View metadata, citation and similar papers at [Link].

uk brought to you by CORE

provided by Research Papers in Economics

Does Inflation Uncertainty

Increase with Inflation?

By John E. Golob

ne of the most important costs of inflation —which is independent of the level of inflation—

O is the uncertainty it creates about future

inflation. This uncertainty clouds the deci-
sionmaking of consumers and businesses and
are not well understood. Nevertheless, accounting
for the downtrend is important in determining the
true relationship between inflation and inflation
reduces economic well-being. Without this uncer- uncertainty.
tainty, consumers and businesses could better plan This article accounts for the downtrend in
for the future. inflation uncertainty and finds unambiguous evi-
According to many analysts, uncertainty about dence that inflation uncertainty rises with inflation.
future inflation rises as inflation rises. As a result, The first section identifies the consequences of
these analysts argue that the Federal Reserve could uncertainty about inflation and discusses some
reduce inflation uncertainty by reducing inflation. likely causes of the positive relationship between
Other analysts argue that high inflation creates no inflation and inflation uncertainty. The second sec-
more uncertainty than low inflation, as long as tion reviews the results and inconsistencies in
inflation remains stable. As a result, these analysts previous research. The third section presents empiri-
argue that high inflation does not necessarily cal evidence resolving these inconsistencies and
interfere with decisionmaking or reduce eco- pointing to a robust positive relationship between
nomic well-being. inflation and inflation uncertainty.
While most previous studies have found a posi-
tive relationship between inflation and inflation
uncertainty, a few key studies have not. Previous HOW DOES INFLATION UNCERTAINTY
studies may be flawed, however, because they ignore INTERACT WITH THE ECONOMY?
a general downtrend in inflation uncertainty that has
occurred over time. Reasons for the downtrend Whenever expected inflation is a factor in an
economic decision, uncertainty about inflation is
also likely to be a factor. For example, uncertainty
about future inflation can affect both business in-
John E. Golob is an economist at the Federal Reserve Bank
of Kansas City. Carol Manthey, an assistant economist at the
vestment decisions and consumer saving decisions.
bank, and Fabio Silva, a research assistant at the bank, This uncertainty has adverse economic conse-
helped prepare the article. quences that potentially rise with inflation.1
28 FEDERAL RESERVE BANK OF KANSAS CITY

Consequences of inflation uncertainty example, inflation uncertainty can cause employers

and employees to be uncertain about future wages,
Uncertainty about inflation has two types of eco- and landlords and tenants to be uncertain about
nomic effects. First, inflation uncertainty causes future rents. To the extent that taxes are not indexed
businesses and consumers to make economic deci- to inflation, inflation uncertainty also implies uncer-
sions that differ from the ones they would make tain tax rates. For example, capital gains taxes are not
otherwise. Analysts refer to these effects as ex ante, indexed, so inflation uncertainty implies that entre-
because the decisions anticipate future inflation. The preneurs will be uncertain about the tax rates on
second category of effects takes place after the deci- their capital gains. Also, the value of depreciation
sions have been made, or ex post. These effects occur deductions will be uncertain, affecting the way
when inflation differs from what had been expected. profits are calculated and taxed. This spread of
Ex ante effects. Uncertainty about inflation can uncertainty to other economic variables interferes with
affect the economy ex ante through three channels. the ability of consumers and businesses to make
First, inflation uncertainty affects financial markets informed decisions.
by raising long-term interest rates. Second, inflation Uncertainty about interest rates and other eco-
uncertainty leads to uncertainty about other vari- nomic variables can reduce economic activity.
ables that are important in economic decisions. When businesses are uncertain about interest rates,
Finally, inflation uncertainty encourages businesses wages, tax rates, and profits, they may choose to
to spend resources avoiding the associated risks. delay hiring, production, and investment decisions
The first channel through which inflation un- until some of the uncertainty is resolved. Invest-
certainty affects the economy is by increasing long- ment is most vulnerable because investment is so
term interest rates. An important determinant of costly to reverse.
long-term rates is the return required by investors. Uncertainty about interest rates also encour-
If inflation is uncertain, the return on nominal long- ages businesses and consumers to finance invest-
term debt will be riskier. As a result, investors will ment with long-term fixed-rate debt to avoid the
require higher expected returns, which imply higher risk of increases in short-term interest rates. But
long-term interest rates. Higher rates, in turn, imply since fixed long-term rates are typically higher than
that businesses will invest less in plant and equip- short-term rates, using long-term debt increases
ment, and consumers will invest less in housing and financing costs and thereby reduces investment.
other durable goods. The purchase of a home mortgage provides an
Some economists believe inflation uncertainty example of this effect. A consumer who is uncertain
has been an important factor in explaining high about future inflation will be uncertain about future
long-term interest rates in the 1980s and 1990s. interest rates as well. To eliminate the risk of future
Before the high inflation of the 1970s, the spread increases in interest rates, the consumer may choose
between short-term and long-term rates was usually a fixed-rate over a variable-rate mortgage. But this
much lower than in recent years. Concern about choice could lead the consumer to take out a smaller
another episode of high inflation is one possible mortgage than otherwise because interest rates are
reason the term premium remains high today. typically higher in the first years of fixed-rate mort-
The second channel through which inflation gages. So inflation uncertainty could limit the size
uncertainty affects the economy is by causing un- of the mortgage and therefore the size of the home
certainty about interest rates and other economic that the consumer purchases.
variables. When the payments in a contract are not In the third channel through which inflation
indexed to inflation, inflation uncertainty causes the uncertainty affects the economy, businesses spend
real value of future payments to be uncertain. For resources avoiding the risks of future inflation. For
ECONOMIC REVIEW • THIRD QUARTER 1994 29

