary lump-sum taxes and subsidies are not
In this simple model, the cost of inflation depends on how available to the government, then raising revenue through an
much the demand for cash varies with the nominal interest inflation tax—seigniorage—may be no less desirable than
rate. The cost will be positively related both to the rate of other forms of taxation which distort economic behaviour.(7)
inflation (which will be reflected in the nominal interest rate) Inflation will still be costly for the economy, and this will
and to the sensitivity of the demand for cash to the interest limit the extent to which it should be used as a source of
foregone as a result of holding cash. Using such an revenue, but its optimal rate may still be positive.
approach, some estimates have suggested that even fully
anticipated inflation may have large welfare costs.(2) But Finally, the existence of tax systems that are not fully
these estimates are very sensitive to the specification of the indexed and of contracts set in nominal terms (as, for
money demand function and to the chosen definition of example, for most mortgage borrowing) leads to further
money (in most developed economies, cash in domestic distortions from perfectly anticipated inflation. The true cost
circulation is a fairly small proportion of national income). of inflation in this respect is the cost of adapting the tax
system or financial contracts so that they are fully indexed, if
In addition, the demand for money approach is incomplete, that is possible, rather than the costs arising from a
since real income, real wealth and the real rate of interest are combination of inflation and non-indexation; but if
assumed to be unaffected by inflation. And the approach is non-indexation persists then inflation could be extremely
limited to a comparison of various rates of inflation damaging to an economy.
(different nominal interest rates) rather than of different
growth paths of the economy. In reality, inflation probably Unanticipated inflation
affects savings behaviour and capital accumulation, not least
because any change in the real demand for cash represents a Redistribution costs
switch in the portfolio of assets held in the economy.
Unanticipated inflation leads to redistributions of income
and wealth—in particular from creditors to debtors, when
The implications of these points have been considered in a contracts are less than fully indexed, and from those with
number of theoretical models, which generate a variety of fixed nominal incomes to those who pay them. Such
outcomes.(3) In a simple neo-classical growth model, redistributions may be very costly for certain individuals and
introducing real cash balances as the only alternative form of sectors of the economy. They may also undermine
wealth can lead to the conclusion that higher inflation will confidence in property rights. The difficulty of measuring
be associated with increased physical capital but a slower the overall welfare costs here—not least because for every
rate of growth of output per head, as the economy moves immediate loser there is an immediate gainer—should not
towards its steady state. But if the level of wealth is made to obscure their importance.
depend on saving behaviour, which is in turn influenced by
inflation, the result can be reversed.
Costs for decision-taking
The conclusions also depend on the assumptions made about Uncertainty about future price levels is likely to distort the
the role played by money within the economy—for example allocation of resources in a number of ways. First, in the
on whether money is included in consumers’ utility absence of index-linked assets, increased uncertainty may
functions directly, or whether there is assumed to be a increase the attractiveness of real (as opposed to nominal)
cash-in-advance constraint, so that purchases can take place assets because they give a hedge against inflation. Second,
only using money balances held for some time in advance.(4) uncertainty is likely to discourage agents from entering into
And in models with more elaborate sectoral distinctions, long-term monetary contracts, thereby removing the
inflation can result in an inefficient allocation of productive assurance provided by longer-term contracts. This is likely
capital to the private financial sector.(5) to inhibit investments where the return is a long time ahead,
and thus to reduce companies’ investment rates and lead to
But it is difficult to identify an intuitively-appealing role for investment in shorter-lived assets (which may represent a
money within traditional growth models, and equally less efficient form of investment). Third, savers and lenders
difficult to provide a rationale for the existence of money in may respond to uncertainty by demanding a risk premium,
a hypothetical economy in which the only distortion is the so increasing the real cost of funds for borrowers. Fourth,

(1) As argued by Friedman (1969) in his derivation of the ‘optimum’ quantity of money.
(2) Bailey (1956) provides an early example of this, using Cagan’s (1956) data on hyperinflation episodes. Fischer (1981a) and Lucas (1993) offer more
recent analyses. The estimated cost of exceptionally high rates of inflation can be up to 50% of GDP, but even with low inflation a one
percentage point increase in inflation is estimated to generate a welfare cost equivalent to up to 0.1% of output.
(3) These are described in the survey article by Orphanides and Solow (1990).
(4) Again, estimates of the cost of inflation can be derived from such models. They vary widely, but are generally smaller than those derived from
measuring the area under a money demand curve; see the survey in Gomme (1993).
(5) For example, in Ireland (1994).
(6) As discussed by Hahn (1971, 1973).
(7) A point made by Phelps (1972, 1973). Cooley and Hansen (1991) calibrate a cash-in-advance model to assess the optimal distribution of tax on
money, goods, labour and capital. But Kimbrough (1986a, 1986b) argues that since money balances are an intermediate good they should not be
taxed.

34
 Costs of inflation

capital will be misallocated if savers and investors form But some caution should be applied in interpreting these
different expectations of inflation and thus different views of results. They depend not only on whether the equation for
the ex ante real rate of interest.(1) forecasting inflation is specified using a univariate or a
structural model (and on whether models or survey data are
Uncertainty about future rates of inflation is likely to be used) but also on whether it allows for structural changes in
greater at higher rates of inflation.(2) During a period of low the inflation process. In general, the relationship between
inflation, the public may be reasonably sure that the inflation uncertainty and the level of inflation appears to be
authorities will be content with the situation and will attempt strongest when there is a change in the trend rate of
to prolong it. Following an inflationary shock, however, if inflation—or when there is uncertainty about the possibility
the public are unsure about the preferences of the authorities, of such a change—rather than when there is shorter-term
they will be uncertain how far the shock will be variability in inflation around an unchanged trend rate. This
accommodated through monetary policy and therefore about is important if, as suggested above, inflation is most costly
the future rate of inflation.(3) The authorities themselves may when the period of uncertainty stretches over a number of
be unsure what to do in a period of high inflation, because years rather than over shorter periods.
they face a dilemma: they would like to disinflate, but may
be reluctant to do so because of the recession that would There may also be a causal link between the variability and
probably result over the short term. Disinflation the average level of inflation, at least if monetary policy is
will probably occur eventually, but its timing will be accommodating. For example, variable inflation might lead
uncertain.(4) risk-averse workers to negotiate a nominal wage that
incorporated a premium in case the price level proved higher
But future inflation may be uncertain even when current than expected. This would tend to push up nominal and real
inflation is low. The public may be unsure whether any wages. Moreover, if unanticipated inflation generated the
future exogenous inflationary shocks (for example, adverse illusion of a real increase in company profits—or even an
shifts in the terms of trade) will be accommodated by the actual increase, perhaps because firms had financed
authorities, or whether they might deliberately create an themselves earlier by borrowing at fixed nominal interest
unexpected bout of inflation in an attempt to boost output rates below current market levels, or were paying below
and employment in the short term, perhaps for electoral current market rents on old leases—then firms might be
reasons. prepared to concede higher wages.

