Donald Trump’s surprising candidacy can instill a premonition of greater uncertainty—and a larger error term—in 2016 than normal.
Good morning Austin:
Pinch yourself, Texas.
Donald Trump, the Republican candidate for president of the United States, the only Donald Trump ever to run for president and the least couth major party nominee since Andrew Jackson, is devoting a precious day of general election campaigning to our less than humble state, and, best of all, treating Austin, Texas, which, when it comes to Donald Trump, has done nothing to earn it, with the special favor of a genuine, bona fide Donald Trump mass rally at the Expo Center – the best political entertainment value to hit Texas since Pass the Biscuits Pappy O’Daniel.
Here, if you want to prep for the rally, is last night’s rally in Akron.]
Right now, that view would make Trump an outlier – the general view being that the Trump campaign is headed to a bad end – but who knows. I call your attention to the quote atop today’s First Reading – that Donald Trump’s surprising candidacy can instill a premonition of greater uncertainty -—and a larger error term—in 2016 than normal – and no, that premonition of greater uncertainty, is not, so far as I know, a borrow from Tennessee Williams, but rather an original coinage by two academics – the larger error term is the giveaway – namely Columbia University political scientist Robert Erikson, and University of Texas government professor Chistopher Wlezien.
Erkson and Wlezien are two of the most prominent and most accurate election forecasters, and while their forecast is that Clinton is most likely to be our next president, they are not alone in believing that Trump is not a forecaster’s friend.
University of Virginia political scientist Larry Sabato has been compiling and publishing forecasts on his Crystal Ball site.
Here is the Crystal Ball Electoral College projection as of late last week.
Let’s suppose Trump gains steam in the fall, maybe after some entertaining or overpowering debate performances. (This is a hypothetical, not a prediction.) Where could he grab states currently in our Democratic column? Three big states that are perfectly capable of voting Republican stand out: Florida, North Carolina, and Ohio. To this let’s add Iowa. So far Trump is faring better in the Hawkeye State than in other competitive swing states. It’s not difficult to see why. Almost half of Iowa’s electorate will likely consist of non-college whites, while minority voters will probably comprise less than 10%. Clinton has never been a favorite at Iowa’s caucus time, though she secured a paper-thin victory in February 2016 after a third-place finish in 2008. Democrats will have to work hard to gain these six EVs in November.
Returning to ground level, however, suppose Trump wins all the Romney states (206 EVs, which includes North Carolina and NE-2) and he adds Florida, Ohio, and Iowa. Trump will be at 259, still short 11 EVs. It isn’t at all obvious where the extra 11 would come from, though the easiest path might be Nevada (six EVs) and New Hampshire (four) to produce a 269-269 tie. Presumably, the House of Representatives will remain Republican and at least 26 states will have a unit vote in favor of the GOP and Trump. Presumably. Or will there be defections in a few strategically placed states?
And wait — didn’t we just change New Hampshire to Likely Democratic? Daydreams and nightmares don’t last long in the August hothouse of 2016.
And here are the forecasts Sabato has published so far.
On the face of it, two models predict a Democratic victory, and two predict a Republican victory.
Last night, Chris Wlezien provided me with the updated, and final version of their projection, which relies on both Leading Economic Indicators and trial-heat polls and comes, with a slightly higher level of confidence, to the same projection.
From Erikson and Wlezien’s final projection:
The value of the model is early prediction. For the past five presidential elections, we have used our model (as updated at that time) to predict the November vote using only Quarter 13 cumulative LEI growth and either presidential approval or trial-heat polls from Quarter 15. Our public forecasts have been close, picking the correct popular vote winner each time with an average absolute error of 1.6 percentage points of the two-party vote.[i]
[i] Our one major prediction error was 2000, an election that foiled all forecasters. We overestimated the Gore vote by 5.2 points. (In that year, we used presidential approval as our indicator of public opinion; using trial-heat polls, as we have since, the forecast would have been a smaller, but still sizable 3.7 points.) Our other forecasts using LEI were much better, producing an absolute error of under 1.0 in three years – in 1996, 200 and 2012 – and a middling error of 1.5 points in 2008.
