Repasar los
conceptos básicos del curso de introducción a la economía. Por
ejemplo, en el libro, "Principios de Economía" de Gregory
Mankiw, McGraw Hill, 1998
Para las secciones a, b, y c: capítulo 33, Pág. 663-673
Para las sección d: capítulo 28, Pág. 563-570
- 1. Existe una relación inversa entre
inflación y desempleo (más inflación menos desempleo y viceversa).
- 2. La curva de Phillips no es vertical
- 3. Los bancos centrales pueden reducir permanentemente el desempleo
permitiendo más inflación.
- Curva de Philllips: Es la relación entre la tasa de
inflación y la tasa de desempleo.
- A. W. H. Phillips (1958) mostró que
existía una
relación inversa entre la tasa de crecimiento de
los salarios y la tasa de desempleo en el Reino Unido
en el periodo 1861-1957 (con la excepción del periodo
de alta volatilidad de la inflación entre las dos
guerras mundiales). Con bajo desempleo los salarios
aumentaban rápidamente. Con alto desempleo salarios
aumentaban lentamente
- Explicación de la curva de Phillips.
Cuando la tasa de desempleo es baja, el trabajo es escaso
y los empresarios tiene que ofrecer salarios mayores. Cuando
la tasa de desempleo es alta ocurre lo contrario. La
curva de Phillips representaba representaba la relación
media entre el desempleo y los salarios en el ciclo
económico. Cuando la economía se expande , las empresas
aumentan los salarios por encima de lo "normal" para una
tasa de desempleo determinada; cuando la economía se está
contrayendo las empresas aumentan los salarios más
despacio de lo "normal".
- Evidencia empírica sobre la curva de
Phillips. La estimación empírica de la curva
de Phillips se hace estudiando la relación entre
desempleo e inflación (en lugar de aumento de salarios).
Tras la aportación de Phillips la evidencia
empírica
para los países desarrollados mostraba que existía
una curva de Phillips estable.
Curva de
Phillips para Estados Unidos 1961-69
La relación entre inflación y desempleo no parece
variar a lo largo del ciclo económico.
- La curva de Phillips como "menú" para la política
económica. Si la curva de Phillips es estable parecía
lógico pensar que representaba un menú o disyuntiva para
la política económica. El gobierno podía descender el
desempleo a costa de una mayor inflación.
Tras la aportación de Friedman
y Phelps las conclusiones son:
- 1. Los individuos forman
expectativas sobre la inflación.
- 2. No hay relación inversa a largo plazo entre inflación y
desempleo, es decir, la curva de Phillips a largo plazo es
vertical
- 3. Sí hay relación a corto plazo entre inflación y desempleo.
La curva de Phillips a corto plazo no es vertical
- 4. A largo plazo el desempleo tiende a la tasa natural de paro
Friedman y Phelps de forma independiente
pusieron en duda la base teórica de la curva de
Phillips. Señalaron que trabajadores y empresarios
racionales tendrían en cuenta a la hora de tomar
sus decisiones sólo el salario real (es decir ,
el salarios nominal ajustado por la inflación).
El salario real se ajustaría hasta igualar la oferta y y
la demanda de trabajo a un nivel de desempleo que
denominaron "la tasa natural de desempleo".
¿Por qué no hay relación inversa entre inflación y desempleo? Si
la tasa de desempleo es igual a la tasa natural el salario real es
constante ya que los trabajadores impiden la perdida de poder
adquisitivo demandando un aumento de salarios igual a la tasa de
inflación. Si el gobierno realiza una política monetaria o fiscal
expansiva el aumento de demanda hace aumentar la inflación
por encina de lo previsto. Las empresas están dispuestas a
contratar más trabajo a un menor salario real. Los trabajadores no
se dan cuenta inmediatamente de que el salario real a caído sino
que solo observan un aumento del salario nominal y estarán
dispuestos a ofrecer más trabajo (ilusión monetaria). Con el
transcurso del tiempo los trabajadores anticipan
correctamente la inflación y demandaran aumentos salariales que
mantengan su salario real por lo que el salario real vuelve a su
antiguo nivel y la tasa de paro vuelve a la tasa natural de paro.
El corto y el largo plazo. En el análisis
de Friedman-Phelps la curva de Phillips a corto y largo plazo es
diferente. A largo plazo-el tiempo suficiente para que las
expectativas de los trabajadores sobre la inflación se hayan podio
ajustar- la curva de Phillips es una línea vertical que parte de
la tasa natural de desempleo, mostrando que la política expansiva
sólo puede reducir el desempleo durante un periodo corto de
tiempo. A corto plazo existe una relación inversa entre inflación
y desempleo cuando la inflación permanece aproximadamente
constante.