The earliest empirical work on the relationship between Impact on relative price movements
inflation and uncertainty used fairly simple measures of The relationship between inflation and changes in relative
inflation variability, such as standard deviations around the prices has also been much studied. Changes in relative
average level of inflation over periods or across countries. prices give signals which guide resource allocation in market
The results suggested that there was a positive relationship economies. It is therefore an important question whether
between the variability and the level of inflation.(5) But there higher inflation makes it more difficult to perceive and to
can also be considerable variability of inflation around a near react to changes in relative prices, or causes relative price
zero mean, as was the case in many countries before the changes which would not otherwise occur.
Second World War.(6)
Misperceptions of relative price changes are usually
Moreover, variability and uncertainty are not the same thing. analysed using models that assume that expectations are
Inflation might be highly variable, but if the processes rational and that the market-clearing process always
generating it were understood there might be little associated functions smoothly, whereas unanticipated changes in
uncertainty; and the costs of variable inflation will be lower inflation and increased relative price variability both result
if the variations are predictable. Attempts have been made from unanticipated changes in the money stock.(8) In such
to construct measures of uncertainty by adjusting measures models, a fully perceived change in the money stock has no
of variability to take account of this, using either effect on relative prices and there is no confusion between
econometric models or survey data on inflation expectations aggregate and relative price changes. However, a
to compare inflation outturns with predicted values.(7) Most misperceived change leads to movements in prices in
of the results suggest a positive relationship between the rate individual markets which participants regard, at least in part,
of inflation and these measures of ‘conditional’ uncertainty, as changes in relative prices. Assuming demand and supply
particularly for uncertainty over longer time horizons. elasticities in individual markets differ, these perceived

(1) There have been relatively few attempts to model formally the impact of uncertainty on economic welfare. Rankin (1994) surveys this area and
suggests a model in which uncertainty about the future money supply—and so inflation—has a detrimental effect on the economy because
risk-averse workers push up money wages and thus unemployment, and because of the increased variability of future prices and output.
(2) As argued by Friedman (1977) and Okun (1971).
(3) This is a familiar result in the ‘rules versus discretion’ literature, surveyed in Fischer (1990) and Cukierman (1992).
(4) Ball (1992) presents a model along these lines.
(5) See, for example, Foster (1978), Jaffee and Kleiman (1977), Logue and Willet (1976) and Okun (1971).
(6) Backus and Kehoe (1992) cite mean rates of inflation for the United Kingdom of 0.09%, -1.07% and 6.92% for the periods 1870–1914, 1920–39
and 1950–83 respectively, with standard deviations of 2.37, 6.86 and 5.05.
(7) Including Ball and Cecchetti (1990), Brunner and Hess (1993), Cukierman and Wachtel (1979), Engle (1983), Evans (1991), Evans and
Wachtel (1993), Jansen (1989), Joyce (1994), McTaggart (1992) and Pagan, Hall and Trivedi (1983).
(8) Models of this type have been developed from Lucas (1973) and include Cukierman (1982 and 1984). It could be argued that the costs arising in
these types of model represent the cost of money surprises rather than of inflation, but in practice the link between money and the aggregate price
level is a convenience rather than a necessity, and the basic results carry across to any model in which market participants cannot distinguish
perfectly between relative price changes and movements in the aggregate price level.

35
Bank of England Quarterly Bulletin: February 1995

relative price changes result in changes in actual relative favour of modest inflation. But his approach treated the
prices, which in turn cause a misallocation of resources. asymmetry of price adjustment as exogenous whereas, as just
Such a misallocation arises from unanticipated inflation or outlined, in the menu-cost model of Ball and Mankiw it may
disinflation. be an endogenous response to inflation. In these models,
inflation is costly because it creates inefficient relative price
As with uncertainty about future inflation, there may also be variability without any offsetting benefit.
an intertemporal misallocation of resources. For example,
where imperfect information creates a variety of expectations If, however, the existence of menu costs means that prices
of inflation, resources will be inefficiently allocated over adjust more rapidly at higher rates of inflation then the
time because real rates of return are misperceived by at least impact of certain types of shocks on quantities (such as real
some agents. output and employment) could be mitigated through higher
inflation. But this is an argument in favour of price
A second type of model has been built on the notion of flexibility rather than inflation itself; and given its other
‘menu costs’. This type of model was intended to explain costs, it is unlikely that inflation is the best means of
the non-neutrality of money (why changes in the money achieving such flexibility.
supply may have effects on the real economy), at least over
the short term, by providing an explanation for nominal price The impact of unanticipated inflation on relative prices is
rigidity.(1) Menu costs—including the costs of changing also crucial to the effect of inflation on what Laidler (1990)
price labels, gathering information about markets and taking terms the social productivity of money. Even low rates of
decisions to change prices—cause prices to be adjusted inflation may be costly because they undermine the
infrequently, but more rapidly in response to large shocks usefulness of money as a unit of account and as a store of
than to small ones. Models of menu costs usually assume value, while high rates of inflation may also undermine its
that firms adjust their prices either at fixed intervals(2) or usefulness as a means of exchange. This cost cannot, by
whenever their relative prices move too far away from their definition, be assessed using a model which implicitly treats
correct levels.(3) In either case, the price level will usually the contribution of money to the functioning of an economy
not adjust immediately to a monetary shock, so money may as negligible. If the use of money confers only small
not be neutral. When inflation increases, prices are changed benefits then any damage that inflation might do must
more frequently, but not frequently enough to maintain the necessarily be minor. But if the social productivity of the
previous dispersion of relative prices. As a result, relative monetary system is high, the disruption of that system by
prices move out of line, leading to a misallocation of inflation is potentially much more serious.
resources. The menu-cost approach relates increased relative
price variability to inflation or deflation itself, rather than to Empirical work suggests that relative price variability and
unanticipated inflation, and so suggests a way in which even the rate of inflation have been positively related over a wide
fully anticipated inflation entails costs. range of countries and over time in individual countries, and
that relative price variability is positively related to the
Such models generally imply that the optimal rate of extent of unanticipated inflation.(5)
inflation, if one exists, is zero.(4) For example, Ball and
Mankiw (1994b) presented a model in which firms change Again, however, questions about the direction of causation
their prices only when induced to do so by a sufficiently arise. Relative price variability might be exogenous, in
large shock: they tolerate limited deviations of actual from which case an asymmetric response of prices to shocks could
desired prices. Positive inflation will then cause firms’ lead to inflation if some prices are inflexible downwards.
relative prices to decline automatically between price This effect would diminish as inflation increased, unless
adjustments. So when a firm wants to lower its relative price there were some reason why prices rose more easily than
it may not need to pay the full menu cost, because inflation they fell in relation to some core or expected rate of inflation.
does some or all of the work. By contrast, inflation will Price inflexibility might then lead to both higher inflation
widen the gap between desired and actual prices when a firm and resource misallocations.
wants to increase its relative price. So shocks that raise
firms’ desired prices cause larger price responses than shocks Similarly, inflation and relative price variability might be
that lower desired prices. (The opposite would be true if the positively related if both were affected by major supply
general price level was falling—a firm wanting a lower shocks, if speeds of adjustment or short-run supply
relative price would have to pay the menu cost to jump ahead elasticities varied across industries, or if an accommodating
of the falling prices charged by other firms.) monetary policy allowed major price shocks to lead to higher
inflation. Or government policy might cause both inflation
Tobin (1972) used an assumption of downward price and and relative price variability. For example, higher
wage rigidity to suggest that a positive rate of inflation could government spending might both increase inflation and
be optimal—a variant of the ‘oiling the wheels’ argument in change the composition of final demand and so relative
(1) Ball and Mankiw (1994a) provide a broad overview of the role of menu costs and of the various approaches adopted towards modelling them.
(2) Examples of this type include Blanchard (1983), Blinder (1991) and Taylor (1979).
(3) Models of this type include Ball and Mankiw (1994b), Barro (1972), Caballero and Engel (1992), Caplin and Leahy (1991), Caplin and
Spulber (1987), Danzinger (1984 and 1987) and Sheshinski and Weiss (1977 and 1983).
(4) However, as discussed below, the non-neutrality of money may have implications for the costs of moving from positive to zero inflation.
(5) See Clare and Thomas (1993), Fischer (1981b), Jaffee and Kleiman (1977), Parks (1978), and Vining and Elwertowski (1976).