example, when inflation uncertainty is high businesses S&L industry. If the inflation of the 1970s had been
may spend more resources improving their forecast less of a surprise, the taxpayer bailout of the indus-
of inflation. In addition, some businesses may try to try might have been avoided.
hedge against unexpected inflation using specialized
financial instruments, known as derivatives. But both
forecasting and hedging activities imply that resources Why inflation uncertainty might increase
are diverted from other more productive business with inflation
purposes. And while these strategies reduce the risk
of unexpected inflation, they do not eliminate risk.2 While the costs of inflation uncertainty are
Furthermore, forecasting and hedging are not prac- relatively easy to identify, explaining why inflation
tical for most small businesses and consumers. uncertainty increases with inflation is more diffi-
Ex post effects. The other effects of inflation cult. The most appealing explanation involves the
uncertainty—the ex post effects—occur when infla- response of monetary policy to inflation.3 When
tion differs from what had been expected. Unex- inflation is low, monetary policymakers try to keep
pected inflation leads to a transfer of wealth it low. To the extent they are successful, inflation
whenever the payments in a contract are specified remains low and stable. When inflation is high,
in nominal dollars. When inflation is higher than however, monetary policymakers are more likely to
forecast, the real value of nominal payments is adopt disinflationary policies. These policies, by
lower than expected. A fixed-rate mortgage provides lowering the inflation rate, increase inflation vari-
one example where unexpected inflation implies a ability. Moreover, the policies create inflation
transfer of wealth from the lender to the borrower. uncertainty because the timing and short-run impact
If inflation is unexpectedly high, the real value of of policy on inflation are uncertain.
the mortgage payments to the lender is less than had The timing of disinflationary policy actions is
been expected. Similar effects occur in wage and uncertain, in part, because of short-run tradeoffs
rent contracts. When wages and rents are fixed in among the goals of monetary policy. Although the
nominal dollars, employees and landlords are hurt long-run goal of monetary policy is to make prog-
by an unexpected increase in inflation. ress toward eliminating inflation, the Federal Reserve
Because a wealth transfer implies that someone also tries in the short run to moderate the depth of
wins while someone else loses, it is difficult to economic downturns. When inflation is high at the
measure aggregate ex post effects. But if the unex- same time the economy is in a slump, it is not
pected inflation is large enough, the effect can be obvious which goal should take immediate priority.
felt throughout the economy. The crisis in the Thus, uncertainty arises about the timing of policy
savings and loan industry provides a striking exam- actions to reduce inflation.
ple of an inflation-induced wealth transfer. In this The impact of monetary policy on inflation is
industry, S&Ls used short-term deposits to make also uncertain (Holland 1993b). In particular, the
long-term loans. When inflation rose unexpectedly effects of policy take time to work their way through
in the late 1970s, the real value of the payments on the banking system, to the real economy, and even-
fixed-rate mortgages declined. Meanwhile, as tually to inflation. Moreover, the speed with which
short-term nominal interest rates rose with infla- monetary policy actions are transmitted to inflation
tion, S&Ls were forced to pay higher rates to their varies widely over time. Thus, the complexity of
depositors. By paying higher rates on deposits than predicting how much and how quickly prices will
they were receiving on loans, many S&Ls went respond to monetary policy creates inflation uncer-
bankrupt. Thus, the unexpected inflation of the tainty, even if the stance of monetary policy were
1970s led to a massive transfer of wealth out of the known with certainty.4
30 FEDERAL RESERVE BANK OF KANSAS CITY

