Does real exchange rate volatility affect sectoral trade flows?
Caglayan, Mustafa ; Di, Jing
1. Introduction
A review of the empirical and the theoretical literature that spans
the period after the breakdown of the Bretton Woods agreement reveals
that there is no consensus on the impact of exchange rate volatility on
trade flows. Several theoretical studies arrive at the conclusion that
exchange rate volatility can have a negative impact on trade flows. (1)
Equally, several others conclude that the effect is uncertain or
positive. (2) Interestingly, one cannot reach a firm conclusion from
empirical studies, either. Results are conflicting and sensitive to
various factors. (3)
When we focus on the recent empirical literature, we come across
several possible reasons why researchers have reached conflicting
conclusions. Early empirical research, which concentrated on aggregate
U.S. or (37 data, suggests that exchange rate uncertainty may have a
positive or negative effect on trade flows. (4) Recent research that
focuses on bilateral rather than aggregate trade data of advanced
countries concludes that exchange rate volatility has no or little
effect on trade flows. (5) In this study, we utilize a broader data set
that contains both advanced and emerging top trade partners of the
United States. Hence, we avoid the narrow focus on the United States and
the advanced country data that have characterized much of the
literature.
We should point out that the inclusion of advanced and emerging
countries in our investigation is important, as recent research suggests
that exchange rate volatility has a significant negative impact on trade
flows of emerging countries. For instance, Grier and Smallwood (2007)
conclude that while real exchange rate volatility has a significant
negative impact on international trade for emerging countries, there is
no such effect for the advanced economies. Several other researchers
also report similar findings for different sets of emerging countries on
the linkages between exchange rate volatility and trade flows. (6)
Although one can claim that the presence of a significant relationship
may be due to the lack of proper financial tools in emerging countries
that firms can use to hedge against exchange rate fluctuations, Wei
(1999) cannot find empirical evidence to that end. In this article, we
utilize data from nine advanced and five emerging countries.
Although the use of country-specific bilateral trade data is an
improvement over aggregate trade data, sectoral trade data can help us
further disentangle the linkages between exchange rate volatility and
trade flows that may exist across sectors but not in bilateral data.
However, there are only a handful of articles that use sectoral data to
investigate the impact of exchange rate uncertainty on sectoral trade
flows. Also, the early literature that used sectoral data summarizes the
impact of exchange rate volatility on sectoral trade flows in one
coefficient as researchers implement panel data methodologies. In
contrast, we focus on country sector-specific bilateral relationships
and investigate dozens of models. (7) Our data are organized with
respect to bilateral sectoral trade flows between the United States and
its top 13 trading countries. Our 14-country data set includes the
United States, Japan, Germany, the United Kingdom, France, Italy, the
Netherlands, Ireland, Canada, South Korea, Singapore, Malaysia, China,
and Brazil and covers the period between 1996 and 2007 on a monthly
basis.
Another important factor that may affect the results in this
literature is the method that one uses to generate a proxy for real
exchange rate volatility. (8) Generally, the early research has used a
moving average standard deviation of the past monthly exchange rates or
variants of ARCH methodology to generate a proxy for exchange rate
volatility. We utilize daily spot exchange rates to proxy for exchange
rate volatility employing a method proposed by Merton (1980). This
method, also used by researchers including Baum et al. (2004) and
Klaassen (2004) in similar contexts, exploits daily exchange rate
movements to proxy for monthly exchange rate volatility. Furthermore,
both studies indicate that this approach yields a more representative
measure of volatility avoiding problems associated with proxies derived
from ARCH methodology or moving standard deviations. In particular, the
Merton (1980) methodology avoids potential problems, including high
persistence of shocks when moving average representations are used or
low correlation in volatility when ARCH/GARCH models are applied.
Last but not least, our empirical model takes the form of a simple
distributed lag model where we allow each variable to affect trade flows
up to six lags, which is shown to be adequate to capture the explanatory
variables' impact. We keep those models that yield a stable dynamic
relationship and discard the remaining models, which are dynamically
unstable. In total, we scrutinize over 200 models where we discuss the
impact of volatility measures across sectors and countries. To address
an interesting suggestion raised by Baum et al. (2004), we also allow
for income volatility and an interaction term between income and
exchange rate volatilities in our model. They suggest that higher
volatility of foreign income may signal greater profit opportunities
inducing entry into the market or delaying exit from the market. Also,
the interaction term between foreign income and exchange rate
volatilities may help capture indirect effects emanating from any of
these variables, which may capture the impact of the expansion or
retention of trade flows as foreign income and the exchange rate
fluctuate while addressing the presence of nonlinearities in the model.
Our results provide evidence that exchange rate uncertainty has
little effect on sectoral trade flows. We find that the impact of real
exchange rate volatility on trade flows is significant in about only 6%
of the models at the 5% significance level, where the effect is
positive. Furthermore, although this relationship is slightly stronger
for the emerging countries, our findings do not support earlier findings
that exchange rate volatility plays an important role for emerging
country trade flows. Overall, our results show that there is little
effect of exchange rate volatility on sectoral trade flows of advanced
and emerging economies.
When we investigate the effects of income volatility and the
interaction term between exchange rate volatility and income volatility
on trade flows, we come across some interesting observations. It turns
out that the interaction term is significant in almost all cases when
exchange rate volatility plays a significant role in the model.
Furthermore, it takes the opposite sign to that of exchange rate
volatility, reversing the impact of exchange rate volatility on trade
flows. From this perspective, it is apparent that omitting the
interaction term from the analysis would lead to wrong policy
prescriptions. When we observe the role of income uncertainty, we see
that this variable significantly affects trade flows in 6% of the models
at the 5% level, while its sign is generally the same with that of
exchange rate volatility. This variable seems to play a more important
role when we concentrate on exports of the United States to its trading
partners. This is not surprising, as the income of the trading partners
over the period under investigation was much more volatile than that of
the United States.
We finally check for the robustness of our findings by implementing
a semirestricted model to test those effects arising from exchange rate
and income volatilities and their interaction. Our investigation
provides support for our earlier conclusion that exchange rate
uncertainty has negligible impact on trade flows.
The reminder of this article is organized as follows. Section 2
outlines the model, discusses our volatility measures, and provides
information on the data. Section 3 reports the empirical results, and
section 4 concludes.
2. Model Specification
Most of the early research that concentrated on the impact of
exchange rate volatility on trade flows used country-level aggregate or
bilateral trade flow data. However, as Bini-Smaghi (1991) indicates,
because sectoral data do not constrain income and price elasticities
across sectors, one should employ sector-specific data when exploring
the linkages between trade flows and exchange rate movements. Yet there
are only a handful of studies that utilize sectoral data. (9) These
studies follow an Armington (1969) approach and estimate both price and
output elasticities. In particular, to capture export flows from country
i to j, the model takes the form
[X.sub.ijt], = f([P.sub.ijt], [Y.sub.jt], [[sigma].sub.ijt]), (1)
where [Y.sub.jt], [X.sub.ijt], [P.sub.ijt], and [[sigma].sub.ijt]
denote income of country j and exports, relative price, and exchange
rate volatility from country i to country j, respectively. The price and
output elasticities (coefficients associated with relative prices and
output) are estimated in a panel context using sectoral trade flow data
for each sector. Naturally, this approach yields a single
sector-specific price and output elasticity along with the impact of
exchange rate volatility, which is then compared across sectors.
