David Cooper of the Economic Policy Institute replies to our reply and gets it all wrong

David Cooper of the Economic Policy Institute wrote a rather upset counterpoint on Tuesday to my counterpoint criticizing his study comparing the economic performances of Minnesota and Wisconsin since Governors Dayton and Walker took office in January 2011. This is unfortunate. With calmer reflection Mr. Cooper might have seen his errors.

Regarding unemployment, I wrote that “the EPI fails to mention that in the race to these pre-recession levels of unemployment, Wisconsin was starting from a rate of 8.1 percent and Minnesota was starting from a rate of 7.1 percent.” Mr. Cooper says that these numbers were included in his report. But they play no part whatsoever in his analysis.

The data from the Bureau of Labor Statistics show that from December 2010 – the last month before the Governors took office – to March 2018 – when Mr. Cooper’s report came out – Wisconsin’s unemployment rate fell by more percentage points than Minnesota’s and to a lower level – by 5.2 percentage points to 2.9 percent for Wisconsin and 3.9 percentage points to 3.2 percent for Minnesota. In short, the unemployment rate has fallen faster and further in Wisconsin under Gov. Walker than it has in Minnesota under Gov. Dayton.

Yet, in his report, Mr. Cooper claims that “Minnesota was arguably more effective at reducing unemployment than Wisconsin was throughout the recovery”. How does he argue this? As he writes on page 9 of his report,

Minnesota was back at its prerecession (December 2007) unemployment rate of 4.7 percent by September 2013, fewer than three years after Governor Dayton took office. In contrast, it took until December of 2014—15 months later—for Wisconsin to reach its pre-recession unemployment rate of 4.8 percent.

Note that in a report comparing the performance of two governors who assumed office in January 2011, Mr. Cooper has slipped into talking about “throughout the recovery” and “pre-recession unemployment rate(s)”, which are different time spans.

So the point stands. In judging Wisconsin under Gov. Walker a failure on unemployment, Mr. Cooper is simply misrepresenting the data.

When it comes to GDP growth, Mr. Cooper writes

Quarterly GDP data are not “more precise.” Quarterly GDP data are more volatile than annual data — that’s why the federal Bureau of Economic Analysis issues revisions to the quarterly data for several quarters after each release. As anyone with a knowledge of statistics would know, a larger sample leads to a more precise estimate — so if you want to understand trends more precisely, comparing larger periods increases the sample and smooths out volatility in the data.

I wrote that using quarterly GDP data allows for “a more precise starting point”. And this matters. To compare the economic performances of the two governors, we want to compare the GDP when they took office with the GDP now. The closest we can get to figures for GDP ‘then’ and ‘now’ are the quarterly ones. The Bureau of Economic Analysis’ figures for annual GDP in a given year are simply an average of that year’s quarterly numbers. So, the figures Mr. Cooper is using include the first quarter of 2010, a year before they took office, and in which Wisconsin’s GDP contracted by 2% and Minnesota’s by 1%.

The greater volatility of a data series does not make the data points in it less precise, as Mr. Cooper mysteriously claims. Neither does the point about sample size hold up. Each quarter’s GDP data are not based on separate random data sources or surveys but on the same methodology and similar sample set repeated four times each year. And, while it is true that the quarterly series is more volatile than the annual ones, it is also irrelevant. That fact remains that according to the BEA, in the three months before Dayton and Walker took office, GDP was $275.5 billion in Minnesota and $255.6 billion in Wisconsin (in 2009 dollars). In the last three months of 2017, Minnesota’s GDP was $305.4 billion, an increase of 11%, and Wisconsin’s was $286.1 billion, a rise of 12%.

State comparisons have their uses and their limitations too. For each similarity the states have, there are differences. Neither Minnesota nor Wisconsin — yet — is beating the other hands down over the last eight years. To say otherwise is to go against the data, which is why the Mr. Cooper had to go to such lengths to make that argument.

John Phelan is an economist at the Center of the American Experiment.