36
 Costs of inflation

prices or, through indirect taxation, it might change relative output per head relative to the steady-state position. If all
prices directly.(1) Fischer (1981b) demonstrated that the countries had the same steady-state position, poorer
relationship between inflation and relative price variability countries (defined by their initial stock of physical and
in the United States was mainly the result of food and human capital) would grow more rapidly until they caught
energy price shocks after 1973, and that much the same had up with the richer countries. But this ‘convergence’
been true in West Germany and Japan. He also found a hypothesis would not necessarily hold if national
strong contemporaneous correlation between inflation and steady-state positions differed.
relative price variability, with no clear sequence. Studies
using UK data show that, apart from oil price shocks, the More recent growth models have focused on the
major determinant of relative price variability has been determinants of technical progress, which include
changes in indirect taxation.(2) investment and the level and effectiveness of research and
development expenditure. The rate of inflation (and other
So it is difficult to reach any firm conclusion that higher factors) may, by influencing these, be important in
rates of inflation necessarily lead to greater relative price determining both the steady-state rate of growth and the path
variability. Inflation and changes in relative prices could along which an economy approaches it.
have a common cause; or relative price variability could be
the cause of inflation. In each case, it would be wrong to Empirical work here has been conducted on a number of
conclude that higher inflation was itself the cause of bases. First, some studies have used time-series data for
increased relative price variability, or even that relative price individual countries, whereas others have adopted a
variability necessarily involved a welfare cost. cross-country approach in which the data for each country
are averaged over extended periods. A further method
Indexation is to use a panel of data which combines these two
approaches.
If many of the costs of inflation could be avoided using
index-linked contracts, why is indexation not widespread? Second, some studies include inflation as the only
Institutional arrangements have tended not to adapt to take determinant of growth, while others include a range of other
account of inflation (except in some countries experiencing possible determinants, either to take account of one-off
very high inflation). This may be because the costs of disturbances (such as oil price shocks) or in an attempt to
indexation are relatively high. Indexation may also be model growth more comprehensively by testing the
inefficient because it inhibits changes in relative prices significance of other possible influences. Third, although
which would otherwise be desirable. And as it becomes a growth per head is usually chosen as the dependent variable,
feature of an economic system, excess demand and other some studies have focused on factor productivity or
inflationary pressures will tend to be transmitted into prices investment. And some have used both the rate of inflation
more rapidly. So even if an indexed economy suffers fewer and its variability (absolute or relative to a prediction) as
of the costs of inflation, it also tends to have higher inflation explanatory variables.
(since indexation may also reduce the counterinflationary
resolve of the authorities). As a result, those costs which Although a few studies have found no relationship between
cannot be removed through indexation (in particular those inflation and the growth rate, the general consensus is that
relating to relative prices and to anticipated inflation) may growth is significantly and negatively related to inflation. In
become more severe.(3) some cases, the correlation is estimated to be quite large,
suggesting that a one percentage point reduction in inflation
The effects of inflation on growth could be associated with an increase in the rate of growth by
something between 0.1 and 0.5 percentage points. But it is
The arguments presented above suggest that inflation and
recognised that this relationship is unlikely to be monotonic:
inflation uncertainty lead to a misallocation of resources;
the results do not imply that a move from stable prices to
they are therefore likely to reduce the rate of growth of an
deflation would increase the growth rate. (There have
economy. Attempts have been made to estimate by how
been so few instances of sustained falling prices that the
much.
available data do not permit any reliable assessment of their
effect.)
One possible starting-point is a neo-classical growth model.
In this, an economy’s growth rate is determined by technical
Time-series approaches
progress and the growth rate of labour supply, both of which
are assumed to be exogenous, but the level of output per Beginning with single country time-series analysis, the
‘effective’ worker in steady-state equilibrium depends on a simplest approach is to regress output growth on current and
set of variables—which might include the rate of inflation— lagged inflation. Grimes (1991) ran such equations for
that determine the efficiency with which labour and capital 21 industrialised countries, including terms of trade and
are used. If returns to capital are diminishing, the growth year-specific dummy variables to pick up supply-side
rate will be slower the higher is the initial level of real shocks. He found a significant negative relationship for 13
(1) These possibilities are discussed in Fischer (1981b).
(2) See, for example, Mizon and Thomas (1988).
(3) Patinkin (1993) offers a vivid account of the problems faced by Israel under a system of widespread indexation.

