Rick Scott of Florida has a second job at the moment besides serving as his state’s junior senator. He is also the chairman of the National Republican Senatorial Committee and therefore largely responsible for ensuring the Republican Party moves from the minority to the majority in the Senate. And on that matter, Scott professes a great deal of optimism.
“We’re going to get 52 Republican senators. We have to win here,” he said Thursday at an event in North Carolina, where the party is working to hold the seat of Sen. Richard Burr. In fact, he said, “I think we can get 53, 54, 55.”
Scott certainly has reason to think his party has a decent shot at winning a majority in the Senate. After all, it’s a 50-50 split at the moment, and a shift to the Democrats over the summer appears to have moved back in the other direction. But 55? Could Senate Republicans end up with a 10-vote margin over Democrats?
Well, yes. But it would mean a lot of things going very, very badly for the opposition.
First of all, the Republicans are defending more seats than the Democrats. There are 14 Democratic seats on the ballot next month and 21 Republican. So a five-seat shift to the GOP means winning 26 of 35 contests.
FiveThirtyEight has a model that incorporates recent polling and other factors to estimate how Senate races are likely to play out. The “lite” version of the model looks only at those polls. If we array the current (as of Thursday) lite-model margins in each of those contested seats from most heavily Democratic to most heavily Republican, it looks like this.
All 35 of them are between a 40-point Democratic advantage and a 40-point Republican one — except Alaska, which broke the scale and is just floating there off by itself, a position the state will find quite familiar.
Notice that there are only two seats that flip between parties: Nevada, going from Democrat to Republican, and Pennsylvania, going the other direction. In other words, this suggests a 50-50 Senate in January 2023 and therefore a Democratic majority (thanks to Vice President Harris’s tiebreaking vote). This is the point at which I note that polls showing candidates essentially tied are not really polls in which we can determine a probable winner. But for the purposes of this exercise, I am going to unwisely set that aside.
FiveThirtyEight also has a “deluxe” model, which incorporates things such as fundraising and other factors. That’s the model that leads its site, so it’s the one I’m going to focus on, too. If we move all of the contested seats into their “deluxe” model positions, you see that things space out a bit differently. But still only two states flip, and it’s still a 50-50 chamber.
What Scott is salivating over is all of those blue squares sitting just above that middle line. So if we assume that there is a national shift in the polls between now and Election Day or that there is a consistent error in the polls that advantages Republicans equally, what would it take for Scott to hit his 55-seat figure?
Well, he needs to flip five seats. Give him Nevada, which is on the bubble. The next-closest is Georgia. Then keep Pennsylvania (which doesn’t count to the five needed flips, obviously) and add in Arizona, New Hampshire and Colorado, in that order. That would require the model’s estimates being off as much as eight points.
There has been a consistent pattern over the past four election cycles: Republicans have overperformed FiveThirtyEight’s modeled results more often than not. Usually, though, it’s more-Republican places that overperform more, not swing states — probably in part because there’s more polling in the latter that can better inform the model.
Of course FiveThirtyEight also answers Scott’s question directly. When it runs 40,000 iterations of Election Day based on the information it has about each race, there are occasions on which the GOP ends up at 55 seats — 3 percent of the time, to be specific. Meaning that the odds of this happening are not zero but not big.
But there is good news for Scott in FiveThirtyEight’s estimates: About a third of their modeled elections have the GOP ending up with at least 52 seats. For that to happen, he needs far less to go wrong for the Democrats.