The State’s Spin Machine

Every few months, Marquette’s Law School sponsors an opinion survey of a sample of Wisconsin voters. Question 7 in this month’s poll asked:

Do you think Wisconsin is creating jobs at a faster rate than most other states, at about the same rate, or is Wisconsin lagging behind other states in job creation?

Rather than asking for an opinion, this question tests voters’ economic knowledge. It has a right answer. As can be seen from the chart below, the correct answer is “lagging behind.”

Clearly, the data doesn’t support either of the other two possible answers. Even if you restricted the data to private-sector jobs, the numbers would change but Wisconsin would still lag behind. Wisconsin’s lagging job growth has been well-publicized both in the press and by partisans of the other side.

Yet almost half (47.5%) of voters willing to hazard an answer gave a wrong one. Of those identifying themselves as Republicans, three-fourths (73.6%) got it wrong. And about a fifth of Republicans believed that Wisconsin is creating jobs faster than most other states.

Why did so many voters—especially Republican voters—pick a wrong answer? Much of the explanation, of course, comes from our human tendency to embrace facts that support our belief system and filter out those that challenge it. Perhaps the placement of the question, immediately following one that asked whether voters approved of Scott Walker, influenced the responses: they may have reinterpreted the question as asking whether Walker was good for the economy.

But the responses may also reflect disinformation offered every month on Wisconsin’s job performance by the state agency charged with releasing the jobs data, the Wisconsin Department of Workforce Development. In its cover letter accompanying the data release, the department tries to select data that puts Wisconsin’s performance in the most favorable light. A few quotes from the June report illustrate the approach:

• “Our labor market economists note the seasonally adjusted total nonfarm job gain from May to June of 17,500 is the largest month-to-month gain for any month since April 1992….” This is an example of taking advantage of data volatility stemming from small sample size. Using the same approach, one could argue that the job loss from March to April of 21,900 is the largest month-to-month loss since April 2009, the height of the Great Recession.

• “Wisconsin ranked 21st in total private sector jobs created from December 2011 to 2012 with a gain of 32,282.” Using absolute numbers, rather than percentage gains, moves Wisconsin up in the rankings since most states are smaller. Also not mentioned is that this job gain represents a percentage increase of 1.01%, compared to 1.77% for the average American state.

• “Wisconsin’s total private sector job gain outpaces Iowa and jumped 11 spots in percentage of private sector job growth to 33rd highest at 1.4 percent, higher than neighboring Illinois.” Since January 2013, Iowa has gained jobs at a higher rate than Wisconsin, but it is smaller so using absolute numbers gives the advantage to Wisconsin. Illinois has gained at about the same rate as Wisconsin; because it is larger, its absolute gain is larger.

• “… Wisconsin gained over 62,000 private sector jobs in the last two years. The private sector job gains under Governor Walker are the best two-year gains under any Governor in over a decade.” The comparison is a mystery to me. There are several two-year periods that beat this number. For example the two-year period that includes Doyle’s last year and Walker’s first had a private sector gain of 63,200.

There are several tricks used here. One is using numbers of jobs when comparing Wisconsin to smaller states such as Iowa and switching to percentage changes when comparing Wisconsin to larger states like Illinois. Another is avoiding any comparison to national job gains. Finally there are numerous comparisons to job losses during the Doyle administration, with no hint that Doyle had the bad luck to be governor when the Great Recession hit. It’s cherry picking of data.

Some may take the poll as evidence of conservatives’ weakening grip on reality, similar to the widespread insistence that Mitt Romney was about to win the last election. It is hard to know without looking at states with Democratic governors and weak job growth. For example, Illinois’ job growth has been very similar to Wisconsin’s over the past two and a half years but has a Democratic governor. Would a majority of Democrats insist that job growth meets or beats that of other states? Does the Illinois jobs agency, in reporting monthly results, spin them to favor the governor?

One question is whether a state agency, funded by taxpayers, should be in the business of spinning data for partisan purposes. While I appreciate the secretary’s desire to make his boss look good, there are number of organizations in the private sector that would be happy to take up the task. Instead they let the department do the work for them.

Another is whether people in politics have an obligation to be straight with the people who support them. In this example, those most likely to be misled are supportive of Walker. Among other factors, those less supportive likely get their information from elsewhere. Is it ethical to reward one’s most faithful followers by misleading them?

Finally, ignoring facts can lead to bad public policy. As Daniel Patrick Moynihan observed, “Everyone is entitled to his own opinion, but not his own facts.” In last year’s presidential debates, Mitt Romney had his own version: “You’re entitled, Mr. President, as the president to your own airplane and your own house, but not to your own facts.” If each party feels free to create its own facts, effective solutions to problems such as how to encourage job growth become even more intractable than they already are.