Unions and the Democratic Vote

First off, some housekeeping.  The posting here has been extremely light as of late.  The reason for this is that I've recently joined the ranks at the Washington Examiner's Opinion Zone blog.  Thus, much of what has been running through my head has been appearing over there.  I'm going to work hard to keep posting over here as well.

The big story of the past few weeks has been the debate about the role of unions, specificially public sector unions, in our political system.  As numerous states, most notably Wisconsin, try to fix budget deficits, public sector unions have been called upon to increase their contributions to health care, pension, and other benefit packages.  While Wisconsin unions have agreed to these concessions, legislation that would take away their collective bargaining rights have set off a firestorm, resulting in large scale protests and a crippled state government.

In the commentary surrounding this, the role of unions more broadly in our politics has gotten quite a bit of attention.  I've written about this here and here.  To summarize, unions have been a key component of the Democratic Party's electoral coalition even as their membership has been declining and has been becoming more public than private sector oriented.  To get a sense of this, I thought I'd run an experiment whereby I rank states according to their level of unionization and see if there is a rough correlation with their statewide vote in presidential contests.  Thanks to the good folks over at the Bureau of Labor Statistics, one can get this data quite easily.

I've put together the following chart which looks at the past five presidential elections.  For each year I've ranked the 50 states plus DC according to the percentage of the population that belongs to a union--represented in the second column for each year.  I've then color coded the state for how it voted.

One thing to note is that while there is some variation as to the relative rankings of the states as to their level of unionization, by and large the state orderings remain constant.  High unionization states include New York, Hawaii, Michigan, Alaska, Washington, and California while low unionization states include the states of the deep south.  As you can clearly see, highly unionized states have tended to vote Democratic (and vice versa for low unionization states).  This obviously isn't a huge surprise but it is nice to be able to view the data this way.  Consider that in 2008, Barack Obama won 23 of the 27 most unionized states in the country.  He also, interestingly, won 2 of the 4 least unionized. 

As I was creating this chart, the Clinton elections of 1992 and 1996 jumped out as having seemingly less of a correlation between state unionization and the vote.  Clearly there was some regional appeal taking place with his wins in Georgia, Louisiana, and Arkansas (wins that no subsequent Democratic nominee has been able to replicate).  However, at the top of the scale, Clinton did just as well as Obama.  Thus, equating national or statewide electoral outcomes simply with union presence is a mistake.  There are a variety of other variables at play--a point made the other day by Nate Silver in an excellent post on the subject.  Especially when we have elections that verge on entering "landslide" territory, isolating one variable and assigning causality is likely to steer us in the wrong direction.

Nonetheless, it shouldn't surprise us that the proposals unleashed over the past weeks by Republican governors across the midwest--where presidential elections are ultimately decided--have set off the reaction they have.  Unions clearly see these bills not just as part of a debate about fiscal policy, but as a more fundamental attack on their role in society and politics.