PREVIOUS RESEARCH spondents to provide a range of values over which

inflation might fall.6 For example, one respondent
Research on inflation uncertainty goes back might expect inflation of 3 to 4 percent, while
over 20 years. In the first study on the issue, Okun another one might expect inflation of 2 to 5 percent.
found that countries with high inflation also had Because the second respondent identified a wider
more variable inflation. He interpreted the greater range of possible outcomes, this respondent is pre-
variability as an indication of greater uncertainty. sumed to be more uncertain about future inflation.
Since Okun’s initial work, over 20 empirical papers The second approach to estimating inflation
have been published on inflation uncertainty. The uncertainty from surveys is based on the variability,
greatest flurry of activity occurred in the early or dispersion, of inflation expectations across sur-
1980s, after a decade of unusually high inflation in vey participants. Unlike the first approach where
the industrialized countries. uncertainty can be estimated for an individual, the
The vast majority of the research—17 of the 21 variability estimate of uncertainty requires several
papers since Okun’s study—has concluded that high survey participants. When survey participants have
inflation leads to more inflation uncertainty. However, similar expectations of future inflation, uncertainty
four papers have been unable to find this relation- is presumed to be low. But if they disagree about
ship. In addition to these four papers, some research the inflation outlook, uncertainty is presumed to be
on exchange rate regimes is relevant. When infla- high. For example, if 90 percent of participants’
tion uncertainty is examined across exchange rate median inflation forecasts are between 3 and 4
regimes, the evidence suggests uncertainty does not percent, uncertainty is lower than if only 60 percent
rise with inflation. Thus, although substantial evi- of the forecasts are in this range.
dence suggests that inflation leads to more inflation When inflation uncertainty is estimated from
uncertainty, the evidence is not unanimous. surveys, researchers consistently find that uncertainty
In recent research, two different strategies have is high during periods of high inflation. Eight papers
been used to estimate inflation uncertainty.5 The reach this conclusion using data from four surveys.7
first strategy uses surveys and the second uses
forecasting models. Since different estimates of
uncertainty can lead to different empirical results, The forecasting model strategy
this section discusses both research strategies. Re-
searchers using survey estimates of uncertainty The second strategy for estimating inflation
consistently find a positive relationship between uncertainty uses economic forecasting models. In
inflation and inflation uncertainty. Researchers us- this approach researchers use an econometric model
ing forecasting models do not find this relationship of inflation to forecast future inflation. Large fore-
as consistently because different models yield dif- cast errors from the model imply high uncertainty,
ferent results. while small forecast errors imply low uncertainty.
Results from the forecasting model strategy are
less consistent than those from the survey strategy.
The survey strategy While most researchers find large forecast errors
during periods of high inflation, some do not. One
The first strategy for estimating uncertainty reason for this inconsistency is a lack of consensus
about inflation uses surveys of economists and con- about the best way to forecast inflation. Forecasts
sumers. Analysts estimate inflation uncertainty are typically based on previous values of a variety
from the surveys using two different approaches. of economic variables, such as wage inflation,
One approach estimates uncertainty by asking re- money growth, unemployment, import price
ECONOMIC REVIEW • THIRD QUARTER 1994 31

