top of page
  • Writer's pictureOren Levin-Waldman

Labor Market Demographics May be a Better Predictor of Who Wins an Election

As the battle for the Democratic party nomination for President intensifies, much attention is paid to public opinion polls. In fact, it has become an obsession. As important as traditional public opinion polling may be, understanding economic transformations that have occurred in particular geographic areas should be as important. The reality is that pollsters missed the 2016 election, and they appeared to be at all attuned to the effects of the labor market on the political environment. While the polls could gauge preferences with regards to candidates, they failed to capture the anxiety felt by many voters, particularly traditional blue-collar voters over economic transformations due to our global economy. Following the 2016 election, I analyzed data from the Current Population Survey (CPS) for Michigan, Pennsylvania, and Wisconsin. The analysis concluded that economic changes, particularly the loss of manufacturing and other high paying blue collar work, left many displaced with a feeling that the political elite directed little attention to their plight.

Let’s for instance consider the following: If craftsmen can be defined as skilled labor, there was a 5.9 percent decrease in Michigan; a 11.7 percent decrease in Pennsylvania; and a 7.8 percent decrease in Wisconsin. In the U.S. the percentage decrease in craftsmen was 11.5 percent. Still, there were fewer craftsmen in the U.S. as a whole than in any of these three states. Operatives might also be another category of skilled workers. Again the number of operatives in all of these three states is higher than in the U.S. as a whole in both 2008 and 2016. But the percentage decrease in operatives was greater in these states than in the U.S. as a whole.

Meanwhile, in Manufacturing, which traditionally was where middle class and unionized jobs could be found, there was a .8 percent decrease in Michigan; a 14.5 percent decrease in Pennsylvania; and an 8.1 percent decrease in Wisconsin. It should be noted that much of the manufacturing base in Michigan declined in the years prior to 2008. Still, the percentage decrease in manufacturing was in the U.S., although lower than in Pennsylvania, but higher than in Michigan and Wisconsin. For the exception of Michigan in 2008, there was more manufacturing in these states than in the U.S. as whole. And yet, what appears to be most interesting is the percentage increase in Mining in both Pennsylvania and Wisconsin.

Between 2008 and 2016, those working in Mining increased by 14.3 percent in Pennsylvania and by 50 percent in Wisconsin. Mining relative to other industries is still low. Only in 2008 did Michigan have more mining than the other states and the U.S. as a whole. Michigan also had a 66.7 percent decrease in mining between 2008 and 2016. The state with the least mining was Wisconsin, but it had the greatest percentage increase.

When seen in the context of a Democratic candidate who suggested to mine workers that they need to not only get used to doing something else, but if elected she would pursue policies aimed at eliminating their jobs, it might well have been a contributing factor to these states flipping red. Not only was the Democratic candidate not addressing the anxiety of workers displaced by global market forces, but she was heightening them.

The outcome of the election clearly demonstrated the myopia of the Clinton campaign in not bringing that message to these voters. Although public opinion polls are useful, the 2016 election clearly demonstrated their limitations. Conclusions from data analysis support the importance of campaigns having access to the most recent labor market demographics and the commensurate effect on voter choices.

For those running for President, a similar demographic analysis should be conducted for every state in the union based on the most recent data. The typical poll will simply ask voters how they feel about certain issues. But a serious demographic profile can help the pollster ask the correct questions in the first place. This is the type of information that could in the first instance be critical to framing the questions for more traditional opinion polls. It is this vital information that can be obtained from this type of profiling which has the potential to significantly inform campaigns as they develop their strategies for future elections.

13 views0 comments


bottom of page