“The world is full of obvious things which nobody by any chance ever observes.”
So said Sherlock Holmes in “The Hound of the Baskervilles,” and the great detective’s observation may well apply to the upcoming midterm elections.
One of the great lessons of 2016 was the failure of the political class to understand the difference between probability and possibility. On Oct. 17, 2016, Hillary Clinton had a 7-point lead over Donald Trump in the RealClearPolitics average of national polls. By the Friday before the election, Nov. 4, that advantage had slipped to 2 points.
Although most politicos argued it was unlikely that Trump could win the presidency trailing Clinton going into the last weekend before the election, they forgot or chose to deny that it was still theoretically possible that voters could deliver a surprise outcome. Many are still in denial.
Reading the political tea leaves three weeks out from next month’s elections is as risky today as it was in 2016. But one thing is clear: Both parties can claim one dubious victory so far. Democrats and Republicans have successfully turned their opponents into objects of dislike and disdain in the minds of many voters, especially independents.
Also Watch: Trump Heads West to Campaign — And a Lot of Senators Do, Too
For readers of this column, my belief in the importance of the “independent factor” is not news. It’s a trend we saw in the waning weeks of the 2016 election that may well be becoming a more permanent political reality as we see undecided voters, once again, still struggling to make a choice and define a direction.
Complicating this decision-making, as the midterm campaign season has droned on, the two parties’ unfavorable ratings have gotten worse. In a recent Winning the Issues survey, 55 percent of independents who are undecided on the generic ballot have an unfavorable view of both parties. These voters seem to be saying, “A pox on both your houses,” much as they did with the presidential candidates two years ago.
Unfortunately, what these voters are hearing mostly through candidates’ negative advertising is not helping them choose. When all voters were asked to define the Democratic agenda, for example, 66 percent said it was stopping President Trump and his policies, and only 21 percent said it was a clear set of policy initiatives. Among independents, only 14 percent said it was a clear set of policy initiatives.
Things were not much better for Republicans. Only 35 percent of voters knew that the last year’s tax reform gave every taxpayer a rate cut — the central piece of congressional Republicans’ signature legislation. Even more discouraging, only 31 percent of independents said the law contained tax cuts, while more said tax cuts for every working American weren’t included. Republicans appear to have lost the messaging battle to Democrats, who have shamelessly told voters that the tax law was created to help only the rich.
Given this environment and with just a few weeks to go, Democrats and the media seem convinced this election is a “done deal.” It could be. But a brief examination of a few House races might give us some better insight not only on what is deemed probable, but also what is possible.
How they lean
Over the past few weeks, only one event has seemed to affect the current environment — the Kavanaugh confirmation, which has changed the motivation levels of many Republicans. We know that from a number of national polls.
What isn’t as clear, thanks to wide differences in national polls, is just where independents are leaning. For example, an Oct. 3-9 Reuters/Ipsos poll showed independents backing Democrats by a 17-point margin. Yet an Oct. 7-9 Economist/YouGov survey gave Republicans the lead among independents by 8 points.
Sufficiently confused? Don’t feel alone. National polls today are often difficult to assess because many organizations don’t share their methodology in enough detail to make a thorough assessment possible.
But there are other ways to get a handle on the midterm state of play. To test possible scenarios, I recently overlaid the New York Times Upshot/Siena College surveys of House races (available online) with Charlie Cook’s House race ratings, both as of Tuesday, Oct. 16. One caveat: I don’t necessarily agree with all of the Cook ratings or the Times’, but together they provide a framework to examine the probable and the possible.
I decided to use the Times/Siena surveys because, along with offering up district-by-district polling data, they did something interesting. They created a number of election models that allow us to test what those numbers might mean to different outcomes.
Looking at the seats in play according to Cook — the “Likely” and “Lean” seats for each party — my calculations put Democrats at +14 seats — nine short of what’s needed to win the House. So the majority comes down to the 31 Toss-up races in play — 2 Democratic and 29 Republican.
Let’s say the parties split the two Democratic Toss-up races. At this point, one of the Times/Siena surveys shows a significant Republican lead in one of those contests, so that would reduce the Democrats to +13.
Now we get to the heart of the matter — the 29 Republican Toss-up races, of which Democrats need 10 to gain a majority in the House.
The Times has not yet done surveys in six of the 29 races. Of the remaining 23, Republicans lead in 15, Democrats in seven, and in one race is even. If this is the state of play for the House, what happens when you plug this data into three turnout models — The New York Times’ turnout projection, the 2014 election and the 2016 presidential election?
First, the results of the Times’ turnout model. Of the Toss-up races with a 3 point lead or more, Republicans are winning 10, Democrats three. Of the remaining 16 (six of which have not yet been surveyed), Democrats would need to win seven for control.
Next, the presidential election turnout model, which conventional wisdom says favors Democrats. Of the 29 Toss-up races, Republicans lead by 3 or more points in 12 districts; Democrats lead in four districts. Of the 13 remaining, including the six without survey data, Democrats would need to win six.
Under the 2014 election turnout model, Republicans have a 3-point lead in 18 races; Democrats in one. Of the remaining 10 races, including the six without surveys, Democrats would have to win nine.
It is interesting to note that for the 43 districts surveyed at this point by the Times/Siena, the 2016 presidential year model is more favorable to Republicans than the poll’s current model in 30 of the districts.
I am not trying to predict one outcome over the other — but rather show that there are still a range of possible scenarios. With a sizable number of undecided voters in most of these races, the outcomes in these contests are simply not clear — in either direction. So, while things unquestionably still look challenging for Republicans, data does show possibilities.
If independents break hard one way or the other as they did in 2006 and 2010, that would likely change everything.
The difference between probable and possible in 2016 was the difference between President Clinton and President Trump. What that difference will deliver in 2018 remains to be seen.
David Winston is the president of The Winston Group and a longtime adviser to congressional Republicans. He previously served as the director of planning for House Speaker Newt Gingrich. He advises Fortune 100 companies, foundations, and nonprofit organizations on strategic planning and public policy issues, and is an election analyst for CBS News.