Author Archives: Matthew Dickinson

Why Trump Won: It Was Fundamental!

In life, they say, timing is everything.  When I began doing my election-themed talks in late summer, after it was clear who the general election candidates were, Hillary Clinton consistently held a lead in the various aggregate polling results, such this one by Huffington Post, by about 5%-8%.

pollster-2016-general-election-trump-vs-clinton-5

Nonetheless, I assured my audiences that there were good reasons to expect this race to tighten during the next two months.   As evidence, I cited the political science forecast models which, again looked at in the aggregate, seemed to indicate that this race was going to be a dead heat.  As long time readers know, these models attempt to predict the two-party presidential popular vote as a function of the “fundamentals” – that is,  how well the economy is doing, whether the country is at war, and how long the incumbent party has held the White House, to name some of the most frequently utilized variables.  While not perfect, and keeping in mind that they differ in the particulars, and thus in the final forecasts, these models nonetheless provide a decent template for understanding the context which both candidates then try to exploit in their favor.   Simply put, when things are going well, the incumbent party candidate should try to run a clarifying campaign, to use Lyn Vavreck’s term, while the opponent will seek to focus the message on something else less favorable to the party in power.  Assuming candidates make proper use of these fundamentals, the forecast models issued by Labor Day are a reliable, if not perfect, indicator of how the race will turn out.  With that in mind, the key slide in my talks, which never failed to elicit a crowd reaction, was this one:

poll-of-poll

This was based on the political science forecast models available at the time – subsequent ones changed this aggregate forecast slightly at the margins, but the essential point remained:  this election was too close to call, and could go either way.  That, of course, was not what most of my audiences wanted to hear.  As it turned out, however, the political science models – looked at in the aggregate, (which is how I typically made my prediction every four years) – were spot on in 2016.  As of today, with votes still coming in, Clinton has won about 50.2% of the two-party popular vote – or almost exactly what the political science median forecast predicted.

How, then, to explain those polls showing that Clinton was in the lead?  Early in the campaign season, I told my audiences that, assuming Trump and Clinton ran effective campaigns – that is, that they made effective use of the fundamentals in crafting their respective messages, the polling gap between the two should close.  Indeed, there is extensive evidence from previous elections, as documented by Erickson and Wlezien, that as the campaign progresses, partisans come home to roost in a way that tends to lead to a tightening in the polls. However, as the election droned on, it became increasingly clear that in my talks I had to address the 800-pound hairdo in the room:  Trump was not closing the gap with Clinton nearly as quickly as I anticipated. This was surprising, because as Drew Linzer and others have demonstrated, and as the graph below shows, election polls typically get increasingly accurate as the potential voters begin tuning in and become more informed regarding which candidate comes closer to their partisan leanings. (The y-axis in the graph is a coefficient showing how much the polling aggregate predicts the final Democratic popular vote share.)

polling

The result, as political scientists have documented in previous presidential elections, is that as the campaign heads toward the finish line, and partisans come home to roost, the polls should prove an increasingly accurate indicator of the final vote. Indeed, Drew Linzer had correctly forecast the Electoral College vote in 2012 by updating one particular forecast model (Abramowitz’s Time for a Change model) using only state-based polls. However, in 2016 his state-level poll-based forecast consistently showed a likely Clinton victory all the way up to Election Day.  Indeed on Election eve, he predicted a Clinton Electoral College victory of 323-215.  And he wasn’t the only one to do so – other analysts who had made accurate polls-only predictions in the past, such as Sam Wang, were also forecasting an Electoral College victory for Clinton.

So why wasn’t Trump closing the gap so that polls came closer to the political science fundamentals-only forecast?  I could think of two explanations.  One was that the state polls were somehow off, and were underestimating Trump’s support. In all my talks I reminded my audience that the polls-only forecast depended on the polls being right.  (Not surprisingly, many audience members have forgotten that!)  The second was that he was running a sub-optimal campaign, one in which his continual missteps made it more difficult for him to capitalize on the fundamentals that predicted this was a 50/50 race. In the end, I went with Trump running a suboptimal campaign.  That was probably the wrong choice.  But it’s worth explaining why I made it.