Our approach differs from the above specification, as we model the
impact of exchange rate volatility for each sector- and country-specific
trade flow separately. Given that we have 14 countries where data are
ordered with respect to i) sectoral exports of 13 countries to the
United States and ii) sectoral exports of the United States to the same
set of countries, the maximum number of models that we can estimate is
260. However, because of a lack of data on exports from Ireland to the
United States for sectors 4 and 5, we estimate 258 models. Of these 258
cases, we discard 28 models, as they fail the dynamic stability
conditions, rendering us with 230 models to scrutinize. Our model takes
the form
[X.sup.i[right arrow]j.sub.k,t] = f ([Y.sub.j,t], [s.sub.t],
[[sigma].sub.s,t-n], [[sigma].sub.Y,t-n], [[sigma].sup.s,t-n] x
[[sigma].sub.Y,t-n]), (2)
where i [right arrow] j implies exports from country i to country
j, k stands for the sector, and t denotes the time. We introduce the
real exchange rate, s, and real exchange rate volatility and income
volatility ([[sigma].sub.s] and [[sigma].sub.Y], respectively) in our
model. The joint impact of the two volatilities as suggested by Baum et
al. (2004) is captured by [[sigma].sub.s] x [[sigma].sub.Y]. In our
investigation, we are interested in the sign and the significance of the
coefficients associated with exchange rate and income volatilities as
well as that of the interaction term between income and exchange rate
volatilities, [[sigma].sub.s] x [[sigma].sub.Y], and we report and
compare the effects of these variables across sectors and countries. All
variables are allowed to have up to n lags, which is set to six in our
empirical investigation.
Prior to providing information on our data and the empirical model
that we use, in the next subsection we explain how we generate a proxy
for exchange rate and income volatilities. We first provide details of
the Merton (1980) methodology that we implement to derive a proxy for
exchange rate volatility. We then discuss the approach that we use to
generate income volatility. The interaction term in the model is the
product of the two volatilities.
Generating Exchange Rate Volatility
To generate a proxy of exchange rate volatility, one can pursue
different methodologies. One of the most commonly employed methods to
proxy for exchange rate volatility is the moving standard deviation of
exchange rate changes. As this methodology includes the past 12 or 24
months of data, the proxy may contain substantial correlation.
Alternatively, it is possible to use ARCH/GARCH models to generate such
a proxy. This approach may find weak persistence of shocks, and the
generated proxy will be very much model dependent. In this study, we
adopt a measure of risk proposed by Merton (1980). (10) This measure
considers the daily changes in the exchange rates between each pair of
countries in our data set to calculate monthly exchange rate volatility.
Given that traders export their products to several countries, the
exchange rate volatility perceived by an exporter in a sector will
differ across the countries that it trades with by design.
To implement Merton's methodology, we calculate the daily real
exchange rate series ([s.sup.d.sub.t]) for the countries in our data
set. Hence, we first compute daily prices by interpolating the relative
prices for all countries within the month while taking into account the
intervening business days. Then we generate the daily real exchange rate
series by multiplying the daily spot exchange rate series with the
exporting country to domestic country price ratio. Finally, we calculate
the squared first difference of the log real exchange rate series and
deflate it by the number of elapsed days between observations,
[[??].sup.d.sub.t] (100 [DELTA][s.sup.d.sub.t]/[square root of
[DELTA][phi]]t), (3)
where the denominator ([DELTA][[phi].sub.t]) captures the calendar
time difference between each successive observation on the s process.
For our case, [DELTA][[phi].sub.t] [member of] [1, 5] because of
weekends and holidays. The value we compute in Equation 3 is the daily
volatility faced by the exporter. We then define the monthly volatility
as [[PHI].sub.t][[S.sub.t]] = [square root of
[[summation].sup.T.sub.t=1] [[??].sup.d.sub.t]], where the time index
for exchange rate volatility is at the monthly frequency.
The price series for each country are taken from the Main Economic
Indicators published by the Organization for Economic Cooperation and
Development (OECD), and the exchange rate series are downloaded from the
Pacific Exchange Rate Service, which is provided by the University of
British Columbia's Sauder School of Business.
Generating Income Volatility
Our empirical investigation requires a proxy for real income
volatility for the importing countries on a monthly basis. Given that we
will be exploring the behavior of sectoral trade flows, we believe that
it would be preferable to use monthly industrial production series. Our
choice is appropriate, as most of the trade between countries is
intrasectoral. We should note that some researchers interpolate gross
domestic product to monthly frequency when they use aggregate data.
However, this process may add significant noise into the process, in
particular, for the case of emerging countries.
To generate a measure of monthly income volatility, [sigma]y , we
first test whether the first difference of real income series exhibit
time-varying heteroscedasticity. Observing that all the industrial
production series exhibit time-varying conditional heteroscedasticity,
we use ARCH methodology to generate a proxy for income volatility. (11)
The Dynamic Model of Exports
In our empirical investigation, we concentrate on the log
difference of deseasonalized sectoral real exports, [x.sub.t], of
country i to j and employ a dynamic distributed lag model to capture the
effects of exchange rate volatility [[sigma].sub.s] along with income
volatility [[sigma].sub.y], and the interaction of income and exchange
rate volatility, [[sigma].sub.s] x [[sigma].sub.y], on sectoral trade
flows. (12) As explained earlier, in total we investigate 230 models and
focus on the significance of coefficients associated with exchange rate
and income volatilities as well as the interaction between the two
series. Each model includes the standard variables, such as the change
in log importing country real income, [y.sub.t], and change in log real
exchange rate, [s.sub.t], as well as the lagged dependent variable. Our
model takes the following form:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (4)
where k [member of] [1, 2, ..., 10] denotes sector and [delta] is a
fixed coefficient. The two additional terms in our model--the impact of
foreign income volatility on trade flows and the interaction between
foreign income and exchange rate volatility--have been suggested by Baum
et al. (2004) to capture the impact of the expansion or the retention of
the trade flows as foreign income and the exchange rate fluctuates. Such
an approach, according to Baum et al. (2004), requires a simultaneous
consideration of the behavior of the exchange rate, foreign income, and
the risks that can be captured through the interaction between income
and exchange rate volatilities. Although they find mixed results on the
effect of income volatility on trade flows, a subsequent analysis by
Grier and Smallwood (2007) provides evidence in support for the
importance of income volatility. (13)
Prior to estimating Equation 4, we must determine the maximum lag
that the variables should take. Earlier research suggests that empirical
models that embody 6 to 12 lags successfully capture the potential
effects regarding the agent's decision to purchase and complete
their transactions. (14) In general, seeking new opportunities to
expand, establish, retain, or shut down the business in a market
requires suppliers not to react instantaneously to changes in market
conditions when faced with high short-term profits or losses. This
behavior seems reasonable, as any change in a business model requires
substantial resource allocation problems and implies that
exporters' reactions to exchange rate or income volatility should
be modeled with several lags. In our empirical analysis, to be
parsimonious, we report results while allowing variables to take up to
six lags. (15) We also set the lag parameter [delta] to a specific value
to ensure stability of the dynamic relationship. Particularly, we report
our results setting [delta] = 0.3 for stability reasons. We should note
that we also experimented with linear weights giving higher weights to
more recent observations. This modification did not lead to any
significant changes in the results.
Given the vast number of models that we consider, in Tables 6-8, we
depict the country-industry pairs and the sign of the impact of exchange
rate, income uncertainty and the interaction between the two, as the
associated coefficient with those variables attains significance at the
1%, 5%, or 10% level. We present two panels per table. While the upper
panel presents our summary results for the sectoral exports of 13
countries to the United States, the lower panel concentrates on the
sectoral exports of the United States to the same set of countries.
These tables also reveal the differences across emerging versus advanced
countries.