37
Bank of England Quarterly Bulletin: February 1995

countries, which implied that a sustained increase in inflation A further problem with these results is that much of the
from 0% to 9% would lead to a full percentage point negative correlation between inflation and output or
reduction in annual growth rates. In contrast, Stanners productivity growth depends on a relatively small number of
(1993) found only a weak (but negative) correlation between observations, in particular in the years immediately
inflation and growth using time-series data for nine following the oil price shocks of 1972–73 and 1979, when
industrialised countries. inflation was relatively high and output and productivity
growth relatively low. If these years were excluded, the
Simple equations regressing growth on inflation cannot, results presented by Rudebusch and Wilcox would become
however, be expected to generate unbiased results. For one less significant. So the conclusions depend heavily on how
thing, in almost all countries there is a positive relationship, the evidence of the oil price shocks is interpreted. As the
at least over the short run, between growth and inflation, box on page 39 discusses, data for the United Kingdom
with the direction of causation running from higher growth reveal a similar difficulty.
(at least in relation to productive potential) to higher
inflation. In addition, single-country time-series The results are also based on a limited range of explanatory
observations that exhibit a negative correlation may be variables (partly because there are a limited number of data
picking up the results of the authorities’ reactions: a period observations). This means that the estimated equations do
of high inflation (or inflationary pressures) is likely to not allow for the influence of many possible determinants of
provoke a tightening of monetary policy, which in turn growth other than the rate of inflation, which may distort the
will cause both inflation and (in the short run) growth to results. One response—adopted by McTaggart (1992) and
decline. Smyth (1994)—is to estimate a production function
including inflation as an argument. Both studies found that
Similar difficulties afflict studies which use time series from inflation had a negative impact on growth, but neither could
single countries to estimate the influence of inflation on identify the channels at work. Again, these regressions may
productivity growth.(1) Some of them—for example have picked up short-run effects, rather than longer-term
Rudebusch and Wilcox (1994)—attempted to allow for the determinants of the growth rate. The results were broadly
short-run trade-off between inflation and growth, and for consistent with those for other time-series studies: for
reaction function considerations. Without any allowance for example, Smyth found that a one percentage point increase
cyclical factors, they estimated a significant negative in inflation reduced the growth rate of private-sector output
relationship between inflation and productivity growth in the by 0.2 percentage points.
United States, with a one percentage point reduction in
inflation associated with an increase in annual productivity Some time-series studies have also assessed the importance
growth of 0.35 percentage points.(2) In addition, they found of inflation variability. McTaggart (1992) found that
that inflation ‘Granger causes’ productivity growth (that is, inflation variability had a positive effect on the growth rate,
productivity growth can be ‘explained’ in a statistical sense but Jansen (1989) found that although inflation had a
by lagged inflation terms, but inflation cannot be ‘explained’ significant negative relationship with output growth,
by lagged productivity growth). But once a cyclically attempts to measure the effect on growth of the conditional
adjusted productivity growth series was used, the estimated variance of inflation yielded insignificant results.
relationship became much weaker. An alternative method of
allowing for cyclical factors—by including the growth of It is also unclear why a change in inflation (or inflation
real output as an additional variable—weakened the original uncertainty) should have as rapid an impact on output or
results far less, so that a statistically-significant negative productivity growth as some of the results suggest. It seems
correlation between inflation and productivity growth more plausible that productivity or output growth should
remained; but Sbordone and Kuttner (1994) have shown that respond favourably to a regime of low inflation (and low
including the US federal funds rate as a further additional uncertainty about future inflation) extending over a much
variable eliminates this correlation. longer period, closer to a decade than a single quarter or
year.
Jarrett and Selody (1982) also attempted to isolate the effects
of policy reactions, in their case by including capacity Cross-country approaches
utilisation as an additional explanatory variable. Their The other main method of estimating the effect of inflation
results, using Canadian data, were very close to those on growth is to use cross-country data. The use of such data
derived by Rudebusch and Wilcox using cyclically was helped by the work of the World Bank and of Summers
unadjusted data: a one percentage point reduction in and Heston (1988), who developed a database on growth
inflation was associated with a 0.3 percentage point increase rates and their possible determinants for 130 countries from
in productivity growth. However, an updating of their study 1950. This work has encouraged other researchers to
by Fortin (1993) found that although inflation had a negative construct consistent series for additional explanatory
impact on productivity growth over a longer sample period, variables with a similar coverage. By averaging the data for
the result was no longer statistically significant. each country in the sample over a number of years, it is

(1) Including Clark (1982), Jarrett and Selody (1982), McTaggart (1992), Rudebusch and Wilcox (1994), and Smyth (1994).
(2) They reported similar results for Canada and the United Kingdom, but much smaller and generally insignificant results for Japan, France, Germany
and Italy.

38
 Costs of inflation

The relationship between UK inflation and productivity growth

Time-series data on inflation and productivity growth in was very high and productivity growth low. (Data for
the United Kingdom and the United States are plotted in these years form a group in the bottom right-hand corner
Charts 1 and 2 below. The charts suggests that in both
countries there has been a negative relationship between Chart 2
the two variables in the post-war period; the line US inflation and productivity growth
Productivity growth; per cent
Chart 1 12
UK inflation and productivity growth
10
Productivity growth; per cent
8
8

6 6

4
4

2
2
+0
_
+
2
_0
4
2 2 _ 0 + 2 4 6 8 10 12 14
Inflation (per cent)

4
0 4 8 12 16 20 24 28 of each chart; these observations might be regarded as
Inflation (per cent) belonging to a different inflation ‘regime’.) In the pre-oil
shock period (before 1972), there was no significant
included in each case is drawn to provide a best fit to the correlation between inflation and productivity growth in
data. In the United Kingdom, a one percentage point rise the United Kingdom—the negative correlation is
in inflation is associated with a reduction in productivity strongest after 1973.
growth of 0.14 of a percentage point. A similar increase
in US inflation is associated with a 0.22 percentage point The effect of adjusting the UK data for cyclical
reduction in productivity growth. influences—by introducing real output as an additional
variable—is to weaken the negative correlation between
The results for both countries are, however, influenced by inflation and productivity growth. The same effect is
the particular conditions during the 1970s, when inflation obtained when adjusting the US data.

possible to avoid many of the problems of short-run different country groups. For OECD countries, they found
trade-offs and policy reactions which arise when using high strong negative correlations between growth and the share of
frequency data. However, statistical tests on the direction of government spending in national income, and between
causation cannot be applied to cross-sectional data. growth and the variability of inflation, but no significant
relation between growth and inflation. Elsewhere, the only
One of the earliest cross-sectional studies was by Kormendi significant relation between inflation and growth was a
and Meguire (1985). Using data for 47 countries over the negative association in the African countries; and inflation
1950–77 period and a wide set of explanatory variables— variability had a significant negative relation with growth in
each averaged over six-year periods—they found that the Asian countries.
inflation had a significant negative correlation with output
growth, apparently because of the negative association De Gregorio (1992, 1993) used cross-sectional data for 12
between inflation and investment. Their results suggested Latin American countries to test the implications of an
that one percentage point higher inflation was associated endogenous growth model in which the level and efficiency
with a half-point reduction in the annual growth rate. of investment are related negatively to the rate of inflation.
He found that both inflation and its variance were negatively
Grier and Tullock (1989) used pooled time series (five-year correlated with growth; the effect appeared to arise mainly
averages) and cross-sectional data between 1951 and 1980 because of a reduction in the efficiency of investment. His
for 113 countries to assess the impact of a range of variables results suggested that a halving of the inflation rate in these
on real output growth. They found that a single empirical countries between 1950 and 1985—from 34% to 17%—
model could not explain differences in growth among these might have increased their annual growth rates by half a
countries and therefore presented different results for percentage point. However, he used only a limited set of

