If there’s some strategy afoot to have Democrats raid the May 16 Republican primary to have a preferred candidate win that nomination since their field is so crippling weak, rank-and-file voters of that party aren’t cooperating.
Speculation has risen about the impact of party registration, or raiding, with the reinstalment of closed primaries for congressional contests in Louisiana. With the new rules such as they are, there’s not much incentive for unaffiliated voters to pick a major party label as they can choose which party primary (carrying through with the choice if a runoff emerges) in which to participate.
But with incumbent Republican Sen. Bill Cassidy under duress for nomination, some observers have wondered whether he would make explicit appeals to non-Republicans to vote for him. He apparently already has done that in one media appearance.
Regardless of whatever efforts his campaign has made along these lines, a rough test can be performed of whether there are movements afoot for increased unaffiliated voter registration and GOP registration at expense of Democrats by use of registration data past and present. This investigation makes sense with the presence of closed primaries and the recognition that a Democrat nominee will lose handily to whoever wins the Republican version.
This may be tested by analyzing data from federal elections since 2008. Louisiana has had closed primaries for presidential nominations for several decades, so by reviewing these from the past two decades it can establish a baseline as to how the presence of a looming nomination contest can affect registrations. There also is a slight point of comparison in 2008 and 2010, as in those cycles were the state’s previous closed primaries for congressional elections.
However, note that the dynamics of the presidential contest will drive the process whenever the two levels of offices overlap: voter registrant motivations in 2008 for most individuals likely overwhelmingly came from a desire to have a voice in nominating a presidential candidate in two competitive contests. Still, the 2010 Senate contest that featured a pair of semi-competitive nomination races can provide data worth extra scrutiny.
By reviewing the presidential nomination contests, a baseline of change may be derived as a sort of normal registration pattern. Averaging them out can provide a smoothed estimate taking into account that they vary in degree of competitiveness, with really only the 2008 election the exception to just one-party competition in presidential nominations as this featured no previous officeholder (Hillary Clinton in 2016 served as a surrogate) plus spirited competition.
The longitudinal analysis also acknowledges the long-term trend that has afflicted Democrats of steady loss of registrants at the expense of the Republicans and unaffiliated registrants. The analysis looks at registration totals, on the date the state reports these publicly, between 30 and 60 days prior to the nomination election (as registration closes 30 days prior) and then at the appropriate date three months prior. That starts the date for the analysis at November, 2007, at which point the Democrats had just under 1.5 million registrants, with Republicans at just under 700,000 and almost 70,000 fewer unaffiliated voters than those with the GOP.
Those numbers held fairly steady through 2010, after which the bottom dropped out for Democrats who saw a precipitous decline as a result. The last data point, the numbers at the start of this month, show Democrats well below 1.1 million with Republicans within thousands of reaching parity. Unaffiliated registrants have encroached well above 800,000. Thus, observing the nine data points between and inclusive of 2008 to 2024 (two in 2008 because of differing primary dates, none in 2018 because in that midterm election was not a Senate race) and taking the mean averages the Democrats’ loss and gains by the others can indicate whether 2026 appears exceptional.
For each data point, the difference between the two totals prior to the primaries three months apart was compared for each category and mean differences calculated for presidential (six observations) and midterm (three cases) elections. Doing so produces the following:
Presidential years – Democrats: -0.20 percent; Republicans: +0.96 percent; unaffiliated: +0.67 percent
Midterm years – Democrats -0.23 percent;
Republicans: +0.42 percent; unaffiliated: +1.03 percent
These aren’t surprising. Considering presidential election years, Democrats lost consistently in the three-month span, likely as they were switching to be able to participate in the Republican primary. In fact, out of all nine, in only one period did they increase (although they barely declined in three of the presidential ones): the 2014 cycle where Democrats unsuccessfully tried to save Democrat former Sen. Mary Landrieu’s seat from Cassidy. But their 0.36 percent gain trailed both the Republicans’ 0.51 percent boost and very much behind the 1.69 percent increase by unaffiliated voters (to clarify, the state reports in this category every voter who is not registered as a Republican or Democrat).