-En la década de los setenta se
confirmó el punto de vista de Friedman y Phelps.
Una mayor inflación estuvo asociada con un mayor
desempleo. La tasa media de inflación pasó del 2.5% en los
1960s al l7% en los 1970s.
Expectativas de una inflación mayor desplaza la
curva de Phillips hacia arriba
El desplazamiento del la curva de Phillips
en USA. Traducido de j-bradford-delong
En los años
1960s la inflación y el desempleo en los estados Unidos se
movieron a lo largo de una curva de Phillips situada en una
situación muy favorable. Sin embargo a finales de los 1960s,
la curva de Phillips se desplazó a la derecha como
consecuencia de: la pérdida de confianza en el compromiso de
la Reserva Federal para luchar contra la inflación y del
choque del petróleo de 1973 que llevaron a a un gran aumento
de la inflación esperada.
Para 1986 la
mayor parte del aumento de la inflación esperada desde 1960s
se ha invertido, y los niveles de inflación y desempleo se
empezaron amover una vez más a lo largo de una curva de
Phillips favorablemente localizada.
Hoy es totalmente aceptado el punto
central de análisis de
Friedman-Phelps.
Existe una tasa natural de paro (también llamada NAIRU "tasa de paro
que no acelera la inflación") que es compatible con una tasa
constante de inflación.
La respuesta a la
pregunta de si menores tipos de interés pueden reducir el desempleo
es por tanto:
(Explicar esta respuesta con el gráfico)
(Analizar en el gráfico de abajo la política
monetaria de USA en el período 1960-2002)
Crecimiento del PIB y tipo
de descuento (discount window borrowing rate) (área sombreada recesión)
Monetary
Policy (Bank of England)
One of the Bank of England's
core
purposes is 'maintaining the integrity and value of the
currency'. The Bank pursues this core purpose primarily
through the conduct of monetary policy. Above all, this
involves maintaining price stability, as defined by the
inflation target set by the Government, as a precondition for
achieving a wider economic goal of sustainable growth and
employment. High inflation can be damaging to the functioning
of the economy. Low inflation - price stability - can help to
foster sustainable long-term economic growth.
The Bank aims to meet the Government's inflation target -
currently 2.5% - by setting short-term interest rates.
Interest rate decisions are taken by the
Monetary Policy Committee (MPC) of the Bank. Monetary
policy operates by influencing the cost of money. The Bank
sets an interest rate for its own dealings with the market and
that rate then affects the whole pattern of rates set by the
commercial banks for their savers and borrowers. This, in
turn, affects spending and output in the economy, and
eventually costs and prices. Broadly speaking, interest rates
are set at a level to ensure demand in the economy is in line
with the productive capacity of the economy. If interest rates
are set too low, demand may exceed supply and lead to the
emergence of inflationary pressures so that inflation is
accelerating; if they are set too high, output is likely to be
unnecessarily low and inflation is likely to be decelerating.
Core Purposes
In pursuing its goal of maintaining a stable and
efficient monetary and financial framework as its contribution
to a healthy economy, the Bank has three core purposes;
achieving them depends on the work of the Bank as whole. This
part of the website describes and explains each core purpose
and some of the work that is undertaken to achieve them. This
material adds to that provided on the 'About the Bank' main
page. Other parts of the website provide more information
about each of the Bank's activities. Click on the main site
headings or Selected Topics.
Core Purpose 1
- Maintaining the integrity and value of the currency
Core Purpose 2
- Maintaining the stability of the financial system, both
domestic and international
Core Purpose 3
- Seeking to ensure the effectiveness of the UK's financial
services
- Variaciones en precios medios y
relativos
- Incertidumbre sobre el futuro.
- Distorsiones adicionales: coste de suela de zapatos y de menú
- El sistema impositivo. Efecto nocivo sobre el ahorro y el
crecimiento
La tasa óptima de inflación
1. Existe acuerdo unánime al aceptar
que la inflación dedos dígitos tiene un alto coste (citar Barro). No
hay acuerdo en si es deseable una pequeña tasa de inflación (del
orden del 2%) a una inflación cero.
2.
Inflación óptima igual a cero: Los que defienden un objetivo de
inflación cero, alegan dos motivos:
- Motivo 1. Incluso una
inflación baja puede tener un alto coste (Ver Feldstein NBER.
Impacto negativo sobre el ahorro)
-
Motivo 2. La inflación cero produce una mejor asignación de recursos.