39
Bank of England Quarterly Bulletin: February 1995

explanatory variables. Peng (1993) offered supporting Other recent examples of this approach focus on a variety of
evidence in a study of three Latin American countries, but similar variables. Knight, Loayza and Villanueva (1993)
found no significant relationship in three Pacific Basin considered both the openness of a country’s trade policies
economies. He attributed this difference to the persistence of (since the external trade sector can serve as a vehicle for
high inflation in the Latin American countries. Alexander technology transfer, a channel for promoting efficiency and a
(1994) used a combination of time-series and cross-sectional source of foreign exchange) and the stock of public sector
data for 11 OECD countries, and found a significant negative infrastructure (which again could improve the efficiency
relationship between growth and both the level and the rate with which factors of production are used). King and Levine
of change of inflation, even having allowed for the growth in (1993) concentrated on the role of financial institutions in
capital and labour. A one percentage point reduction in the evaluating prospective entrepreneurs and funding the most
rate of inflation would, according to these results, be promising ones, on the assumption that this would lead to a
associated with a quarter-point increase in real annual more efficient use of capital and increase the probability of
growth. successful innovations. And Easterly (1993) looked at
various types of price distortion, such as subsidies on input
Fischer (1993) reported a study of the impact of inflation on prices and investment goods, ceilings on nominal interest
growth using cross-sectional and panel data for 80 countries. rates and the black market premium on foreign exchange.
He presented tests for the importance of macroeconomic Each study found a number of proxies for the relevant
stability—of which inflation is just one indicator—as a variables to have a significant correlation (with the expected
determinant of growth, and found that inflation was sign) with the rate of growth.
significantly negatively correlated with growth and also
negatively, but less statistically significantly, related to the The apparent importance of a wide range of other factors
rate of capital accumulation and productivity growth. The (even if the results in relation to them are no more robust
results suggested that a one percentage point increase in than those for inflation variables) makes it more difficult to
inflation was associated with a decline in annual output gauge the significance and magnitude of the impact of
growth of 0.04 percentage points. But the effect varied with inflation on growth. This is particularly so given that many
the level of inflation: it was higher at lower rates of inflation of the other factors are likely to be related to inflation—
(in the range of 0%–15% inflation, a one percentage point either causally or because both are influenced by a third
increase in inflation was associated with a reduction in factor. But at the broadest level, the available evidence
annual output growth of 0.125 points). The negative supports the view that well-run and well-governed
relationship was obtained using data both before and after economies with strong and efficient productive structures
1973, when supply shocks became more important. tend to exhibit both low inflation and high growth, though it
is extremely difficult to identify and estimate the separate
Barro and Sala-i-Martin (1994) presented tests based on influence of inflation.
panel data for almost 100 countries, where the variables were
averaged over the periods 1965–75 and 1975–85. They used There are, however, a number of reasons to treat all of these
a number of variables—including schooling, health and life results with some caution. First, some later studies have
expectancy—to capture the initial stock of physical and found that earlier results are sensitive even to fairly small
human capital in each economy, together with a wide variety changes in the sample period, the sample of countries used,
of ‘environmental’ variables, which may be thought of as the definitions of the variables and the specification of the
determining the steady-state level of output per ‘effective’ estimated equation.(1)
worker in a neo-classical growth model. The implication is
that the higher this steady state, the more rapid will be a Second, inflation and growth may be determined
country’s rate of growth from a given starting-point. So far endogenously. One response to this possibility is to use an
as the ‘environmental’ variables were concerned, they found instrumental variable which is sufficiently correlated with
that higher government consumption as a proportion of inflation to generate reasonable estimation results and which
national income had a negative impact on the growth rate, as is exogenous. Cukierman et al (1993) used a measure of
did proxies for the extent of market distortions in an central bank independence as an instrument for inflation, but
economy, such as the size of the black market premium on other research has suggested that central bank independence
foreign exchange and of tariffs imposed on external trade. and inflation may not be well enough correlated to justify
Similarly, a number of variables designed to pick up the this. Barro and Sala-i-Martin (1994) used lagged values of
impact of government-induced actions were statistically their ‘environmental’ variables, but admitted that this may be
significant with the expected sign. These included a imperfect because their starting-point values may be
negative impact of political instability (used as a proxy for determined by past growth performance, and because the
the security of property rights) and a positive impact of prospects for growth in the future and the manner in which
proxies for the rule of law. In subsequent research, not yet an economy is managed can become mutually reinforcing.
published, Barro has found a significant, negative relation
across countries between inflation and growth when a variety Even studies that use cross-sectional data could be
of other influences are held constant. invalidated if growth and inflation were determined by a

(1) See in particular, Levine and Renelt (1992) and Levine and Zervos (1993).

40
 Costs of inflation

third variable. Researchers commonly cite supply shocks as will be permanent, whereas any costs of doing so will be
a candidate for this, in particular the oil price shock in the temporary, the costs could still outweigh the benefits,
early 1970s which lowered growth and raised inflation in depending on the rate of discount to be applied to the
most countries.(1) Some attempts have been made to allow benefits. Theoretically, it might be better to accept a
for this by including terms of trade changes in the estimated permanent modest rate of inflation than to pay the costs of
equations(2) but this may not be an adequate proxy for all reducing inflation to an even lower rate, although Okun
supply shocks. Fischer (1981b) concluded that ‘since the (1971) doubted whether inflation could be held permanently
inflation rate is not an exogenous variable to the economy, at a modest level, describing this as the ‘mirage’ of steady
there is some logical difficulty in discussing the costs of inflation.
inflation per se rather than the costs and benefits of
alternative policy choices’. It is also important to consider the pace of adjustment,
particularly because the costs of disinflation may be related
Such considerations have led a number of commentators to to its speed. In some models of disinflation, the optimal
express scepticism about the value of empirical work on approach is to reduce inflation slowly, because a sharp
inflation and growth (and indeed about tests of the reduction may generate greater uncertainty—which is
determinants of growth generally). For example, Solow costly.(5) However, this result depends on assumptions about
(1994) commented that although various political-economic the information available to economic agents and the manner
factors ‘might easily affect the growth rate if the growth rate in which they form their expectations about inflation. And
were easily affected I do not find this a confidence-inspiring gradualism may not be an appropriate response to very high
project. It seems altogether too vulnerable to bias from inflation, because the costs of a sharp reduction may be
omitted variables, to reverse causation, and above all to the lower than those of continuing high inflation.
recurrent suspicion that the experiences of very different
national economies are not to be explained as if they
The costs of reducing inflation are no easier to calculate than
represented different points on some well defined surface . . .
the benefits of lower inflation. Indeed, the cost of reducing
The introduction of a wide range of explanatory variables
inflation by a percentage point may not remain constant as
has the advantage of offering partial shelter from the bias
the rate of inflation falls towards zero. It may be, for
due to omitted variables. But this protection is paid for. As
example, that the credibility gained through achieving low
the range of explanation broadens, it becomes harder and
inflation in the past (or through central bank independence)
harder to believe in an underlying structural, reversible
reduces the costs of subsequent policies designed to reduce
relation . . .’.
inflation. Sargent (1982, 1983) suggested that since
contracts denominated in nominal terms will be adjusted
However, this scepticism may itself be overdone. Despite a
more promptly and more fully if the authorities announce a
number of shortcomings, the available evidence provides credible policy of disinflation, then such an announcement
support for a negative relationship between inflation and could reduce the rate of inflation with little cost in terms of
growth, consistent with the predictions of the theoretical output or employment. And Chadha, Masson and Meredith
literature. But it would be still more convincing if a (1992) argued—on the basis of both theoretical
structured, micro-level testing of the hypotheses generated considerations and a multicountry model developed by the
by the economic theory of the costs of inflation could be IMF—that the output costs of a disinflationary policy will be
undertaken—and if this confirmed the negative relationship smaller: if the policy is announced in advance and is phased
between inflation and growth. in gradually; the more credible is the policy; and the greater
are the importance of expected future inflation in
Costs of disinflation determining current wage and price setting, and the
There are likely to be costs, in lost output and employment, responsiveness of wages and prices to demand conditions.
attached to reducing the rate of inflation. For example, there
are many different kinds of nominal rigidity—especially in But others have suggested that the cost of reducing inflation
the labour market, but also in debt contracts—which imply is likely to be higher at lower rates of inflation. For
that it takes time for economic agents to adjust their example, as discussed earlier, Lucas (1973) argued that since
behaviour if inflation is at an unexpected rate.(3) Unless inflation variability is likely to be lower at low rates of
economic agents anticipate inflation reductions—and have inflation, movements in prices are more likely to be regarded
time to adjust their contracts accordingly—disinflation will as relative price movements than as changes in the general
lead to lost output and employment, at least over the short price level. This may lead to larger adjustments in output
term. And it has been argued that such costs may persist and employment as individual firms respond to the price
over the medium term—if not necessarily permanently— changes which they observe. Thus the short-run Phillips
particularly in the labour market.(4) Thus even though the curve may be flatter at lower and less variable rates of
efficiency gains from moving to an optimal rate of inflation inflation. Alternatively, Gray (1978) presented a model in