In adjudicating between the two explanations, I compared what I was seeing at Trump rallies with what prior research had shown about polling accuracy at the state level. As I’ve documented in many previous posts, Trump’s rallies were huge and enthusiastic.  And in talking to his supporters, it became clear that the vast majority were not the xenophobic racists that pundits (and not a few of my colleagues) thought they were.  But I worried that focusing on rallies did not give me nearly as accurate a view of the entire electorate as did the polling numbers.   By the end of the campaign, I was concluding my talks by saying this was going to be a close race – one closer than the polls indicated – but unless those polls were completely wrong (and they hadn’t been in the past), Clinton was likely to win the election.  I summarized the Trump-as-poor-campaigner in this slide that suggested the election WAS rigged – by Trump!:

rigged

As it turned out, however, the political science forecast models had it right, and the state-level polls did not.  This is not to say that the national polls were wildly inaccurate. Indeed, as Sean Trende suggests, they were, on the whole, about as accurate as the national polls were in 2012, which on average understated Obama’s final victory margin of 3.9% by about 3.2%. It’s probably worth repeating that, as of this moment, Clinton has a popular vote lead of about 700,000 votes, or about .5% and that could grow to about 1% by the time all the votes are counted.  That’s less than the RCP final four-way average which gave her about a 3.3% margin. But that’s a difference that’s actually a tad smaller than the 2012 RCP error margin. Keep in mind as well that due to the random error associated with statistical sampling, polls in the aggregate don’t usually exactly match the final vote total, even though they typically do approach that total, as I pointed out in my lectures.

But as Linzer showed, in 2016 the state-level polls consistently underestimated Trump’s support, and that miss proved crucial in forecasting key swing states, including Wisconsin, Michigan, and Pennsylvania, which was enough to throw the Electoral College projections in the polls-only models off.  Generally speaking, the polls got Clinton’s support right, but they underestimated Trump’s. This graph by Linzer reveals the extent to which the state-level polls underestimated Trump’s support.

And this Washington Post table shows how much the polls missed Trump’s support in those key states.

I will devote future blog posts to examining why the state-level polls were wrong, as I expect Linzer and others will do as well.  But for now the important takeaway is that, once again, in the aggregate the political science forecast models got this right – exactly right, as it turns out (which undoubtedly will again elicit remarks about how smug we are).  And they did so because this was an election governed by dynamics that were largely unchanged from previous presidential elections, as Larry Bartels points out.  Bartels shows this by comparing Trump’s state-by-state performance with Romney’s 2012 results. As you can see, Trump did well in states in which Romney did well (with Utah a notable exception!) and not so well in the states in which Romney struggled.

The fact that Trump’s performance was both predicted, and that it doesn’t suggest a significant realignment in the electorate, is probably something that political pundits, whose professional existence depends on creating the impression that elections can and usually do change in reaction to every campaign event (Comey cost Clinton the election!), and that the outcome represents something new (and therefore newsworthy) may not want to hear.  But it’s what the evidence suggests.

Yes, some of my colleagues are expressing mea culpas for relying too heavily on the polls in making their final projections.  I understand that sentiment – by the end of the campaign season I was also telling my audiences that although the race would be close, the polls were usually pretty accurate, and that they seemed to suggest a high probability of a Clinton victory. But let’s be clear:  political science got this election exactly right, even if some political scientists (like me!) weren’t smart enough to realize it.  And here’s the proof, again provided courtesy of Linzer:

And this is a reminder that if you want to know who is going to win the presidential election, polls are (usually!) a pretty reliable indicator, although they certainly were not as accurate at the state level this time around as they have been in previous years.  But if you want to know why Trump won, the political science forecast models issued by Labor Day are a good place to start.  And they suggest Trump’s victory was, in large part, fundamental.

