State and national politics are not as different as you might think
To many followers of politics, myself included, a Democrat’s victory in Louisiana’s gubernatorial election came as something of a surprise. The surprise was not so much what happened on Election Day; by then, the polls had consistently shown John Bel Edwards well ahead of Republican David Vitter for weeks and there was little reason to doubt the polling.
The surprise came weeks earlier when it first became clear that Vitter was in trouble. Vitter’s baggage was well known, but he had successfully – even easily – survived reelection to the U.S. Senate in 2010.
For me, the question is not simply: 'Why did a Republican candidate struggle in a Republican state?'
Instead, the truly intriguing question is: 'Why did a Republican candidate who coasted to victory in 2010 struggle in a Republican state in 2015?'
The most tempting explanation - one I have heard quite a bit both during and after the campaign - is that 2010 was a U.S. Senate race and 2015 was a governor's race. The idea, as I understand what I have heard, is that voters think about state politics differently than they think about national politics and, therefore, do not make the same decisions in these two kinds of elections.
I am skeptical of this argument. While I readily concede that there are some important differences between state and federal elections, I do not think it is because voters are making different kinds of judgments in the two kinds of elections. Voters are pretty darn consistent in supporting their party in presidential, congressional, and gubernatorial elections - not to the same degree, but we're talking about the difference between 90% partisan support and 85% partisan support. Most of the time, voters tend to end up in the same place regardless of whether it is an election for state or federal office.
But, obviously, that did not happen in Louisiana in 2015. Do Louisiana voters treat state and federal elections differently? Two empirical tests say 'nope'.
Test #1: Legislative vs Presidential Contests
Let's start by comparing the most localized state government elections (the state legislature) to the most national (president). If voters come at state and national politics with different considerations, then we should definitely see that in this comparison. To do this, I pulled the 2012 presidential election results for each legislative district in Louisiana (available from the hard-working folks at Daily Kos) and matched them against the outcome of this year's legislative elections.
The partisan overlap in the outcomes is widespread. When the new legislative term begins in January, the Louisiana House of Representatives will include only 12 Democrats from districts carried by Mitt Romney in 2012 (Districts 10, 18, 28, 30, 32, 33, 38, 42, 50, 54, 60, and 75). The House will include just one Republican from a district carried by Barack Obama (District 92). Out of a total of 105 seats, 90 will be held by the same party as the presidential candidate who carried the district (86% of the chamber). [The remaining two seats will be held by independents in districts carried by Romney.]
The share is even higher in the state senate, where 92% of seats are held by the same party as the presidential candidate who carried the district. Only three senate districts carried by Romney will be represented by a Democrat (Districts 19, 28, and 38), and no senate districts carried by Obama will be represented by a Republican.
In other words, electorates within legislative districts that tend to be Republican in presidential elections also tend to be Republican in elections for state representatives and state senators; likewise for Democrats. This does not mean that voters were thinking about Obama or Romney when selecting legislators in 2015, but this degree of consistency casts serious doubt on the idea that voters treat elections for state and federal offices differently.
Test #2: What Could Have Been in 2016… Might Not Have Been So Different than What Happened in 2015
The second test comes from a study that I did in collaboration with Martin Johnson, a colleague of mine at LSU’s Manship School of Mass Communication. We conducted an experiment during the final days of the gubernatorial campaign (before Vitter announced he would not seek reelection to the U.S. Senate) to see how voters might react to Vitter if he sought reelection next year.
We asked approximately 350 registered voters how they might vote in next year’s senate race, but we did not ask everyone the question in the same way. We randomly assigned some of our participants to answer version 1: “Next year, Louisiana will have an election for the U.S. Senate seat currently held by David Vitter. If David Vitter does NOT seek re-election to the U.S. Senate, do you plan to vote for a Republican or a Democrat in the 2016 U.S. Senate election?” Other participants answered version 2: “Next year, Louisiana will have an election for the U.S. Senate seat currently held by David Vitter. If David Vitter seeks re-election to the U.S. Senate, do you plan to vote for Vitter (who is a Republican) or a Democrat in the 2016 U.S. Senate election?”