Data
In our investigation, we utilize deseasonalized monthly data on
sectoral bilateral real exports, in each direction, over the period
January 1996 and September 2007 between the United States and its top 13
trading countries. Nine of the countries in our data set, including the
United States, Japan, Germany, the United Kingdom, France, Italy, the
Netherlands, Ireland, and Canada, have highly advanced economies. The
remaining five countries, namely, South Korea, Singapore, Malaysia,
China, and Brazil, are considered emerging economies. Given the earlier
findings that exchange rate volatility has a significant impact on the
trade flows of emerging rather than the advanced economies, our data
set, which contains advanced and emerging countries, can help us find
out if this observation holds true for sectoral trade flows.
Furthermore, the use of sectoral data can help us determine if the
significant effects of exchange rate volatility on emerging country
trade flows are artifacts of data aggregation. In particular, our data
set includes trade flows gathered from 10 sectors and is available from
the Foreign Trade Division of the U.S. Census Bureau. The sectors are
(i) food and live animals; (ii) beverages and tobacco; (iii) crude
materials; (iv) mineral fuels, lubricants, and related materials; (v)
animal and vegetable oils, fats, and waxes; (vi) chemicals and related
products; (vii) manufactured goods; (viii) machinery and transport
equipment; (ix) miscellaneous manufactured articles; and (x) commodities
and transactions.
The sectoral trade data are in current U.S. dollars, which are then
converted into local currency units using the spot exchange rate in
relation to the U.S. dollar. Then we deflate the sectoral trade data by
the export price index for both advanced and emerging countries to
obtain real trade flows. (16) AS we discussed earlier, the real exchange
rate data are constructed using the spot rate and the local and the U.S.
consumer price indices. Spot daily exchange rates are obtained from the
Pacific Exchange Rate Service. Consumer price indices for the United
States and the remaining countries are obtained from the Main Economic
Indicators published by the OECD. Export price indices are extracted
from the International Monetary Fund's International Financial
Statistics. Finally, deseasonalized industrial production series, which
we proxy for the income of a country, are extracted from the Main
Economic Indicators published by the OECD.
3. Empirical Findings
Descriptive Statistics
Given that we will be investigating the linkages between sectoral
trade flows and real exchange rate and real income variations, we first
provide some statistics on the common features, as well as the
dissimilarities, of these series. Table 1 presents the real exchange
rate volatility correlations among those countries that we have in our
data set. These correlations show that similar real exchange volatility
patterns are experienced by many of the advanced countries, except for
Japan, perhaps reflecting these countries' sizable exports to the
United States. High correlations between these countries may also
reflect the agreements between the European countries that eventually
led to the launch of the euro. When we turn our attention to the
correlations between the real exchange rate volatility measures of the
emerging countries, we observe some similarities, but the correlations
are not as strong as that between the European countries. Table 1 shows
that the real exchange rate volatility measures across advanced and
emerging countries are very different from one another. This observation
gives the impression that the impact of exchange rate uncertainty on
trade flows could differ between advanced and emerging economies.
We next focus on descriptive measures of foreign income volatility
and the interaction term that we introduce in our model. The
correlations of foreign income volatility measures and that of the
interaction term--the product of the exchange rate volatility and
foreign income volatility--for our exporting countries are presented in
Tables 2 and 3, respectively. Inspecting Table 2, we do not detect much
comovement of income volatility between the countries in our data set.
Similarly, as shown in Table 3, we detect no systematic relationship
across countries with respect to the interaction term.
To evaluate how the exchange rate and income volatility measures
can affect the sectoral bilateral trade of advanced and emerging
countries, in Tables 4 and 5, we present sectoral export flow
correlations for Germany and China, respectively. Table 4, which gives
the correlation matrix for Germany, does not reveal any significant
sector-specific trade flow correlations. This observation can be
explained by the fact that Germany has a well-developed economy whose
sectoral exports to the United States are not much affected by movements
in the export volume of one sector or another. However, Table 5, which
provides information on Chinese sectoral exports to the United States,
shows high correlations between most sectoral trade flows. This finding
can be explained by the acceleration of sectoral trade flows from China
to the United States over the past 10 years. Similar patterns can be
observed for the other emerging countries as well.
Given the information presented in the correlation tables, it seems
reasonable to conjecture that the intensity of development could be
important regarding the role of exchange rate uncertainty on trade
flows. For emerging countries where international trade is consistently
improving and where trading partners or exportable products are not
diverse, significant effects of exchange rate volatility on trade flows
should not be too surprising, whereas for countries whose economies are
well developed and have established trade links, the impact of exchange
rate volatility may be insignificant. We finally check if there are any
sector-specific correlations across countries but find no systematic
associations. (17)
We must note that prior to estimating our model, we test each
series for a unit root using the augmented Dickey-Fuller (see Dickey and
Fuller 1981) and the Phillips and Perron (1988) unit root tests. These
tests verify that each series that enters the model is stationary. We
also check for the presence of autocorrelation and normality of the
error terms. Breusch-Godfrey and Q-tests show that the model's
error term does not suffer from autocorrelation. It is possible to check
for normality of the errors using visual methods or numerical methods.
While graphical methods are intuitive and easy to interpret, numerical
methods provide an objective means to examine normality. Inspection of
the graphs for several series leads us to believe that the errors are
normally distributed. Then we subject these series to the Shapiro-Wilk
W-test, which is the ratio of the best estimator of the variance to the
usual corrected sum of squares estimator of the variance (Shapiro and
Wilk 1965). The test results confirm our visual inspection that the
errors are normally distributed. Finally, to avoid problems that may
arise from heteroscedasticity, we report robust standard errors.
In the next section, we investigate the role of exchange rate and
income volatilities and the interaction between exchange rate and income
volatilities on trade flows. Given that we are working with dozens of
models to understand sectoral bilateral trade flows between the United
States and its 13 trading partners, we provide summary statistics on the
significance of those coefficients broken down into sectors and the
destination of exports (exports to and from the United States) for the
full sample and the advanced countries. We must also note that the other
variables in our model (lagged dependent variable, income and exchange
rate) take the expected signs for all country pairs that we investigate.
In that, the lagged dependent variable is always significant, while the
other two variables are significant for the preponderance of the models.
Results
In what follows, we first discuss the impact of exchange rate
volatility on sectoral trade flows. Next, we examine the effect of
income volatility and the interaction term on trade flows.
The Role of Exchange Rate Volatility. We first focus on the sign
and the significance of the coefficient associated with exchange rate
volatility, [[beta].sub.3], which is coming from Equation 4. The number
of significant effects detected for sectoral exports to and from the
United States are reported in Table 6. A quick look at the table reveals
that exchange rate uncertainty has a significant effect on sectoral
trade flows only for a handful of cases. When we concentrate on sectoral
exports to the United States, we see that there are 4 (9) out of 110
possible models where [[beta].sub.3] is significantly different from
zero at the 5% (10%) level. The tally when we turn to the significance
of 133 for the exports of the United States is similar in nature; 9 (12)
out of 120 models are significant at the 5% (10%) level. That is,
overall for about 6% of the cases does exchange rate uncertainty have a
significant impact on sectoral trade flows to and from the United States
at the 5% significance level. When we scrutinize sign of the impact, we
find that exchange rate uncertainty has a slight positive effect at the
median.
Given earlier findings that exchange rate volatility has a
significant negative impact on the trade flows of emerging countries, it
is important to investigate the regression results for our set of
emerging countries closely. To that end, we find that at the median,
this effect is positive yet small. When we consider all possible models
for the emerging economies, we come across eight significant models out
of possible 92 cases, corresponding to 9% of all cases at the 5% level.
Overall, the effect of exchange rate uncertainty on trade flows between
the United States and its emerging trade partners is stronger in
comparison to that of advanced countries. However, this is, too, a small
number of significant cases in comparison to earlier studies, setting a
serious doubt on the claim that exchange rate uncertainty affects
emerging country trade flows.