The Most Nationalized House Election Since Eisenhower?

It has become fashionable of late, particularly among liberal pundits, to argue that the future of the Republican Party depends on its leaders severing all connections with Trump and his movement.  For a number of reasons that I will discuss in future posts, I think this is profoundly stupid advice for Republicans to follow.  But whatever one’s views on the topic, one thing is clear: if you are House Speaker Paul Ryan, you want to Donald Trump to do well tonight – bigly, even.

The reason is that in recent years, House elections have become increasingly nationalized.  That is, the outcome in any particular House election increasingly is affected by factors outside that district, including how well the House candidate’s party does in the presidential election. There are a number of reasons why this is the case.  A big factor is party sorting, which has made the Democratic Party and the Republican Party increasingly homogeneous in terms of ideology.  That means the party brand name serves as a more important cue for voters during House elections.  A second reason is developments in how campaigns are funded, with an increasing percentage of candidate funds coming from outside the House district as congressional candidates are more likely to look outside their own districts for funding by tapping into more ideologically-oriented issue activist.. There are other reasons that I’ve discussed elsewhere.

Several decades ago, of course, the legendary House Speaker Thomas “Tip” O’Neill proclaimed that all politics is local.  But that has long ceased to be the case, at least when it comes to the House.  Just how nationalized are House elections?  One crude way to estimate the relative proportion of the House vote that can be explained by national and local forces is to regress the House vote in any district against the presidential vote in that particular district as well as the prior House vote. The coefficient for the presidential vote can be viewed as a proxy for national forces, while the House coefficient represents the local component.  I’ve presented previous versions of this data before, but Middlebury College students Tina Berger and Martin Naunov have updated it through the 2014 midterms.  Here’s a chart that shows the relative influence of local and national forces dating back to 1954 for midterm elections.  (Note that I have skipped elections immediately following decennial redistricting, since it is impossible to calculate the prior presidential or House vote for that district.)  These tables only include contested House races.

picture1

And here are the results for presidential election years.

picture3

As you can see, there has been a steady increase in the relative impact of the presidential vote on the House vote for both presidential and midterm House elections, reflecting the increasing nationalization of House races more generally. In fact, 2014 saw the most nationalized House elections dating back to Eisenhower’s presidency. I have no reason to believe that this trend will reverse itself in today’s House races.  For better or for worse, then, the fate of the Republican House majority rests in large part on Donald Trump’s somewhat tiny hands.  Fortunately for Republicans, the much-discussed (by pundits) electoral landslide for Clinton does not seem imminent, at least if Drew Linzer’s forecast model is correct (and it has been consistently the best forecast model out there.)  While Linzer is predicting a relatively comfortable victory for Clinton of about 52% in the two-party popular vote, that’s probably not going to be enough to tip the House to the Democrats. Of course, the better the Donald does tonight, the better for the Republican House majority – at least if recent trends hold.  If I’m Paul Ryan, then, I’m secretly hoping that the Donald does very very well tonight.

I’ll be on a bit later tonight on this site to blog tonight’s election results live from the [Karl Rove] Crossroads Café at Middlebury College.  Hope you can join in!

