Democrats Have Numbers on Their Side in Battle for the House. Republicans Have the Map.

Dozens of House races remain extremely close in the closing days of the midterms, according to New York Times Upshot/Siena College polls, making it easy to envision a Democratic blowout or a district-by-district battle for control that lasts for weeks.

The difference between the two outcomes will depend on whether enough Democratic candidates get over the top in the long list of Republican-leaning areas they’ve put into play. The Democratic gains in predominantly white, well-educated suburbs have stretched the Republican majority exceedingly thin. Fighting against this is partisan polarization, which could allow Republican incumbents to narrowly hold on to districts carried by the president.

The uncertainty isn’t just about hedging. It’s a reflection of the sheer number of highly competitive districts, and the limited data available about each one. There are typically only one or two polls of the most competitive races. It is a very different information environment than a presidential election, when there are dozens of polls of a handful of highly competitive states.

The polls can be wrong, of course, as they were in crucial states in 2016. But in a presidential race, we at least know what the polls say about a state like Pennsylvania or Florida. This time, we don’t know what the polls say in most districts, and the polls could be wrong even if we did know.

But taken in the aggregate, the dozens of Times/Siena polls paint a clear picture of a still uncertain race.

The Democrats appear positioned to net at least 14 of the 23 districts they need for control, since they seem plainly favored in 16 districts held by the Republicans, while the Republicans seem favored in two seats held by the Democrats.

Those 16 districts are rated as “lean Democratic” by the Cook Political Report, and recent Times/Siena polls of those districts show the Democrats ahead by an average of 7 points. They’re an interesting mix of districts, but many are either ones where the current Republican retired and Democrats found a particularly strong recruit, or predominantly white, well-educated suburban districts that voted for Hillary Clinton.

After that, things get murky fast. The Democrats have put a long list of districts into play. There are 29 races rated as “tossup” by the Cook Political Report and 26 considered “lean Republican.” The Democrats would need to win a mere nine of those 55 districts to take control.

The Democrats are considered clear favorites to win control because of the sheer number of opportunities they have to do so. They’re competitive in so many races because they’ve recruited many strong candidates, raised unexpectedly large amounts of money and benefit from a national political environment that continues to favor their party.

All the conditions for a “wave” election remain in place. It’s a midterm year, when the president’s party typically struggles; the president’s approval rating is beneath 50 percent; and the Democrats hold a wide high-single digit lead on the generic congressional ballot, which asks voters whether they’ll vote for Democrats or Republicans for the House.

But this year, such an election doesn’t guarantee anything like the 63 seats a previous wave election brought to the Republicans in 2010. The Democrats don’t have nearly as many good opportunities to pick up seats, because of partisan gerrymandering and the tendency for Democrats to win by lopsided margins in urban areas. Alone, it’s enough to give the Republicans a chance to survive a wave that would otherwise give the Democrats a huge majority.

Indeed, the map remains the Democrats’ big challenge. They have clearly put a long list of districts into play. But most of them lean Republican in presidential elections, and it is not obvious whether the Democrats actually lead in a lot of them.

The results in the races classified as “tossup” have been within 1 point in nine of the last 12 such districts we’ve polled, all over the last 10 days, and within 3 points in 11 of the 12. Individual House polls are fairly noisy, but the aggregated picture is clear: The tossups are really close.

The Democrats need to win only about one-third of the tossups to win the House. So if the polls we’ve taken to this point were exactly right (they’re not) the Democrats would take control by a modest margin. Maybe they’d pick up 27 to 30 seats. It would be a wave, on paper, given the limited opportunities that Democrats have on favorable terrain. But it wouldn’t necessarily feel like a wave. The race might not be called for days, as California and Washington count late mail ballots.

It’s not a comfortable margin for Democrats, given how close these races appear. Republicans wouldn’t need many lucky breaks in races that were quite close.

The view that the Democrats are more overwhelming favorites depends on one of two interpretations of the data.

One argument is that the Democrats are likely to win at least a handful of “lean Republican” races. There aren’t many polls in these races; some haven’t been polled at all. If, hypothetically, Democrats won only three or four out of the 26 races characterized as “lean Republican” by the Cook Political Report, it would make the Republican path to victory exceedingly narrow.

In our polling, the “lean Republican” races live up to their label: on average, Republicans led in those races by a 6-point margin over eight polls taken over the last 10 days. But Democrats did narrowly lead in one of those races, Virginia 5. And there are many more races that we won’t poll at all. It’s hard to say how many of these races the Democrats can realistically win.

The second argument is that the race is likelier to break toward the Democrats at the end, presumably because undecided voters will break for previously unknown Democratic challengers against well-known Republican incumbents.

Obviously, polls don’t predict how voters will move as we approach the election. But there’s nothing about our polling that would rule out this theory. Many Republican incumbents are stuck around 45 percent, and in several districts the undecided voters are predominantly young, nonwhite and Democratic. And Democratic challengers are spending millions in the closing weeks to increase their name recognition. The national political environment, which still favors the Democrats, would give them a bit of wind at their backs as well.

If there’s any truth to this theory, many on the long list of extremely close tossups could fall to the Democrats. They would also break through in a number of “lean” Republican districts where the Democrats are in striking distance. The “blue wave” wouldn’t just materialize — it would feel like a blue wave, too.

But there’s still one big force on the GOP side: the Republican-lean of the battleground. If Republicans can polarize the electorate along the lines of recent presidential elections, they will lock the Democrats into that disadvantageous map. Over the last month, the race has trended somewhat in that direction. Republicans have pulled ahead in red-state Senate contests. If this holds, it’s a big shift from the last year’s special election results, when Democrats routinely fared extremely well in Republican-held areas. It would be more like last year’s Virginia general election, when results were highly correlated with the presidential race.

Fully 18 of the 29 “tossup” districts voted for Trump in 2016, and so did 24 of the 26 “lean Republican” races. So a more polarized electorate would help the Republicans run the table in the “lean Republican” contests and stay in the game in the tossups.

Alone, a more polarized electorate wouldn’t be enough for the Republicans to win. Too many Republicans have retired. Too many gerrymanders have been weakened by the courts since 2014. But it would keep them in striking distance. It would mean a late election night, and they wouldn’t need too many more lucky breaks to hold on.