Overall, our findings confirm that the effect of exchange rate
volatility on trade flows is negligible and that the sign of the effect
is ambiguous for both emerging and advanced trading partners of the
United States. While this effect is more pronounced for emerging
economies, the significant models are not more than a handful, where
almost an equal number of positive and negative impacts are observed.
The Role of Income Volatility. We discuss the observed effect of
income volatility, captured by [[beta].sub.4] in our model, on
exporters' behavior. Table 7 provide the number of significant
coefficients for exports to and from the United States. When we consider
the impact of income volatility on exports to the United States, we
observe that [[beta].sub.4] is significantly different from zero in only
4 (9) cases out of 110 models at the 5% (10%) significance level.
Perhaps the low significance of the U.S. income volatility on trade
flows reflects the fact that the U.S. economy over the period of our
investigation did not experience much variation. However, when we turn
to understand trade flows from the United States, we see that the effect
of income uncertainty becomes somewhat more noticeable; we record 10
(15) significant cases out of 120 possible models at the 5% (10%) level.
This difference can be explained by the fact that the trade partners of
the United States have experienced much more volatile income patterns
than that of the U.S. over the period of investigation. The impact of
income volatility can be equally positive or negative, where the median
effect happens to be negative yet small. Nevertheless, these results do
not provide convincing evidence that income volatility is an important
determinant of sectoral trade flows.
The Role of the Interaction Term between Income and Exchange Rate
Volatilities. We finally explore whether the interaction term (captured
by [[beta].sub.5]) between the real exchange rate and industrial
production (IP) volatility has any effect on sectoral trade flows. As in
the previous two subsections, we provide summary information on the role
of the interaction for exports to and from the United States in Table 8.
Considering exports to and from the United States, we see that the
effect can be positive or negative, where we observe 18 (37) significant
cases at the 5% (10%) significance level. When we focus on exports to
the United States, we observe that only 7 (17) out of 110 cases have the
significant effect and are mainly negative at the 5% (10%) significance
level--five of those seven significant cases are realized for emerging
countries: China, South Korea, Brazil, and Malaysia. When we observe the
results for U.S. exporters, 11 cases are significant, and six of them
are registered for emerging countries.
Given these observations, one may conclude that the interaction
term has a minor role in the determination of trade flows. However,
considering Tables 6 and 8 together, we see that the interaction term is
generally significant if the corresponding coefficient for exchange rate
volatility is significant. Moreover, the sign of the interaction term is
the opposite of that of exchange rate volatility, negating the impact of
exchange rate uncertainty on trade flows. This is an interesting
observation, one that is not reported in the earlier literature, and
implies that, depending on the relative size of exchange rate volatility
and income volatility, the impact of exchange rate volatility would be
nullified. Models that do not incorporate this interaction term are
clearly misspecified, and interpretations regarding the impact of
exchange rate uncertainty on trade flows that are based on these models
will yield erroneous conclusions.
Robustness Check Using a Semirestricted Model
The distributed lag model that we estimate in the previous section
constrains the effects of our variables on trade flows to a single
coefficient per model while placing more weight on the most recent lags.
In this subsection, we relax the restrictions that we place on exchange
rate, income volatilities, and their interaction while keeping the
structure of the other variables the same as before. This modeling
approach, which we call a semirestricted model, although it sacrifices
the parsimony of the earlier model, allows the data to determine the
sign and the significance of the variables' coefficients that are
of most interest to us up to six lags. Hence, we can check the
robustness of our claims regarding the role of exchange rate and income
uncertainty and their interactions on trade flows. The semirestricted
model takes the following form:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]. (5)
In this new model, the coefficients associated with exchange rate
and income volatility and their interaction, where n [member of] [1, 2,
..., 6] denotes lags, are now allowed to take a different value for each
lag. For the other variables, we keep the same structure as before,
where [delta] = 0.3. Estimating all possible combinations, we report
summary results gathered from 233 models (25 models failed the dynamic
stability conditions) in Tables 9-14. Observing these tables, we see
that the coefficients [[beta].sub.3,n], [[beta].sub.4,n], and
[[beta].sub.5,n] take significant values at the 5% level for various
lags for different sectors. However, we do not see a systematic pattern
of significance across sectors or countries. Furthermore, the sign of
the coefficients can be positive as well as negative. To gain a better
view of the impact of the variables of interest, we compute their joint
effect over the six lags. In the case of exports (see Tables 9 and 10)
from the United States to its partners, we find that exchange rate
volatility affects trade flows in eight cases across six countries,
including the United Kingdom (9), France (8), Italy (5), Brazil (4, 6),
Canada (10), and Malaysia (6, 10), where affected sectors are shown in
brackets. In the reverse case, when we inspect the exports to the United
States by its trade partners, we find that exchange rate volatility
affects trade flows in five cases for five countries, including Japan
(8), the United Kingdom (5), South Korea (11), the Netherlands (6), and
Malaysia (3).
A similar exercise is carried out to understand the impact of
income volatility on trade flows, where Tables 11 and 12 present the
results for Equation 5. Similarly, income volatility, too, has
significant effects at various lags for different sectors with no
systematic pattern across sectors or countries. When we consider exports
from the United States to its partners, joint effects of the six lags
are found to be significant for seven cases across five countries,
including the United Kingdom (9), Italy (5), Brazil (4, 6, 7), Canada
(10), and Malaysia (2). When we turn to investigate the joint
significance of the coefficients for exports to the United States by its
trade partners, we find four significant cases, including Japan (8),
South Korea (11), France (3), and Malaysia (3).
Finally, Tables 13 and 14 present results for the interaction term.
Here, too, we observe that various lags of the interaction term take
significant coefficients without a systematic pattern. The joint impact
of the interaction term when we consider exports from the United States
to its trade partners is significant for nine cases across seven
countries, including the United Kingdom (9), South Korea (10), France
(8), Italy (5), Brazil (4, 6), Canada (10), and Malaysia (6, 10). When
we concentrate on the exports to the United States by its trade
partners, the joint significance of the coefficients is observed for
four cases, including Japan (8), South Korea (11), the Netherlands (6),
and Malaysia (3).
Overall, these results verify our earlier findings that exchange
rate volatility, income volatility, and the interaction of these two
variables do not play a significant role in explaining the trade flows
between the United States and its top trading partners. Furthermore,
results from the semirestricted models imply that coefficients are model
dependent, yet our conclusion regarding the impact of exchange rate and
income volatilities and the interaction term on trade flows are similar
across the two approaches. (18)
4. Conclusion
In this article, we investigate the impact of exchange rate
volatility on sectoral bilateral trade flows between the United States
and its 13 top trading countries over the period between 1996 and 2007.
Our monthly data set includes both emerging and advanced economies,
allowing us to avoid the narrow focus on the United States or the G7
country data that has characterized much of the literature. Furthermore,
concentrating on the behavior of sectoral trade flows, we avoid
potential biases that may arise because of the use of aggregate data.
Overall, we investigate bilateral trade flows for dozens of
sector-country pairs separately to shed a broader view on the linkages
between the variables of interest. In our investigation, we also
entertain an idea suggested by Baum et al. (2004) that income volatility
and its interaction with exchange rate volatility may have an impact on
trade flows.
Our results provide evidence that exchange rate uncertainty has
little effect on sectoral trade flows. We find that the impact of real
exchange rate volatility on trade flows is significant in about only 6%
of the models at the 5% significance level, where the effect is yet
positive. Furthermore, although this relationship is slightly stronger
for the emerging countries, our findings do not allow us to confirm
earlier findings that exchange rate volatility plays an important role
for trade flows of emerging countries. Overall, our results show that
there is little effect of exchange rate volatility on sectoral trade
flows, and this holds for both advanced and emerging economies.