2. Inflación óptima (2%-3%).
Los que defienden un objetivo de
una pequeña inflación no nula, alegan tres motivos:
- Motivo 1. Los salarios
nominales son rígidos a la baja (evidencia de Brookings
Institution ) /Discutible (analizar datos)
- Motivo 2. Los
tipos de interés no pueden caer por debajo de cero
- Motivo 3. El
IPC exagera la inflación
Economics focus
How low can you go?
Nov 9th 2000
From The Economist print edition
Central banks disagree on the rate of inflation to aim for
ALL
central bankers agree that price stability should be the
primary goal of monetary policy. But what exactly does “price
stability” mean? This question was debated last week at a
conference organised by the European Central Bank (ECB).
The issue is more controversial than most central bankers
would have you believe. Some economists argue that very low
inflation rates can curb economic growth.
The
ECB defines price stability as a year-on-year
increase in consumer prices of less than 2% over the medium
term. The central banks of Australia, Britain, Canada, New
Zealand and Sweden all have inflation targets with mid-points
between 1.5% and 2.5%. America’s Fed has no explicit target,
but the country’s current inflation rate of 3.5% suggests that
it is happy to tolerate a somewhat higher rate than the
ECB. So is the ECB’s
inflation goal too restrictive?
One
reason that no central bank actually aims for zero inflation
is that official price indices tend to overstate true
inflation, partly because they fail to take full account of
improvements in the quality of products. But there are two
potentially more important reasons why some economists favour
an inflation rate of 3-4%, say, rather than the
ECB’s target of below 2%.
The
first is that wages tend to be sticky downwards. Workers may
be prepared to accept zero wage increases if inflation is 3%,
implying a 3% real pay cut. But they are reluctant to accept a
pay cut in money terms. So if inflation is zero, real wages
cannot easily fall in declining regions or industries, and
unemployment will rise. Inflation, the argument runs, greases
the wheels of the labour market and lowers unemployment.
But
inflation also throws sand into the economic gears by
distorting price signals and causing a misallocation of
resources. When inflation is high, people find it hard to tell
if a rise in the price of a product reflects general inflation
or an increase in demand for that product. High rates of
inflation are therefore bad for growth and jobs. This suggests
that the optimal rate of inflation depends on the balance
between the effects of the grease and the sand.
A few
years ago, a classic study* of
America’s labour market concluded that an inflation rate of
around 3% is best for growth and jobs. But what about Europe,
in which labour markets tend to be more rigid? A paper
presented at the ECB conferenc by
Charles Wyplosz, an economist at the Graduate Institute of
International Studies in Geneva, examines how the structural
rate of unemployment in European economies has varied with the
level of inflation over the past 30 years.
After
adjusting for changes in other factors that affect joblessness,
such as unemployment benefits and job protection laws, Mr
Wyplosz reckons that, in the big EU
economies, the rate of unemployment does tend to be higher at
very low rates of inflation. Indeed, at inflation rates of
0-2%, the ECB’s target, the adverse
impact on unemployment seems to be at its greatest, lifting
the structural unemployment rate by 2-4 percentage points. Mr
Wyplosz concludes that low rates of inflation carry a
significant unemployment cost, and that the optimal rate may
be higher in Europe than it is in America. He suggests that 4%
might be a sensible inflation goal for the euro area.
If this
is correct, the ECB is making a costly
mistake. But other economists at the conference cast doubt on
the reliability of Mr Wyplosz’s results, since his data
included few periods when inflation was actually below 2%. If
inflation remains low for an extended period, workers’
resistance to nominal wage cuts might fade. Moreover, so long
as productivity is rising, firms can cut unit labour costs
even without pay cuts. This suggests that concerns about
nominal wage rigidities can be overdone. Look at America’s
experience: its inflation fell to an average of only 2% in
1997-99, yet unemployment there has continued to fall.
Pushing on a string
The
second concern about low inflation is that monetary policy may
become ineffective. Nominal interest rates cannot fall below
zero. So when inflation is zero, there is no way of achieving
negative real interest rates should they be required to pull
an economy out of deep recession.
But the
need for negative real interest rates can also be exaggerated.
A paper presented by Jose Viñals, an economist at Spain’s
central bank, pointed out that real short-term interest rates
have been negative in America and the euro area only once in
the past 30 years: during the mid 1970s. And that was a
mistake, resulting in an overly lax policy. Mr Viñals has also
carried out simulations using different assumptions about
inflation, the level of real equilibrium interest rates, and
the likely impact of demand and supply shocks. He concludes
that an inflation target of 1-2% avoids most of the problems
likely to arise from the zero floor for interest rates.
Japan’s
recent experience suggests that an economy may sometimes need
negative real interest rates. Yet if the Bank of Japan had had
an inflation target in the 1990s, its economy might be in
better shape now. A symmetric inflation target demands that a
central bank fights deflation as aggressively as inflation.
