III. THE DYNAMIC RELATIONSHIP BETWEEN U.S. EXPORT GROWTH AND FOREIGN ECONOMIC GROWTH
Although the analysis in the previous section provides a convenient way to characterize the static relationship between U.S. export growth and contemporaneous foreign GDP growth, it does not capture the dynamic perspective showing how U.S. export growth and foreign growth move together over a relatively longer period. Taking into account the interactions of these variables over time produces a richer picture of their relationships. These interactions are important because they include both the direct relationship (how a growth change in one region is directly related to U.S. export growth) and the indirect relationship (how each region’s growth is related to other regions’ growth and, therefore, is indirectly related to U.S. export growth). The analysis in this section shows that the relationship between various regions’ growth and U.S. export growth differs both directly and indirectly, and both in the short term and in the long term.[22]
The reason for conducting such a dynamic analysis is to track how growth in different regions, exchange rates between foreign countries and the United States, and U.S. export growth interact with each other over time. In particular, such an analysis can show how an increase in one region’s growth rate in a particular quarter is associated with growth in U.S. exports, growth in other regions, and exchange rate movements over the next few quarters.[23] This method separates U.S. export growth that is directly related to one region’s growth from U.S. export growth that is indirectly related to that region through its interaction with growth in other regions. Movements in exchange rates are also taken into account in the model. The analysis is conducted through the use of a vector autoregression (VAR), a method that simultaneously estimates a set of regression equations relating current and past values of the key variables.[24]
This dynamic analysis yields three results regarding the relationship between GDP growth in different regions and U.S. export growth.
First, U.S. export growth is most sensitive to a growth increase in Europe during the first four quarters following the initial growth increase in the region. Following a 1.0-percentage-point increase in the European growth rate, the U.S. export growth rate increases by 1.8 percentage points over the first four quarters (Table 2, column 1). Canada has the second-strongest correlation with U.S. export growth over the same period, followed by Asia and the rest of the world.[25] The order of these regions is the same as in the regressions in the static analysis (Table 1, column 2).
Second, in the short term, U.S. export growth is more sensitive through indirect channels to growth in Canada and Europe than other regions (Table 2, column 1). These two regions exhibit a strong positive correlation with growth in other regions, implying that a growth increase in these regions is associated with a relatively larger increase in growth in other regions and thus a larger increase in U.S. export growth. In contrast, the simulation results show that Mexico’s growth usually is negatively correlated with growth in other regions.[26] Therefore, increased growth in Mexico typically is associated with reduced growth in other regions and thus is associated with a decline in U.S. export growth. This negative indirect relationship is not offset by a positive direct relationship, explaining the overall negative relationship between growth in Mexico and U.S. export growth in the first four quarters (Table 2, column 1).
Third, for most regions, the positive relationship between their growth and U.S. export growth over the first year is subsequently offset — but only partially—by a correction over the course of the following year. On net, the relationship over the first two years is still positive, however. In other words, following a positive relationship in the first four quarters, most regions have a negative relationship with U.S. export growth in the next four quarters (Table 2, column 2).[27] The negative relationship does not mean that U.S. export growth turns negative, but only that U.S. exports grow at a rate below their historical average.[28]
Thus, following an increase in growth in a foreign region, U.S. exports initially grow at a faster-than-average pace in the first four quarters and then at a slower-than-average pace in the next four quarters (though the deviation below average in the latter quarters is less than the initial deviation above average).
The slower pace in the second four quarters is primarily because a growth increase in one region usually is followed by some decline in its future growth, which also may be correlated with declines in other regions’ growth in the second four quarters. This pattern is especially pronounced for Europe and Asia. In contrast, Canada’s growth decline in the second four quarters is not correlated with a growth decline in other regions, leading Canada to have the strongest positive correlation with export growth in the first eight quarters.