What will Turnout look like in Louisiana's U.S. Senate race?
Among the people who write about Louisiana politics, this year’s U.S., Senate race has been an underwhelming event. Folks have called it "sleepy", "quiet", and "sedate". I’ve been right there with them, calling it "late". It’s not just us, a double digit share of undecided voters with just two weeks to go in recent polls is a sure sign the campaign has failed to capture voters’ imaginations.
Nevertheless, it is on its way to one of the highest voter turnouts the state has seen in a U.S. Senate race during the past 40 years…only to be followed by the largest drop in turnout from the November primary to the December runoff that Louisiana has ever seen in a statewide race.
Presidential elections drive turnout up, way up
The presidential contest is the big game when it comes to voter turnout, and many voters show up to cast ballots. Since 2000, the national percentage of the voting eligible population (VEP) that cast ballots in presidential elections is nearly 20 points higher than the percentage in midterm elections. More than 1.9 million Louisiana voters cast ballots in each of the last three presidential contests. Only about 1.5 million voted in the 2014 U.S. Senate race.
Higher turnout in presidential elections spills over into other contests that share the ballot. Each U.S. Senate seat will share the ballot with a presidential election one out of every two times it comes up for a vote (i.e., every twelve years). That has not always been the case in Louisiana. In the past 40 years, senate races in Louisiana have almost never appeared on the same ballot as the presidential contest – even when they have been held in presidential election years.
This is because of a series of changes to the state’s election system starting in the mid-1970s. At that time, the state moved to its unique ‘jungle primary’ and set the dates for the first round of voting in U.S. Senate and House elections about a month before federal Election Day (the Tuesday after the first Monday in even numbered years). Runoffs, if necessary, fell on federal Election Day.
Four of Louisiana’s seven U.S. Senate elections fell on a presidential election year during this period: 1980, 1984, 1992, and 1996. The rules in place, however, meant those primaries were held in September or October instead of November. Turnout in these primaries was 22.5 points lower, on average, than national turnout on federal Election Day. [The 1996 race resulted in a runoff held on the date of the presidential election, more on this below.]
In the late 1990s, we moved the calendar back: The first round of voting would be on federal Election Day and runoffs would be in December. We used this system for three senate races (1998, 2002, and 2004) before changing again. For the 2008 and 2010 senate elections we used an entirely different system that followed the rules used in just about every other state: Party primaries in the spring or summer followed by a general election on federal Election Day. By 2014, however, we were back to a ‘jungle primary’ on federal Election Day.
Looking at the 2004 primary and the 2008 general election in Louisiana - the only two held on the day of the presidential election - it is clear how much of an impact a presidential election has on turnout. They had the highest turnout, by far, of any U.S. Senate elections in the state during this period.
Turnout in this year’s senate primary will also see high turnout because it too will share the ballot with the presidential election.
Louisiana really, really likes voting for president
In presidential elections, Louisiana generally meets or exceeds the national turnout rate. Even more interesting, Louisiana exceeds the turnout rate we should expect given the state’s demographic characteristics. States in the south generally have lower turnout than the nation as a whole (although the gap has shrunk in recent elections); this is largely due to lower levels of educational attainment and household income, which are strongly tied to turnout.
In recent decades national turnout for presidential elections ranges from a low of 52 percent of VEP in 1996 to a high of 62 percent in 2008. Louisiana’s presidential turnout ranges from a low of 54 percent in 1980 to a high of 61 percent in 2008.
There is some debate about whether we should expect lower or higher turnout this year than the typical presidential election. For example, the Pew Research Center report on the 2016 election shows that satisfaction with the candidates is low, which may indicate less engagement in the election. On the other hand, the same report also shows more interest in the election, more concern about who will win, and greater attention to the campaign that other recent elections.
However, it is important to remember that ‘low turnout’ and ‘high turnout’ are relative terms. Even a ‘low turnout’ presidential election would have significantly more voters than an election without a presidential contest on the ballot.
December will be an entirely different story
Louisiana’s senate race appears to be headed for a runoff in December. When that happens we will almost certainly see the largest drop in turnout from primary to runoff for a statewide election in state history.
There is not a consistent trend in turnout across the two stages of major statewide elections (governor and U.S. Senate races). Gubernatorial elections have seen an increase in the number of voters from the primary to the runoff, but these elections do not fall on federal Election Day at either stage.
The trend for U.S. Senate elections appear far more sensitive to issues of timing relative to federal Election Day. The number of voters in the 1986 election rose by 15 percent from the September primary to the runoff held on federal Election Day. The boost was even larger in 1996 when the runoff fell on the same day as the presidential election; 38 percent more voters cast ballots in the later round than had in the first round.
The other two U.S. Senate runoffs – when the primary was held on federal Election Day and the runoff held in December – saw fewer voters show up in the second round than in the first round. But even these are not a good gauge for how much turnout will drop this year between the primary and the runoff. Neither the 2002 nor the 2014 primary fell on the same day as a presidential election, but the 2016 primary will - so it has much more room to drop.
This scenario - a senate runoff following a month after a primary that was on the day of a presidential election - would be unprecedented in Louisiana history. So, there is no direct example to use as a model for expectations this year.
There are, however, three races for seats in the U.S. House of Representatives that had primaries on the day of a presidential contest and then runoffs in December: the 2004 election in the Third District, the 2004 election in the Seventh District, and the 2012 election in the Third District.
In each case, the number of voters decreased dramatically. Just half as many voters cast ballots in the 2004 Seventh District runoff as had in the primary; 69 percent fewer voters cast ballots in the 2012 runoff for the Third District than had in the primary.
U.S. House races often draw less attention that U.S. Senate races, so these drops are probably larger than what we will see this year. They should be taken as an upper limit or ceiling for the range of possible shifts.
More likely, if presidential turnout is at typical levels, the percent drop in the number of voters will be in the thirties (based on the reverse of the 1996 boost or the difference between the number of voters in the 2012 presidential election and the 2014 U.S. Senate runoff). If presidential turnout is on the lower end this year, then the percent drop in the number of voters from primary to runoff in the senate race could be in the twenties.
Either way, it will almost certainly be the largest drop on record for a U.S. Senate race in the state.
Of course, there is always the chance – albeit an extremely small one – that the December runoff could determine which party controls the U.S. Senate. That would require having both a Democrat and a Republican make the runoff and either Clinton winning the presidency and Democrats holding exactly 49 seats in the U.S. Senate after November 8th or Trump winning the presidency and Republicans holding exactly 49 seats.
In that case, the election will be anything but sleepy.