That would have forced the Bank of Japan to loosen policy
earlier, and it would have helped to persuade people that
prices would not keep falling, thereby reducing the need for
negative interest rates.
In any
case, monetary policy is not impotent at a zero inflation rate.
Even if interest rates cannot go below zero, a monetary
expansion can still boost the economy by pushing down the
currency.
The
awkward truth is that, for all the popularity of inflation
targeting, we do not know what the optimal rate is. Zero is
too low, but an inflation target of 2% may allow enough room
for both wage flexibility and low interest rates. If the
ECB were to draw any lessons from its
conference, it should not be that it needs to raise the
ceiling of its medium-term goal from 2%, but that it needs to
fix a floor at 1%, say, rather than zero.
*
“The Macroeconomics of Low Inflation”, by George Akerlof,
William Dickens and George Perry. Brookings Papers on Economic
Activity 1, 1996.
- Inflación y crecimiento 1955-73
- Inflación y crecimiento 1974-91
- Inflación y tipos de interés reales
- Inflación y desempleo
III. Low Inflation and Real Interest Rates
Bradford DeLong
The end of moderate inflation in the
United States in the early 1980s also saw a substantial
increase in real interest rates--an increase in real interest
rates that has now persisted for a decade and a half.
The real interest rates plotted in the
figure above are naively constructed by subtracting the rate
of inflation over the previous twelve months from the nominal
interest rate. They are thus very imperfect measures of
changes in real interest rates in the short term. To the
extent, however, that investors believe that inflation is
persistent and that changes in inflation are nearly
unforecastable, this figure will nevertheless manage to
provide a reasonable guide to differences in real interest
rates across decades.
The three most striking features of the
figure are (i) the downward trend in real interest rates from
the 1960s into the inflationary 1970s, (ii) the upward jump in
real interest rates to what were (for the United States)
extraordinary levels during the Volcker disinflation, and (iii)
the continued high level of real interest rates since. Real
interest rates today are some one hundred basis points higher
at the short end and at least one hundred fifty basis points
higher at the long end than in the early 1960s.
IV. Low
Inflation, Productivity Growth, and Utility Bradford DeLong
Economists'
faith that low inflation is a goal worth pursuing rests in the
end on a belief that low inflation is a source of higher real
productivity and real material standards of living. Yet this
association appears to be surprisingly hard to document
empirically.
Alesina and
Summers (1993) showed that there is no evidence at all that
independent central banks that pursue low-inflation policies
are sacrificing any other worthwhile macroeconomic objective
in the long run. But that is only half of what needs to be
demonstrated. Rudebusch and Wilcox (1994) found striking
correlations between productivity growth and inflation, but
could not convincingly show causation. After all, if total
nominal demand is predetermined then a strong correlation
between high productivity and low inflation is guaranteed by
the identity that quantity times price equals expenditure.
Here economic
history is of no help. As long as inflation remains moderate,
there is no chance of teasing out of the data any convincing
causal chain at the macroeconomic level running from lower
inflation to faster productivity growth.
Nevertheless,
economists' confidence that it is there remains strong. Rates
of inflation low and stable enough that nobody has to worry
about confusing overall changes in the nominal price level
with real changes in relative prices reduce the magnitude of
the problem economic agents have in interpreting the price
signals they see. With one less thing to worry about, the
organizational time and effort that had gone into forecasting
inflation and interpreting news in an inflationary environment
can be devoted to analyzing other things instead--and at least
some of those other things should raise economic productivity.
In the absence of convincing evidence to the contrary,
economists' priors will remain centered on the belief that low
inflation is a source of stronger economic growth.
But the case
for pursuing and welcoming low enough inflation to be called "price
stability" does not have to rest there. Politicians welcome
low inflation for a reason: they believe that to come out
against low inflation is electoral death. It was Arthur Okun
in the mid-1970s who popularized the "misery index"--the sum
of the annual inflation and unemployment rates. It proved an
effective rhetorical weapon in the presidential campaign of
1976 against Gerald Ford, and in the presidential campaign of
1980 against Jimmy Carter.
Robert Shiller
(1997) has explored the reasons for people's distaste for
inflation, and came up with two broad conclusions. First, that
the public believes that inflation is a sign that all is not
right with economic policy--it is a signal of possible
incompetence on the part of economic policy makers. Second,
that the public finds inflation to be one additional source of
risk in an already risky world--moreover, a source of risk
that they cannot easily assess or understand without learning
much more about macroeconomics than they would wish. Thus a
distaste for inflation appears to be in the utility function.
And if a
distaste for inflation is indeed in the utility function,
economists should recognize that low inflation is an
appropriate policy goal for that reason alone.