changes, and overall inflation. But there are many extent that the assumptions in these models are
ways of building a model of inflation from these inappropriate, the results from the models are also
variables, and the relative performance of different suspect.
models depends on the time period being consid- Inflation uncertainty across exchange rate
ered. For example, growth in the M1 money supply regimes. If the only evidence against a relationship
has often been used in inflation forecasting models. between inflation and inflation uncertainty came
But changes in the behavior of M1 in the early from restricted-uncertainty models, analysts might
1980s caused the performance of these models to discount this limited contrary evidence. But re-
deteriorate. Since economists have different opin- search on exchange rate regimes also suggests that
ions about how to forecast inflation, they have inflation uncertainty does not increase with infla-
different interpretations of the forecasting model tion. The United States has been in two distinct
evidence on inflation uncertainty. regimes since the end of World War II. Exchange
Two branches of research using forecasting rates were fixed during the Bretton Woods period,
models fail to find the positive relationship be- which ended in 1973. Since the collapse of the
tween inflation and inflation uncertainty. In the Bretton Woods system, U.S. exchange rates have
first branch, researchers use a highly restrictive been allowed to float.
model structure to investigate the link between By examining inflation uncertainty across ex-
inflation and uncertainty. In the second branch, change rate regimes, researchers have indirectly
inflation uncertainty is examined across exchange provided evidence on the link between inflation and
rate regimes. inflation uncertainty. This indirect evidence exists
Restricted-uncertainty models. Of the four because the average level of inflation is higher in
papers in the literature that do not find more uncer- the floating-rate regime. If inflation uncertainty
tainty when inflation is high, three use “restricted- increases with inflation, uncertainty should also be
uncertainty” models.8 These models were originally higher in the floating-rate regime.
developed to analyze financial data, where volatility Inflation uncertainty has been examined across
often changes over time. Since inflation volatility exchange rate regimes by Meltzer (1985, 1986,
also appears to change over time, researchers have 1988) and Meltzer and Robinson.9 The forecasting
adapted these models for analyzing inflation. The models use quarterly data for the GDP deflator and
restricted-uncertainty models typically constrain annual data for the producer price index. In all of
uncertainty to change slowly over time. these papers, inflation uncertainty was about the
Although researchers have found restricted-un- same or slightly higher during the fixed-rate regime
certainty models useful for financial data, the as- as during the post-1973 floating-rate regime. Yet
sumptions may be inappropriate for inflation inflation was only about half as high during the
uncertainty. Specifically, any rapid change in infla- fixed-rate regime as during the floating-rate regime.
tion uncertainty would be inconsistent with the This finding suggests that inflation uncertainty may
constraints typically imposed in these models. For not increase with inflation.
example, after Iraq’s invasion of Kuwait in 1990, In summary, researchers who estimate inflation
oil prices increased rapidly and uncertainty about uncertainty from survey data consistently find that
their impact on overall inflation was high. However, inflation uncertainty rises with inflation. But results
the quick resolution of the resulting war in the Gulf from forecasting models are not as conclusive. Re-
led to a rapid decline in both oil prices and uncer- sults from restricted-uncertainty models are mixed,
tainty about inflation. Such a rapid decline in uncer- and results from exchange rate research suggest no
tainty would be inconsistent with the assumptions relationship at all. To reconcile the disagreement,
in a typical restricted-uncertainty model. To the the next section takes another look at the evidence.
32 FEDERAL RESERVE BANK OF KANSAS CITY

Chart 1
Inflation and Inflation Uncertainty

Percent Percent
12 2.5

10
2.0
8 Inflation
uncertainty
6 (right scale) 1.5

4
1.0

2 Inflation
(left scale)
.5
0

-2 0
1954 ’58 ’62 ’66 ’70 ’74 ’78 ’82 ’86 ’90

Note: Inflation is the average of the Livingston Survey 12-month inflation forecasts. Inflation uncertainty is the dispersion of
the Livingston Survey 12-month inflation forecasts.
Source: Livingston Survey, Federal Reserve Bank of Philadelphia.

EMPIRICAL EVIDENCE episodes of deflation that could complicate the

analysis. To ensure robustness and overcome the
This section presents empirical evidence of an criticisms of the individual measurement techniques,
unambiguously positive relationship between infla- the analysis uses evidence from a survey and from
tion uncertainty and inflation. The analysis also forecasting models.
reveals that, independent of the level of inflation,
inflation uncertainty has been trending down over
time. This downtrend explains the apparent incon- Survey evidence
sistencies in both the restricted-uncertainty models
and the research on exchange rate regimes. Uncertainty in CPI inflation is estimated using
The empirical analysis in this paper is based on the Livingston survey, the only survey conducted
the 1954-93 period. The 1954 starting date is typical continuously since 1954.11 In the Livingston survey,
of research on inflation uncertainty.10 Although ear- approximately 50 economists are surveyed twice a
lier data are available, it is desirable to avoid the year, in June and December.12 Inflation uncertainty
influences of World War II, the Korean War, a price is estimated as the standard deviation of the partici-
control period in the early 1950s, and occasional pants’ inflation expectations. The standard devia-
ECONOMIC REVIEW • THIRD QUARTER 1994 33

Table 1
Regression Results Using the Livingston Survey, 1954-93

Dependent variable Constant Time trend Lagged uncertainty Expected inflation

Inflation uncertainty .443** -.011** .432** .137**

(6-month horizon) (.096) (.003) (.098) (.025)

Inflation uncertainty .479** -.011** .395** .110**

(12-month horizon) (.133) (.004) (.146) (.027)

Note: Inflation uncertainty is estimated as the standard deviation of inflation forecasts from the Livingston Survey.
Standard errors are in parentheses.

