STEPHEN P.
MAGEE
University
of Chicago and
Brookings Institution
U.S.
Import
Prices
in
the
Currency-
Contract
Period
EVENTS IN
INTERNATIONAL
MONEY MARKETS since 1971 have aroused
considerable
interest in
the
effects of
changes
in foreign exchange rates
on
trade
patterns.
In
any
theoretical
approach,
the
prices
of
traded
goods
are
crucial
to
economic
activity following
devaluation. Because
the
quantities
of
exports
and
imports
may
be
inflexible
for a
time following
a
devaluation,
price changes determine
the
movement in
the trade balance
in
the short
run. "Currency-contract
analysis" deals with the
first round, or impact,
effect
of
devaluation
on
the
prices
of internationally traded goods
that
cross
national boundaries after devaluation but that were contracted
for
before it took
place.'
The
crucial determinant
of
this effect
on the
trade
balance is whether these contracts
are
denominated
in
home currency
or
in
foreign currency.
My
earlier
paper
in this
journal
stressed
that the initial decline
in
the
trade balance
that countries
sometimes
experience
following
devaluation
Note:
I
am
indebted
to
the Rockfeller Foundation
and the
National Science
Founda-
tion for
research
support;
to Rudiger Dornbusch,
George
N. Ecklund, Norman
S.
Fieleke, Otto
Kiehn,
Max Lechter,
Colleen
Ledgerwood,
Stephen
Nyschot, Paul
Won-
nacott,
and
participants
in the Brookings
panel
and in the
Workshop
in
International
Economics
at the
University
of
Rochester for
helpful
comments; and to
Deborah
DuBourdieu,
Jeffrey
J. Schott, and especially
Ellen Hahn
for research assistance.
1. See
Stephen
P.
Magee,
"Currency Contracts, Pass-through,
and
Devaluation,"
Brookings
Papers
on
Economic Activity (1:1973),
pp.
303-23.
117
118
Brookings
Papers
on Economic Activity, 1:1974
(referred
to
as the "J-curve"
in the press)
is
not
a
theoretical
inevitability.
For example,
if U.S.
import
contracts
are denominated
in
dollars
and
U.S.
exports
are
denominated
in
foreign
currency,
the U.S.
trade balance would
increase
rather
than decrease
immediately
following
a
devaluation
of the
dollar,
and
the J-curve would
not appear.
The
reason
is that
the
value
of
U.S.
import
contracts
would remain
constant
in
dollars;
however,
since
devaluation
implies
a
higher
dollar
value
of
foreign
currency,
outstanding
U.S. exports
contracted
in
foreign
currency
would
yield
a
higher
price
in
dollars.
The
J-curve
result-that is,
an
immediate
decline
in the trade
bal-
ance
following
devaluation-will always
ensue
if the
proportion
of con-
tracts
denominated
in
foreign
currency
is
higher
for
imports
than for
exports (given
an initial
trade
deficit).
What
little empirical
evidence
there
is
on
currency
contracts
is
consistent
with
the J-curve.
Grassman
found
that
in
1968,
66
percent
of
Swedish
ex-
ports
were
denominated
in kroners
while
25
percent
were
in the purchasing
country's
currency;
59
percent
of the
import
contracts
were
denominated
in
the
selling
country's
currency
while
26
percent
were
denominated
in
kroners.2
The
symmetry
noted
by
Grassman
is
important:
roughly
two-
thirds
of contracts
were
in the
seller's
currency
and
one-fourth
in
the
purchaser's.
With
the
proportion
of
contracts
denominated in
foreign
cur-
rency
higher
for
imports
than
for
exports,
devaluation
by
Sweden would
be
expected
to
lead
to an
initial
decline
in
that
country's
trade balance.3
Grass-
man
also
investigated
the
bilateral
pattern
of
currency
contracts
between
Sweden
and
a
number
of her
major trading
partners.
He
found
that
94.3
percent
of Swedish
imports
from
the United
States
and Canada
were
in
the
exporting
country's
currency,
while 64.5
percent
of
Swedish
exports
to
the United
States
and
Canada were
in the
importing
country's
currency.
Although
this
pattern
differs
from that of
Sweden's
total
trade,
devaluation
by
Sweden
still
results
in an
initial deterioration
in
Sweden's
bilateral
trade
balance
with
these
two
countries.
Since
the
currency
contracts
of U.S.
trade have
not
been studied
hereto-
fore,
this paper
is a
pilot
study
of
the
currency
of
denomination
and
the
2.
See Sven Grassman,
Exchange
Reserves
and
the Financial
Structure
of
Foreign
Trade
(Lexington
Books,
1973),
and
Sven Grassman,
"A Fundamental
Symmetry
in
International
Payment
Patterns,"
Journal
of
International
Economics,
Vol.
3
(May
1973),
pp.
105-16.
3.
If the
devaluing
country's
trade
balance
is measured
in terms of
foreign
currency,
it declines
as
well,
since
the
capital
gain
on
imports
is
small
relative
to the
capital
loss
on
exports.
Stephen
P.
Magee
119
length of contracts.
Time
and data limited this
study
to
U.S.
imports
from
Japan and West
Germany.
These countries were
selected for
two reasons.
First, they are second
only
to Canada as
trading partners
with the United
States; in 1971, Japan
supplied
16
percent
of
U.S.
merchandise
imports
and
Germany supplied 8 percent.
Second,
as
Figure
1
indicates,
the
parity
changes
of the
yen
and the
deutsche mark
vis-'a-vis
the
dollar
have
been
large
since
1971.4
The
paper
is
organized
in the
following
manner.
