511
Eastern Economic Journal, Vol. 33, No. 4, Fall 2007
Jonathan McCarthy: Macroeconomic and Monetary Studies Function, Federal Reserve Bank of New
York, 33 Liberty St., New York, NY 10045. E-mail: jonathan.mccarthy@ny.frb.org.
PASS-THROUGH OF EXCHANGE RATES AND
IMPORT PRICES TO DOMESTIC
INFLATION IN SOME INDUSTRIALIZED
ECONOMIES
Jonathan McCarthy
Federal Reserve Bank of New York
INTRODUCTION
In most industrialized economies, infl ation rates in the 1990s and 2000s were low
compared to those of the 1970s and 1980s. Further underscoring these differences,
infl ation remained low even in countries—in particular, the United States—that ex-
perienced lengthy economic expansions. In fact, the infl ation rate in the US continued
to decline in the 1990s even as the unemployment rate fell below levels generally
associated with rising infl ation during the previous two decades.
Because of low infl ation and because the relationship between infl ation and eco-
nomic activity in many countries during the past decade was contrary to the standard
paradigm, economists have searched for “special factors” to explain this phenomenon.
1
Among the more-cited special factors have been import prices and exchange rates:
many analysts have pointed to a general decline of import prices in industrialized
economies, partly induced by the 1997-98 Asian crisis, to explain declining infl ation
during the late 1990s. More narrowly, commentators have attributed a signifi cant
portion of the decline in infl ation in the US and UK during the late 1990s to the dis-
infl ationary impact of exchange rate appreciation and import price defl ation.
2
For the
US, some analysts also have suggested that the greater openness of the economy has
increased foreign competitive pressures on domestic fi rms, thus restraining domestic
infl ation to a greater extent than in previous episodes of dollar appreciation.
Clearly then, the extent to which exchange rates and import prices infl uence do-
mestic infl ation is of major concern for monetary policy.
3
If the lower infl ation of the
1990s can be attributed largely to such special factors, then a reversal of such factors
could herald higher future infl ation. For example, many analysts were concerned
that as emerging market economies recovered from the 1997-98 crisis, the resulting
higher import prices would lead to greater infl ationary pressures in the industrialized
economies. In fact, the European Central Bank cited the infl ationary effects of a weak
euro as a factor behind its tightening of monetary policy in 2000 and the disinfl ation-
ary effects of a strong euro as a factor behind the loosening in 2003.
4
Beyond the policy implications, economists long have been interested in the infl u-
ence of exchange rate and import price fl uctuations on domestic infl ation. Accordingly,
this subject has spawned many studies through the years. Most have concentrated
512
EASTERN ECONOMIC JOURNAL
on the pass-through of a country’s exchange rate fl uctuations to its import prices, a
literature that has been surveyed comprehensively by Goldberg and Knetter [1997].
5
There also have been a number of studies on the pass-through to domestic producer
and consumer prices; some examples include Woo [1984], Feinberg [1986; 1989], and
Parsley and Popper [1998].
Although much recent work concentrated on pass-through at the fi rm or industry
level, several recent studies have examined macroeconomic pass-through in the US
and other countries. In one case of the latter, Dellmo [1996] fi nds that the effect of
import prices on the CPI in Swedish data is weak, a surprising result given that Swe-
den is a small open economy. For the large, relatively closed US economy, a number
of papers—for example, Gordon [1998], Stock [1998], Koenig [1998], and Rich and
Rissmiller [2000]—fi nd that import prices explain a substantial portion of the forecast
error and improves forecasts during the 1990s.
Two recent papers in particular pertain to this paper. Campa and Goldberg [2005]
estimate exchange rate pass-through to import prices for OECD countries, a broader
set than used in this paper. For countries common to the two papers, their results
are consistent with those in this paper. However, the model in this paper allows for
analysis further along the pricing chain (to producer and consumer prices) than can
be done in their single equation model. Choudhri, Faruqee, and Hakura [2005] use
similar techniques as in this paper to examine pass-through to import, producer, and
consumer prices in the non-US G7 countries, a narrower set of countries compared to
this paper. Moreover, their model has a less extensive real and monetary sector than
the model in this paper (although it is more extensive in other directions). Where there
are common results, most of their results are consistent with those in this paper.
Returning to this paper, I use a VAR model that permits one to track pass-through
from exchange rate fl uctuations to each stage of the distribution chain in a simple
integrated framework. The model has a similar structure to that of Clark [1999], who
studies responses of prices at different production stages to monetary policy shocks.
However, his model does not explicitly include exchange rates and import prices.
