Election forecast: Democrats have a clear shot at a Senate majority — here's how
The 2020 election is rapidly entering the endgame, which means it’s time to resume the Daily Kos Elections Senate forecast. Each week, we’re going to take an in-depth look at the state of play in the Senate, relying heavily on the aggregated polling data at our 2020 Senate portal. If you’re interested in what’s happening in the Senate, we encourage you to bookmark that page and check on it regularly; you can click on each state on the interactive map, which leads you to Bayesian trendlines for the Senate polls in each state, which are updated every day as new polls come in. But for the slightly-less-obsessive, we’ll be doing weekly summaries of what has changed over each week.
Now, granted, much of the attention this year is going to the presidential race, and deservedly so. We, however, are not doing a predictive model of the presidential race this year, or aggregating presidential polling data. This does not mean that we are treating the presidential race as a foregone conclusion, because it isn’t, and naturally you shouldn’t treat it that way, if for some reason you were thinking “eh, somebody else has got this.” Nevertheless, let’s stipulate that Joe Biden has maintained a clear and consistent lead in both nationwide and swing state polling for many months, and appears to be continuing to do so.
But here’s why the Senate deserves a big piece of attention in the coming six weeks. For one thing, from a probability standpoint, at this point it’s a good bit more uncertain who will control the Senate than who will control the White House. The number of pivotal Senate seats is smaller than the number of presidential swing states, and in an already closely-divided Senate, there’s not a lot of room for error. The problem is compounded by that there are some pivotal Senate races in states that aren’t also presidential swing states (like Maine or Kansas), where we tend not to see very many polls.
The other issue, of course, is that any hypothetical Biden administration that controls the House but not the Senate is going to be constantly spinning its wheels. Their most important tasks may simply involve hosing the last four years of crud out of the stables, which can happen within the executive branch, but progressive legislation that actually moves the ball forward isn’t going to happen without a Democratic-controlled Senate. Beyond that, just as far as the crud-hosing process goes, with Mitch McConnell still in charge in the Senate, they’ll struggle to even get their cabinet and judicial nominations approved. Knowing what’s at stake, let’s dig into the detail on this year’s Senate battlefield.
The relatively good news is that it appears, at least at this point in time, likelier than not that the Democrats will be able take control of the Senate. We aren’t going to assign percentage odds to that likelihood, but Democratic candidates lead in enough separate races in Republican-held seats that, if the eventual results match current polling, there would be a Democratic majority.
You might have heard that story before; if you recall back to 2014 (when the fight was to hold control the Senate), 2016, or 2018, it looked plausible several months out in each of those years that the Democrats could win a majority, but the wheels gradually fell off over the course of the last month of the cycle. The difference this year may be that Democratic candidates seem to have bigger and more consistent leads in the probably decisive races, as well as few places where they’re on the defensive and a broader array of second-chance races if any of the big ones take a late bad turn. While there is plenty of time for things to still go wrong, Democratic candidates in the biggest races have seemed unexpectedly resilient this year, in much the same way that Biden has at the top of the ticket.
Let’s start with the basic math: Democrats (and Dem-aligned independents) currently hold 47 seats, while Republicans hold 53 seats. So, the Democrats would need a net gain of three seats to control the Senate, assuming that they also win the presidential race and a Democratic vice-president is present to break ties. To be on the safe side, you might think of a net gain of four as a better target as insulation against a Republican vice-president (though because of the important role of coattails in downballot races, it’s extremely unlikely we’d be gaining that many Senate seats against the backdrop of a Republican presidential victory).
Another way to think of a four-seat gain, though, would be that’s enough to create not just a majority but a Joe Manchin-proof majority, which could make all the difference between a Senate that merely does no harm versus one that can pass progressive legislation — or even, say, have the votes to eliminate the filibuster in order to even contemplate passing such legislation in the first place. Of course, then, you might start thinking about also needing a five-seat gain to create a Kyrsten Sinema-proof majority, a six-seat gain to create a Tom Carper-proof majority, and so on, but let’s not go wild just yet.
So, there are several ways to achieve the bare minimum three-seat net gain. The simplest way, of course, would be to win three seats while losing none; for instance, those might be Arizona, Colorado, and Maine, three races where the Democratic candidate has consistently led all year, often outside the margin of error. The bad news, however, is that most likely the Republicans will be picking up a seat in Alabama. If anyone is capable of holding that seat, it’s probably Doug Jones — but this has been a widely expected outcome after he won a 2017 special election in one of the nation’s reddest states against a bizarrely flawed Republican opponent, Roy Moore — and polls currently show that the widely expected outcome is indeed likely to happen.
