A better measure of 2020 electability
Elizabeth Warren’s new angling at least takes public opinion into account, but maybe there’s an even better way
When you spot a pattern in the data:
The takeaway: Any Democrat that is even treading water in their favorability ratings is probably “electable” against Donald Trump. But it’s still a long way until we’ll get those final numbers.
Editor’s note:
Thanks for reading my thoughts on this subject. And thanks for subscribing! Your membership adds up and makes all this newslettering possible (reminder: I do all this work independently). Please consider sharing online or with a friend; the more readers, the merrier. Remember that apart from getting special articles subscribers can also comment below each post and participate in exclusive threads.
As always, send me your tips about what you’d like to read about next. Or what you don’t want to read. Or your feedback otherwise. Also, cat pictures are nice, so please send them to me! I’m elliott@thecrosstab.com, or @gelliottmorris on Twitter.
Thanks all,
—Elliott
Let’s just take a break from impeachment for a second.
Elizabeth Warren was in Keene, New Hampshire on Wednesday. She made a case for her electability as the Democratic nominee:
There’s a whole bunch of core issues, like raising the minimum wage, and giving unions more power, and more regulations over financial institutions, and canceling student loan debt, and a wealth tax, that the majority of Americans — not just the majority of Democrats, the majority of Americans — are on board.
She’s right, of course, about public support for her plans. In a February poll from Morning Consult, 61% of registered voters—and 50% of Republicans—approved of Warren’s proposal to tax the wealthiest Americans (those with $50 million) at 2%. A separate poll from Morning Consult found that 56% of voters approve of canceling student loan debt.
Talking about polling is a recent development for Warren. Just last week, she said “I don’t do polls” when asked if she saw new numbers that have her ahead nationally. But this is an especially important development because of how Warren typically talks about electability. Sahil Kapur, a national reporter for Bloomberg, put it this way on Twitter:
I welcome the senator’s usage of polling in assessing who might be “electable” next November. But policy preferences will not decide everything. Voters care about more than just where a candidate stands on a wealth tax, regulating Wall Street and Medicare for All. Among other things, some voters will care about whether the candidate is a man or a woman, black or white, rich or poor and “genuine” or not. And how can we gauge the impact of those factors?
One possible (and maybe too obvious) approach is to devise a predictive model that takes the president and their challenger’s favorability ratings into account. David Shor, a data scientist at Civis Analytics, has done just this. Here are the coefficients for a model he made after the 2016 election.
Let me note of course, that it’s early, so favorability data will change between now and the election. (In fact, I bet that most candidates will get less favorable as the general election continues—but my prior on this guess is not very strong.) Nevertheless, this regression indicates that the candidates who have higher favorability ratings will fare better in the general election (which, duh!). So, if we used current polling data we would conclude that very popular candidates such as Warren and Buttigieg might fare better in November than less popular candidates like Sanders or Biden.
But here’s a better point: electability in 2020 is perhaps less about the Democratic candidate and more about the president. Trump is so unpopular that any mildly generic candidate can probably beat him next November. If we plug our own numbers into the regression above, we can see this steep uphill battle the president might face against any decent Democratic opponent.
The latest polling from The Economsit and YouGov has Trump at a 41% favorability rating and a 51% disfavorability rating, or what we would call a -10 “net favorability” rating. Warren clocks in at 42-36 favorable, or a +6 net favorability rating. Plugging these data into Shor’s formula (0.4829 + 0.02 x 1.12728 + -16 x 0.0021984) yields a prediction for next year’s presidential election.
Based on these numbers, we would predict Trump to win 47% of the two-party vote, +/- 1.2% (assuming GDP growth of 2%). Anyone more popular than Warren does even better against Trump, and anyone treading water in favorability can hold POTUS to 48% of the vote.
Keeping in mind that the Democrats need to win the popular vote by about 2.5% to win a majority in the Electoral College next year (according to my math), these data should be reassuring for progressives. But they shouldn’t hold their breath; predictions like these often have larger margins of error than we think, and Trump could well beat the odds again next time.
But heed this point: any Democrat who is simply treading water on favorability in November is probably favored to defeat the president.