** Indicates significance at 0.01.

tion is a measure of the variability, or dispersion, of the time trend implies that uncertainty has been
their inflation expectations. declining over time.16 This decline in inflation
A chart of the 12-month Livingston forecasts uncertainty over time is independent of the relation-
reveals a positive relationship between inflation and ship between inflation uncertainty and the level
inflation uncertainty (Chart 1).13 Both of these vari- of inflation.17
ables were highest in 1980, when inflation was over The main result in Table 1 confirms what other
10 percent and the standard deviation of inflation researchers have found in survey data—that infla-
expectations was over 2 percent.14 tion uncertainty increases with inflation. The down-
In addition to the positive relationship between trend, however, is a feature of the data that was not
inflation and inflation uncertainty, Chart 1 also reveals observed in previous research. Given the results in
a modest downtrend in uncertainty. In particular, the the literature, finding a positive relationship be-
estimates of inflation uncertainty are generally tween inflation and uncertainty is not surprising. A
above estimates of inflation in the early years and more interesting question is whether the downtrend
generally below inflation in the later years.15 in uncertainty can resolve any of the inconsistent
A formal statistical analysis of the relationship evidence from forecasting models.
between the level of CPI inflation and the dispersion
measure of CPI uncertainty confirms the results
suggested by Chart 1. Table 1 shows the results Forecasting model evidence
when uncertainty is regressed on expected inflation,
time, and last period’s uncertainty. For both six- Inflation uncertainty is estimated in forecasting
month and 12-month Livingston forecasts, the models for two versions of the CPI and for the GDP
coefficient on inflation is positive. This result im- deflator. The two versions of the CPI are the total
plies that higher inflation is associated with more CPI and the core CPI. The total CPI reflects the
inflation uncertainty. The negative coefficient on prices paid by a typical urban consumer, whereas
34 FEDERAL RESERVE BANK OF KANSAS CITY

Table 2
Regression Results from Forecasting Models for Consumer Price Indexes
(Models based on 1957-93 quarterly data)

Dependent variable Constant Time trend Lagged inflation

Uncertainty in .092 -.0012* .28**

core CPI inflation (.053) (.0006) (.035)

Uncertainty in .266** -.0002 .127**

total CPI inflation (.057) (.0006) (.033)

Note: Uncertainty is estimated from forecast errors. Standard errors are in parentheses.

* Indicates significance at 0.05.

** Indicates significance at 0.01.

the core CPI excludes food and energy prices. This uncertainty (as measured by the forecast error) in-
latter index may give a more accurate repre- creases by 0.28 percentage points.
sentation of inflation because the total CPI is dis- The forecasting model results for the total CPI
torted by short-term volatility in its food and differ slightly from those for core CPI (Table 2).
energy components. The analysis uses both of While uncertainty still rises as total CPI inflation
these indexes to see if the downtrend in uncertainty rises, the coefficient on the time trend is no longer
can be attributed to a decline in the volatility of food significant. The insignificant coefficient implies
and energy prices.18 that unlike core CPI, uncertainty is not trending
Uncertainty in the CPI. The forecasting model down in total CPI. This result likely occurs because
used to estimate uncertainty in total CPI and core the volatility of food and energy prices is the domi-
CPI inflation is similar to a model evaluated by nant component of uncertainty in total CPI. Food
Stockton and Glassman. The model uses quarterly and energy volatility has not declined over time, and
data on inflation and assumes that next quarter’s this high volatility obscures the declining uncertainty
inflation depends on inflation in each of the past in the core component of total CPI.21
four quarters.19 Stockton and Glassman have shown Uncertainty in the GDP Deflator. Two different
that the performance of this simple model is compa- forecasting models are used to estimate uncertainty
rable to the performance of more complicated models. in the GDP deflator. Ball and Cecchetti devised the
For uncertainty in core CPI inflation, results first model, which forecasts inflation using past
from the forecasting model approach are similar to values of inflation and past forecast errors. Bol-
results from the survey approach (Table 2). Uncer- lerslev devised the second model, which estimates
tainty increases as inflation rises, but uncertainty uncertainty for the GDP deflator using a re-
declines over time.20 More precisely, for a 1 percent stricted-uncertainty model.22 Inflation still de-
increase in inflation, the regression indicates that pends on past inflation, but the model restricts how
ECONOMIC REVIEW • THIRD QUARTER 1994 35

Table 3
Regression Results from Forecasting Models for the GDP Deflator
(Models based on 1954-93 quarterly data)

Dependent variable Constant Time trend Lagged inflation

Ball and Cecchetti Model

Inflation uncertainty .325* -.0062* .094*
(.044) (.0014) (.039)

Bollerslev Restricted-Uncertainty Model

Inflation uncertainty .219** -.0008* .920*
(.038) (.0003) (.415)
Inflation uncertainty .135** .75
(.026) (.42)

Note: For both the Ball-Cecchetti and Bollerslev models inflation uncertainty is estimated from forecast errors, but
the Bollerslev model imposes restrictions on how uncertainty varies over time. Standard errors are in parentheses.