The
first section
de-
scribes the
sample
of customs
invoices used in
this
study
for
U.S.
imports
from Japan
and Germany
in
two U.S.
fiscal
years,
1971
and
1973.5 About
two-thirds
of
U.S. imports
from
Japan
are denominated in
dollars
while
four-fifths
of
imports
from
Germany
are
in
deutsche
marks. Thus, U.S.
imports
from
Japan
conform
to the
pattern
established
in
Grassman's
study
of those
from
Sweden,
while U.S.
imports
from
Germany
reverse
it.
The next section
reports
calculations
of
the three
lags
between
orders
and
deliveries
of
Japanese
and
German
goods
to
the
United States: the produc-
tion
lag,
the
transportation
lag,
and the
entry lag.
The
sum
of
these
three
lags determines the
length
of
the
currency-contract
period
for
U.S.
im-
porters.
In the
following
section, frequency
distributions
and
cumulative
distributions of the length
of
the currency-contract
period for U.S. im-
porters and foreign exporters,
considering
both currencies, are used to
determine the effects
of devaluation on
U.S. import
prices and on foreign
export prices attributable
in
1971-73 to
the
currency denomination
of
contracts outstanding
at
the time
of
devaluation.
I
then
examine three errors in
the
1971-73 statistics on the
U.S.
balance
of
payments
for the
currency-contract
period
that are caused by the regu-
lations followed by
the U.S.
Bureau of
Customs
in measuring imports.
The
first
is
the
tendency
to
ignore
the
currency
of
the
import contract
and
4.
When
changes
in
exchange
rates are not
large,
it
is
difficult
to detect
any pass-
through of the price effects after
the
currency-contract period. See, for
example, Robert
M.
Dunn, Jr.,
"Flexible
Exchange
Rates and
Oligopoly Pricing:
A
Study
of
Canadian
Markets,"
Journlal
of Political
Economy, Vol.
78
(January/February 1970),
pp. 140-51.
Dunn's study covered the period
of the flexible Canadian exchange rate
and focused on
six
products in which
competition was
less
than perfect. The evidence is
not uniform,
however.
A
study by John
R.
Dominguez, Devaliuation
and
Futures Markets
(Lexington
Books, 1972), found that
anticipation of the 1967 devaluation of British
sterling was re-
flected
in
cocoa futures
in New York
and London.
In
this case, pass-through
of the rela-
tive
price
effects
began
before the
devaluation.
5. Hereafter, these years will
be referred to without the designation
"fiscal year."
When
calendar
years
are
meant,
the text will
so state.
*
t
f
*
gb
Ps~~~~~~~~~0
PoII
gl~~~~~~~~~~~~
3z
00~~~~
0.
.P*~~~~~~~~'
W;4
~ ~ ~ ~ ~ ~ ~ ~ ~ ~
'
~ *
Stephen P. Magee 121
to mark up all imports by the amount of the devaluation. The second arises
from the stickiness in the statutory rate of exchange used to value imports.
The third is the valuation of imports at the time of foreign export rather
than of arrival in the
United
States. In
a
period
in
which the dollar
is
depreciating, the first error causes an overstatement of U.S. imports, while
the second and
the third cause an
understatement.
A
final
section summarizes
the
paper.
The Currency Denomination
of
Contracts
THE DATA
The
data
used in the
study below
are taken from
a
sample
of
a
1
percent
sample of U.S.
customs
invoices
for
imports
from
West
Germany
and
Japan in U.S. fiscal years 1971 and 1973.
The
1
percent sample is catalogued
by
the fiscal
year
in which the documents are
liquidated.6
Because of
the
long
lag in liquidations
for
some items,
a
given fiscal year may
contain invoices
covering exports from the foreign country
to
the United States
that took
place one
to two
years earlier.
Because
of
the time and expense required
to
tabulate the information on
currency contracts and the various lags involved in the currency-contract
period,
the
sample
used in
this
study was,
of
necessity,
small.
For
U.S.
im-
ports
from
Japan, 176 and
173
invoices
were drawn
from
the
1
percent
sample
for 1971 and
1973,
respectively;
for
Germany,
76 and 139 invoices
were drawn.
There were thus 349
observations
for
Japan
and
215
observa-
tions
for
Germany, or
an
overall sample size from both countries
of 564
observations. Because
of the
confidentiality
of some of
the data
on the in-
voices,
I
was
provided
tabulations
only
of
the denomination
of the
con-
tracts, the lags involved,
and
so on,
but
no information
that
would permit
6. The U.S.
Bureau of Customs collects
import statistics in two steps. First, when the
importer obtains
possession of the imported
goods, he must file
all of the documents
re-
lated to the transaction
and pay estimated
duties plus a bond
on the goods being
im-
ported.
The date
of
these
actions
is the "entry date."
All
customs invoices
are then
checked by the bureau
for the accuracy
of the information, the
correctness of the amount
of duties paid, and
so on. The completion
of this checking
marks the "liquidation
date." The time
required for liquidation
ranges from about
one month to
several
years, depending
on whether
there is
a
change
in
valuation,
classification, and other
factors affecting
duty
liabilities at the
time
of entry. (The
mean and
median liquida-
tion
times
were
150
days and
142
days,
respectively, for U.S.
imports from Japan
in
the
1973 sample.) The
goods are reported in
the U.S. import statistics
at the time of entry.
They are revised
only
if
large changes are
made at the time of
liquidation.