Next, we turn our attention to the impact of income volatility and
the interaction term between exchange rate volatility and income
volatility. It turns out that the interaction term is significant in
almost all cases when exchange rate volatility has a significant role in
the model. Furthermore, it takes the opposite sign to that of exchange
rate volatility, reversing the impact of exchange rate volatility on
trade flows in the opposite direction. From this perspective, omitting
the interaction term from the analysis would lead to the wrong
conclusion and inappropriate policy prescriptions. Finally, when we
investigate the impact of income volatility on trade flows, we observe
that income uncertainty has a significant effect in only 5% of the
models. The sign of this coefficient is negative at the median. However,
this variable seems to play a more important role when we investigate
the exports of the United States to its trading partners as the trading
partners experience more volatile income patterns. This is reasonable,
as the income pattern of the trade partners of the United States over
the period of investigation was much more volatile than that of the
United States.
Finally, we check for the robustness of the model by investigating
a less restricted model. Results from this experiment verify our earlier
conclusion that exchange rate, income uncertainty, and the interaction
of these terms do not have a meaningful impact on trade flows. However,
given that we concentrated our investigation on the United States and
its top trading partners' sectoral trade flows, it would be useful
to investigate data from other countries to generalize these findings.
References
Akhtar, Mohammad A., and Spense R. Hilton. 1984. Effects of
exchange rate uncertainty on Germany and U.S. trade. Federal Reserve
Bank of New York Quarterly Review 9:7-16.
Arize, Augustine C., Thomas Osang, and Daniel J. Slottje. 2000.
Exchange rate volatility and foreign trade: Evidence from thirteen LDCs.
Journal of Business and Economic Statistics 18:10-17.
Armington, Paul. 1969. A theory of demand for products
distinguished by place of production. IMF Staff Papers 16:159-78.
Barkoulas, John T., Christopher F. Baum, and Mustafa Caglayan.
2002. Exchange rate effects on the volume and variability of trade
flows. Journal of International Money and Finance 21:481-96.
Baron, P. David. 1976. Fluctuating exchange rates and the pricing
of exports. Economic Inquiry 14:425-38.
Baum, Christopher F., and Mustafa Caglayan. 2010. On the
sensitivity of the volume and volatility of bilateral trade flows to
exchange rate volatility. Journal of International Money and Finance
29:79-93.
Baum, Christopher F., Mustafa Caglayan, and Neslihan Ozkan. 2004.
Nonlinear effects of exchange rate volatility on the volume of bilateral
exports. Journal of Applied Econometrics 19:1-23.
Baum, Christopher F., Mustafa Caglayan, Neslihan Ozkan, and
Oleksandr Talavera. 2006. The impact of macroeconomic uncertainty on
non-financial firms' demand for liquidity. Review of Financial
Economics 15:289-304.
Belanger Denis, Sylvia Gutierrez, Daniel Racette, and Jacques
Raynauld. 1992. The impact of exchange rate variability on trade flows:
Further results from sectoral U.S. imports from Canada. North American
Journal of Economics and Finance 3:888-92.
Bini-Smaghi, Lorenzo. 1991. Exchange rate variability and trade:
Why is it so difficult to find any relationship? Applied Economics
23:927-36.
Byrne, Joseph P., Julia Darby, and Ronald MacDonald. 2008. U.S.
trade and exchange rate volatility: A real sectoral bilateral analysis.
Journal of Macroeconomics 30:238-59.
Clark, B. Peter. 1973. Uncertainty, exchange risk, and the level of
international trade. Western Economic Journal 11:302-13.
Clark, Peter, Natalia Tamirisia, Shang-Jin Wei, Azim Sadikov, and
Li Zeng. 2004. Exchange rate volatility and trade flows C some new
evidence. IMF Occasional Paper. Accessed September 21, 2008. Available
http://www.imf.org/ external/np/res/exrate/2004/eng/051904.htm.
Cushman, David. 1983. The effects of real exchange rate risk on
international trade. Journal of International Economics 15:45-63.
Cushman, David. 1988. U.S. bilateral trade flows and exchange risk
during the floating period. Journal of International Economics
24:317-30.
De Vita, Glauco, and Andrew J. Abbott. 2004. The impact of exchange
rate volatility on U.K. exports to EU countries. Scottish Journal of
Political Economy 51:62-81.
Dickey, David A., and Wayne A. Fuller. 1981. Likelihood ratio
statistics for autoregressive time series with a unit root. Econometrica
49:1057-72.
Doyle, Eleanor. 2001. Exchange rate volatility and Irish-U.K.
trade, 1979-1992. Applied Economics 33:249-65.
Franke, Guenter. 1991. Exchange rate volatility and international
trading strategy. Journal of International Money and Finance 10:292-307.
Gagnon, E. Joseph. 1993. Exchange rate variability and the level of
international trade. Journal of International Economics 34:269-87.
Grier, Kevin B., and Aaron D. Smallwood. 2007. Uncertainty and
export performance: Evidence from 18 countries. Journal of Money, Credit
and Banking 39:965-79.
Hooper, Peter, and Steven W. Kohlhagen. 1978. The effect of
exchange rate uncertainty on the prices and volume of international
trade. Journal of International Economics 8:483-511.
Kenen, Peter B., and Dani Rodrik. 1986. Measuring and analyzing the
effects of short term volatility in real exchange rates. Review of
Economics and Statistics 68:311-5.
Klaassen, Franc. 2004. Why is it so difficult to find an effect of
exchange rate risk on trade? Journal of International Money and Finance
23:817-39.
Klein, W. Michael. 1990. Sectoral effects of exchange rate
volatility on U.S. exports. Journal of International Money and Finance
9:299-308.
Koray, Faik, and William D. Lastrapes. 1989. Real exchange rate
volatility and U.S. bilateral trade: A VAR approach. Review of Economics
and Statistics 71:708-12.
Koren, Miklos, and Adam Szeidl. 2003. Exchange rate uncertainty and
export prices. Unpublished paper, Harvard University.
Merton, C. Robert. 1980. On estimating the expected return to the
market: An exploratory investigation. Journal of Financial Economics
8:323-61.
McKenzie, D. Michael. 1999. The impact of exchange rate volatility
on international flows. Journal of Economic Surveys 13:71-106.
Pere, Eric, and Alfred Steinherr. 1989. Exchange rate uncertainty
and foreign trade. European Economic Review 33:1241-64.
Peridy, Nicolas. 2003. Exchange rate volatility, sectoral trade and
aggregation. Weltwirtschaftliches Arehiv 139:389-418.
Phillips, Peter C. B., and Pierre Perron. 1988. Testing for a unit
root in time series regressions. Biometrika 75:335-46.
Saito, Mika. 2004. Armington elasticities in intermediate inputs
trade: A problem in using multilateral trade data. Canadian Journal of
Economics 37:1097-117.
Sauer, Christine, and Alok K. Bohara. 2001. Exchange rate
volatility and exports: regional differences between developing and
industrialized countries. Review of International Economics 9:133-52.
Sercu, Piet, and Cynthia Vanhulle. 1992. Exchange rate volatility,
international trade, and the value of exporting firm. Journal of Banking
and Finance 16:152-82.
Shapiro, S. S., and M. B. Wilk. 1965. An analysis of variance test
for normality (complete samples). Biometrika 52:591-611.
Tenreyro, Silvana. 2004. On the trade impact of nominal exchange
rate volatility. FRB of Boston Working Paper No. 03-2.
Thursby, Jerry G., and Marie C. Thursby. 1987. Bilateral trade
flows, the Linder hypothesis and exchange rate risk. Review of Economics
and Statistics 69:488-95.