(1) Easterly et al (1993) present results which suggest that shocks, in particular terms of trade shocks, explain statistically as much of the variance in
growth rates over ten-year periods as do country characteristics and policies.
(2) For example, Fischer (1993).
(3) The vast literature on this subject is surveyed by Blanchard (1990), Dixon and Rankin (1994) and Romer (1993).
(4) As discussed and assessed in Blanchard and Summers (1986) and Fortin (1993).
(5) As argued in Balvers and Cosimano (1994).

41
Bank of England Quarterly Bulletin: February 1995

which lower and less variable inflation increased the optimal disinflation were positively related to the degree of central
length of contracts and reduced the optimal degree of bank independence, but there did not appear to be any
indexation. This in turn increased the extent of nominal significant relationship between central bank independence
wage rigidity and thereby increased the impact of nominal and nominal wage rigidity.
shocks on output and employment.
On the optimal speed of disinflation, Ball (1993) found that
In Ball, Mankiw and Romer’s (1988) menu-cost model, the short-run costs of reducing inflation were inversely
lower inflation leads to less frequent price changes for a related to the speed of adjustment: a short, sharp shock was
given level of adjustment costs, so lessening the the best approach to reducing inflation. However, the results
responsiveness of prices to nominal shocks, again leading to of Walsh (1994) could be interpreted as suggesting the
higher short-run output and employment costs. Finally, opposite, at least if more independent central banks tend to
building on these earlier models, Walsh (1994) argued that introduce more rapid disinflations. And Yates and Chapple
greater central bank independence could increase the costs of (1994) found it difficult to establish any clear relationship
disinflation, not only by creating lower and less variable between the speed and the costs of disinflation. An
inflation but also by reducing expectations of nominal important consideration for the authorities is that too blunt
shocks. The degree of central bank independence might an effort to reduce inflation could undermine public support
reduce the cost of lowering inflation through Sargent’s for price stability and therefore prove self-defeating.
‘credibility bonus’, but this might be more than offset by its
impact on nominal rigidities. Finally, the authorities may be able to influence the costs of
reducing inflation, for example by establishing greater
credibility, or by removing the micro-level rigidities which
The empirical evidence suggests that there is indeed a make the process of wage and price formation unresponsive
significant and negative correlation between the average to deflationary pressures. Even if the extent of nominal
level of inflation and the short-term output: inflation rigidities is itself a function of the inflationary regime, it may
trade-off, although it is not possible to identify the cause of also be responsive to supply-side initiatives introduced by
this correlation. The seminal results are those of Ball, the authorities.
Mankiw and Romer (1988), who found a strikingly large
negative correlation.(1) Ball (1993) found that the negative
correlation was lower when wage-setting was more flexible, Conclusions
and Walsh (1994) reported results (albeit based on a small Economic theory suggests that inflation imposes costs on the
sample of eight European countries) which suggested that economy through a variety of channels. And although the
the magnitude of the short-term output:inflation trade-off is empirical evidence cannot be regarded as conclusive, it is
related negatively to the rate of inflation and positively to the broadly consistent with the theoretical results. This implies
degree of central bank independence. Using data on 17 that there are advantages in achieving and maintaining price
OECD countries, Posen (1994) also found that the costs of stability, even if these are difficult to quantify precisely.

(1) Cozier and Wilkinson (1990) found no such evidence for Canada, but Yates and Chapple (1994) found the negative correlation to be remarkably
robust to changes in functional form, sample size and sample periods. However, although the sign and statistical significance of the negative
correlation were robust, the magnitude of the coefficients on average inflation varied considerably across different specifications.

42
 Costs of inflation

References

Alexander, W R J (1994), ‘Inflation and economic growth: evidence from Caplin, A S and Leahy, J (1991), ‘State-dependent pricing and the
a neo-classical growth equation’, University of Otago, Working Paper dynamics of money and output’, Quarterly Journal of Economics, 106,
9,317. pages 683–708.

Backus, D and Kehoe, P (1992), ‘International evidence on the historical Caplin, A S and Spulber, D F (1987), ‘Menu costs and the neutrality of
properties of business cycles’, American Economic Review, 82, money’, Quarterly Journal of Economics, 102, pages 703–25.
pages 864–88.
Chadha, B, Masson, P R and Meredith, G (1992), ‘Models of inflation
Bailey, J M (1956), ‘The welfare cost of inflationary finance’, Journal of and the costs of disinflation’, IMF Staff Papers, 39, pages 395–431.
Political Economy, 64, pages 93–110.
Clare, A D and Thomas, S H (1993), ‘Relative price variability and
Ball, L (1992), ‘Why does high inflation raise inflation uncertainty?’, inflation in an equilibrium price misperceptions model’, Economic Letters,
Journal of Monetary Economics, 29, pages 371–88. 42, pages 51–7.

Ball, L (1993), ‘What determines the sacrifice ratio?’, NBER Working Clark, P K (1982) ‘Inflation and the productivity decline’, American
Paper 4,306, March. Economic Review, 72, pages 149–54.

Ball, L and Cecchetti, S G (1990), ‘Inflation and uncertainty at short and Cooley, T F and Hansen, G D (1991), ‘The welfare costs of moderate
long horizons’, Brookings Papers on Economic Activity, 1, pages 215–54. inflations’, Journal of Money, Credit and Banking, 23 (2), pages 483–503.

Ball, L and Mankiw, N G (1994a), ‘A sticky-price manifesto’, NBER Cozier, B and Wilkinson, G (1990), ‘How large are the costs of
Working Paper 4,677, March. disinflation in Canada?’, Bank of Canada, Working Paper 90-6, November.

