I’m giving a presentation on what drives Wisconsin exports. Here’s an interesting finding: Since 2018, exports of real goods have underperformed what would be expected based on the GDP of the rest of the world and the real value of the US dollar.
Here is Wisconsin’s (from BEA) goods exports deflated in real terms using the BLS export price index for all commodities, plotted with the trade-weighted real GDP index for the rest of the world.
Figure 1: Wisconsin goods exports in billions of $ 2000, SAAR (blue log scale, left) and real GDP of the rest of the world, 2005 = 100 (brown log scale, right). NBER has defined peak-to-low recession dates shaded in gray. Source: BEA / Census, BLS both via FRED, Dallas Fed DGEI, NBER and author’s calculations.
I use an error correction model involving exports, the GDP of the rest of the world and the real exchange rate, estimated from 1996-2017. The long-run elasticity of Wisconsin’s commodity exports relative to the GDP of the rest of the world is about 1.1, the exchange rate elasticity about 0.40. I use this equation to dynamically predict Q1 2018 onwards and I get this (note: units are in millions, monthly rate):
Figure 2: Wisconsin Freight Exports in Millions of $ 2000, seasonally adjusted (black, log scale) and dynamic forecast from ECM (blue), +/- 1 standard error confidence interval (gray lines). NBER has defined peak-to-low recession dates shaded in gray. Source: BEA / Census, BLS via FRED, NBER and author’s calculations.
Increasing forecasting errors are associated with the onset of the trade war. In other words, the historical correlations obtained up to 2017 resulted in continued, albeit slow, growth in real exports through 2019, consistent with the growth of the world economy, and a strong dollar.
Quantitative importance? In the two years of the trade war before covid-19 began, forecast errors averaged about $ 1 billion (in $ 2,000). With real exports in 2021 at around $ 14.6 billion, that’s on the order of a 0.7% reduction.
The prediction error could be attributed to an incorrect model specification or the influence of other important determinants that have been omitted from this three-variable specification (the corrected R2 is about 36%). It is always risky to argue on the basis of residuals, but at first glance, given what we know about the retaliation imposed by our trading partners against the fluctuating state, the trade war and retaliatory actions do not appear to be an implausible source of the forecast error.
So (again), thank you Trump!