PREDICTING 2016 USING LEADING INDICATORS AND QUARTER 14 POLLS
What does our model suggest for 2016? For 2016, cumulative LEI growth is 0.22, just slightly below the average of 0.23 over the sixteen elections between 1952 and 2012. Using RealClearPolitics.com, Quarter 14 Clinton-vs.-Trump trial-heat polls average 54.2 percent for Clinton. Plugging in our model from Table 1 with trial-heat polls measured for Quarter 14 yields 52.2 percent as Hillary Clinton’s predicted share of the two-party vote in 2016, which is only a little more than we predict (51.8 percent) with LEI growth alone. Based on the standard forecast error, the estimate implies a 76 percent chance of a Clinton popular vote victory.[i]
In recent election years, the national party conventions were held in August or even into September. In 2016 they occurred in July, the earliest in the calendar since 1960. Historically, the conventions have considerable impact, with the leader in the polls afterward almost always winning the election (Erikson and Wlezien, 2012b). This can be seen in Table 2, which displays poll shares for the incumbent party candidate from one week before the first convention and then two weeks after the second convention, along with the final vote. [i] There we can see that the leader in the polls before the conventions ultimately won the popular vote in 11 of the 16 elections; after the conventions, the leader won the popular vote in every year, bearing in mind that the polls were tied in 1980.
[i] The pre-convention measure is for the week ending the Monday before the start of the first convention. The post-convention measure is for the week starting the second Tuesday after the second convention. Only live-interviewer polls are included. Where data are missing for some years (no polling in the designated week), we substitute the most recent poll (pre-conventions) or the next poll (post-convention). The data of the poll is always the midpoint of the reported polling period.
So, what did our model say about 2016 before the conventions began? Our 2016 pre-conventions trial-heat reading, from RealClearPolitics.com is 51.5 percent of the two-party vote for Clinton. Plugging into our pre-conventions equation in Table 3 predicts 51.8 percent for Clinton, with a probability of winning of .72. This is only slightly less than what we forecasted in early-June, as discussed above.
What about after the conventions? For the dates August 9-16, fully two weeks after the Democratic convention ended, the one available poll (from PEW) indicates 52.6 percent for Clinton.[i] Inserting the number into our post-conventions equation in Table 3 predicts 52.0 percent for Clinton, with a probability of victory of .82. Once again, this is little changed from what we predicted prior to the conventions and also earlier, in June. Our electoral expectations have remained quite stable.
[i] Per our past practice, we only consider live-interview polls, which rules out internet polls and others, which are slightly more favorable for Clinton, e.g., the RealClearPolitics average on August 22, 2016, implies a 53.1% share.
And then, the Trump caveat and the premonition of greater uncertainty.
We close with an obvious caveat about forecasting the presidential vote in the unique election of 2016. The theoretical underpinning of forecasting models is bolstered by arguments that each party runs a typical campaign that is supported by party elites. Donald Trump’s surprising candidacy can instill a premonition of greater uncertainty—and a larger error term—in 2016 than normal. Our model partially captures a Trump effect by the incorporation of trial-heat polls, which reflect Trump’s support at the moment. With trial-heat polls in the equation, the error term represents the effects of cumulative campaign shocks from the date of the poll to Election Day. The possibility of greater campaign effects than we typically observe should constrain our confidence in the early predictions presented here.
For Emory University political scientist Alan Abramowitz, the premonition is so strong as to lead him to rethink the usefulness of his forecast.
Here is Abramowitz’s explanation of his model.
The Time for Change forecasting model has correctly predicted the winner of the national popular vote in every presidential election since 1988. This model is based on three predictors — the incumbent president’s approval rating at midyear (late June or early July) in the Gallup Poll, the growth rate of real GDP in the second quarter of the election year, and whether the incumbent president’s party has held the White House for one term or more than one term. Using these three predictors, it is possible to forecast the incumbent party’s share of the major party vote with a high degree of accuracy around three months before Election Day.
And here is what his model would forecast.
Based on a net approval rating for Barack Obama of +6 in the Gallup weekly tracking poll for the week of June 27-July 4, an estimated second quarter change in real GDP of 1.2% according to the Bureau of Economic Analysis, and the fact that Hillary Clinton is seeking a third consecutive Democratic term in the White House, the Time for Change Model predicts a narrow victory for Donald Trump — 51.4% of the major party vote to 48.6%.
And here is his giant Trump caveat, which essentially leads Abramowitz to declare his forecast null and void thanks to Trump.
