As the events in Wisconsin have been transpiring over the past weeks, several of my more Democratically inclined friends and relatives have been asking “How did we ever elect Scott Walker?”
As I’ll suggest, they might not like the answer.
First, it should be noted that they elected him quite easily. Statewide, Walker bested Democratic nominee Tom Barrett by about six points. Compared to the size of Barack Obama’s win in Wisconsin two years earlier, this was a massive turnaround. To help understand how it happened, we need to dig into the numbers a bit.
Over the past few days I’ve been playing around with the Wisconsin election results to try and figure out just what went on.
Woody Allen famously said that “90% of life is just showing up.” Elections are pretty much the same folks. For those bewildered by Walker’s success, they need to first think long and hard about turnout.
When we look at who actually showed up to vote last November we can get a sense of where his victory came from. There are a few ways that we can think about and measure turnout. Michael McDonald at George Mason University, in looking at those eligible to vote found that Wisconsin turnout was 51.7%, about ten points better than the national average. In my look at the numbers, I decided to look at turnout as a proportion of registered voters.
Registration in Wisconsin is among the easiest in the country, with same day on site registration allowed. Among those registered, turnout was nearly 58%. For a midterm election this level of turnout is extremely high no matter how it’s measured.
But who really turned out?
Breaking things down a bit, turnout varied across the state from a low of 39% in tiny Menomonee County to a high of 71% in suburban Waukesha County. On first blush, one might guess that turnout was highest in the most Republican counties and that this produced Walker’s win. In part, this is true. The heavily Republican Waukesha, Washington, and Ozaukee counties which ring Milwaukee ranked first, seventh, and eighth in turnout.
In fact, the congressional district in which these counties lie had the highest turnout of any district in the country. Furthermore, these are relatively large counties as well, providing about 13% of the votes cast in November. So obviously Walker did really well in turning out the people most inclined to vote for him.
However, when one looks at the entirety of the state’s counties one doesn’t find a correlation between turnout and how it voted. Walker’s victory took place in both high and low turnout counties. The same can be said for his opponent Barrett.
Even though Barrett won only 13 counties, he won the two largest in the state, Milwaukee and Dane (Madison), by large margins. This is something that every successful Democratic nominee must do in order to win the state. However, it wasn’t as if turnout in these areas was low. In fact, both Milwaukee (62.7%) and Dane (65.7%) counties had turnout well above the state average. Unlike defeats that can be attributed to your voters not showing up, that wasn’t quite the case here.
With all of this data I tried to create a scenario in which Walker could have lost as a result of greater Democratic turnout. As I’ll explain this was extremely difficult. With the help of recently released Census data I calculated the number of non-voters in each county (the difference between the population of those 18 and older and the number of votes actually cast). Then I ran an experiment to figure out how much more turnout would have been needed in either Milwaukee or Dane counties to give the Democrats the governor’s mansion.
Consider…if every adult in Milwaukee County who did not vote (370,000 people) did in fact vote, (assuming they voted as the rest of the county did—62% Barrett / 37% Walker), Walker still would have won the election with over 36,000 votes to spare.
Now one might rightfully argue that those who didn’t vote were probably more Democratic than those who did. Turnout, especially in midterms tends to be lower for minority and low income voters. In Wisconsin, most of these voters could be found in Milwaukee County. Also, much discussion was given to the “enthusiasm gap” between Republican and Democratic voters.
While we can’t know for sure how the non-voting population broke down, we can nonetheless calculate how Democratic the non-voters votes would have needed to be. Specifically, if everyone who didn’t vote in Milwaukee decided to show up, they would have had to give Barrett at least 68% of their vote for him to become governor. That’s a pretty high hurdle to overcome, even in a highly Democratic area like Milwaukee.
Looking at Dane county, the state’s second largest and the one which gave Barrett 68%, the math is in fact even more difficult. For Walker to have lost, all of Dane County’s non-voters (about 162,000 people) would have needed to split their vote 89%/11% in favor of Barrett!
So focusing on turnout only gets us so far in explaining Scott Walker’s victory.
Rather, what seems to matter perhaps more than turnout is the fact that a lot of people who voted for Barack Obama in 2008 turned around and voted for Scott Walker last year. This phenomenon was hit upon the other day by Nate Silver. As he notes, most of the non-Milwaukee and non-Madison parts of Wisconsin are highly competitive. Especially in the close 2000 and 2004 elections, Al Gore and John Kerry were able to point to these outlying areas as the reason they were able to squeak out victories. Lots of these voters would not identify themselves as strong Democrats but have nonetheless shown a tendency to vote that way quite regularly.
Further evidence of this can be seen by comparing exit poll results from 2008 and 2010 and looking at a few key groups of voters. Among independents, for example, there was a dramatic shift as the support for both Obama and Walker. Each received 58% of independents’ votes. In the most recent polling on the budget repair bill, 68% of independents wanted Walker to compromise with Democrats on collective bargaining rights.
Similarly, white voters with less than a college degree supported Obama with 52% in 2008 yet turned around and gave Walker 58% last November. Thus, the fluidity among this part of the electorate must be noted to help us further complete the picture of how Walker was able to win.
Who might some of these white non-college voters be? Some of them surely belong to unions. As Craig Gilbert of the Milwaukee Journal Sentinel noted as the protests in Madison began to heat up, Scott Walker won 37% of the union vote. Though this is 2 points less than they gave John McCain in 2008 it is a sizable chunk of the electorate.
Thus, as the Wisconsin protests move into their next phase of court challenges and recalls, many of those who have been the loudest and most active may want to spend some time reckoning with an underlying truth about how Scott Walker became Wisconsin’s governor.
To quote the great cartoon Pogo, “We have seen the enemy and he is us.”