Ball, L and Mankiw, N G (1994b), ‘Asymmetric price adjustment and Cukierman, A (1982), ‘Relative price variability, inflation, and the
economic fluctuations’, Economic Journal, 104, pages 247–61. allocative efficiency of the price mechanism’, Journal of Monetary
Economics, 9, pages 131–62.
Ball, L, Mankiw, N G and Romer, D (1988), ‘The new Keynesian
economics and the output-inflation trade-off’, Brookings Papers on Cukierman, A (1984), Inflation, stagflation, relative prices and imperfect
Economic Activity, 1, pages 1–65. information, Cambridge: Cambridge University Press.

Balvers, R J and Cosimano, T F (1994), ‘Inflation variability and Cukierman, A (1992), Central bank strategy, credibility and independence:
gradualist monetary policy’, Review of Economic Studies, 61, pages 721–38. theory and evidence, Cambridge Mass: MIT Press.

Barro, R J (1972), ‘A theory of monopolistic price adjustment’, Review of Cukierman, A, Kalaitzidakis, P, Summers, L and Webb, S (1993),
Economic Studies, 39, pages 17–26. ‘Central bank independence, growth, investment and real rates’,
Carnegie-Rochester Conference Series on Public Policy, 39, pages 95–140.
Barro, R J and Sala-i-Martin, X (1994), Economic growth, New York:
McGraw-Hill. Cukierman, A and Wachtel, P (1979), ‘Differential inflationary
expectations and the variability of the rate of inflation: theory and
Blanchard, O J (1983), ‘Price asynchronisation and price level inertia’, in evidence’, American Economic Review, 69, pages 595–609.
Inflation, debt and indexation, Eds Dornbusch, R and Simonsen, M,
Cambridge Mass: MIT Press. Danziger, L (1984), ‘Stochastic inflation and the optimal policy of price
adjustment’, Economic Enquiry, 22, pages 98–108.
Blanchard, O J (1990), ‘Why does money affect output? A survey’, in
Handbook of monetary economics, Vol II, Eds Friedman, B M and Hahn, Danziger, L (1987), ‘Inflation, fixed cost of price adjustment, and
F H, Amsterdam: North Holland. measurement of relative price variability: theory and evidence’, American
Economic Review, 77, pages 704–13.
Blanchard, O J and Summers, L H (1986), ‘Hysteresis and European
unemployment’, NBER Macroeconomics Annual, 1, pages 15–89. Dixon, H and Rankin, N (1994), ‘Imperfect competition and
macroeconomics: A survey’, Oxford Economic Papers, 46, pages 171–99.
Blinder, A S (1991), ‘Why are prices sticky? Preliminary results from an
interview study’, American Economic Review, 81, pages 89–96. Driffill, J, Mizon, G E and Ulph, A (1990), ‘Costs of inflation’, in
Handbook of monetary economics Vol II, Eds Friedman, B M and Hahn,
Brunner, A D and Hess, G D (1993), ‘Are higher levels of inflation less F H, Amsterdam: North Holland.
predictable? A state-dependent conditional heteroscedasticity approach’,
Journal of Business and Economic Statistics, 11, pages 187–97. Easterly, W (1993), ‘How much do distortions affect growth?’, Journal of
Monetary Economics, 32, pages 187–212.
Caballero, R and Engel, E M R A (1992), ‘Price rigidities, asymmetries,
and output fluctuations’, NBER Working Paper 4,091. Easterly, W, Kremer, M, Pritchett, L and Summers, L H (1993), ‘Good
policy or good luck?’, Journal of Monetary Economics, 32, pages 459–83.
Cagan, P (1956), ‘The monetary dynamics of hyperinflation’, in Studies in
the quantity theory of money, Ed Friedman, M, Chicago: University of Engle, R F (1983), ‘Estimates of the variance of US inflation based on the
Chicago Press. ARCH model’, Journal of Money, Credit and Banking, 15, pages 286–301.

43
Bank of England Quarterly Bulletin: February 1995

Evans, M (1991), ‘Discovering the link between inflation rates and inflation Howitt, P (1990), ‘Zero inflation as a long-term target for monetary policy’,
uncertainty’, Journal of Money, Credit and Banking, 23 (2), pages 169–84. in Zero Inflation: The Goal of Price Stability, Ed Lipsey, R G, Toronto:
C D Howe Institute.
Evans, M and Wachtel, P (1993), ‘Inflation regimes and the sources of
inflation uncertainty’, Journal of Money, Credit and Banking, 25 (3), Ireland, P N (1994), ‘Money and growth: an alternative approach’,
pages 475–511. American Economic Review, 84, March, pages 47–65.

Fischer, S (1981a), ‘Towards an understanding of the costs of inflation, II’, Jaffee, B and Kleiman, E (1977), ‘The welfare implications of uneven
in The costs and consequences of inflation, Eds Brunner, K and Meltzer, A, inflation’, in Inflation Theory and Anti-Inflation Policy, Ed Lundberg, B,
Carnegie-Rochester Conference Series on Public Policy, Vol 15, Bolder, Colorado: West View Press.
Amsterdam: North Holland.
Jansen, D W (1989), ‘Does inflation uncertainty affect output growth?
Fischer, S (1981b), ‘Relative shocks, relative price variability, and Further evidence’, Federal Reserve Bank of St Louis Review, July/August,
inflation’, Brookings Papers on Economic Activity, 2, pages 381–431. pages 43–54.

Fischer, S (1990), ‘Rules versus discretion in monetary policy’, in Jarrett, J P and Selody, J (1982) ‘The productivity-inflation nexus in
Handbook of monetary economics, Vol II, Eds Friedman, B M and Hahn, Canada, 1963–1979’, Review of Economics and Statistics, 64,
F H, Amsterdam: North Holland. pages 361–67.

Fischer, S (1993), ‘The role of macroeconomic factors in growth’, Journal Joyce, M A S (1994), ‘Modelling UK inflation uncertainty: the impact of
of Monetary Economics, 32, pages 485–512. news and the relationship with inflation’, Bank of England mimeo,
September.
Fischer, S (1994), ‘Modern central banking’, in The future of Central
Banking, Capie, F, Goodhart, G, Fischer, S and Schnadt, N, Cambridge: Kimbrough, K P (1986a), ‘The optimum quantity of money rule in the
Cambridge University Press. theory of public finance’, Journal of Monetary Economics, 18,
pages 277–84.
Fischer, S and Modigliani, F (1975), ‘Towards an understanding of the
real effects and costs of inflation’, Weltwirtschaftliches Archive, 114, Kimbrough, K P (1986b), ‘Inflation, employment and welfare in the
pages 810–33. presence of transaction costs’, Journal of Money, Credit and Banking, 18,
pages 127–40.
Fortin, P (1993), ‘The unbearable lightness of zero-inflation optimism’,
University of Quebec, Working Paper 15, June. King, R G and Levine, R (1993), ‘Finance, entrepreneurship and growth’,
Journal of Monetary Economics, 32, pages 513–42.
Foster, E (1978), ‘The variability of inflation’, Review of Economics and
Statistics, 60, pages 364–50. Knight, M, Loayza, N and Villanueva, D (1993), ‘Testing the
neo-classical theory of economic growth: a panel data approach’, IMF Staff
Friedman, M (1969), ‘The optimum quantity of money’, in The optimum Papers, 40, pages 512–41.
quantity of money and other essays, Chicago: Aldine.
Kormendi, R C and Meguire, P G (1985), ‘Macroeconomic determinants
Friedman, M (1977), ‘Nobel lecture: inflation and unemployment’, of growth: cross-country evidence’, Journal of Monetary Economics, 16,
Journal of Political Economy, 85, pages 451–72. pages 141–63.

