Quarterly state GDP figures for the second quarter were released yesterday. Estimated Wisconsin GDP is flat relative to 2015Q4; 2016Q1 GDP was revised down by 1.7%.
The first point is placed in context in Figure 1.
The relevant comparison is where Wisconsin GDP is relative to where we expect it to be. One way to make this comparison is to use historical correlations. I use Wisconsin GDP back to 1988 (interpolated from annual data upto 2004) and US GDP. I run a regression of log first difference Wisconsin GDP on a contemporaneous and lagged log first difference of US GDP over the 1988-2010 period.
Δ yWIt = 0.0007 + 0.665 Δ yUSt + 0.173 Δ yUSt-1
Adj. R2 = 0.38, SER = 0.006, n=92. bold face denotes significance at 5% msl using HAC standard errors. Q(4), Q(8), Q(12) statistics fail to reject null of no serial correlation.
Figure 3 depicts the dynamic forecast (red line) against the actual outcome, along with the 90% prediction interval.
Figure 3: Log real Wisconsin GDP (black), and out-of-sample dynamic forecast (red), and 90% prediction interval. Source: BEA state GDP release, and BEA GDP, 2016Q3 second release, and author’s calculations (see text).
Reported GDP is 3.9% lower than that expected using historical correlations up to 2010.
How does the 2016Q2 level of Wisconsin GDP compare against the Walker Administration’s forecast? I don’t know, since the Administration has ceased to publish the Wisconsin Economic Outlook. The last published forecast from May 2015 forecasted average 2016 GDP 4% higher (in log terms) than actual 2016Q2 levels.
Update, 12/9, 7pm Pacific: Manufacturing as a share of GDP in Wisconsin is about the same as in Michigan, but appears less cyclically-sensitive.
Figure 4: Manufacturing value added share of GDP in Michigan (blue), in Wisconsin (red). BEA and author’s calculations.