Wei, Shang-Jin. 1999. Currency hedging and goods trade. European
Economic Review 43:1371-94.
Mustafa Caglayan, Department of Economics, University of Sheffield,
9 Mappin Street, Sheffield S1 4DT, United Kingdom; E-mail
mcaglayan@shef.ac.uk; corresponding author.
Jing Di, Department of Economics, University of Sheffield, 9 Mappin
Street, Sheffield S1 4DT, United Kingdom; E-mail
ecp07jd@sheffield.ac.uk.
We would like to thank Sarah Brown and Monica Hernandez for
productive conversations and comments. The standard disclaimer applies.
Received September 2008; accepted December 2009.
(1) See, for instance, Clark (1973), Baron (1976), Pete and
Steinherr (1989).
(2) Franke (1991) and Sercu and Vanhulle (1992) show that exchange
rate volatility can have a positive or an ambiguous effect on trade
flows. Barkoulas, Baum, and Caglayan (2002) claim that the types of
shocks that firms are exposed to will determine the relationship, which
may be positive, negative, or ambiguous.
(3) Although researchers implementing gravity models consistently
conclude that exchange rate volatility has a negative impact on trade
flows, Clark et al. (2004) indicate that this finding is not robust to a
more general setting that embodies the recent theoretical advances in a
gravity model.
(4) For instance, while Cushman (1983, 1988), Akhtar and Hilton
(1984), Kenen and Rodrik (1986), and Thursby and Thursby (1987), among
others, find negative effects, Hooper and Kohlhagen (1978), Koray and
Lastrapes (1989), and Gagnon (1993) report insignificant effects.
(5) See, for instance, Baum, Caglayan, and Ozkan (2004) and Baum
and Caglayan (2010), who use the same bilateral trade flows data from 13
advanced countries while implementing different empirical methodologies.
(6) See also Arize, Osang, and Slottje (2000); Clark et al. (2004);
Peridy (2003); and Sauer and Bohara (2001).
(7) One problem with our approach is the lack of monthly data on
sectoral export prices, rendering us to use monthly aggregate export
prices to construct monthly sectoral real export data.
(8) Although generally researchers consider the effect of real
exchange rate variability on trade flows, nominal exchange rate
variability has also been used in the past. For instance, Tenreyro
(2004) shows that nominal exchange rate volatility does not affect trade
flows.
(9) See Klein (1990), Belanger et al. (1992), McKenzie (1999),
Doyle (2001), Peridy (2003), De Vita and Abbott (2004), Saito (2004),
and Byrne, Darby, and MacDonald (2008).
(10) Researchers use Merton's (1980) methodology to generate
proxies for exchange rate, interest rate, (monetary) policy, or stock
market volatilities. See, for instance, Baum et al. (2006) for an
implementation of Merton's method on stock returns.
(11) Details are available on request from the authors.
(12) Sectoral trade series are seasonally adjusted using seasonal
dummies.
(13) Koren and Szeidl (2003) suggest that exchange rate volatility
should affect trade volumes through the covariances of the exchange rate
with the other key variables.
(14) See Baum et al. (2003) and Baum and Caglayan (2009) on this
issue.
(15) Results (available on request from the authors) that allow
variables to take up to 12 lags do not differ from those that we present
here.
(16) Converting nominal monthly sectoral trade flows into real
flows by using monthly aggregate export prices rather than sectoral
export prices is a weakness of our study that we cannot rectify because
of a lack of data, although researchers who use annual sectoral data
(see, e.g., Byrne et al. 2008) in their investigation are not
constrained to that effect.
(17) Sector-specific correlation tables are not provided for space
considerations but are available on request.
(18) In fact, results for the semirestricted model are worse than
those from the restricted model.
Table 1. Real Exchange Rate Uncertainty Correlations across
Countries
JPY DEM GBP FRF ITL
JPY 1.0000
DEM 0.1064 1.0000
GBP -0.0937 0.8468 1.0000
FRF 0.0782 0.9985 0.8553 1.0000
ITL -0.0479 0.9761 0.8817 0.9853 1.0000
NLG -0.0323 0.9760 0.8409 0.9847 0.9944
IEP -0.2181 0.9111 0.8489 0.9283 0.9734
CAD -0.2581 0.7248 0.7254 0.7444 0.7999
SGD 0.6604 0.5051 0.2487 0.4651 0.3249
KRW 0.1081 0.5373 0.3538 0.5386 0.5428
MYR 0.5363 0.4232 0.0778 0.3898 0.2771
CNY 0.7466 0.2023 -0.0755 0.1571 -0.0043
BRL 0.2805 0.5262 0.4353 0.4971 0.4051
NLG IEP CAD SGD
JPY
DEM
GBP
FRF
ITL
NLG 1.0000
IEP 0.9691 1.0000
CAD 0.7795 0.8541 1.0000
SGD 0.3259 0.1701 0.1045 1.0000
KRW 0.5569 0.5774 0.6524 0.3550
MYR 0.3021 0.1917 0.1487 0.8784
CNY 0.0118 -0.1814 -0.2892 0.8915
BRL 0.3720 0.2866 0.4023 0.7497
KRW MYR CNY BRL
JPY
DEM
GBP
FRF
ITL
NLG
IEP
CAD
SGD
KRW 1.0000
MYR 0.5891 1.0000
CNY 0.0239 0.7465 1.0000
BRL 0.2827 0.5448 0.5602 1.0000
The currencies are ordered for Japan, Germany, the United Kingdom,
France, Italy, The Netherlands, Ireland, Canada, Singapore, South
Korea, Malaysia, China, and Brazil.
Table 2. Income Volatility Correlations across Countries
U.S. JP GE U.K. FR
U.S. 1.0000
JP -0.0706 1.0000
GE -0.1263 0.6271 1.0000
U.K. -0.2816 0.4538 0.3979 1.0000
FR 0.2812 0.1368 0.2023 0.2334 1.0000
IT 0.2951 0.1442 0.2086 0.2687 0.9686
NL -0.4920 0.3470 0.6116 0.5979 0.0588
IE -0.1477 0.2359 0.8126 0.1669 0.0129
CAN -0.1308 0.2400 0.7488 0.3371 0.0681
SG -0.2071 -0.0045 -0.0228 0.0002 -0.1867
KR -0.2115 0.0638 0.0758 0.0894 -0.0827
ML 0.2546 -0.1144 -0.2948 -0.2777 -0.1639
CN -0.1692 0.3295 0.1468 0.2909 -0.1520
BR -0.2265 -0.2429 -0.1329 -0.1918 -0.1937
IT NL IE CAN SG
U.S.
JP
GE
U.K.
FR
IT 1.0000
NL 0.0890 1.0000
IE 0.0133 0.6339 1.0000
CAN 0.0922 0.6498 0.8866 1.0000
SG -0.1929 0.0111 -0.0045 0.0191 1.0000
KR -0.0989 0.0787 0.0702 0.1008 0.5972
ML -0.1686 -0.3506 -0.2458 -0.2378 0.0352
CN -0.1347 0.2656 0.0683 0.1379 0.0815
BR -0.2018 0.0185 -0.0255 -0.0808 0.2127
KR ML CN BR
U.S.
JP
GE
U.K.
FR
IT
NL
IE
CAN
SG
KR 1.0000
ML 0.0212 1.0000
CN 0.2119 -0.0226 1.0000
BR 0.2805 0.0662 -0.0430 1.0000
The countries are ordered as the United States, Japan, Germany, the
United Kingdom, France, Italy, The Netherlands, Ireland, Canada,
Singapore, South Korea, Malaysia, China, and Brazil.
