Featured Post

The Anchor and the Waterline

Given that midterm elections have long been thought of as referendums on the president’s party, it seems like a worthwhile venture to take a look at Donald Trump’s job approval rating in the states holding senate and gubernatorial contests this November. By examining the results of past contests, and comparing them to state-level presidential job approval, patterns emerge that offer insight into what to expect after the dust settles on November 6.

For the purposes of this article, exit polling from the last four midterm elections was used to determine the president’s job approval rating in any given state (2014, 2010, 2006, 1998). Because of issues with the data at the time it was taken, exit polling from the 2002 midterm election is excluded.

To begin, three important rules emerge from looking at presidential job approval ratings as compared to midterm senate and gubernatorial results:

  1. When the president’s job approval rating is under 45% in a particular state, there’s a strong chance (85%) that the non-presidential party will either flip that seat to their side, or at least hold the seat for their party.
  2. When the president’s job approval rating is between 45-49% in a particular state, the odds of the non-presidential party flipping or holding the seat are about even (56% of the time).
  3. When the president’s job approval rating is at 50% or greater, the odds of the non-presidential party flipping or holding the seat are only a little better than one in three (38% of the time).

To see how I arrived at these conclusions, consider the tables below, which document the number of senate and gubernatorial contests that fell into one of these three ranges of presidential job approval in the last four midterm elections.

The Anchor and the Waterline

The fifth table combines the data from the previous four midterm cycles into one, making for larger sample sizes. The overall trend holds across a total of 210 exit-polled senate and gubernatorial contests.

The Anchor and the Waterline

As you can see, over the last four midterm cycles, there were 89 exit-polled senate and gubernatorial contests in which the president’s approval rating was under 45%, and the party out of presidential power won 76 of them. There were 43 exit-polled contests in which the president’s approval rating fell between 45-49%, and the party out of presidential power won 24 of them. There were 78 exit-polled contests in which the president’s approval rating was at 50% or higher, and the opposing party only won 30 of them.

What are the implications for Republican prospects in senate and gubernatorial contests this November, given the opposing party’s performance within a range of presidential job approval over the past four midterm cycles? That depends on Donald Trump’s state level job approval rating.

The table below documents just that – the president’s average approval rating in every state holding a senate or gubernatorial contest this November.

The Anchor and the WaterlineThe Anchor and the Waterline

There’s some obvious bad news here for President Trump. His approval rating is currently under 45% in more than half (56%) of the senate and gubernatorial contests to be held in November. Why is that bad for the president? Because the party out of presidential power wins senate and/or gubernatorial midterm contests 85% of the time when the president’s approval rating is under 45%.

The good news for President Trump? It could be worse. Take the 2014 midterm election, for example. Barack Obama’s job approval rating was under 45% in 70% of the exit-polled senate and gubernatorial contests that year. Republicans ended up gaining 9 senate seats and 2 governor’s mansions. Likewise, George W. Bush saw his approval rating under 45% in 65% of exit-polled senate and gubernatorial contests in 2006, earning Democrats 6 senate seats and 6 governorships.

The table below illustrates presidential job approval in exit-polled senate and gubernatorial contests from the 2014, 2010, 2006, and 1998 midterms, and compares that to President Trump’s job approval rating in states holding contests this November. As you can see, Barack Obama in 2014 and George W. Bush in 2006 were very unpopular in the majority of exit-polled contests. Bill Clinton in 1998 was quite the opposite. President Trump’s job approval rating, as of today, falls in between them.