* Indicates significance at 0.05.

**Indicates significance at 0.01.

fast uncertainty can change over time. a positive relationship between inflation and infla-
Results from both GDP forecasting models are tion uncertainty is not found in some previous re-
similar to results from the analysis of core CPI search. The failure of previous researchers to
uncertainty. In the first model in Table 3, estimated recognize the downtrend may have biased their
uncertainty is larger when inflation is higher, but results. While higher inflation in the second half
uncertainty declines over time. Similar results are of the sample tends to raise uncertainty, the down-
obtained with the restricted-uncertainty model.23 trend works in the other direction. Thus, the failure
Table 3 also shows the results of a second regression of previous studies to detect a relationship between
with the restricted-uncertainty model. When time inflation and uncertainty arises because the two
is removed from the regression, the coefficient on effects on uncertainty tend to counteract each other.
past inflation is no longer significant. This result The downtrend also explains the results from
illustrates how excluding time from the analysis research on exchange rate regimes. Recall that al-
can change the interpretation of the results.24 though inflation was almost twice as high in the
floating-rate as in the fixed-rate regime, there was
no corresponding increase in inflation uncertainty.
Reconciling the evidence In this case, the combined effects of inflation and
time neutralized each other, so there was little
The downtrend in uncertainty may explain why change in uncertainty across the two regimes.
36 FEDERAL RESERVE BANK OF KANSAS CITY

Recognizing that inflation uncertainty has evidence is found both in restricted-uncertainty

trended down over time substantially increases the models and in exchange rate research. This article
weight of evidence that uncertainty increases with reaffirms the positive relationship between inflation
inflation. While some studies from the literature and inflation uncertainty, and offers an explanation
seem to be inconsistent with this result, the vast for the inconsistent results in previous research.
majority of these studies were based on either re- The article provides evidence of a downtrend
stricted-uncertainty models or exchange rate re- in inflation uncertainty, and shows how this down-
gime research. When these analyses are reexamined trend can conceal the positive relationship between
in the light of this article’s evidence of a downtrend inflation and inflation uncertainty. Both survey and
in inflation uncertainty, the conclusion that high forecasting model estimates of uncertainty confirm
inflation is associated with high uncertainty is even the downtrend. When inflation uncertainty research
more compelling. is reexamined in the light of this downtrend, the
conclusion that uncertainty increases with inflation
is unambiguous.
SUMMARY The results in this article have a clear implica-
tion for monetary policy. To minimize the disrup-
Most research on inflation uncertainty finds tions to economic decisionmaking caused by
high uncertainty during periods of high inflation. inflation uncertainty, the Federal Reserve should
But this conclusion is not universal, and contrary continue to work toward price stability.

ENDNOTES
1 Several researchers find that inflation uncertainty has 4 Uncertainty about the impact of monetary policy is likely
negative effects on economic activity. Holland (1993a) and to contribute more to inflation uncertainty, at least in the
Golob give summary discussions of this research. short run, than uncertainty about monetary policy itself.
Most evidence suggests monetary policy takes six months
2 A few companies have encountered problems using com- to a year to have an impact on inflation. Consequently, a
plicated hedging strategies with derivatives. In an ironic change in monetary policy today will have only a limited
twist, the strategies have inadvertently led to greater rather impact on forecasts for inflation over the next six months to
than less risk. The strategies are vulnerable to two problems. a year. The near-term outlook for inflation will, however,
First, the strategies are so complicated that even alleged continue to be clouded by uncertainty about the impact of
experts find it difficult to anticipate all possible contingen- past monetary policy actions.
cies. Second, strategies can require frequent trading when
markets are moving, which requires that markets exist for 5 Early researchers assumed that inflation variability was a
each financial instrument used in a strategy. Unfortunately, good measure of uncertainty, but limitations of this approach
when long-term interest rates moved rapidly in early 1994, were quickly recognized (Foster). A basic weakness is that
even dealers were unable to establish prices of some exotic some variations in inflation can be predicted, so variability
derivatives, so these markets essentially shut down. This led does not always represent uncertainty. Because of this weak-
to a failure of the hedging strategies that depended on the ness, the variability approach has not been used much over
closed markets. the last decade.