Table 3. Interaction Terms Correlations across Countries
JP GE U.K. FR IT
JP 1.0000
GE 0.1466 1.0000
U.K. 0.1931 0.7495 1.0000
FR 0.2938 0.9378 0.8096 1.0000
IT 0.2109 0.9697 0.7417 0.9510 1.0000
NL 0.3033 0.8813 0.8187 0.9144 0.8881
IE -0.0555 0.7681 0.4495 0.6242 0.7616
CAN -0.1155 0.1566 0.1059 0.2115 0.1887
SG 0.2602 0.0514 0.0398 0.0860 0.0555
KR 0.0201 -0.0371 -0.0916 -0.0916 -0.0401
ML 0.2847 -0.0587 0.0280 -0.0269 -0.0466
CN 0.1804 -0.2374 -0.0228 -0.1802 -0.2264
BR 0.1667 -0.1360 -0.2362 -0.1777 -0.0719
NL IE CAN SG
JP
GE
U.K.
FR
IT
NL 1.0000
IE 0.5196 1.0000
CAN -0.0159 0.2045 1.0000
SG -0.0213 -0.1049 0.2038 1.0000
KR -0.1088 -0.0526 -0.0275 0.3807
ML 0.0167 -0.2206 0.0048 0.3114
CN -0.1476 -0.2358 -0.0037 0.4431
BR -0.2254 0.0233 -0.0262 0.1858
KR ML CN BR
JP
GE
U.K.
FR
IT
NL
IE
CAN
SG
KR 1.0000
ML 0.2457 1.0000
CN 0.2365 0.3475 1.0000
BR 0.3226 0.3643 0.0756 1.0000
See notes to Table 2.
Table 4. Correlations of German Sectoral Exports to the
United States
1 2 3 4 5
1 1.0000
2 -0.0110 1.0000
3 0.1844 0.0837 1.0000
4 0.2883 0.2522 0.2306 1.0000
5 0.2070 0.0245 -0.0198 0.1274 1.0000
6 0.4710 -0.0356 0.2515 0.2348 -0.0111
7 0.4567 0.1890 0.3502 0.2888 0.0696
8 0.3392 0.2891 0.2323 0.2047 0.0853
9 0.4199 0.2923 0.2759 0.2507 0.0590
10 0.3350 0.2999 0.2738 0.1751 0.1480
6 7 8 9 10
1
2
3
4
5
6 1.0000
7 0.8028 1.0000
8 0.6720 0.8732 1.0000
9 0.7241 0.9034 0.8958 1.0000
10 0.4826 0.7320 0.7519 0.7284 1.0000
Numbers 1 to 10 denote sectors, namely, (i) food and live animals;
(ii) beverages and tobacco; (iii) crude materials; (iv) mineral fuels,
lubricants, and related materials; (v) animal and vegetable oils,
fats, and waxes; (vi) chemicals and related products; (vii)
manufactured goods; (viii) machinery and transport equipment; (ix)
miscellaneous manufactured articles; and (x) commodities and
transactions.
Table 5. Correlations of Chinese Sectoral Exports to the
United States
1 2 3 4 5
1 1.0000
2 0.6077 1.0000
3 0.8411 0.5932 1.0000
4 0.4425 0.3647 0.4488 1.0000
5 0.0131 0.0258 -0.0979 0.1517 1.0000
6 0.7979 0.6743 0.9014 0.4671 0.0031
7 0.8167 0.6961 0.9311 0.4348 -0.0571
8 0.7964 0.6781 0.9065 0.4496 -0.0170
9 0.8282 0.6765 0.9369 0.4585 -0.1051
10 0.7660 0.6663 0.8642 0.4353 -0.0208
6 7 8 9 10
1
2
3
4
5
6 1.0000
7 0.9246 1.0000
8 0.8978 0.9309 1.0000
9 0.9260 0.9566 0.9456 1.0000
10 0.9054 0.9157 0.8993 0.9075 1.0000
See notes to Table 4.
Table 6. Significant Exchange Rate Volatility ((Ts,) Effects
10% 5% 1%
Export to the
United States
+ CN (2) BRA (5) CN (9)
MAY (2, 8) GE (6)
- CAD (3, 6) ITA (5)
Export from the
United States
+ BRA (4) KR (7) BRA (8) SG (8)
CAD (7) GE (6) FRA (8)
- FR (9) CN (3) U.K. (8) BRA (3) MAY (5)
Sectoral indices are given in brackets. See notes to Table 4
for sector names.
Table 7. Significant Income Volatility ([[sigma].sub.y]) Effects
10% 5% 1%
Export to the
United States
+ CN (2, 9) GE (6)
- CN (3) U.K. (2, 10) ITA (5)
IR (11) CAD (6)
Export from the
United States
+ CN (4) GE (6) FR (8, 11)
NL (2) SG (8)
- U.K. (2) FR (7, 9) CN (3) KR (5) MAY (4)
SG (6) U.K. (8) CAD (9)
See notes to Table 6.
Table 8. Significant Interaction ([MATHEMATICAL EXPRESSION NOT
REPRODUCIBLE IN ASCII] x [[sigma].sub.y]) Effects
10% 5% 1%
Export to the
United States
+ CN (3) U.K. (2) ITA (5)
IR (11) CAD (3)
- CN (2) KR (8) SG (2) KR (3, 9) CN (3) GE (6)
CAD (4) MAY (2, 6) BRA (5) MAY (8)
Export from the
United States
+ U.K. (8) KR (5, 8) CN (3) MAY (4) BRA (3)
FR (6, 9) SG (6) GE (6) GR (7)
MAY (11) CAD (9)
- GE (10) BRA (4) KR (7) BRA (8) FR (I1) NL (2)
SG (8)
See notes to Table 6.
Table 9. Significant Exchange Rate Volatility ([MATHEMATICAL
EXPRESSION NOT REPRODUCIBLE IN ASCII]) Effects for the
Semirestricted Model: Exports to the United States
Index Developed Economies Emerging Economies
1 GE (L6,-) CAD (L3,+) KR (L5,-) BR (L5,+)
2 JP (L2,-) U.K. (L2,-) CN (L1,-)
3 GE (L5,+) U.K. (L2,-) KR (L3,-) BRA (L3,-)
(L3,+) SG (L5,-)
4 GE (L4,-) NL (L3,+) KR (L2,-) SG (L2,+)
(L6,-) (L4,-)
5 GE (L1,+) CN (L2,-) MAY (L1,+)
(L6,+) (L3,-)
(L4,+)
(L5,+)
6 JP (L3,+) GE (L5,+) CN (L1,-) BRA (L5,+)
FR (L1,+) CAD (L3,+) KR (L1,+)
(L5,-)
(L6,+)
7 FR (L3,+) ITA (L2,-) KR (L1,+) SG (L2,+)
(L4,+) (L5,-)
IR (L5,-) CAD (L3,+)
8 U.K. (L2,-) NL (L2,-) MAY (L4,+) SG (L3,+)
(L6,-) (L5,-)
CN (L1,+)
9 GE (L5,+) FR (L4,+) BRA (L5,+)
NL (L2,+) CAD (L3,+)
10 FR (L4,+) IR (L3,+) KR (L4,-) SG (L3,+)
Significance at the 5% level only. Index denotes the sector. See
Table 2 for country codes. The sign and the lag (L1-L6) when
significance is attained are given in the brackets.
