Why was the New Jersey gubernatorial race so close?
Two state governors were up for election Tuesday. In Virginia, Glenn Youngkin, a former Republican businessman, defeated former governor and longtime Democratic Party member Terry McAuliffe. The collapse of McAuliffe’s support in the final days of the race, in a state Joe Biden won by nearly ten percentage points last year, and Youngkin’s channeling of the right-wing frenzy on the theory criticism of the breed, sparked a wave of media coverage and analysis. Tuesday’s result was treated as a sort of national political biopsy, a targeted procedure that taught us something about the body as a whole.
In New Jersey, Phil Murphy, the incumbent Democrat, defeated Jack Ciattarelli, his Republican challenger, and became the first Democrat to win re-election as Governor of the Garden State in forty-four years. But the result was much closer than expected. In 2017, Murphy won his first gubernatorial election by fourteen points. In 2020, Biden won the state by sixteen points. On election night this week, the race between Murphy and Ciattarelli was too close to be announced. “You know, we just had the most experience in New Jersey,” Murphy said, in a victory speech he gave Wednesday, after the Associated Press finally called the race. “I was on my way somewhere, and it took us longer to get there than expected.” Friday afternoon, with ninety-five percent of the vote, Murphy’s lead was just over two points.
On Friday, I spoke with Danny Franklin, a longtime pollster for Murphy, and asked him what he thought of the results. Our conversation has been edited for length and clarity.
Can you tell me what the expectations were, prior to Tuesday, and how did those expectations live up to what happened?
No one expected a rash. We weren’t going to expect a twenty, fifteen point win. We expected to win a little more comfortably than we had felt on Tuesday night. I will say that as more votes are counted the margin could end up being close to three points. Which is a different result from what we felt on Tuesday night. But even this margin is a bit narrower than expected. And that night, when we watched the returns come in, I’m not going to lie, it was a surprise.
The public poll suggested Murphy could win by eight points, ten points, something like that. Do you have any idea, at this point, what the poll did not take into account?
So, yeah, the answer to what happened is actually quite simple. There have been historic increases in turnout in Republican counties. And not in democratic countries. Just to give a few details: in Ocean County, where I think the governor did some polls in the thirties and there are a good number of votes, the turnout compared to 2017 has jumped by 30 %. Other smaller Republican counties — Sussex, Warren, Cape May, Monmouth — all places the governor got forty percent of the vote or less, all jumped twenty percent or more in turnout from 2017. Votes are still being counted, but I’d be surprised if one of the major Democratic counties – Essex, Hudson, Mercer, Union – jumps more than ten percent.
Technically, how could the polls have explained this increase?
That’s a really good question. Polls struggle to measure late surges in turnout among a slice of the electorate. We saw it in 2016, in 2010, in 2006. When a wave breaks on one side of the aisle, the polls can miss it. This is more likely to happen in out-of-year elections, in midterm elections, than in presidential elections, because during these years, with lower turnout, slight changes in the electorate can have a very significant impact. Take this election. Ultimately, I think the turnout will be around forty-three percent, 2.7 million votes. If you have a hundred thousand Republican voters who weren’t supposed to vote, and a hundred thousand Democratic voters who were supposed to go out but stayed at home, each of those groups is only four percent of the electorate, but you put these together, there is your difference between a two point win and a ten point win. And this change can happen under extreme circumstances, like now, when there is a labor shortage, inflation, a supply chain crisis and we are facing a pandemic in Classes. We are very focused on the binary question of the Republican versus the Democrat. I think we don’t give enough credit to the broader macro forces.
Well, that’s my next question. In that race, Ciattarelli criticized Murphy over tax rates, a recurring problem in New Jersey. He also criticized Murphy’s handling of the pandemic, an issue very specific to this year. And then, overlooking it all, the national debates on abortion, critical race theory and what’s going on in Washington. How do you analyze what is local versus what is national in terms of what happened on Tuesday?
The short answer is that elections are driven by anger, and the other side had all the gas this year. And it does happen. It is difficult to quantitatively distinguish how different forces have contributed to this, as voters themselves often do not know. Voters will say, for example, “I voted for Ciattarelli because of the property taxes.” Truly? We had high property taxes in 2017 as well, and we gained fifteen points. This problem has always existed. Was it that, or was it the fact that we are in a time where it feels like nothing is working, and we were promised that a pandemic that has dominated our lives for a year and a half will be over? And are we just cranky?
Are there ways to compare Tuesday’s results with the ballot, to try to understand what happened?
So we don’t know much about the election results. We know how many people voted, where they voted and who they voted for. In six months, hopefully sooner, voter registers will be updated. And we’ll know how many Democrats and Republicans voted – not who they voted for, of course, but who came to the polls. We’ll know how many young people voted, how many seniors, and we can start comparing that to our data and see where we were right and where we were wrong. Virginie had exit polls. New Jersey did not have exit polls because no one expected it to be close. So in the absence of that, you’re kind of looking at it from a top of a mountain.
So what can we see now? We can go back to our polls, county by county, and see where we were right and where we were wrong. And we can look at the predictive models of turnout, and see where they were right and where they were wrong. And we can start to isolate where the errors occurred, and make assumptions about it. But then, to your question about the impact of Congress on this, how much was Delta, how much was inflation, that’s a much more difficult thing, because you don’t have good attitude data on this. who was on people’s minds when they voted. You just know what they voted for. I will say, though, that it’s not sort of, you know, a “Home” diagnosis, which takes a full fifty minutes of a TV show to figure out. Republican counties turned out. Democratic counties don’t have that many.