3 This explanation is similar to a formal economic model 6 Only the Survey of Professional Forecasters measures
developed by Ball. In Ball’s model policymakers have dif- the uncertainty of individual respondents. Respondents are
ferent attitudes toward inflation, some will disinflate while asked to assign specific probabilities to different ranges
others will not. Since the public is uncertain about who will for inflation. For example, a respondent could assign a prob-
control policy in the future, the public is uncertain about ability of one-half to the range from 2 to 3 percent and a
whether high inflation will be reduced. probability of one-half to the range from 3 to 4 percent.
ECONOMIC REVIEW • THIRD QUARTER 1994 37

This survey was first conducted by the American Statistical of inflation avoids the controversy about the time horizon of
Association and the National Bureau of Economic Re- the Livingston survey that is discussed in note 12.
search, and is sometimes referred to as the ASA-NBER
survey. It is currently conducted by the Federal Reserve 14 Chart 1 also reveals that inflation uncertainty is more
Bank of Philadelphia. variable than expected inflation. For example, although ex-
pected inflation was approximately 6 percent in 1976 and
7 Most of this research is based on U.S. data, but Australian 1977, uncertainty declined from 1.7 to 1.1 percent over these
researchers find a similar result using the Morgan Poll. two years.

8 The technical description of these models is “conditional 15 Uncertainty was particularly high in the early part of the
heteroskedasticity,” which is often designated by the acro- sample. But an econometric analysis reveals that the down-
nyms ARCH (autoregressive conditional heteroskedasticity) trend exists even when the 1954-60 period is excluded.
and GARCH (generalized ARCH).
16 The regressions were corrected for serial correlation in
9 Inflation uncertainty in the United States since World War the residuals using a maximum likelihood approach (Hall,
II is only one aspect of the research on uncertainty across TSP User’s Guide). Other specifications for the regression
exchange rate regimes. The research also considers uncer- were also explored. But coefficients were not significant for
tainty in other economic variables, other countries, and time lagged expected inflation or for higher order lags on uncer-
periods back to the 1800s. The general conclusion from this tainty. To allow for a possible unit root in inflation, a regres-
research is that economic uncertainty in the United States sion was also conducted on differenced data. In this
was much higher at the turn of the century when the United regression, inflation uncertainty increased with inflation and
States was on the gold standard. Uncertainty was higher in a negative constant suggested a downtrend over time. But
inflation, nominal GDP, real GDP, and money. Results for the constant coefficient was not statistically significant.
other countries are similar, although not as consistent across
different economic variables as in the United States. 17 While there is evidence of a downtrend in inflation
uncertainty, the reasons for this decline are not immediately
10 Several researchers have noted that inflation was signifi- apparent. Lower uncertainty could reflect structural changes
cantly more volatile before 1954 (Cosimano and Jansen). in the economy or more knowledgeable forecasters. An
Most of the results in this article are robust to changing the example of structural change in the economy would be a
starting and ending dates of the analysis. change in how consumers spend their income. For example,
compared with 40 years ago, more of the typical budget is
11 The survey is named after the late Joseph Livingston, who spent on health services and less is spent on durable goods.
started the survey in 1946 when he was a columnist with the If the price of health services is less variable than the price
Philadelphia Enquirer. The survey is currently conducted of durable goods, the increased weight of health services
and published by the Federal Reserve Bank of Philadelphia. might make overall inflation less variable and more predict-
able. Better forecasting is another reason that uncertainty
The Survey of Professional Forecasters includes estimates may have declined. That is, forecasters may be using ad-
of both the GDP deflator and the CPI, but this survey did not vances in computer and communication technology to de-
begin until 1968. The University of Michigan Survey did not velop better models of the economy. However, with the
begin asking for estimates of CPI inflation until 1966. Pre- limited evidence shown above, it is not yet possible to
viously, participants had only been asked whether prices diagnose definitively the true cause of the downtrend in
were going up or down. inflation uncertainty.

