Princeton Election Consortium

Innovations in democracy since 2004

Outcome: Biden 306 EV (D+1.2% from toss-up), Senate 50 D (D+1.0%)
Nov 3 polls: Biden 342 EV (D+5.3%), Senate 50-55 D (D+3.9%), House control D+4.6%
Moneyball states: President AZ NE-2 NV, Senate MT ME AK, Legislatures KS TX NC

Can Third-Party Candidates Help Save The Republican Downticket?

August 10th, 2016, 8:36pm by Sam Wang

On Monday, the Princeton Election Consortium got its 10 millionth view since it became a WordPress site in 2008. Traffic in July 2016 was over 50 times larger than July 2012. Thank you, both old and new readers!


The Presidential cake is baking. Hillary Clinton’s lead over Donald Trump has increased in national polls to around 7%. Since January, the Clinton-over-Trump national margin has been 4.7 ± 2.3 % (average ± SD). The corresponding SD for January-August 2012, a re-election year, is 1.4%. Before that, Wlezien and Erikson’s data on Presidential campaigns from 1952 to 2008 show higher values for SD in all years, as high as 11.0% (1964). So 2016 is a very stable year so far. As much as 2012, it appears that voter attitudes are extremely well set. This fits with the fact that even after making increasingly incendiary statements, Trump’s support is still around 40%.

Clinton’s advantage in the Electoral College is smaller, as evidenced by a Meta-Margin of 3.7%, which is defined as how much current state poll margins would have to move in order to create an electoral tie. 3.7% may not sound like much, but Clinton’s Meta-Margin has not gone outside the 2.5%-4.5% range since May, when our calculation for 2016 began. All in all, measured in terms of public opinion, 2016 is a contender for the most stable Presidential race since World War II. Of course, we still have nearly three months for movement to occur.

A focus on downticket races is in the air. House Speaker Paul Ryan has raised the possibility that his party could lose its majority in November. He is verbalizing the fear by Republican insiders that Donald Trump could drag the rest of the ticket down with him. National surveys and state-by-state comparisons with 2012 suggest that 20-30% of Republican voters won’t vote for Trump. If those disaffected Republicans stay home in November, that could doom many of their party’s Senate and House candidates.

This is where Evan McMullin, who even political junkies may not know, comes in. McMullin stepped down as chief policy director of the House Republican caucus to run for President. He’s an insider. I’m certain he is not running because he expects to become President. Instead, he’s probably trying to give Republican voters a reason to turn out. If he can get even a small number of Republican voters to show up who otherwise wouldn’t, that might be enough to save a seat. Increase that, and he could put a sizable dent in Democratic gains.

To put it quantitatively: an increase of 1% in the GOP vote share nationally translates to about 6 House seats*. McMullin would only need to make a modest showing in marginal districts to make an impact.

Existing third-party candidates do not seem to matter for the Presidential outcome. National polls (RealClearPolitics) currently show Clinton +7.7% in a two-way race, Clinton +7.1% when Libertarian Gary Johnson is added, and Clinton +7.3% when Green Jill Stein is also added. Trump is lagging worse in Republican states than Clinton in Democratic states, and these candidates may be soaking up undecided voters. If so, their net effect is to give Republican voters somebody to vote for at the top of the ticket. At this point, the national Republican party should want as many Presidential candidates as possible.


Since 1992, two-party Presidential voting and the national Congressional vote have been closely linked:

From 1948 to 1988 (black symbols), the average difference between the Presidential and Congressional popular margins was 13.0%. From 1992 to 2012 (red symbols), that difference averaged only 2.9%. This convergence arose at the same time that the parties became polarized and their policy positions became nationalized. Think of Newt Gingrich’s 1994 revolution and the Contract With America.

In this context, Republicans cannot be pleased to see three major indicators moving in the same direction: (1) the Clinton-versus-Trump Presidential race, (2) President Obama’s approval rating, and (3) the generic Congressional ballot.