What is the optimal rate of inflation? As
inflation has fallen from earlier high levels toward something
approaching price stability, the question of what is the optimal
inflation rate has become more important. Indeed, a growing number
of central banks have adopted an explicit and numerical target for
inflation. Milton Friedman (1969) argued that anticipated inflation
should, on average, be negative. Steady deflation.at a rate equal to
the real rate of interest.is optimal because only at a zero nominal
interest rate is the marginal opportunity cost of holding cash equal
to its marginal production cost (close to zero in practice).
Other considerations suggest that a changing
price level.whether inflation or deflation.creates costs. These
include the distortionary effects
of an unindexed tax system, especially on capital income, and
increased menu costs as prices have to be adjusted more frequently.
As a result, many have argued for the objective of pure price
stability, that is zero measured inflation (for example, Feldstein,
1996). One problem with the objective of zero inflation is that the
official indices used to measure inflation are subject to biases of
several kinds. Most studies suggest that these measures overstate
the true rate of inflation by an amount that could lie in a
range from 0.5 percent to 2 percent per year.
The Boskin Commission (1996) produced a central
estimate of the overstatement of inflation in the U.S. consumer
price index of 1.1 percent per year. Such estimates are not
uncontroversial and there is no reason to presume that the bias
remains constant from year to year. Moreover, there is no unique
price index to measure general inflation in a world in which
relative prices move around. When average inflation is high, the
differences in inflation recorded by different indices are small.
But when overall inflation is low, differences between indices are
more apparent. For example, Johnson, Small, and Tryon (1999) found
sizeable discrepancies between alternative inflation measures in the
U.S. since 1975. The Bank of England discusses a number of measures
of inflation in its quarterly
Inflation Report.
No one measure fully captures all of the information that is
relevant to the setting of monetary policy. Asingle measure and a
single target for inflation are useful in terms of the transparency
of the objectives of policy and the accountability of those
responsible for decisions. But the need to examine different
measures of inflation highlights the difficulty of identifying
precisely an .optimal. rate of inflation. Nevertheless, concern
about the measurement bias problem has led to suggestions that the
optimal measured rate of inflation is positive.
Yet, other economists have argued that an
inflation rate well above zero is desirable because it leads to
higher output and employment. Krugman (1996), for example, proposed
a long-run inflation target of 3 to 4 percent. Two reasons, in
particular, have been advanced for aiming at a positive inflation
rate. The first concerns the significance of downward nominal
rigidities in wages and prices. If nominal wages and prices are
inflexible downward, then a higher rate of inflation might enable a
faster adjustment of real relative wages and prices, which would
improve efficiency. Second, the fact that nominal interest rates
cannot fall below zero may constrain monetary policy in a time of
recession. Both arguments have attracted some support recently, and
I consider them in turn.
Downward nominal rigidities 2(i)a
In a provocative and much-cited paper, Akerlof,
Dickens, and Perry (1996) claimed that .targeting zero inflation
will lead to a large inefficiency in the allocation of resources, as
reflected in a sustainable rate of unemployment that is
unnecessarily high.. They studied how downward nominal wage rigidity
affects the optimal inflation rate. Their contribution was twofold.
First, they reported the empirical evidence on the frequency of
nominal wage cuts in the United States. Second, they argued that the
existence of downward nominal wage rigidity implied that, at low
rates of inflation, there is a permanent trade-off between inflation
and unemployment.a trade-off whose existence many of us expend a
great deal of energy denying. It is not surprising that downward
nominal rigidity in wages and prices means that zero inflation will
be costly for unemployment. But is such rigidity theoretically
plausible? And does theory imply that inflation would be a cure? The
assumptions required to generate downward nominal rigidities, for
which inflation would be a cure, are complex. For example, it is
commonly thought that if wage earners were subject to .money
illusion,. then positive inflation would provide room for periodic
real wage cuts without necessitating cuts in nominal wages or
undesirable increases in unemployment. There is, indeed, some
evidence that supports the existence of money illusion.
For example, Shiller (1996) found that 59 percent
of his respondents stated that they would be happier with higher
money wages though unchanged real wages. Even 10 percent of
economists displayed this kind of money illusion. However, money
illusion is not by itself sufficient to generate downward nominal
rigidities whose effects could be mitigated by inflation (see Yates
[1998]). Money illusion means that people care about nominal wages
in addition to real wages. But it does not explain why people care
more about a fall in nominal wage growth from 0 percent to .3
percent than a change from 3 percent to 0 percent.
Akerlof, Dickens, and Perry argued that the
proportion of salary earners accepting nominal pay cuts could be as
low as 2 to 3 percent. The evidence on the frequency of
nominal wage cuts is not so clear cut if we look at other studies3.
