Your District is More Flippable Than You Think

In our series on voter turnout, we previously looked at the congressional abstention rates (casting a ballot but not voting for a congressional candidate) in New Jersey Municipalities in the 2016 election. After accounting for the statewide trend of higher congressional abstention in municipalities with lower incomes, we found a pretty strong district level trend where turnout was higher in districts that ended up being more competitive races. Since voters can't know the outcome of the election before it happens, any direct effect on turnout must be due to perceived competitiveness. This made us wonder about how perceptions of competitiveness can be manipulated to change the actual competitiveness of a race. With this in mind, in this post we will look at New Jersey's 11th district, which has been held for over two decades by Republican Rodney Frelinghuysen. 

Frelinghuysen has won reelection very comfortably for his whole career, most recently defeating Democrat Joe Wenzel and winning 58% of the vote in 2016. Republicans generally dominate state and local public offices in the district as well, and the conventional wisdom is that the district is a fairly safe Republican seat. As far as we can tell however, the safety of Frelinghuysen's seat depends on his campaigns consistently turning out a Republican base that is not nearly as large as Frelinghuysen would like opponents to believe. This job is made even more difficult because he needs to convince substantial numbers of these core supporters to consistently vote for the Republican party against their own policy interests. 

Let's jump into some numbers, a poll of registered voters who had voted in 2016 and 2014 was conducted by NJ11 for Change, an citizens advocacy group in the district. This poll design will tend to skew results towards the preferences of the more conservative voters in the district who turn out more reliably for mid term elections, but the results can still be reasonably interpreted as representative of registered voters in the district. Here is a rundown of opinions on several policies which are at the heart of the Republican's national agenda, with a 3.5% margin of error:

  • 4.8% think there are no circumstances when abortion should be allowed
  • 21.2% want to see New Jersey's comparatively strict gun laws relaxed
  • 21.5% approve of cuts to Medicaid
  • 22.9% believe Planned Parenthood should be barred from receiving federal funding
  • 24.4% approve of a 30% cut to the Environmental Protection Agencies superfund program
  • 25.5% want and outright repeal of the Affordable Care Act
  • 34.3% want a border wall with Mexico

So how is any Republican candidate able to get 58% of the vote in a district with such abysmally low approval for Republican policies? The answer is because 58% of the vote corresponds to 35.8% of everyone who was registered and therefore eligible to vote in the 2016 NJ11 congressional race. When we consider everyone who did not vote in the election, we see low voter turnout makes the size of the Republican base in the district appear much larger than actually is.

While the poll did not cover all of the issue that NJ11 constituents care about, there is a clear that discrepancy between Frelinghuysen's support at the ballot box and support for Republican policies. This could be explained by a polling error that is more than double the average state level polling error from the 2016 election, but it is more likely that Frelinghuysen is picking up a pretty substantial chunk of voters based on partisan identity, name recognition, and incumbency advantage rather than actual policy preference. He is also benefitting from promoting the perception that he will be the inevitable winner, which as we found previously, could be depressing turnout among opponents in the district.

It's also possible that he is simply flying under the radar of many of the district's voters. There are some striking patterns in congressional abstention in NJ11 municipalities if we look by county. NJ11 contains pieces of Morris, Sussex, Passaic, and Essex counties, and abstentions are much higher in Passaic and Essex than in Morris and Sussex. This probably has something to do with name recognition. The Frelinghuysen family is very well known in Morris, particularly in the Southern part of the district near his hometown of Harding. In addition, parts of Essex and Passaic county were drawn in to the 11th district after the 2010 redistricting, and so it's likely that constituents in these places are less familiar with Frelinghuysen and his record. The congressional turnout divided by the presidential turnout for each municipality in the 11th is mapped below. The pre 2010 district lines are shown in red, and darker shading corresponds to fewer people casting a ballot but abstaining from the congressional race.

It clear that these patterns in turnout correspond to a lot of votes, but we'd like to go a step further and try to estimate a concrete number. We can do this with a regression model that is similar to the models we have used in previous posts. The goal is to determine what the turnout rate would be in Passaic and Essex municipalities if those municipalities followed the turnout patterns of municipalities in Morris and Sussex. In order for this to work, we need to have some method for determining which municipalities are comparable to one another across counties. Income is strongly associated with turnout, so that's what we'll use to determine what a municipality's turnout rate "should" be, assuming that Morris and Sussex set the standard. 


