Tuesday, August 31, 2021

Is turnout going to be decisive in the recall? A Deep Dive into the data and what we can learn from California's voting patterns

(Sorry, but I better put a link for the book up top -- and it's also available to read for subscribers on Taegan Goddard's Political Wire).

Throughout the Newsom recall campaign, there has been a focus on the importance of voter turnout. There is an expectation that, due to the fact that it is a special election, there will be low turnout, which would be an advantage for the recall proponents (who are already motivated to come out). It is true that many legislative and local recalls have low turnout, but this has not been the case for past gubernatorial (or even high-profile mayoral) recalls. In fact, in each of the three gubernatorial recalls, we saw a turnout boom. We've also seen that officials are slightly more likely to survive a special election recall vote than they are one that takes place on a general or primary election day.

How would decreased turnout help the recall proponents and increased turnout help Newsom? Let's look at the numbers and how it played out in 2003 to see what numbers both sides may need to hit to win.

Play the numbers, play the odds

In 2003, the vote went 55.4-44.6% against Davis. We have one big exit polls from that election to work with, the Edison/Mitofsky Poll. Among voters, 88% of Republicans voted for the recall, while only 76% Democrats voted to oppose it. Effectively, Democrats were twice as likely to vote in favor of the recall as Republicans were to oppose it.  No Party/Others were in favor of the recall 53-47%. If voter turnout was the same across the board, based on registration numbers below, under those percentages Davis would have done three points better, though still losing with a 47.4% margin. Instead, the Republicans’ clear turnout advantage pushed down those numbers.

In the 18 years since, voter registration numbers have changed radically. Democrats now make up 46.5% percent of registered voters and Republicans only 24.1%. The Democrats do not have an absolute majority, as 23.3% list No Party preference and 6.2% are in the Other category.  How do these numbers compare to 2003 and what does it mean for the recall? Well, Democrats haven’t done so much better in the years since – in 2003, they were at 44.1%. But Republicans have cratered, dropping more than 11 points – down from 35.3%.

If we use the 2021 registration numbers, and take the same percentage breakdown as in the exit poll, we can see how much how much has changed and how turnout can matter. Assuming equal turnout, we have an almost five point swing in the Democrats’ favor, with Newsom winning with 52% of the vote. But again, turnout wasn't equal, and Davis lost three points due to differential turnout. If the same thing were to happen here, that could be the difference maker in a tight race.

Low Cards are what I've got, but I'll play this hand whether I like it or not

But that part may be academic. I suspect we’re very unlikely to see a similar partisan breakdown in a heightened political environment, which means that Democrats are more likely to vote for Newsom and Republicans are more likely to vote against. 

Looking at the recent polls, we see a range of numbers on this: The Berkeley IGS poll, which has the race at 50-47, suggests that 97% of Trump voters support a recall and 86% of Biden voters oppose one. The Emerson poll had it at 48-46% with a seemingly rosy bi-partisan view -- Republicans split 80/16 in favor of a recall, and Democrats 23/73. Independents actually push the recall forward with 54/34 numbers, with an enormous undecided number there. There are two blow out polls: Survey USA shows Newsom losing 40-51%, with the Republicans up 8:1 in favor of the recall and Newsom only at 3:1 among Democrats (mimicking the exit polls from 2003). Change Research has Newsom up 57-42 (though they don’t seem to have a partisan breakdown). The CBS Poll, which has the race at 50-47, sees 92-8 recall vote from Republicans, 15-85 against the recall from Democrats and 49-51 against from independents. 

All use a likely voter number that benefits the pro-recall forces and all show that the Republicans are more likely to vote for the recall than the Democrats are to vote against it -- which frankly feels correct based on those exit polls and the more hunkered down nature of the political environment. But the more polarized the electorate, the better for Newsom. If we use that 2003 baseline, he has a pool of 24 points from his party to try to gain support from. Republicans can only gain from those 12 percent of nos.

One possible thought that could comfort Republicans is that the party’s numbers have fallen so much, primarily to the No Party/Other preference because, of social pressure – people do not want to publicly identify as Republicans in a Democratic region. If this is the case, perhaps the Republicans can do better with the non-Democrats. But recent election numbers don’t seem to bear out any such thoughts from voters. Thanks to California’s top-two system, and looking at the presidential races, we have a close approximation of the recall and how a no-party or independent voter casts their ballot. And they seem to split close to the middle, arguably with a slight lean to the Democratic side.

In 2014, Jerry Brown won 60-40. The registration numbers were a bit better for Republicans that year -- 43.4% Democrat, 28.2% Republican. In 2018, Newsom won 62-38, and the registration numbers show Republicans' fall -- 43.8 percent Democrat, 24.5 Republican. But in each case, the two Democrats outperformed what we may think based solely on registration numbers.

So either Democrats had better turnout, Republicans were more likely to vote Democrat or those independents split in favor of the Democrats. Republicans are believed to have better turnout in California than Democrats. If so, than it would suggest that no-party/other breaks even more for Democrats.

But the recent general election numbers may mean something as well. I’d prefer to do this as a table, but it seems challenging on the blogging program, so bear with me. We’re going to look at the presidential and the gubernatorial numbers. Turnout for the presidential race is always going to be the highest.

Let's look at the actual numbers as well, including the presidential totals. The 2020 presidential election saw a voting boom on both sides of the aisle. That is almost certainly represents the ceiling of what the candidates can hope for from just their supporters. Biden got 11,110,250 and Trump got 6,006,429. The 2016 election saw Clinton at 8,753,788 and Trump at 4,448,3819.  In both of these elections, we saw about a 30 point gap in the Democrats favor.

Gubernatorial races, especially during the top-two era, show a similar picture. 2018 actually saw very high turnout, with Republican John Cox topping Trump’s 2016 numbers. Cox got 4,742,825, though that was a drop in the bucket compared to Newsom’s 7,721,410 – Newsom won by almost 24 points. 2014, the first year of top-two, saw much lower voter numbers, as Brown won 60-40, though the raw numbers were 4,388,368-2,929,213.  

If the presidential race is really a ceiling, the six million is probably a key figure. Pro-recall forces need to get those voters out, while Democrats have quite a bit more of a cushion, as Biden got over 11 million and Newsom got 7.7 million. The challenge for the Republicans is whether Newsom's number can drop enough for them to win. 

Can persuasion work? I haven’t done any study, but from memory, the focus on turnout over persuasion seems to have taken off in the post-2000 world – no surprise that this happened after an election decided by 537 votes. I’m skeptical of this overemphasis on turnout for two reasons (in general elections. Primaries are a different animal). For one, increasing turnout tactics can run into Newton’s third law – there is a reaction (admittedly not necessarily equal) and result in a turnout increase on the other side. 2020 is a good example of this. Turnout increases seem to work for at best one election, and that’s usually only if the other side doesn’t have voter anger on their side. The other reason is that persuasion is really valuable. In a zero-sum election, persuading a voter is worth 2 points – one for you, minus one for your opponent, versus turnout, which is just one for you.

The Moral of the Story is Plain to See

During the 2012 recall of Wisconsin Governor Scott Walker, we saw a turnout boom from four million to four and half million. But the results were nearly a mirror-image of the 2010 election, with Walker gaining less than a point. Turnout was not decisive. Persuasion, specifically keeping the independents who voted for Walker the first time, was what worked. I would say the same thing happened with Davis and Frazier in 1921. 

But California today is very different. Wisconsin is evenly divided. California is most certainly not. Turnout can be the difference maker here. Tomorrow, I hope to look at what we can glean from the information about the ballots already handed in.

No comments:

Post a Comment