Returning to presidential cycles, Republicans gained in all those plus in the other and unaffiliated voters gained in all except 2020, plus all of the other. If Democrats fled their previous label because they felt the party no longer represented them and so they wanted to influence matters in the party that did, the switch to the GOP would make sense. Likely most of the unaffiliated pickup and some of the GOP’s came from new voters disproportionately choosing these statuses over being a Democrat (until this upcoming election, party primaries were completely closed for presidential nomination).
Reviewing the off-year results, the increase for Republicans on average was less than half than during the presidential cycle, indicating that there was greater desire to participate in a presidential primary with less necessity of switching if left only to the blanket primary. The near-equal proportion of Democrat decrease as with presidential election years would indicate a similar lack of enthusiasm for the party but with increased mobilization efforts in the quadrennial cycle able to offset some switching. Add to that the much higher unaffiliated voter increases would argue that, without a primary, new voters in particular disproportionately chose that designation without the allure of a nomination for either major party to settle.
Using these numbers compared to the 2026 changes can provide some circumstantial evidence as to whether Democrats are switching to raid into the GOP primary. That would be indicated if the registration of Democrats change decrease is significantly greater than the average for a midterm election and if the increase for the other two categories higher and more like the average seen for Republicans like that of presidential years and even more elevated for the unaffiliated, as perhaps the unaffiliated being more attractive since under the new law this can be a halfway house for Democrats who want to raid into Republican contests yet reserve the right to participate in matchups of Democrats here and there.
In fact, the Apr. 1, 2026 data verify half of the equation. Democrats’ registration dropped dramatically from Jan. 1, 0.90 percent, the largest of all ten observations. However, they didn’t seem to go much of anywhere. Republican registration increased only 0.35 percent, or below the typical midterm change, while unaffiliated voters increased only 0.62 percent, even further below the typical midterm change.
This created an unusual situation for this period. In the runup to elections, typically an increase in registrations occur (which then declines after the election). But the registrant total from January to April of this year, because of the smaller-than-usual increases for Republicans and the unaffiliated and the highest decrease for Democrats, actually fell 0.03 percent. In these ten observations, for the entire electorate that happened only once – perhaps not coincidentally, in 2010, the only other year where there were party primaries for congressional candidates without an attached presidential election, falling the same percentage.
And the main driver then again was large falloff of Democrats. That year, their total fell 0.56 percent, exceeded only in 2024 slightly which would be understandable given the lack of enthusiasm among Democrats for reelecting their president and (eventually substituting in) vice president two years ago. In 2010, the GOP gain proportion was identical to 2026, and the unaffiliateds’ were a bit higher then, but still well below the midterm average.
This means that while there has been no great exodus of Democrats registering as Republicans or as unaffiliated in order to raid into the GOP primary and affect that outcome, it does mean that Democrats are atypically discouraged and simply dropping out of the electorate, through a combination of disproportionately lower new registrants, lower former registrants having moved and reregistering, and of them moving out of state. Except for those who vacate the state (and evidence is actually higher-income individuals, who less likely are registered Democrats, disproportionately leave Louisiana), these others likely are less-attentive voters so with the low-quality Democrat field plus no desire to support any Republican, for those who move in state they disproportionately don’t bother to reregister and as-yet unregistered voters see no compelling reason to register.
It’s important to note limitations that exist with this approach. It relies on aggregate data, or averages of the whole, as opposed to a panel study where each individual voter at the poles of the three-month span were would be asked whether he had switched and even why, so some imputation is occurring. That increases the chances that some idiosyncratic extraneous factor could be at work affecting these results.
That aside, if things like the Cassidy campaign wooing Democrats to switch registrations, organizations allied with Democrats behind closed doors working to do the same, or even attentive Democrats on their own thinking to change registrations at least temporarily to aid Cassidy or at least to keep the most conservative candidate in the primary, GOP Treas. John Fleming, from being nominated, it doesn’t seem to be happening in any significant way. If designed to allow genuine partisans to have the major say in who gets nominated, the system seems to be working.
No comments:
Post a Comment