The regression trendline is shown in red along with the turnout data (congressional votes per registered voter) for each municipality. The trendline represents what we expect the turnout rate to be in a municipality in Morris or Sussex given it's level of income. We know the incomes of the municipalities in Essex and Passiac, so we can use the trendline to estimate what the voter turnout would have been if those municipalities followed the turnout patterns of Morris and Sussex. The difference between the actual turnout and the Morris-Sussex trend is about 22,000 votes, just shy of 7% of all votes cast in the congressional race. While that's not enough to tip the election and of course many of those potential Essex and Passaic voters would have voted for Frelinghuysen, that's a pretty astonishingly large number of votes not cast considering that the reason behind it may simply be that fewer people in Passaic and Essex are familiar with Frelinghuysen and his votes. When taken together with other factors like the lack of support for many Republican policies in the district, Democratic challengers building momentum, and huge opportunities to increase turnout by engaging more middle and working class constituents, Frelinghuysen is in trouble. Then the icing on the cake it that the GOP tax reform proposal has the potential to hit Frelinghuysen's tax weary voter base.

So how is Frelinghuysen dealing with all this? He is without a doubt very familiar with all the information pointing to his vulnerability that we've discussed here. He has a tricky job though, since he needs to communicate urgency to his base in order to keep turnout among his supporters up, while simultaneously discouraging that anyone else turnout by telling any potential opposition to just not bother. In order to manage these conflicting messages, he's hired Mike Duhaime, a career political consultant who was a top strategist for NJ Governor Chris Christie. Governor Christie is of course famous for several brilliantly strategic moments in public relations such as calling a guy a big shot while holding nachos, as well as calling a guy a big shot while holding ice cream, so we know that Frelinghuysen has hired top notch talent.

Berating constituents while holding junk food is not quite on brand for the more soft spoken Frelinghuysen, but we can see these dual messages at work. Compare the two statements below. The first is from a fundraising letter sent to one of Frelinghuysen's local donors, which became public due to a handwritten signature outing one of the NJ11 for Change "Ringleader" Saily Avelanda who was an employee of the recipient and ultimately resigned from her job as a result of the letter. The second is Mike Duhaime's response to the poll conducted by NJ11 for Change, which in addition to the policy questions we already discussed, showed Frelinghuysen trailing an unnamed Democratic challenger.

From the closing paragraph of Frelinghuysen's fundraising letter:

"We cannot let the Washington-controlled special interests dictate the will of the people in this district. It does not matter who the competition is, I have and never will take my election for granted. That is why I am counting on you. I need resources to fend of any attack I might receive. This fundraising season needs to be my strongest yet and I want to start it early."

And from Duhaime, disputing the polling results

"This partisan democrat front group (NJ11 for Change) only selectively released one ballot question taken only after giving a series of negative push questions against Rodney Frelinghuysen.  This gives a deliberately false impression to trick donors and reporters into thinking a race is closer than it is" and "Rodney will defeat any candidate who faces him next year."

To be clear, these "negative push questions" were questions on Frelinghuysen's yea vote on the house version of Trumpcare and on his refusal to hold in-person town halls. Therefore these questions are only negative if constituents disagree with his behavior, but that's beside the point. A perfectly reasonable response to the poll would have been to point out that early polls are not very predictive of election outcomes and that many respondents were undecided. However that's not good enough when the last thing on earth you want is donors to spend their money on a challenger and reporters to cover the race, even if it's just out of curiosity. More attention just means Frelinghuysen runs the risk of more people finding out about his terrible voting record and that there are more than enough people in the district to vote him out. It's much better for him to publicly present himself as the inevitable winner and present any signs of credible opposition as the result of trickery from interests outside the district, while he quietly tells his ever shrinking group of supporters that the sky is falling.

Every day more constituents are recognizing that Frelinghuysen is not representing them. It's likely that many of those people assume that popular support for the Republican agenda will let Frelinghuysen carry the district regardless of how unhappy they are. Getting the word out that such popular support simply does not exist will be a key factor for bringing change to the 11th in 2018 and beyond. We also suspect that NJ11 is not unique, and that there are many federal, state, and local races around the country where the illusion of popular Republican support discourages would be challengers. 

You can find the source code for the analysis in this post on GitHub.