That makes a fourth pickup necessary. Luckily, the Democratic candidate in North Carolina, Cal Cunningham, is also routinely leading in the polls. In fact, it wouldn’t be outlandish to think that North Carolina is a better bet than Maine, if you’d prefer; while the polling average in Maine has a showier lead, the lead in North Carolina has been very stable and is based on many, many more polls. The Maine number, in fact, is heavily driven by last week’s Quinnipiac poll, which is the only poll this race has seen with a double-digit lead for Democratic candidate Sara Gideon.
On top of that, there are near ties in Alaska, Kansas, South Carolina, and Georgia’s regularly scheduled race, along with races within arm’s reach in Iowa, Montana, and Texas. Compounding the Republicans’ problems, there is only one other competitive race where they’re on the offense (but still losing), in Michigan. (If we ran a thousand permutations, you’d probably get a few “five minus two” instances where we, say, lost Alabama and Michigan but gained Arizona, Colorado, Maine, North Carolina, and Iowa, and thus the majority. That’s too marginal a case to really plan for, though.)
To help visualize this, as we did in 2018, we’re going to arrange the competitive Senate races in something of a “totem pole,” with the likeliest Democratic wins at the top and the likeliest Republican wins at the bottom. To keep the scope manageable, we aren’t going to include the races that we rate as “Safe” for one party or the other — with a couple exceptions, which I’ll mention later. It won’t increase your situational awareness to know that we’re on track to win the Virginia Senate race by 20 points, for instance. Slicing through the middle of the totem pole is the Red Line of Death, at the 50 D/50 R seat mark. To win the Senate along with winning the White House, there need to be either three Democratic pickups above the line and no losses below it, or four Democratic pickups above the line and one loss below it. (Well, it’s not actually red; you’ll have to use your imagination.)
|STATE||D CAND.||D AVG.||R CAND.||R AVG.||DIFF.||FLIP?|
|NEW HAMPSHIRE||Shaheen (I)||53||Messner||35||+18|
|MAINE||Gideon||51||Collins (I)||40||+11||D FLIP|
|ARIZONA||Kelly||49||McSally (I)||41||+8||D FLIP|
|COLORADO||Hickenlooper||49||Gardner (I)||43||+6||D FLIP|
|NORTH CAROLINA||Cunningham||45||Tillis (I)||41||+4||D FLIP|
|RED LINE||RED LINE||RED LINE|
|SOUTH CAROLINA||Harrison||46||Graham (I)||46||0|
|ALABAMA||Jones (I)||35||Tuberville||52||-15||R FLIP|
As you can see from the table, the Democrats are doing the latter: they’re picking up four seats and giving up one. And that’s a net gain of three without counting the two races that are, for the moment, precisely tied in our averages. We’d still need at least one more of those, however, to have a Joe Manchin-proof majority (or, less appealingly, a Mike Pence-proof majority).
There are two races on here, as I mentioned, that are currently rated Safe Republican in Daily Kos Elections’ qualitative ratings, but which I’m including in this chart anyway (largely because I’d be asked about them all the time anyway, if I didn’t). One is the race in Mississippi, where, in a rematch of the 2018 special election, Democratic ex-Rep. Mike Espy is running against appointed incumbent Cindy Hyde-Smith. Espy lost that race by a surprisingly close eight points, and he’s coming even closer in the current polls. The reason we have this race at Safe R, however, is that because Mississippi has the highest African-American percentage of any state, it’s actually not unusual for a Democratic candidate to hit the low-to-mid-40s here, despite Mississippi’s usual dark-red status. However, that’s because of its heavily racially-polarized voting, so, in an “inelastic” state with few swing voters, the final ascent to the 50% mark is brutally difficult.
The other is Kentucky, which is high on most people’s wish lists and where Democratic candidate Amy McGrath is raising money at an astonishing clip. The unfortunate reality, though, is that McGrath is nevertheless trailing Mitch McConnell by double digits, in another state where it’s not hard for a Democrat to break 40 (thanks more to a lot of white “ancestral Democrats,” rather than a large African-American base) but getting to 50 in a federal race is a Herculean feat. (As a caveat, Kentucky did elect a Democratic governor last year — the son of a popular ex-Governor, against a terrible incumbent, who won by only a fraction of a point.) Think of McGrath’s job more as to raise a ton of money and keep McConnell pinned down and unable to help in closer races.
