2nd Quarter 2014

Open Flash version | pdf | trade report

Jones College logo

Estimating continued

by Steven G. Livingston | 1 | 2 | 3 | 4 | 5 | 6

One alternative is to assume a particular industry
exports at the same rate no matter where it is located.
Then if we know the percentage of an industry that
is located in a state, we know the percentage of U.S.
exports from that state. They're the same. This, for
example, is what the Brookings Institution does in its
"Export Nation" reports. In a way this throws out the
baby with the bathwater. If we assume the same export
propensity across America, we can't really verify this.

Let's go another direction and make use of the
industry production statistics in a different way. If we
know the size of an industry in a major port state, we
can estimate its exports as if it exported at the national
average. We can then compare that number to its actual
reported exports. The difference is the combination of
the port state bias and any particularities of the export
industry in that state. It's realistic to assume industries in major port states export at higher rates than inland
industries. We could account for that, leaving the port
bias. We could estimate a port state's exports if its firms exported at 110% of the national average, 120%, and
so forth. This would adjust for the likely greater export
propensities in a port state. Reported exports higher than
that estimate are credibly the result of a bias or an over-allocation due to documentation or other port issues. If we reallocate that bias or over-allocation to the exports of inland states, we obtain an export figure that may more closely correspond to a state's actual export performance.

Let's take an example, the forging and stamping industry. Two-thirds of industry exports go through ports in just six states. These six states are reported as being the source of just under one-third of all U.S.exports while being home to just under one-quarter of U.S. industry (based on value of shipments). This means recorded exports from these port states are much higher than we would expect if the propensity to export were equal across the U.S. For this industry, these six states actually "over-export" by a factor of a third (the ratio of reported to expected exports is 1.34). One possibility is that this is true. For whatever reasons forging and stamping operations in these state sare simply much more export oriented. More plausible is a port state bias in export documentation. How big is that bias? Our strategy will be to take a plausible range. We will presume the export propensity of a port state industry is at least 110% of the national rate and unlikely to exceed 120% of the national rate.