Now, it is possible that all of these trends simply arise from differential response, i.e. the possibility that Democratic-leaning voters are more likely to answer surveys, as a consequence of watching a good convention in late July; or that Republican-leaning voters didn’t like what they saw at their convention. The differential-response view is a popular topic among polling nerds, and it such an “afterglow” effect has been suggested to contribute to the days of a post-convention bounce. But it has been 2-3 weeks since the conventions, beyond the suggested afterglow period.

It is also the case that opinion can shift. The graphs above show a persistent effect. If you are a Republican, you are probably not ready to laugh off the whole thing as an illusory shift.

Note that a third type of measure, the Senate estimator and Meta-Margin, has also moved toward Democrats, but not as much. Senate polls are still a bit scarce, and many of those races have not heated up yet. It’s a bit early to tell what is happening there.

In recent elections, the final House popular vote has ended up within 3 percentage points of the last stretch of generic Congressional preference polls. From 1992 to 2012, the House popular vote has differed from the Presidential popular vote by a median of 2 percentage points. In an election held based on today’s national polls (Clinton +7%; generic Congressional preference, Democrats +5%), one could reasonably expect House Democratic candidates to win the popular vote by 5-8%. That is in the necessary range for retaking the House. So you can see why Ryan is concerned.

If Hillary Clinton continues to lead Donald Trump, it is in the Republican Party’s interest to sever its downticket candidates from Trump. However, they have to do so gingerly. After Ted Cruz gave his non-endorsement at the Republican national convention, his favorability numbers immediately dropped by 10 points. Even if 1 out of 5 Republicans want Trump to drop out of the race, that means 4 out of 5 are okay with him. That’s the Republican rank and file today, and Republican officeholders are beholden to them.

Therefore any Republican running for office probably wants to avoid the subject of Donald Trump if possible. This is especially true for candidates who still have primaries ahead, such as Arizona Senator John McCain. Conversely, Democrats should want to tie the entire Republican ticket to Trump. Democrats’ best ally may be Trump himself. His policy speech Monday contained some effort to do become a more conventional candidate. The more Trump does that, the better Democrats should like it.

Comments are closed.

*This value is corrected from the original statement, which said “3 seats.”

Tags: 2016 Election · House · President · Senate

24 Comments so far ↓

  • A New Jersey Farmer

    The Republicans will need to decide in September whether to stick with Trump or to shift significant funds to Senate and House races. The first debate will likely decide that.

    • bks

      First debate is 26 Sept. just six weeks before the elections. Some states may already be taking absentee ballots by then(?).

    • Will Hutchinson

      The Koch brothers have already publicly stated they will sit out the Presidential race and are directing all their firepower to the House and Senate. I suspect other Republican SuperPAC’s are doing the same thing, just more quietly.

    • Craigo

      Virginia and Minnesota early voting is 45 days before the election, Ohio is 35 days. Some states are mailing absentee ballots already.

    • Matt McIrvin

      Massachusetts starts early voting this year, but it’s only an 11-day period (and there’s no mail-in voting without giving a reason).

  • Some Body

    National polls haven’t shown that much stability, even when you ignore the conventions (early July, and now early August, saw a good deal of movement, apparently linked to events outside the conventions themselves). The meta margin moved more slowly, but that might have had to do with the relative scarcity of state polls. In short, I have my doubts about saying this cycle has been all that stable so far.

  • Matt McIrvin

    It’s too late for this guy to get on the ballot in more than a handful of states, which reduces the odds that he can accomplish much.

    • Iain Roberts

      It’s not too late for McMullin to get on the ballot in Utah, which has an unexpectedly close race between Trump and Clinton. AIUI the longshot scenario is that McMullin, a Mormon, wins Utah; the Electoral College is deadlocked; and the election is thrown to the House of Representatives — where McMullin could even win.

      It’s theoretically possible, but in practical terms the GOP might as well be praying for Trump to take a vow of silence and retire to a monastery.