12 When the surveys are collected in June and December, 18 Since data on the core CPI are available only from 1957,
official data are only available through April and October, this analysis starts three years later than the other work in
respectively. This has led Carlson to conclude that the fore- this article.
casts actually cover the subsequent 8-month and 14-month
periods. Many analyses of the Livingston data recognize the 19 In the Stockton and Glassman model, inflation depends
Carlson adjustment. on four lagged values. Standard model selection criteria (see
Wei) indicate that this is a good model. For the core CPI, a
13 Inflation is the average of inflation expectations across model with three lagged values is better.
survey participants. The results of the analysis are the same
when the ex post measured inflation rate is substituted for 20 Past inflation is taken as the average over the last six
the expected rate from the survey. Using the survey estimate months to smooth out the short-term fluctuations in quarterly
38 FEDERAL RESERVE BANK OF KANSAS CITY

data. This approach will be used in regressions of uncertainty using a maximum likelihood technique. This regression is
on past inflation for all of the forecasting models. not as robust as the other results in this article. Results from
other models were robust to changing the starting and ending
21 The regressions in Table 2 provide evidence that uncer- dates, and to eliminating the 1973-81 period when oil price
tainty is higher in total CPI than core CPI. The constant term shocks were important. In contrast, results with the restricted-
is almost three times higher for total CPI than for core CPI. uncertainty model are more sensitive to such adjustments.

22 In Bollerslev’s model, inflation depends on inflation over 24 Researchers have proposed alternative explanations for
the previous four quarters. The conditional heteroskedasticity the inconsistent results in inflation uncertainty research.
(uncertainty) is assumed to follow a GARCH(1,1) process. Brunner and Hess suggest that asymmetry is an important
issue in the failure of GARCH models to find a relationship
23 The dependent variable in Table 3 is the conditional between inflation and inflation uncertainty. Ball and Cec-
variance from the restricted-uncertainty model. Since this chetti suggest that inflation uncertainty should be separated
number is very small in a model that forecasts quarterly into short-term and long-term components and that the level
percentage changes in the GDP deflator, all the coefficients of inflation is more closely related to long-term uncertainty
in Table 3 are shown as 10,000 times the actual values. The than to short-term uncertainty.
regressions were corrected for serial correlation in the errors

REFERENCES
Ball, Laurence. 1992. “Why Does High Inflation Raise Infla- __________. 1993b. “Uncertain Effects of Money and the
tion Uncertainty?” Journal of Monetary Economics, June, Link Between the Inflation Rate and Inflation Uncertainty,”
pp. 371-88. Economic Inquiry, January, pp. 39-51.
__________, and Stephen G. Cecchetti. 1990. “Inflation and Meltzer, Allan H. 1988. “On Monetary Stability and Reform,”
Uncertainty at Short and Long Horizons,” Brookings Pa- in Yoshio Suzuki and Mitsuaki Okabe, eds., Toward a World
pers on Economic Activity, pp. 215-54. of Economic Stability. Tokyo: University of Tokyo Press.
Bollerslev, Tim. 1986. “Generalized Autoregressive Condi- __________. 1986. “Some Evidence on the Comparative
tional Heteroskedasticity,” Journal of Econometrics, May, Uncertainty Experienced under Different Monetary Re-
pp. 307-27. gimes,” in Colin D. Campbell and William R. Dougan, eds.,
Carlson, John A. 1977. “A Study of Price Forecasts,” Annals Alternative Monetary Regimes. Baltimore: Johns Hopkins
of Economic and Social Measurement, Winter, pp. 27-56. University Press.
Cosimano, Thomas F., and Dennis W. Jansen. 1988. “Esti- __________. 1985. “Variability of Prices, Output, and
mates of the Variance of U.S. Inflation Based upon the Money under Fixed and Fluctuating Exchange Rates: An
ARCH model,” Journal of Money, Credit and Banking, Empirical Study of Monetary Regimes in Japan and the
August, pp. 409-21. United States,” Bank of Japan Monetary and Economic
Foster, Edward. 1978. “The Variability of Inflation,” Review Studies, December, pp. 1-46.
of Economics and Statistics, August, pp. 346-50. __________, and Saranna Robinson. 1988. “Stability under the
Golob, John E. 1993. “Inflation, Inflation Uncertainty, and Gold Standard in Practice,” in Michael D. Bordo, ed., Money,
Relative Price Variability: A Survey,” Federal Reserve History, and International Finance: Essays in Honor of
Bank of Kansas City, Working Paper 93-15, November. Anna J. Schwartz. Chicago: University of Chicago Press.
Hall, Bronwyn H. 1992. TSP User’s Guide, Version 4.2. Palo Stockton, David J., and James E. Glassman. “An Evaluation
Alto: TSP International. of the Forecast Performance of Alternative Models of
Holland, A. S. 1993a. “Comment on: Inflation Regimes and Inflation,” Review of Economics and Statistics, pp. 108-17.
the Sources of Inflation Uncertainty,” Journal of Money, Wei, William W. S. 1990. Time Series Analysis. New York:
Credit and Banking, August, pp. 514-20. Addison-Wesley.