Product markets also exhibit a prevalence of nominal price cuts. For
example, toward the end of 1998, more than 25 percent of the
components of the U.S. CPI were falling. Broadly, the same was true
for the RPI index in the UK. Moreover, it is difficult to believe
that any downward inflexibility of nominal wages would be unaffected
by changes in inflation. As low inflation becomes the norm,
resistance to nominal wage cuts could well disappear. In Japan,
money wages have been falling since the beginning of 1998. And trend
increases in productivity leave scope for changes in relative real
wages, without reductions in the level of nominal wages. For example,
an inflation target of 2 percent per year and productivity growth
also of 2 percent per year, mean that nominal wages would rise at an
average rate of 4 percent per year, leaving scope for reductions in
relative real wages without cuts in nominal wages. It is important
to focus not only on the frequency of price or wage cuts at any one
time, but also on how the distribution of prices and wages evolves
over time. If the world was characterized by downward nominal
rigidities, we would expect to find that the skewness of price
changes increases, with more zero changes, as the inflation rate
falls. Charts 3 and 4 suggest that this does not happen. As
inflation falls, so the proportion of the index that is falling goes
up. The evidence from more formal regression studies is also broadly
unsupportive of the downward nominal rigidity theory (see Yates
[1998]).
Finally, the most casual, but at the same time
the most striking, piece of evidence relates to recent experience.
Akerlof
et al argued that, at
inflation rates below 3 percent, the existence of a permanent trade-off
meant that unemployment would rise. In fact, since their paper was
presented to a Brookings Panel in March 1996, there have been only
four months when the recorded annual inflation rate in the United
States was above 3 percent. Yet, during that period, unemployment
has continued to fall. No doubt, there are many reasons why this
might have happened. But at least one of them is that any downward
nominal rigidity is too small for the Fed to worry about.
A new Akerlof
et al
study is in the pipeline
to be presented in the autumn. Until that is available, I remain
unconvinced that nominal rigidities mean we should abandon the
pursuit of price stability.
Zero bound on nominal interest rates 2(i)b
A second argument for targeting moderate
inflation rather than price stability is that nominal interest rates
cannot fall below zero. Given the existence of this lower bound, the
ability to reduce interest rates in response to large and persistent
negative demand shocks is likely to be constrained if the average
level of interest rates, and hence inflation, is low. This is no
theoretical curiosum. In Japan, official interest rates have now
been below 1 percent since September 1995 and have been virtually
zero since February 1999. And in Europe, where the average inflation
rate is at present close to 1 percent per year, interest rates have
been reduced to 2.5 percent, a level not seen even in Germany for
more than 20 years. The experience of Japan, in particular,
poses a serious challenge to central bankers and economists alike in
how to think about monetary policy in a world of low inflation.
The proposition that the inability to reduce
interest rates below zero might create problems for monetary policy
was emphasized by Keynes (1936) in the 1930s, later by Vickrey
(1955), and, more recently, by Phelps (1972), Summers (1991), and
Fischer (1996). For most of the post-war period, those problems
seemed to belong to the past. But the return to price stability
raises the question of whether such concerns may be more pressing in
future. The significance of the zero lower limit on nominal short-term
interest rates hinges on whether monetary policy becomes impotent at
the point when the constraint begins to bind. In other words, can a
.liquidity trap. render monetary policy ineffective? I return to
this question in Section 3.
The welfare analysis of the optimal inflation
target depends on both (i) the probability that nominal interest
rates will be constrained at zero, and (ii) the cost of that
constraint should it bind. In the rest of this section, I focus on
(i), because if the constraint is unlikely tobind (ii) is redundant.
The cost of the constraint depends critically upon whether monetary
policy is impotent at that point and is discussed in Section 3.
There are few historical episodes that throw
light on the question. In the 1930s,
nominal
short-term interest rates
were close to zero in a number of countries, including the United
States, for a decade or more. And the same was true of Switzerland
in the 1970s. History can, however, shed light on the frequency of
negative
real interest rates
in past cycles. Is it common for real rates to be negative? That is
of interest because the lower limit on nominal interest rates
implies a bound on real interest rates equal to minus the expected
rate of inflation. The lower the expected rate of inflation, the
higher the lower bound on real interest rates. In the limit, if
prices are expected to be stable, then real interest rates too
cannot become negative.
