The old conventional wisdom: What if Democrats take back both chambers of Congress this fall?

The new conventional wisdom: What if Republicans never lose another election?

Seriously, it’s a cinch that the GOP will lose House seats this fall. The growing mystery is whether they’ll lose enough to actually return Nancy Pelosi to power. Yesterday CNN found the Democratic advantage on the generic ballot down to three points. Today Reuters finds it down to one. Given the feelgood spectacle this morning of Trump and Pence greeting American hostages on their return home from North Korea, it’s quite possible that a poll next week will show the GOP even — or ahead! — on the generic ballot. On Christmas Day 2017, they trailed by 13 points on average.

Three words: Speaker. Kevin. McCarthy.

Well … no, probably not. Despite the shrinking margins in the CNN and Reuters polls, RCP elections analyst Sean Trende remains skeptical.

The RCP generic-ballot average today is Dems by 6.1 points, which is another way of saying that even with CNN and Reuters baked in, the balance of polling still shows a comfortable blue advantage in November. If the election were held today, whoever prevailed would probably end up with only a narrow House majority, unable to do much of anything legislatively.

Trende’s second point is also well taken and worth thinking about as the Senate outlook for Republicans this fall improves a bit. The immediate goal is to hold the Senate by any margin, if only to give Trump a fighting chance in getting his nominees confirmed. But a 51/49 GOP advantage, although better than the alternative, would leave the party in dire shape for the next two cycles. The 2020 and 2022 Senate maps are as gruesome for Republicans as this year’s map is for Democrats; it’s almost exclusively Team Blue that’s defending vulnerable seats this year but that’ll start to reverse two years from now. If the GOP’s going to hold on to the upper chamber as a legislative and confirmation hedge against a Democratic president in 2021, it’s important this year to net some Senate seats and provide itself a cushion for future losses.

And not to be too eeyorish, but there are reasons to question the Reuters numbers. For one thing, their sample is 40D/39R, an unusually tight spread given the electorate’s traditional Democratic lean. For another, Reuters typically pegs Democratic support far lower than most other major pollsters do. Scroll through RCP’s list of generic ballot polls and you’ll find dating back to last year that they’re the *only* major pollster to consistently have Dems hovering as low as 39-40 percent support.

I think that’s because they offer a “don’t know” option that other pollsters don’t when asking people which party they prefer to see win in November. On the one hand, the fact that so many are choosing “don’t know” in Reuters surveys suggests that a significant number who choose “Democrats” in other polls are soft in their preference, and might be dissuaded before the midterms. On the other hand, a “soft” vote for the Dems is still a vote for the Dems. Based on the steady Democratic lead in other polls, there’s every reason to believe that most of the “don’t know” crowd in Reuters’s survey would tilt blue on Election Day. At least right now. Stay tuned.