The Anchor and the Waterline

 

Based on what we know from past midterm exit polling, what does Trump’s state-level job approval ratings tell us about what to expect on election night? In a nutshell, it confirms what we already knew. If presidential approval polling was our only metric to go by in predicting the results of upcoming senate and gubernatorial contests, it suggests that Democrats will pick-up a whole bunch of governorships, and will lose three senate seats (making the composition of the Senate 54 Republicans, 46 Democrats/Independents). This is roughly in line with most forecasts, if not slightly less optimistic for Democratic chances of taking the upper chamber than actual state polling averages of senate and gubernatorial contests.
The Anchor and the WaterlineThe Anchor and the WaterlineThe Anchor and the Waterline

As you can see from the table above, current senate polling averages suggest that if the election were held today, Republicans would end up with 49 senate seats, Democrats would end up with 50 seats, and 1 seat would be a toss-up. The polling suggests that Democrats would flip Arizona, Nevada, and Tennessee. Republicans would flip North Dakota. Missouri would flip from blue to toss-up.

Regarding gubernatorial contests, the table above suggests that the Republican’s 33 seat majority would drop to just 24 governorships if the election were held today. The number of Democrats in governor’s mansions would increase from 16 to 25. There would be one toss-up. Current polling suggests Democrats would flip 9 state governorships – Nevada, New Mexico, Kansas, Iowa, Wisconsin, Illinois, Michigan, Georgia, and Florida. The Republicans would pick-up one seat in Alaska. Maine is presently a toss-up given the only poll we’ve had of the race.

The bottom line is that with President Trump currently under the critical 45% approval mark in over half of the contests taking place this November, Republicans have a lot of exposure. While the current composition of the Senate doesn’t look as though it will change all that much, even a minor shift could give Democrats control of the chamber. And the partisan composition of state governor’s mansions is likely to look very different on Wednesday, November 7. The Trump Admin has two things on their side – time, and the fact that they are not currently in as bad of shape, from a job approval perspective, as Barack Obama and George W. Bush were at this point in 2014 and 2006, respectively. But time can be a double-edged sword. Things could get better for Republicans, or worse.

Ask if you would like access to the data used in compiling Donald Trump’s state-level job approval ratings, or if you would like access to the data used to compile state averages. Data for the latter mostly comes from public sites such as Real Clear Politics, Wikipedia, and Dave Leip’s US Election Atlas.

Please do be so kind as to share this post.
Share

18 thoughts on “The Anchor and the Waterline

  1. The Democrats will probably not flip the Senate but I think there are some real chances for a dark horse.

    One of the things I have said over the past few months is that I don’t think the GOP and/or GOP-sympathizing people realize how pissed off Democrats and/or liberals are. There is a lot of anger and rage and fed-up ness on the left side and these are pretty good motivators to get people to the pols. Democratic loathing of Trump exceeds anything under Bush II by several magnitudes. It is extending to the GOP too. People used to refer to the House Freedom Caucus and now I am seeing it written as House “Freedom” Caucus. Meaning left-leaning Americans and voters are launching challenges at the right for having a monopoly on how to define “freedom” and “liberty”

    The other aspect that might increase Democratic turnout is the Kavanaugh hearing and the handling of Dr. Ford’s accuations via outright dismissal.

    The Senate map is highly favorable to the GOP but the most at-risk Democrats are doing very well and Democrats look like they might be able to pick up a red seat or two. Beto O’Rourke might still be an underdog but he is giving Cruz a run for his money. Cramer had his Aiken moment last week.

    The GOP refuses to get this stuff though.

    Report

      • You could be right. My fear is that this becomes a close but no cigar election for Democrats. So losing 49-51 percent to Ted Cruz is an improvement but still a loss. We have to wait and see though.

        Kevin Drum had some polling showing that support for the Republican party among all women has collapsed except rural women and this was pre-Kavanaugh.

        Report

  2. I can’t believe I didn’t see this a couple of weeks ago.

    I go to 538’s webpage talking about Trump’s approval and jump back and forth between “All Polls” and “Polls of Likely or Registered Voters”.

    I look at the page for the Senate. The page for the House.

    I think about what we knew two years ago.

    What we knew a month ago.

    I think about “undecideds”. I wonder who in the heck would say “undecided” even though it’s 2018.

    Report

Leave a Reply

Your email address will not be published. Required fields are marked *