Table 10. Significant Exchange Rate Volatility ([MATHEMATICAL
EXPRESSION NOT REPRODUCIBLE IN ASCII]) Effects for the
Semirestricted Model: Exports from the United States
Index Developed Economies Emerging Economies
1 GE (L5,-) ITA (L3,-) KR (L2,+)
(L6,-)
CAD (L3,+)
2 FR (L2,-) ITA (L3,-) BRA (L1,-) CN (L1,-)
(L5,+) (L2,+; L5,-)
3 JP (L5,-) NL (L3,-) CN (L2,+)
(L6,+)
4 NL (L3,-) JP (L2,+) CN (L4,+)
(L4,-) (L6,-)
ITA (L5,+) U.K. (L6,-)
5 JP (L5,-) GE (L6,+) KR (L4,+) BRA (L5,-)
ITA (L5,+) CAD (L6,-) SG (L6,-)
(L6,-)
6 GE (L2,-) ITA (L2,-) SG (L3,-) MAY
NL (L3,-)
7 GE (L1,+) ITA (L5,+) SG (L3,+) BRA (L2,+)
(L3,-)
8 JP (L2,-) GE (L3,-)
FR (L2,-) NL (L6,-)
CAD (L5,-) ITA (L1,+)
(L6,-)
9 GE (L5,+) U.K. (L5,+)
(L6,-)
10 JP (L6,+) GE (L1,-) BRA (L2,+)
NL (L5,+) (L4,-)
See notes to Table 9.
Table 11. Significant Income Volatility ([[sigma].sub.y]) Effects
for the Semirestricted Model: Exports to the United States
Index Developed Economies Emerging Economies
1 JP (L6,-) FR (L4,-) KR (L5,-) MAY (L3,-)
CAD (L3,+) BRA (L1,-)
(L5,+) (L2,+)
(L5,+)
(L6,-)
2 JP (L1,+) FR (L3,+) CN (L1,-) SG (L2,+)
(L2,-) (L4,-) (L5,-)
GE (L3,+) U.K. (L2,-)
3 GE (L5,+) U.K. (L2,-) KR (L1,+) BRA (L2,-)
(L3,+) (L3,-)
(L5,+)
SG (L5,-)
4 JP (L5,-) GE (L4,-) KR (L6,-)
NL (L3,+)
(L4,-)
5 GE (L1,+) CAD (L4,-) KR (L4,-)
(L6,-)
6 JP (L3,+) CAD (L3,+) CN (L1,-) BRA (L5,+)
GE (L3,+) FR (L1,+) KR (L1,+)
(L5,+) (L6,+) (L5,+)
(L6,+)
NL (L1,+) IR (L2,+)
7 FR (L3,+) ITA (L2,-) SG (L2,+)
IR (L1,-) CAD (L3,+)
(L5,-)
8 GE (L1,-) U.K. (L2,-) MAY (L3,+) CN (L1,+)
(L4,+) (L3,-)
(L6,+)
FR (L5,+) ITA (L1,-)
9 GE (L5,+) U.K. (L6,-) KR (L1,+) BRA (L5,+)
NL (L1,-) CAD (L3,+)
(L2,+) FR (L4,+)
10 IR (L1,-) KR (L4,+)
(L3,+)
See notes to Table 9.
Table 12. Significant Income Volatility ([[sigma].sub.y]) Effects
for the Semirestricted Model: Exports from the United States
Index Developed Economies Emerging Economies
1 JP (L4,-) U.K. (L1,-) BRA (L6,+)
(L4,+)
FR (L5,-) ITA (L3,-)
NL (L1,+) CAD (L1,+)
2 FR (L2,-) CN (L1,-) MAY (L5,-)
(L3,+) (L5,-) (L6,-)
BRA (L3,+)
3 GE (L2,+) NL (L3,-) MAY (L1,-) CN (L2,+)
(L6,+)
4 JP (L6,-) U.K. (L6,-) CN (L2,+)
NL (L4,-) (L6,-)
5 JP (L3,+) KR (L3,+) BRA (L5,-)
(L5,-) SG (L6,-)
6 GE (L5,-) FR (L3,+) SG (L3,-)
NL (L5,-) CAD (L1,+) (L6,-)
(L3,-)
7 JP (L4,+) ITA (L4,-) SG (L3,+) MAY (L6,-)
CAD (L1,+)
8 JP (L2,-) FR (L2,-) MAY (L3,+)
NL (L3,+) (L5,-)
9 JP (L3,+) FR (L5,+)
CAD (L4,+)
10 FR (L1,+) NL (L5,+) MAY (L1,-)
(L3,+)
See notes to Table 9.
Table 13. Significant Interaction ([MATHEMATICAL EXPRESSION NOT
REPRODUCIBLE IN ASCII] x [[sigma].sub.y]) Effects for the
Semirestricted Model: Exports to the United States
Index Developed Economies Emerging Economies
1 JP (L6,-) GE (L6,+) KR (L5,+) BRA (L1,+)
FR (L4,+) CAD (L3,-) (L5,-)
2 JP (L1,-) U.K. (L2,+) CN (L1,+) SG (L2,-)
FR (L3,-) (L5,+)
3 JP (L2,-) GE (L5,-) BRA (L3,+) SG (L2,-)
(L5,+)
4 JP (L4,-) NL (L3,-)
(L5,+) (L4,+)
GE (L4,+) U.K. (L5,-)
5 GE (L1,-) CAD (L4,+) KR (Ll,-) MAY (L2,- L4,-)
(L6,+) (L6,-) (L3,+)
(L4,-)
CN (L2,+)
6 JP (L3,-) IR (L2,-) BRA (L5,-) CN (L1,+)
GE (L3,-) FR (L1,-)
(L5,-) (L6,-)
7 FR (L3,-) IR (L1,+) BRA (L5,-) KR (L1,-)
(L4,-) (L5,+) (L5,+)
ITA (L2,+) CAD (L3,-) SG (L2,-) MAY (L3,-)
8 GE (L1,+) U.K. (L2,+) CN (L1,-) SG (L3,-)
ITA (L1,+) (L3,+) (L5,+)
KR (L1,-)
9 GE (L5,-) FR (L4,-) BRA (L5,-) CAD (L3,-)
NL (L2,+) MAY (L4,-)
10 FR (L4,-) IR (L1,+) MAY (L1,+)
(L3,-) (L2,-)
See notes to Table 9.
Table 14. Significant Interaction ([MATHEMATICAL EXPRESSION NOT
REPRODUCIBLE IN ASCII] x [sigma].sub.y]) Effects for the
Semirestricted Model: Exports from the United States
Index Developed Economies Emerging Economics
1 JP (L4,+) U.K. (L1,+) KR (L2,-)
FR (L5,+) ITA (L3,+)
NL (L1,-) (L4,+)
2 GE (L4,-) FR (L2,+) BRA (Ll,+) CN (L1,+)
ITA (L5,-) (L3,-) (L3,-)
(L5,+)
3 JP (L5,+) GE (L5,+)
ITA (L2,-) NL (L3,+)
(L5,-) (L6,-)
4 JP (L6,+) U.K. (L6,+) CN (L6,+)
ITA (L6,-) NL (L4,+)
5 JP (L3,-) ITA (L1,-) KR (L4,-) BRA (L5,+)
(L5,+) (L5,-)
NL (L3,-) SG (L6,+)
(L5,-)
6 GE (L1,-) ITA (L4,+) SG (L3,+)
(L2,+) (L5,-) (L6,+)
FR (L3,-) NL (L5,+)
7 JP (L4,-) ITA (L1,-) BRA (L2,-)
(L4,+)
(L5,-)
8 JP (L2,+) GE (L4,+) MAY (L5,+)
FR (L2,+) ITA (L1,-)
CAD (L5,+) (L3,-)
9 JP (L3,-) GE (L1,-)
FR (L5,-)
10 GE (L6,+) FR (L1,-) KR (L5,+)
ITA (L5,+) NL (L5,-)
See notes to Table 9.