6
In addition, I estimate the model for several industrialized economies and examine
whether the factors affecting pass-through that have been identifi ed in the industry-
level studies also explain cross-country differences.
To preview the results, the impulse response functions indicate that exchange
rate shocks have modest effects on domestic infl ation in most of the countries in the
sample, while import price shocks appear to have a larger effect. Pass-through appears
to be larger in countries with a higher import share of domestic demand as well as in
countries with more persistent exchange rates and import prices. Variance decompo-
sitions suggest that the role of exchange rate and import price shocks in explaining
consumer price fl uctuations is relatively modest.
The remainder of this paper proceeds as follows. The next section discusses in-
fl uences on pass-through that have been identifi ed in previous studies and that may
explain cross-country differences. The next two sections describe the model and its
empirical implementation, as well as the data. The following section discusses the
results and then there is a concluding section.
513
PASS-THROUGH TO DOMESTIC INFLATION
INFLUENCES ON PASS-THROUGH
Even within a simple supply-demand model where the law of one price holds, there
can be cross-country variation in the pass-through of exchange rate fl uctuations to
domestic prices. In a large country, the infl ationary effect of a currency depreciation
on domestic prices is counteracted by a decline in the world price (because of lower
world demand), reducing the measured pass-through. For a small country, a currency
depreciation would have no effect on world prices, and thus pass-through would be
complete in the simple model. Therefore, even within the confi nes of this model, pass-
through should be greater in smaller economies.
Still, pass-through appears to vary more—across countries and time as well as
across industries within a country—than can be expected in the simple model. Con-
sequently, many studies have examined fi rms’ adjustment of markups in response to
exchange rate fl uctuations. A theoretical basis for these studies is Dornbusch [1987],
who applied industrial organization models to explain pass-through in terms of
market concentration, import penetration, and the substitutability of imported and
domestic products. Utilizing these principles, Feinberg [1986; 1989] found exchange
rate pass-through to domestic producer prices in the US and Germany to be greater
in industries that were less concentrated and faced greater import penetration. More
generally, Goldberg and Knetter [1997] concluded that the pass-through to import
prices is smaller in more segmented industries where fi rms are able to engage in third
degree price discrimination.
What do these results imply for cross-country differences in pass-through? If a
country’s import share can be assumed to be a good proxy for the import penetra-
tion faced by fi rms, then a country with a larger import share should have greater
pass-through of exchange rate and import price fl uctuations to domestic prices.
7
In
addition, both because of a direct effect as well as through a greater pass-through,
exchange rates and import prices should be more important in explaining domestic
price fl uctuations as the import share increases.
Recent studies investigating the “pricing-to-market” hypothesis of Krugman [1987]
and Marston [1990] suggest additional infl uences on pass-through. Pricing-to-market
behavior occurs when exporters base their foreign currency export prices on competitive
conditions in their foreign markets. As such, exporters allow profi t margins, rather
than foreign currency prices, to fl uctuate in response to exchange rate fl uctuations.
Knetter [1993] fi nds that a fi rm’s industry matters more than its nationality for pric-
ing-to-market behavior. This suggests that cross-country differences in pass-through
may refl ect differences in industrial composition. Also, if fi rms pay less attention to
pricing strategies in smaller markets, pricing-to-market may occur less and pass-
through should be larger in smaller economies (as measured by GDP).
Using the pricing-to-market principles, Mann [1986] discusses some macroeco-
nomic variables that may affect pass-through. One is exchange rate volatility. Greater
exchange rate volatility may make importers more wary of changing prices and more
willing to adjust profi t margins, thus reducing pass-through. Wei and Parsley [1995]
and Engel and Rogers [1998] have provided some empirical evidence confi rming this
514
EASTERN ECONOMIC JOURNAL
hypothesis at the sectoral and product level.
8
Thus pass-through should be less in
countries where the exchange rate has been more volatile.
In a similar vein, if fi rms expect exchange rate or import price shocks to be
persistent, they are more likely to change prices rather than adjust profi t margins
in response to changes in the exchange rate or import prices, which would increase
pass-through.
9
Taylor [2000] developed a model that formalizes this intuition and
provided some evidence from the US supporting his model. Thus pass-through should
be greater in countries where fl uctuations in exchange rates and import prices have
displayed greater persistence.
Another macroeconomic variable discussed by Mann [1986] is aggregate demand
uncertainty. Aggregate demand shifts in conjunction with exchange rate fl uctuations
will alter the profi t margins of importers in an imperfectly competitive environment,
thus reducing measured pass-through. If this hypothesis is true, pass-through should
be less in countries where aggregate demand (which will be proxied by the output
gap) is more volatile.
