For the early birds.

Photographer: Stephen Maturen/AFP/Getty Images

Polls Tell Us a Lot, But They Don’t Tell Us the Results

Jonathan Bernstein is a Bloomberg View columnist. He taught political science at the University of Texas at San Antonio and DePauw University and wrote A Plain Blog About Politics.
Read More.
a | A

We’re one week away from the elections. Yes, plural. The nation is holding thousands of elections on Nov. 8, and all of them are important, whether the people who win them are local judges or the next president of the United States.

I should say: thousands of elections concluding on Nov. 8, because voting started weeks ago (at least in some states), and some 25 million ballots have already been cast. 

As has been the case at least since spring, the most likely outcome will be a good night for the Democrats. Or it’s still quite possible it will turn out to be a great night for the Democrats. Or a good night for the Republicans. Here’s an overview of where everything stands now: 

President: Hillary Clinton isn’t guaranteed a victory, but she’s in good shape, if you average the current projections from polling-based prediction systems. That indicator says there’s a 91 percent chance she will win. The lowest projection, from Nate Silver’s FiveThirtyEight, says her chances are at 75 percent. And Silver still estimates she has a five-percentage point lead, and is almost as likely to win by a double-digit landslide as she is to lose the total vote.

Big caveat: Silver’s model (but not the others) gives Donald Trump a fairly realistic chance of winning the Electoral College, even while losing the total nationwide vote.

Another big caveat: While it’s unlikely, it isn’t impossible that the polls are systematically wrong this time and wrong in a way that consistently underestimates Trump’s vote -- or overestimates it.  

Senate: Democrats need to gain four seats if Clinton wins, and five if Trump wins, to regain control (since in a 50-50 Senate, the vice president breaks the tie). They are still favored to gain seats in Illinois and Wisconsin. Only one Democratic seat, in Nevada, is vulnerable, although Democrat Catherine Cortez Masto appears to have a slim advantage in the race to replace Senator Harry Reid.

This leaves at least five states with Republican senators and competitive races: New Hampshire, Indiana, Pennsylvania, North Carolina and Missouri. In a good Democratic cycle, Florida and perhaps even Arizona could be close. The five poll-based projection systems average out to a 68 percent chance for a Democratic majority. (Important: Especially in the Senate, every seat counts; there’s a big difference between a 53-47 majority and a 51-49 split.)

House: The Republicans will lose some seats from the historical high they have now, but Democrats have little chance of picking up the 30 seats they need for a majority. Looking at the Cook Political Report ratings, Democrats would have to win all of the seats where they currently have an advantage, plus all of the toss-up races, and would still need to win almost all of the “lean Republican” contests to win a majority. Unlikely -- unless Clinton wins in a landslide.

Governors: Only 12 races are contested this year, eight currently held by Democrats and four by Republicans. It appears that the two most likely to see flips are currently held by Republicans: the open contest in Indiana (with the current governor, Mike Pence, running for vice president instead) and in North Carolina (where the incumbent, Pat McCrory, has been unpopular). Though the Democrats have a chance to pick up these two, it’s not impossible that Republicans could gain in four other states.

State legislatures: The Republicans did so well in 2014 in capturing state assemblies and senates -- now holding 68 of the 99 legislative chambers -- that they will likely fall back, even if they have a good night. Thirteen of Ballotpedia’s 20 competitive chambers currently have Republican majorities. Governing magazine rates one Republican-majority body, the Nevada Assembly, as leaning to Democratic and an additional six as toss-ups, with two Democratic-held chambers rated as toss-ups.

These contests are of national importance. Election law, from voter-identification rules to automatic voter registration and more, has been contested in many states over the last few years. And then there is the redistricting after the 2020 census, when the legislatures and governors in many states set new lines for both state legislative and U.S. House districts.

Ballot measures: While marijuana has been the most talked-about subject for state propositions this year, perhaps the most important group of measures has to do with the rules for how elections are conducted. Also important: Minimum-wage increases are up for a vote in several states; two states are voting on the death penalty, and Washington is voting on a carbon tax.

As voters make all their decisions next Tuesday, one factor we can’t reliably count on this presidential cycle is the normal relationship between the top of the Republican ticket and the “down-ballot” candidates. Trump’s candidacy is unusual enough that it’s unclear if the normal party-line discipline of recent elections will hold.

What we do know is Trump hasn’t helped down-ballot Republicans. We just can’t be certain yet if he’ll wind up hurting them. 

This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.

To contact the author of this story:
Jonathan Bernstein at

To contact the editor responsible for this story:
Katy Roberts at