Time for Change vs. Time for Trump
Despite the excellent track record of the Time for Change model, there are good reasons to be skeptical about the 2016 forecast. For one thing, the overwhelming majority of national polls during the spring and summer of 2016 have shown Clinton leading Trump. National polls completed shortly before and after the national party conventions gave Clinton an average lead of about five percentage points, and Clinton is up by about eight points now. Beyond the poll results, the Time for Change forecasting model is based on two crucial assumptions — first, that both major parties will nominate mainstream candidates capable of unifying their parties and, second, that the candidates will conduct equally effective campaigns so that the overall outcome will closely reflect the “fundamentals” incorporated in the model.
While the assumptions of the Time for Change model are generally realistic, they will clearly hold to varying degrees in different elections. An examination of the error terms in Table 2 suggests that candidates and campaigns do sometimes have noticeable effects on the outcomes of presidential elections beyond what would be predicted based on the fundamentals. In 1988 and 2000, for example, poor campaigns very likely contributed to smaller than expected vote shares for Democratic candidates Michael Dukakis and Al Gore. And in 2008, the somewhat smaller than expected vote share for Democrat Barack Obama may have reflected the reluctance of some white voters to support the first African-American nominee of a major political party.
The nomination of Trump by the Republican Party in 2016 appears to violate both of the Time for Change model’s key assumptions. Trump is clearly not a mainstream Republican and he does not appear to be running a competent campaign — he has lagged far behind Clinton in both fundraising and grassroots organizing in the swing states, and his rhetoric on the campaign trail has frequently brought sharp criticism from prominent Republicans as well as Democrats. In fact, there has never been a major party nominee like Trump — a reality TV star and wealthy businessman with no longstanding ties to the Republican Party, no political experience, and a penchant for insulting major voting groups. As a result, many prominent Republican leaders, including the last two Republican presidents, and the party’s 2012 nominee have refused to endorse Trump.
In recent months, Trump has received the highest unfavorability ratings of any major party nominee in the history of the Gallup Poll. Clinton also receives high unfavorability ratings from voters; however, Trump’s ratings have generally been far worse than Clinton’s. According to the most recent Gallup data (Aug. 3-9), Trump had a net favorability rating of -31, while Clinton had a net favorability rating of -17.
The question is how much the Republican Party’s nomination of Trump will move the needle away from its slight tilt toward the GOP based on the fundamentals in 2016. There is no way to answer this question until after the election. Based on the results of other recent presidential elections, however, as well as Trump’s extraordinary unpopularity, it appears very likely that the Republican vote share will fall several points below what would be expected if the GOP had nominated a mainstream candidate and that candidate had run a reasonably competent campaign. Therefore, despite the prediction of the Time for Change model, Clinton should probably be considered a strong favorite to win the 2016 presidential election as suggested by the results of recent national and state polls.
We’ll finish up with Nate Silver’s latest election forecast at FiveThirtyEight, from Sunday, with its own big, fat Trump caveat.
Hillary Clinton moved into a clear polling lead over Donald Trump just after the Democratic convention, which ended on July 28. Pretty much ever since, the reporters and poll watchers that I follow have seemed eager to tell the next twist in the story. Would Trump’s numbers get even worse, possibly leading to the first double-digit victory for a Democratic presidential candidate since 1964? Or would Trump mount a comeback? As of last Tuesday, there wasn’t much evidence of an overall shift in the race. Trump was gaining ground in some polls but losing ground in a roughly equal number of them.
Since then, Trump has gotten some slightly better results, with national polls suggesting a race more in line with a 5- or 6-percentage-point lead for Clinton instead of the 7- or 8-point lead she had earlier in August. But state polls haven’t really followed suit and continue to show Clinton with some of her largest leads of the campaign. Trump received some decent numbers in Iowa and Nevada, but his polls in other swing states have been bad.
Overall, Trump has gained slightly in our forecasts: He’s up to a 15 percent chance of winning the Electoral College in our polls-only model, up from a low of 11 percent a week ago. And he’s at 25 percent in polls-plus, up from a low of 21 percent. But the evidence is conflicting enough that I don’t think we can rule out a larger swing toward Trump or, alternatively, that his position hasn’t improved at all.
Let’s start with those national polls. In the table below, I’ve listed every national poll that we’ve added to our database since Tuesday and how it compared to the previous poll from the same pollster, if there was one.1
And here is Lady Bird’s wonderful home movie of the 1941 Senate race LBJ lost to Pappy O’Daniel, who, as Lady Bird puts it, “flashed like a comet across the Texas political scene.”