Gomme, P (1993), ‘Money and growth revisited’, Journal of Monetary Laidler, D (1990), Taking money seriously, London: Philip Allan.
Economics, 32, pages 51–77.
Levine, R and Renelt, D (1992), ‘A sensitivity analysis of cross-country
Goodhart, C A E (1993), ‘The structure of financial systems’, European growth regressions’, American Economic Review, 82, pages 942–63.
Economic Review, 37, pages 269–91.
Levine, R and Zervos, S (1993), ‘Looking at the facts: what we know
Gray, J A (1978), ‘On indexation and contract length’, Journal of Political about policy and growth from cross-country analysis’, World Bank, Working
Economy, 86, pages 1–18. Paper 1,115.

Logue, D and Willet, T (1976), ‘A note on the relationship between the

De Gregorio, J (1992), ‘The effects of inflation on economic growth:
rate and variability of inflation’, Economica, 43, pages 151–58.
lessons from Latin America’, European Economic Review, 36,
pages 417–25.
Lucas, R E (1973), ‘Some international evidence on output-inflation
trade-offs’, American Economic Review, 63, pages 526–34.
De Gregorio, J (1993), ‘Inflation, taxation, and long-run growth’, Journal
of Monetary Economics, 31, pages 271–98.
Lucas, R E (1993), On the welfare cost of inflation, paper prepared for the
March 1993 Hitotsubashi International Symposium on Financial Markets in
Grier, K B and Tullock, G (1989), ‘An empirical analysis of cross-national
the Changing World.
economic growth’, Journal of Monetary Economics, 24, pages 259–76.
McTaggart, D (1992), ‘The cost of inflation in Australia’, in Inflation,
Grimes, A (1991), ‘The effects of inflation on growth: some international Disinflation and Monetary Policy, Reserve Bank of Australia.
evidence’, Weltwirtschaftliches Archive, 127, pages 631–44.
Mizon, G E and Thomas, S H (1988), ‘Relative price variability and
Hahn, F H (1971), ‘Professor Friedman’s views on money’, Economica, inflation in an open economy: a monthly model for the UK 1972–1983’, in
38, pages 61–80. International competitiveness in the UK Economy, Ed McKenzie, G M,
London: Macmillan.
Hahn, F H (1973), ‘On the foundations of monetary theory’, in Essays in
Monetary Economics, Eds Parkin, M and Nobay, A R, New York: Harper Okun, A (1971), ‘The mirage of steady inflation’, Brookings Papers on
and Rowe. Economic Analysis, 2, pages 435–98.

44
 Costs of inflation

Orphanides, A and Solow, R M (1990), ‘Money, inflation and growth’, in Sbordone, A and Kuttner, K (1994), ‘Does inflation reduce
Handbook of Monetary Economics, Vol 1, Eds Friedman, B M and Hahn, productivity?’, Federal Reserve Bank of Chicago Economic Perspectives,
F H, Amsterdam: North Holland. November/December, pages 2–14.

Pagan, A R, Hall, A D and Trivedi, P K (1983), ‘Assessing the variability Sheshinski, E and Weiss, Y (1977), ‘Inflation and the costs of price
of inflation’, Review of Economic Studies, 50, pages 585–96. adjustment’, Review of Economic Studies, 44, pages 287–303.

Parks, R W (1978), ‘Inflation and relative price variability’, Journal of

Sheshinski, E and Weiss, Y (1983), ‘Optimum pricing policy under
Political Economy, 86, pages 79–85.
stochastic inflation’, Review of Economic Studies, 50, pages 513–29.
Patinkin, D (1993), ‘Israel’s stabilisation program of 1985, or some simple
truths of monetary policy’, Journal of Economic Perspectives, 7, spring, Smyth, D J (1994), ‘Inflation and growth’, Journal of Macroeconomics, 16,
pages 103–28. pages 261–70.

Peng, W (1993), ‘The effects of inflation on natural output growth in six Solow, R M (1994), ‘Perspectives on growth theory’, Journal of Economic
Pacific Basin and Latin American developing economies’, International Perspectives, 8, winter, pages 45–54.
Finance Group, University of Birmingham, April.
Stanners, W (1993), ‘Is low inflation an important condition of high
Phelps, E S (1972), Inflation policy and unemployment theory, New York: growth?’, Cambridge Journal of Economics, 17, pages 79–107.
Norton.
Summers, R and Heston, A (1988), ‘A new set of international
Phelps, E S (1973), ‘Inflation in the theory of public finance’, Swedish comparisons of real product and price level estimates for 130 countries,
Journal of Economics, 75, pages 67–82. 1950–85’, Review of Income and Wealth, 34, pages 1–25.

Posen, A (1994), ‘Central bank independence and disinflationary Taylor, J B (1979), ‘Staggered wage setting in a macro model’, American
credibility: a missing link’, Brookings Discussion Papers in International Economic Review, 69, pages 108–13.
Economics, 109, August.
Tobin, J (1972), ‘Inflation and unemployment’, American Economic
Rankin, N (1994), ‘Nominal rigidity and monetary uncertainty’, Centre for Review, 62, pages 1–18.
Economic Policy Research, Discussion Paper 890.
Vining, D R and Elwertowski, T C (1976), ‘The relationshp between
Romer, D (1993), ‘The new Keynesian synthesis’, Journal of Economic
relative prices and the general price level’, American Economic Review, 66,
Perspectives, 7, winter, pages 5–22.
pages 699–708.
Rudebusch, G D and Wilcox, D W (1994), ‘Productivity and inflation:
Walsh, C E (1994), ‘Central bank independence and the costs of
evidence and interpretations’, Federal Reserve Board, Washington, mimeo,
disinflation in the EC’, University of California, mimeo.
May.

Sargent, T (1982), ‘The ends of four big inflations’, in Inflation: Causes Woodford, M (1990), ‘The optimum quantity of money’, in Handbook of
and Effects, Ed Hall, R, Chicago: University of Chicago Press. Monetary Economics, Vol II, Eds Friedman, B M and Hahn, F H,
Amsterdam: North Holland.
Sargent, T (1983), ‘Stopping moderate inflations: the methods of Poincaré
and Thatcher’, in Inflation, Debt and Indexation, Eds Dornbusch, R and Yates, A and Chapple, B (1994), ‘Inflation: how much and how to get
Simonsen, M, Cambridge Mass: MIT Press. there?’, Bank of England mimeo, November.

45