Why New Hampshire Is Trending Trump – and Why It Likely Doesn’t Matter

For most of this fall in my election talks, I have argued that the most plausible road to victory for Donald Trump runs through New Hampshire. Others seem to agree as, apparently, does Trump; media reports indicate he is scheduled to spend election eve in Manchester, at the Southern New Hampshire University arena which is where he held a massive rally on the night of his New Hampshire primary victory. The Clinton team is countering with President Obama, who will also be in New Hampshire on Monday night, while Clinton will come to New Hampshire on Sunday. My mostly Clinton-leaning audiences here in Vermont have usually found my talks reassuring, since polls consistently showed Clinton ahead in the Granite state by comfortable margins.  Thus just five days ago the RCP polling average had Clinton holding a comfortable 5% lead in New Hampshire. And then the polling bottom fell out for Clinton.  In the five polls in the field since Oct. 28 (that include Gary Johnson and Jill Stein), Trump is ahead in three, and is tied with Clinton in two others, pushing him ahead in the RCP poll in New Hampshire by 1.6% – a drop of almost 7% in Clinton’s support in less than a week. (I should note that Drew Linzer’s site, which aggregates polls in a way that makes the results less sensitive to recent polls) still gives Clinton a very strong probability of winning New Hampshire.) Trump’s polling surge has prompted (mostly panicked) Clinton supporters to email me asking whether this means Trump’s odds for winning the election have significantly increased, since I spent so much time explaining the importance of winning New Hampshire for him.

The short answer is no.  Linzer, whose state-level polling-based forecasts proved to be the gold standard in 2012 and again in 2014,  currently has Trump’s probability of winning the presidential election at 10% – only about 4% better than before Trump’s surge (if polls are accurate) in New Hampshire.  Before explaining why Trump’s NH surge isn’t enough to appreciably improve his election chances, it’s worth exploring why he’s doing so well in my neighboring state.  Long-time readers will remember that I spent considerable time at Trump rallies in New Hampshire (I attended three there), exploring the reasons why New Hampshire voters were supporting The Donald.  As I noted in an earlier post, “it quickly became clear that two themes dominated the thinking of Trump supporters.  The first, expressed – unprompted by me – by every person I talked to, was economic anxiety.  Interestingly, that anxiety was not directed so much at their own situation but toward that of their children, or others close to them.” Put most simply, New Hampshire voters expressed a strong belief that the American Dream was slipping away from their children.  The second theme that emerged, and again I’ll quote from my earlier post, “was a deep antipathy toward Hillary Clinton. One man, in his early 40’s, told me he wasn’t voting for Trump as much as he was against Clinton. Almost to a person those I talked to expressed a fundamental belief that she could not be trusted.  At one point in our conversation, the woman whose daughter is having twins lowered her voice to tell me, “I’m a Roman Catholic and a good Christian, but I just have to say this: that woman [Hillary Clinton] is evil.”  For a more detailed glimpse of what Trump supporters believe, I urge you to read the full post. But my sense is that since those visits, stories regarding the increase in Obamacare premiums, and the steady drip-drip of Wikileaks that seemed to confirm their worst fears regarding Clinton’s honesty, have moved more voters into Trump’s column.  It has probably helped that Trump has finally seemed to develop a coherent message highlighting those themes.  Since my last trip to a New Hampshire Trump rally, he has visited the state twice more, and is now running television ads that directly address issues of Clinton’s trustworthiness, and economic anxiety.

Of course, not everyone accepted my analysis regarding what motivated Trump’s New Hampshire followers.  Many readers responded by citing research by our colleagues (see, for example this) suggesting that the bulk of Trump’s support is motivated by racial resentment, and that in fact feelings of economic anxiety are themselves rooted in part by racism. For brevity’ sake, I’ll not repeat our exchange, except to note that I remain concerned that our efforts to measure “racial resentment” are sometimes tapping into more principled opposition, often expressed to me by Trump supporters in NH, against race-based policies. And, while I acknowledge that it is difficult to detect racial sentiments in direct conversations, my primitive efforts at social anthropology left me persuaded there was a lot more to Trump’s support than simple bigotry, at least among New Hampshire voters.