So, how likely is it that negative real interest
rates will be needed? Summers (1991) suggested that .the real
interest rate [in the United States] has been negative in about a
third of the years sinceWorldWar II.. He did not specify the details
of exactly which real interest rate had been negative. Defining the
real rate as the one-year Treasury constant maturity rate less the
actual CPI inflation rate, the ex post
rate was negative for
about 20 percent of the period since 1950. But the relevant concept
is the ex
ante expected short-term
real interest rate. That rate cannot be observed directly. Estimates
using survey-based measures of inflation expectations produce much
lower frequencies of negative real rates than for ex post
rates. Chart 5 shows the ex ante
real rate of interest
in the United States from 1953 to 1998 H2 defined as the one-year
Treasury constant maturity rate less the expected inflation rate
from the Livingston survey. There are only three brief episodes of
negative real rates. These are 1976 H2-1977 H1, 1980 H1, and 1993
H1. So,
ex ante real interest
rates have been negative only rarely in the post-war period. A
similar finding holds for the UK (see Chart 6)4.
Data on the past behavior of real interest rates,
even
ex
ante rates, are not
conclusive. Low inflation, and the associated change of monetary
policy regime, is likely to have altered the cyclical profile of
interest rates. So, theoretical models of monetary policy may throw
further light on the potential importance of the lower
limit on nominal interest rates. There has in the past two or three
years been an explosion of interesting and imaginative
technical research on exactly this question5.
It is helpful to start, however, by considering a back-of-the-envelope
calculation, based on the assumption that the central bank
follows a .Taylor Rule. under which interest rates are raised or
lowered according to whether output is above or below trend
and inflation is above or below its target level. That rule may be
represented by the following equation for nominal short-term
interest rates:
where i is the short-term nominal
interest rate, i*
is the .neutral. nominal interest rate, y and y*
are the logarithms of the levels of actual and trend output
respectively, p is the inflation rate and p*
the target inflation rate. The two parametersl1 andl2 represent how active
monetary policy is in responding to deviations of output from trend
and inflation from its target level. Negative demand shocks mean
that output can temporarily be below trend and inflation fall below
its target. Suppose that the inflation target was 2 percent per year
and the .neutral. real interest rate was 3 percent per year. Then
the .neutral. nominal interest rate would be 5 percent per year.
Imagine a large negative demand shock that led output to fall some 4
percent below its trend level, and inflation to fall from its target
level of 2 percent per year to zero. Suppose that before the shock,
output and inflation were at their desired levels and interest rates
were at their neutral level. The impact of the shock would require a
reduction in interest rates. But by how much? That would depend on
the coefficients in the Taylor Rule. Typical estimated values for
the coefficients l1 and l2 on output and inflation,
respectively, are 0.5 and 1.5. The latter coefficient must exceed
unity in order that the policy response to an inflationary shock is
a rise in real interest rates. In our example, interest rates would
fall by 2 percentage points because of the shortfall of output from
trend, and by 3 percentage points because of the shortfall of
inflation below its target. Hence, interest rates would fall from 5
percent to zero if policy followed the simple rule.
What does this tell us about the likelihood that
interest rates would hit zero? Only shocks that had a large impact
on either output or inflation would create a problem. Such shocks
are not inconceivable but are unusual. The example suggests that
policy would most likely be constrained when demand shocks were
persistent, so that a negative shock to output and inflation
occurred when output and inflation were already below their normal
levels. Suppose that output was 2 percent below trend and inflation
1 percent below target when a negative demand shock occurred. Then
interest rates would already be 2.5 percentage points below their
normal level, and a shock of only 2 percent to output and another 1
percent to inflation would be sufficient to reduce interest rates to
zero. That suggests that, in practice, the constraint is likely to
bind primarily when either shocks are persistent or policy-makers
have failed to react quickly to demand shocks in the first place,
and find themselves with slow growth and inflation below target when
another negative shock occurs. A
pre-emptive
policy
that is symmetric around
the inflation target will help to make less likely the need for extremely low interest
rates.
America's monetary policy
Overflowing March
11th 2004 From The Economist print edition
Enough liquidity: it is time for the Fed to start raising
interest rates
NO ONE expects
the Federal Reserve's open market committee, the policymaking
body of America's central bank, to raise interest rates when
it meets on March 16th. Indeed, a growing number of economists
believe rates will remain at 1%, a 45-year-low, at least until
the end of 2004. Yet the case for an increase is strengthening.
The main reason
why the Fed is expected to hold rates is the feeble state of
America's labour market (see
article). In February, payrolls rose by only 21,000, an
unusually low figure for a briskly growing economy. Not only
would it be politically provocative to raise rates—especially
in an election year—but sluggish jobs growth will hold down
wage pressures and hence inflation. Core consumer-price
inflation (ie, excluding food and energy) is only 1%, a bit
low for comfort, so there appears to be no reason to tighten
policy. That is the view of most American economists. More
important, it is also thought to be the view of Alan Greenspan,
the Fed's chairman.