Comparing responses across the two versions of the question in our study tells us how Vitter performs when voters are primed to think about an election for national office. We measure performance as the difference between the generic Republican's share of support in version 1 and Vitter's share of support in version 2.
If voters just do not want Vitter in the governor's office but are willing to give him a pass as long as he stays in D.C., then we should not see him do worse than the generic Republican in the hypothetical 2016 U.S. Senate race. However, if Vitter under-performs relative to a generic Republican in this experiment, then that suggests he has issues that even an election for national office does not mitigate.
So, what happened? Interestingly, in the overall sample there was not much a difference between the generic Republican candidate's support in version 1 and Vitter's support in version 2. But among Republican voters... my, oh my! Republican voters who were asked version 1 gave the generic Republican 12 percentage points more of their votes than Republican voters who were asked version 2 gave to Vitter. Republican voters like a generic Republican candidate significantly more than they like Vitter in an election for office in the federal government.
Maybe 2010 Was a Weird Election Too
David Vitter’s struggles had a lot more to do with David Vitter than they had to do with a distinction between state and federal elections. In addition to the experiment described above, Martin Johnson and I also looked at voter decision-making in the final weeks of the campaign. In a report from the Reilly Center for Media and Public Affairs released earlier this week, we show that voters evaluated Edwards more favorable than Vitter on personal traits (e.g., honesty, having strong leadership qualities, caring about their needs, and sharing their values) even though they were closer to Vitter ideologically and trusted him more on certain issues. This mattered because voters were far more likely to base their decisions on the candidates' personal traits than they were to base them on the issues that favored Vitter.
Okay, so Vitter suffered in 2015 because he was a weak candidate and the election focused on him rather than on issues where voters may have agreed with him. But none of that explains why Vitter did so well in 2010.
The answer, I think, does have to do with differences between state and federal elections - just not in the way those differences have been characterized. It's not that voters always come to the two kinds of elections with different mindsets - a nationally focused mindset that emphasizes partisanship in even numbered years and a state focused mindset free of partisanship in odd numbered years. If they did, then the presidential election results would not look so darn similar to the legislative election results.
Instead, the difference between 2015 and 2010 probably has to do with how the timing of elections shapes the possibilities for framing the terms of choice. Louisiana holds its state elections in odd years, removed from presidential or midterm election years when most states hold their elections. Because of this timing, Louisiana's state elections are probably more insulated from the national swings that can favor one party over the other when a bunch of U.S. Senate and House races are on the ballot together.
Vitter's successful reelection was not just any year. It was 2010. That was a really, really, really good year for Republicans nationally. It was the first midterm election of the Obama Administration, when the president's party usually takes losses, and the Tea Party was on the rise. The 2010 U.S. Senate election was framed as a referendum on President Obama rather than a referendum on Senator Vitter. Those circumstances do not necessarily arise every federal election year (after all, Democrat Mary Landrieu last won a U.S. Senate race at the same time that Louisiana voters rejected Obama in 2008). Nor, are partisan predispositions confined only to federal elections (after all, Republicans swept the statewide elections in 2011 and took both chambers of the legislature).
To be absolutely clear, I am not suggesting that a Democrat's win in the 2015 gubernatorial election means that Democrats are any more favored in next year's U.S Senate race. Quite the contrary, I am arguing that 2015 was an unusual election which Vitter lost in large part because it was about him. But 2010 may have been a bit unusual too - not because Vitter won a contest that was a slam dunk for Republicans anyhow, but because the circumstances were particularly ripe for him to win in a way that masked his own weaknesses as a candidate. Those weaknesses finally caught up to him five years later when there was no national tide to lift his boat and the focus fell solely on him.