Be that as it may, the broader point is that we shouldn’t be surprised that Trump may have pulled ahead in New Hampshire – his campaign themes, as I found in my interviews there, resonated with a broad cross-section of New Hampshire voters.  So, why doesn’t this improve his chances for winning the presidency?  To understand why, it’s worth analyzing how analysts like Linzer construct their win probabilities. Remember that most state-level polls consistent of random samples of likely voters.  Done properly, those surveys allow us estimate the range of probabilities regarding how the actual vote is distributed in each state, based on the poll.  By using those probabilities to generate multiple possible outcomes across all states, analysts like Linzer can estimate the overall probability that a candidate will emerge as the winner.  Simply put, when you look at the likelihood of each candidate winning particular states, there are a lot more paths for Clinton to get to 270 electoral votes (the minimum needed to win), based on current polling, than there are for Trump to do so.  Or, to put it in the context of my talks, when I laid out the most probable path for Trump to win the presidency, I noted how important it was for him to win New Hampshire.   However, I didn’t spend nearly as much time talking in depth about the many more combination of states by which Clinton could win – even if she lost New Hampshire.

So, under one scenario I discussed in my talks, Trump could eke out a victory by holding all the Romney states from 2012, and flipping five additional states as well as winning Maine’s 2nd congressional district.  This scenario is not impossible, at least not according to current polling, but it requires Trump winning New Hampshire.

picture1But current polls also suggest that even as Trump pulls ahead in New Hampshire, he’s still behind in Nevada and probably Florida, and it’s not clear he can hold onto North Carolina – a state Romney won in 2012.  Moreover, if you play out all the possible Electoral College scenarios in this way, which is what simulations are designed to do, the odds still heavily favor Clinton.  She simply has more paths to victory.  Trump, in contrast, has far fewer electoral options and thus has to hope all the key states break his way – an outcome that, while not impossible, is much more improbable.

As I noted in my last post, history suggested all along that this race would tighten in the final week as Republicans who expressed skepticism about Trump’s candidacy came home to the partisan roost most familiar to them, twitter talk about landslides notwithstanding.  But don’t let this tightening blind you to the Electoral College math which, based on polling at this point three days out, still strongly favors Clinton.

The Race Was Always Going To Tighten

Let’s be clear: if history is a reliable guide, this presidential race was going to finish closer than the polling last Friday indicated – even without the latest Clinton-Comey email imbroglio. Indeed, there was evidence that the race was already tightening before Comey’s letter telling congressional leaders a new batch of emails that were potentially pertinent to the Clinton email investigation had been discovered.  The reason is, as Robert Erikson and Chris Wlezien explain in their exhaustive study of presidential polling,  as campaigns wind down to Election Day, partisan identification becomes an increasingly important predictor of the vote.  This means that voters who may have told pollsters earlier in the campaign that they were considering voting for the other party’s candidate eventually come home to support their own party’s nominee. It is not surprising, then, that in the two weeks prior to the leaking of Comey’s letter last Friday, Clinton’s lead in the RealClearPolling aggregate had dropped by almost 3%.  Moreover, as I have been telling my election audiences, the race would have tightened more quickly had Trump been able to get out of his own way in order to focus his message on the fundamentals – the economy and terrorism – that are of greatest concern to voters.  Instead, he managed (to Clinton’s relief and not without her help) to focus attention on a series of news stories, from his missing tax returns to Alicia Machado to the 2005 audio tape and allegations of sexual harassment (and worse), that collectively steered the campaign coverage toward exactly the wrong issues from his perspective. In this vein, it is worth recalling that almost all the political scientist election forecast models predicted that this would be a tight race – in the aggregate (which is how I make my quadrennial forecast) the forecast models have Clinton winning the two-party popular vote by about 1%.  (This is probably a good time to remind readers that if you want to know who is going to win the election Drew Linzer’s site  is probably the gold standard. However, if you want to know WHY the candidate won, you need to consult the political scientist forecasts models.)