However, some
Fed officials, along with several foreign central bankers, are
starting to wonder whether policy is being kept too loose for
too long. The Fed was right to cut interest rates immediately
after the stockmarket bubble burst; it saved America's economy
from a deeper recession and the risk of deflation. But rates
cannot stay where they are for ever. Monetary policy is the
loosest for 30 years.
There are
several ways to test the tightness of monetary policy. One is
the level of real interest rates, which are currently around
zero (using the core rate of inflation). This is in fact no
lower than in 1993, when the Fed also faced a jobless recovery.
However, some economists argue that a better gauge is a
comparison of interest rates not with inflation but with
nominal GDP growth. As a
rule of thumb, when rates are above nominal GDP growth,
monetary policy is restrictive; when rates are below it,
policy is expansionary.
One way to
interpret this is to see America's nominal GDP
growth as a proxy for the average return on investment in
America Inc. If this return is higher than the cost of
borrowing, investment and growth will expand. Nominal GDP is
rising at an annual rate of 6%, five percentage points above
the federal funds rate (see chart). This gap is wider than at
any time since the 1970s when inflation took off.
Another useful
measure is Goldman Sachs's financial conditions index. This
takes account not only of real interest rates, but also of
bond yields, the dollar and the stockmarket. In February the
index hit its lowest point in the 35 years it covers, implying
that monetary conditions are extremely loose.
It is popularly
argued that central banks should raise interest rates if
inflation exceeds its target, and cut them if inflation falls
below, but otherwise keep them constant. But this is wrong.
Glenn Stevens, the deputy governor of the Reserve Bank of
Australia, neatly explained why in a recent speech focused on
his own country. It is, he argued, the level of interest rates
that matters most for the stance of monetary policy, not the
direction of change. Monetary policy does not stop working
when rates stay constant. A low interest rate will continue to
stimulate the economy and will eventually produce ever-rising
inflation. Apply Mr Stevens's argument to America: even if
rates were raised by a full percentage point, monetary policy
would still be expansionary.
Once the
circumstances that required super-low American interest rates—ie,
the risk of deflation—have passed, rates need to rise towards
what economists call the “neutral” rate of interest. This is
the rate that neither stimulates the economy nor reins it in,
but allows it to expand in line with its underlying productive
potential, keeping inflation constant. The concept can be
traced back to Knut Wicksell, a Swedish economist who a
century ago defined the neutral (or, as he called it,
“natural”) rate of interest as that earned by fixed capital.
If the rate set by central banks is lower, firms will invest
more, boosting growth and, eventually, inflation.
America's
neutral rate of interest is probably around 5%. While the
economy still has spare capacity, interest rates need to be
below that rate. But they probably no longer need to be quite
so low. It is true that there seems to be little risk of
inflation surging in the near future. Higher prices of oil and
other commodities, along with a weak dollar, may push up the
headline rate of inflation, but wage demands remain moderate.
Better still, inflation may stay low because central banks
have done a good job in the past: the Fed's reputation is now
so strong that households and investors have much lower
expectations of inflation than in the past.
Nevertheless,
all the liquidity being pumped out by the Fed's lax policy has
to flow somewhere. It may not be showing up in conventional
inflation, but it could be encouraging asset-price bubbles and
excess credit growth. In the fourth quarter American house
prices rose at their fastest rate for almost 25 years (see
next article); share prices have been booming again; and
interest-rate spreads look indecently low as investors
desperately seek higher returns, seemingly heedless of the
risk.
In recent weeks,
Mr Greenspan and other Fed officials have started to send out
warnings that interest rates will eventually need to rise, but
with no great sense of urgency. In contrast, Stephen Roach,
the chief economist at Morgan Stanley, has called on the Fed
to raise interest rates immediately by two percentage points.
This, he argues, is needed not only to avoid bubbles but also
to give the Fed ammunition with which to fight the next
economic shock. At current interest rates, the Fed would have
no room to cut rates if the economy took a turn for the worse.
Mr Roach's
favoured interest rate of 3% would, in theory, still leave
policy accommodative. But such a sharp rise could easily cause
the next shock, sending the prices of homes, shares and bond
prices tumbling. The Fed would do better to act gradually. But
it needs to start now.
The Fed faces a
difficult choice. A rise in interest rates when employment is
barely growing would provoke sharp criticism. If it triggered
a slide in share or house prices, recession could even ensue.
On the other hand, the longer the Fed leaves rates so low, the
greater the risk that asset prices will inflate to
unsustainable levels—and the greater the risk of a deeper
downturn in the future. In an election year, however, even a
policymaker as officially independent as Mr Greenspan is
unlikely to inflict short-term pain for long-term gain.