Finally, there’s one other race that isn’t on the chart simply because the cake hasn’t even gone in the oven yet. That’s the special election in Georgia, where appointed incumbent Kelly Loeffler faces voters for the first time in an odd format. All candidates from both parties participate on one ballot in the vote on Election Day, with the top two advancing to a January runoff — so what’s happening on Election Day is more like the primary. Unfortunately, what’s happening right now is that the top two slots in the polling average are both going to Republicans: Loeffler at 24 and Doug Collins at 20. The top Democrat, Raphael Warnock, is only narrowly behind at 17 and there are still many undecided voters, so it’s quite possible he (or another Democrat) breaks into the top two. But, really, we won’t know until after Election Day whether this is even a potential pickup or not.
I would also urge you to consider the Daily Kos Elections qualitative ratings as well, which take other factors, like spending by committees like the DSCC or Super PACs like the Senate Majority Project, how red or blue a state is at the presidential level, and how willing a state has been in the past to ticket-split. That could be a helpful corrective in the case of, say, Maine, which we still rate as a “Tossup.” That’s because there have been relatively few polls here, most of which until recently showed a much closer race, as well as Mainers’ long tradition of ticket-splitting and the ease with which Susan Collins has dispatched previous opponents — though a good sign is that whatever magic touch she had in the past seems to have entirely vanished this year.
Similarly, you might consider the reasons we still have South Carolina at “Likely Republican” despite the very close polls there; it’s a reliably-red state with very racially-polarized voting, where undecided voters are more likely to break to the Republican candidate, and while being at 46 is pretty impressive, it’s still a loooong way away from being at 50. (The same could be said of, for example, Alaska or Kansas.) That contrasts with, say, Iowa (which we have at “Tossup”), where a tied race gives a Democratic candidate a pretty good shot at winning, given its long-standing swing state status and tendency to oscillate back and forth between the two parties based on distaste for whichever party is currently in power.
Finally, this week we’re going to briefly veer off into talking about the forgotten step-child of the elections world, gubernatorial races. Back in 2018, there were so many exciting races that we maintained a separate weekly gubernatorial preview. This year, we won’t. Largely, that’s because most gubernatorial races happen in midterm years; only around a dozen states do it during presidential years. And beyond that, there’s really only one (yes, one) gubernatorial race this year where the outcome seems truly in doubt, and it’s in one of the nation’s least populous states.
That’s the Montana gubernatorial race, which we currently rate “Lean Republican.” This would lead to a net pickup of one gubernatorial seat for the GOP, since it’s an open seat where termed-out current Democratic governor Steve Bullock is currently running for Senate. After 16 years of Democratic governors, it appears that Montana is on track to revert to its usual red-state status, which is unfortunate because this year’s GOP nominee is Rep. Greg Gianforte, probably best known for assaulting a reporter during his 2018 House campaign. However, what few polls we’ve seen have shown a close race between Gianforte and Democratic Lt. Governor Mike Cooney, so it’s possible we’ll get lucky and see no change whatsoever. Here’s the table of the few races that Daily Kos Elections qualitatively rates something other than “Safe,” at least where we have polls. (We rate West Virginia “Likely Republican” but have no polls of the current matchup, period.)
|STATE||D CAND.||D AVG.||R CAND.||R AVG.||DIFF.||FLIP?|
|NORTH CAROLINA||Cooper (I)||50||Forest||40||+10|
|NEW HAMPSHIRE||Feltes||32||Sununu (I)||57||-25|
Rather than having a separate gubernatorial preview on a regular schedule, which would bore even the most devoted of elections junkies, we’ll just pop back in and add this table to a future week’s Senate preview if a noteworthy-enough poll drops in any of these races to change its trajectory.
There’s still the matter of the House side of the equation, where the Democratic majority is in little danger and, in fact, is poised to grow slightly; my esteemed colleague Steve Singiser will be doing regular House previews, also to launch in the coming weeks. And, of course, there’s the presidential race. Daily Kos Elections will not be doing a predictive model as we did in 2016 or aggregating presidential polls. Partly, that’s because Drew Linzer, who programmed our 2016 model for us, is currently otherwise occupied running our polling partner, Civiqs. We also feel that there are already a number of trustworthy practitioners (many of whom are named Nate) providing services in that area, with whom you’re probably already familiar.
Beyond that, there are simply the lessons of 2016, which demonstrated that the business of trying to assign a precise percentage likelihood to a black swan event or systematic polling error is something of a fool’s errand. In fact, research suggests that predictive modeling goes beyond mere fool’s errand territory to having actual harmful effects, possibly lowering turnout among voters who misapply the data or misunderstand probability. (Nevertheless, if you want to look at our “totem pole” and say to yourself “it looks like there’s a 50% chance of the Democrats winning a Senate majority” or “no, actually, it looks like there’s a 66% chance of a Democratic majority,” I won’t try to stop you, and in fact those would both be fairly reasonable conclusions.)