polls

So will this latest email controversy make a lasting difference?  Many journalists think it will.  Thus, WCAX’s superb Eva McKend, citing our earlier conversation in which I expressed skepticism that any “October Surprise” would affect this election, chided me in this tweet: “@MattDickinson44 how about this for an October surprise? Sorry Matt, I think you were wrong on this one. It’s definitely real.”  Eva is right that the Comey story will likely affect the polls, at least in the short run.  But I suspect it won’t affect the overall trajectory of this race nearly as much as Clinton’s supporters fear and her critics hope.  As I told Eva, so-called October surprises rarely have a lasting impact on the presidential race unless they occur in the last few days of the campaign.  One example of an October surprise that might have mattered is the 2000 story, released four days before the presidential election, of George W. Bush’s conviction for drunk driving more than two decades earlier.  In his memoirs, Karl Rove laments that the announcement cost Bush about 2% of the popular vote and perhaps 30 Electoral College votes – enough to win the election without Florida.  Because it occurred so late in the campaign, the Bush team didn’t have much time to go on the offensive and bring the campaign narrative back to the themes that had, until that point, put him in the polling lead.

Another potential example is Special Prosecutor Lawrence Walsh’s decision to announce Caspar Weinberger’s indictment on October 30 for Weinberger’s actions regarding the Iran-contra affair – an indictment that cast doubt on President Bush’s claim that he was not involved in the decisionmaking that led President Reagan to swap arms for hostages.  Bush supporters contend that news story blunted the last minute momentum that had nearly pulled him into a tie with Clinton.

However, both these incidents took place within 4-5 days of Election Day.  That’s not the case with the Comey announcement last Friday, which occurred with 11 days to go before final voting.  Already Clinton surrogates, like John Podesta and Tim Kaine, are taking to the airwaves to turn Comey’s behavior and motivation, and not the investigation into the emails itself, into the main story.  And Clinton is attacking Comey head on during her stump speeches, calling his actions highly unusual.  Not surprisingly, my  twitter feed has seen a dramatic reversal, with Clinton supporters now casting aspersions on Comey’s professionalism, whereas a week ago they were praising his decision not to pursue criminal charges against their candidate, while Trump’s backers now see him as the paragon of integrity – a stark reversal from their views prior to last Friday.  No one ever accused the partisan denizens on social media of consistency!

Note that this is exactly the tactic Clinton and her surrogates have been using to beat back the steady drip, drip, and drip of the Wikileaks stories, which in my view are potentially far more damaging to her presidential ambitions, with the window they provide into the Clinton Foundation fundraising for Bill Clinton and private concerns expressed by Podesta and others over the security of her email server.  To date, however, the Clinton campaign has managed to make the WikiLeaks story as much about Russian meddling in the U.S. election, and stolen emails, as it is a window into possible Clinton corruption. In this they have been aided by Trump’s apparent inability to stay on message, thus inadvertently helping Clinton refocus media attention away from the potentially damaging leaks.

In short, I believe that there is enough time for the latest Comey email issue to be fully debated and digested in a way that won’t dramatically change the trajectory of this race, which was already tightening.   At most, it will move the race toward the equilibrium it was already trending toward, which was a closer contest than what we saw last Friday.  To be clear, if forced to guess, I suspect these emails won’t do a whit to change Comey’s original finding that there was no criminal intent on Clinton’s part – only negligence – when it came to handling classified material. But at this point the substance of the emails hardly matters, since Comey’s agents can’t possible complete a review of what is reported to be thousands of documents in 11 days.  And that vacuum will certainly allow pundits and talking heads to do what they do best: engage in uninformed speculation designed to drive ratings.  Already we see that with cable discussion about what the emails might say, or whether Clinton’s associates might have obstructed justice – all in the absence of any information on either point. Conceivably, of course, the wild cable chatter might influence some of the early voting going on now. At the same time, however, the lack of substance allows Clinton’s team to change the conversation into a discussion about Comey meddling in an election.  Moreover, the media cycle is such that this story is likely to recede into the background noise by Election Day in a way that allows the outcome to turn on the fundamentals that typically drive most presidential elections.  That assumes, of course, that there’s not another October Surprise in the interim – and that Trump can avoid stepping on his own campaign message. So far, the latter has proved to be a very unrealistic assumption.

Does the Comey story help Trump?  Undoubtedly, at least in the short run.  Will it change the outcome of this race?  I suspect not.