MODEL AND METHODOLOGY
To examine the pass-through of exchange rate and import price fl uctuations to do-
mestic producer and consumer infl ation across countries, I use a model of pricing along
a distribution chain.
10
Infl ation at each stage—import, producer, and consumer—in
period t is assumed to be comprised of several components. The fi rst component is the
expected infl ation at that stage based on the available information at the end of period
t-1. The second and third are the effects of period t domestic “supply” and “demand”
shocks on infl ation at that stage. The fourth component is the effect of exchange rate
shocks on infl ation at a particular stage. Next are the effects of shocks at the previous
stages of the chain. Finally, there is that stage’s shock.
The shocks at each stage are that portion of a stage’s infl ation that cannot be
explained using information from period t-1 plus contemporaneous information about
domestic supply and demand variables, exchange rates, and infl ation at previous
stages of the distribution cycle. These shocks can be thought of as changes in the
pricing power and markups of fi rms at these stages. Two other features of the model
are worthy of note. First, the model allows import infl ation shocks to affect domestic
consumer infl ation both directly and indirectly through their effects on producer
infl ation. Second, there is no contemporaneous feedback in the model: for example,
consumer infl ation shocks affect infl ation at the import and producer stages only
through their effect on expected infl ation in future periods.
Under these assumptions, the infl ation rates of country i in period t at each of the
three stages—import, producer (PPI), and consumer (CPI)—can be written as:
11
(1)
π π αε αε αε ε
it
m
tit
m
iit
s
iit
d
iit
e
it
m
E=++++
−1123
()
(2)
π π βε βε βε βε ε
it
w
tit
w
iit
s
iit
d
iit
e
iit
m
it
w
E=+++++
−11234
()
515
PASS-THROUGH TO DOMESTIC INFLATION
(3)
π π γε γε γε γε γε ε
it
c
tit
c
iit
s
iit
d
iit
e
iit
m
iit
w
i
E=++++++
−112345
()
tt
c
where
ππ π
it
m
it
w
it
c
,, and
are import price, PPI, and CPI infl ation respectively;
εε ε
it
s
it
d
it
e
,, and
are the supply, demand, and exchange rate shocks respectively;
εε ε
it
m
it
w
it
c
,, and
are the
import price, PPI, and CPI infl ation shocks; and E
t-1
(•) is the expectation of a variable
based on the information set at the end of period t-1. The shocks are assumed to be
serially uncorrelated as well as uncorrelated with one another within a period.
To complete the model, we specify the supply, demand, and exchange rate shocks,
as well as the interaction between monetary policy and other variables in the model.
To identify aggregate demand and supply shocks and exchange rate shocks, we make
the following assumptions. (1) Supply shocks (
ε
it
s
) are identifi ed from the dynamics of
oil price infl ation denominated in the local currency (
π
it
oil
).
12
(2) Demand shocks (
ε
it
d
) are
identifi ed from the dynamics of the output gap (
y
it
) in the country after taking into
account the contemporaneous effect of the supply shock. (3) Exchange rate shocks
(
ε
it
e
) are identifi ed from the dynamics of exchange rate appreciation (
Δe
it
) after tak-
ing into account the contemporaneous effects of the supply and demand shocks.
13
The
equations of this portion of the model then are the following.
(4)
ππε
it
oil
tit
oil
it
s
E=+
−1
()
(5)
yEya
it t it i it
s
it
d
=++
−11
()
εε
(6)
ΔΔeE eb b
it t it i it
s
iit
d
it
e
=+++
−112
()
εεε
Because monetary policy may react to exchange rate fl uctuations and because
policy also eventually affects exchange rates and domestic infl ation, the last portion of
the model consists of a central bank reaction function and a money demand equation
in the spirit of Christiano, Eichenbaum, and Evans [1996].
14
The reaction function
relates short-term interest rates (r
it
) to the previously cited variables in the model as
central banks use the short-term rate as their monetary policy instrument. The money
demand function relates money growth (Δm
it
) to the other variables in the model.
(7)
rErc c c c c c
it t it i it
s
iit
d
iit
e
iit
m
iit
w
i
=++++++
−1123456
()
εεεε εε
iit
c
it
MP
+
ε
(8)
ΔΔmE md d d
ddd
it t it i it
s
iit
d
iit
e
iit
m
iit
w
=++++
++
−1123
456
()
εεε
εε
iiit
c
iit
MP
it
MD
d
εεε
++
7
Finally, I assume that the conditional expectations (E
t-1
(•)) in equations (1)–(8)
can be replaced by linear projections on lags of the eight variables in the system. Mak-