    • Sam Wang

      I do not think that the policy director of the House Republican Conference quit his job in order to give Utah to Hillary Clinton.

  • Amitabh Lath

    I hate to play the “2016 may be different” card but most years the margin is due to independent voters. This year, Clinton’s margin may be due to Republican voters (especially college educated women?) switching sides. If that is the case, they likely would still vote R downballot, wouldn’t they?

    • pechmerle

      Amitabh, we’re just speculating, but if those college-educated Republican women are repelled by Trump, they might also take it out on Republican candidates that they see as linked to the Party’s choice of Trump as the presidential nominee. It wouldn’t take too many of those to make some real differences in some states down-ballot.

  • Ketan

    Is there any correlation between Senate races and House seats?

    Seems like that correlation should be increasing (polarization) but since it wouldn’t be affected by Republicans voting against Trump, it could be used to estimate how voters will vote for House seats.

    Since there’s only D+1% (or so) meta-margin for the Senate, would that suggest that the meta-margin for the House is less than D+5%?

    • Sam Wang

      The total Senate vote is also correlated with House and Presidential national vote. However, Senate polling is still fluid because many nominees are not settled yet. For this reason it is not yet time to make a comparison. However, note that the Senate Meta-Margin is ticking toward Democrats.

  • Frank Schnittger

    Any chance of including a Congressional snapshot figure with the senate snapshot in the banner headline? I know Congressional district polls are hard to find, but an estimated snapshot based on national congressional polls superimposed on district voting patterns in previous elections would give an indicator, even if less reliable than the Senate snapshot. Democrats need to know which seats are in play and where to focus their money. Indeed if the Presidential poll is as settled as you suggest, Hillary should now be focusing her money and attention on winnable Congressional, Senate and Gubernatorial races so that she will have some chance of implementing her policy preferences as President.

  • Jon Radin

    Thank you for your numbers and analysis. I have followed PEC for the last several election cycles and believe you provide the most incisive, impartial, informative analysis available to the public.

  • Scott J. Tepper

    On Politics & Polls#7 you say [Senator] Susan Collins when you mean [Senator] Kelly Ayotte. Collins is on your mind because she didn’t endorse Trump. But she’s not running and Trump didn’t endorse her. Comments there are off.

  • Rachel Findley

    Thank you so much for the careful poll-based analysis. It gives me a place to come without punditry. When I first discovered your (very clever) method for projecting electoral votes without running a gazillion trials, I was hooked. Even though Kerry lost in 2004.

  • George

    Two questions – only indirectly related to your post (but don’t know where else to ask them)
    1. Above you said “Since January, the Clinton-over-Trump national margin has been 4.7 ± 2.3 % (average ± SD)” Does that mean you set your Baysian prior at Clinton +4.7, except with the larger SD your reference elsewhere? If not – what is the Prior set at – or am I totally missing what the Bayesian Prior actually is?
    2. If you choose to – any commentary on what Drew is doing over at DailyKos? Obviously putting in some none-poll data that reduces the Clinton win probability (to 72%, IIRC), but he was pretty accurate for 2012, so sort of hard to argue with the methodology. Just interested in your thoughts, should you be willing to share.


    • Sam Wang

      1. Almost. In fact I use an average that counts different time points according to how well they predict November. Currently that gives Clinton +5.0%, very similar.
      2. Linzer’s approach in 2012 was rigorous and well done. I have yet to look this year’s model over, but I believe he starts with an overall prior that 2016 favors the Republican candidate slightly on grounds of the economy. Against the backdrop of this prior, polls come in. The resulting probability seems reasonable, given the assumptions.

      After November, we could interpret any discrepancy between election outcomes and his prior as the consequence of this year’s special factors: GOP disunity, the quality of the two candidates, and the campaigns they ran.

  • MR

    Though the two maps have the same coloring, the map with area proportional to population seems stuck on 271 “safe” electoral votes for Hillary Clinton, in contrast to the regular map which changes as the projections change.

Leave a Comment