Does Anybody Know Who’s Electable?

Like most Democrats I know, I wish someone could tell me who is electable. If I knew for a fact that one Democratic candidate would beat Trump in the fall, but that all the others would lose, I would absolutely vote for the “electable” one. Bloomberg is currently my least favorite Democrat, but if I were certain that it would come down to either him or Trump, I’d pick him. Bernie? Joe? Elizabeth? Amy? Doesn’t matter. If only one of them can win, sign me up.

And wouldn’t you know it? Lots of people claim they have that information. The problem is that they disagree.

Two theories. There are two basic theories of how Democrats can beat Trump in November:

  • Swing-voter theory. Elections are decided by moderates who swing from one party to the other, depending on who sounds the most reasonable to them.
  • Turnout theory. Non-voters lean Democratic, but they don’t vote because they don’t see politics making a difference in their lives. To get them to turn out, you need to offer bold ideas that clearly would make a difference.

Obama’s 2008 landslide came from doing both: inspiring new voters without scaring off moderates. Doug Jones’ surprising senate win in Alabama followed a similar formula. Jones was a moderate, but turnout was high anyway.

People arguing that Bernie Sanders isn’t electable usually apply swing-voter theory: He’s the most extreme candidate in the Democratic field, so he will alienate moderate voters who otherwise would be ready to vote against Trump. In particular, Trump’s know-nothing style of governing has alienated a lot of educated suburbanites who used to be loyal Republicans. Those votes are available to a centrist Democrat like Biden or Bloomberg, but not to Sanders.

Conversely, turnout theory says that Sanders is the most electable candidate.

In Michigan and Wisconsin, which were decided in 2016 by roughly 11,000 and 22,700 votes respectively, close to a million young people have since turned 18. Beyond the Midwestern trio of states, the demographic revolution has even more transformative potential. Mr. Trump won Arizona, for example, by 91,000 votes, and 160,000 Latinos have turned 18 in that state since then.

Getting those voters to the polls, the theory says, wins not just for Bernie, but for Democrats in general.

Giving voters too much credit. Neither theory is entirely crazy, but both, in my opinion, oversimplify things. Each in its own way gives some group of voters too much credit.

Like iconoclastic political scientist Rachel Bitecofer, I don’t believe in “this informed, engaged American population [of swing voters] that is watching these political events and watching their elected leaders and assessing their behavior and making a judgment.” Similarly, I don’t buy the turnout-theory image of non-voters as disaffected socialists waiting for the clarion call of political revolution.

No doubt there are a few analytic middle-of-the-roaders judiciously weighing each candidates’ positions on the issues, and a few idealistic left-wing radicals who haven’t been voting because see little difference between Nancy Pelosi and Paul Ryan. But in my opinion, the vast majority of swing voters and non-voters are far less impressive examples of American citizenry: They have little interest in politics and little knowledge of it. They’re more likely to be turned off by Bernie Sanders’ hair than by his policies, or they voted for Obama and then Trump because “Yes We Can” and “Make America Great Again” were both good slogans.

These days, knowledgeable people who care about politics have well-defined opinions and show up to vote. Overwhelmingly, the swinging from one party to the other, or from voter to non-voter, is being done by uninformed folks for not terribly intelligent reasons. CNN’s Ron Brownstein observes:

An exhaustive study from the Knight Foundation that examined the roughly 100 million eligible Americans who did not vote in 2016 underscores [Ruy] Teixeira’s point [that non-voters don’t favor either party]. For the study, which was released last week, the foundation commissioned a survey of 12,000 nonvoters nationwide and in swing states, and held focus groups with Americans who habitually do not vote. The results found nonvoters united by their disconnection from the political process and disengagement from the news, but divided quite closely in their views of the two parties. …

[T]he geographic distribution of nonvoters creates challenges for a Democratic strategy centered on mobilizing them, especially in the Trump era. On a national basis, the best evidence suggests, the Americans who are eligible to vote but don’t split about equally between whites without college degrees, who lean Republican, on one side; and minorities and college-educated whites, who lean Democratic, on the other.

Bold liberal ideas are likely to motivate both groups, not just the one.

But the distribution looks very different in the Rust Belt states that tilted the 2016 election. In Wisconsin, Michigan and Pennsylvania, the three states Trump dislodged from the “blue wall,” whites without college degrees represented a clear majority of the adults who were eligible to vote but did not, according to calculations from census data by David Wasserman of The Cook Political Report. The adults who have become newly eligible to vote in those states since 2016, mostly by turning 18, do lean more toward minorities, according to analysis by the States of Change project, which Teixeira directs. But even accounting for those young entrants into the electorate, many Democrats believe that Trump has a bigger universe of potential new voters available to harvest across the Rust Belt than the Democratic nominee does.

Polls. Whenever people argue about electability, they start comparing polls. A few months ago, moderates touted head-to-head polls that had Biden beating Trump by a larger margin than Bernie beat Trump. Lately, progressives have been pointing to polls that either say the opposite or indicate that there’s no real difference.

In either case, the problem is the same: Polls are pretty good at telling you how people will vote tomorrow, because the people they interview are pretty good at predicting their own short-term behavior, as long as nothing important happens between the interview and the election. But polls about what will happen eight months from now are not nearly so enlightening.

They’re especially useless in evaluating candidates most of the public hasn’t formed a firm opinion about yet. My personal intuition (which may not be worth much) is that the Democrat who would run the best race against Trump in the fall is Amy Klobuchar. I have no data to support that opinion, I just think she contrasts well against Trump: She’s sunny where he’s angry. She’s in the prime of life while he’s a fat old man. She’s sharp where he’s confused. She’s a woman while he’s the embodiment of toxic masculinity. And so on.

But whether that’s true or not, I wouldn’t expect to see it in the polls this far out, because most of the country has never really tried on the idea of President Klobuchar. Or President Buttigieg, for that matter. Either one of them (or Kamala Harris or Cory Booker or one of the other longshot candidates who has since dropped out) would look completely different in November than they do now. By November, Nominee Klobuchar would have been at the center of a successful primary campaign, would have given a convention acceptance speech, and stood toe-to-toe with Trump in the debates. What an election would say then just isn’t predictable from a poll taken now.

But OK, let’s consider the possibility that polls can tell us something about the electability of the two best-known Democrats: Joe Biden and Bernie Sanders. As of this morning, the RealClearPolitics average of head-to-head polls of Biden vs. Trump had Biden up 5.4%. The same number for Sanders vs. Trump had Sanders up 4.9%. (The outlier was Emerson, which had Sanders beating Trump by 2%, but Biden losing by 4%.)

That’s a difference of half a percent, with eight months of events still to be processed. And there are more factors to consider. Political scientists David Broockman and Joshua Kalla point out that the Biden voters and Sanders voters are not the same people.

We found that nominating Sanders would drive many Americans who would otherwise vote for a moderate Democrat to vote for Trump, especially otherwise Trump-skeptical Republicans.

Republicans are more likely to say they would vote for Trump if Sanders is nominated: Approximately 2 percent of Republicans choose Trump over Sanders but desert Trump when we pit him against a more moderate Democrat like Buttigieg, Biden, or Bloomberg.

Democrats and independents are also slightly more likely to say they would vote for Trump if Sanders is nominated. Swing voters may be rare — but their choices between candidates often determine elections, and many appear to favor Trump over Sanders but not over other Democrats. Despite losing these voters to Trump, Sanders appears in our survey data to be similarly electable to the moderates, at least at first blush. Why? Mainly because 11 percent of left-leaning young people say they are undecided, would support a third-party candidate, or, most often, just would not vote if a moderate were nominated — but say they would turn out and vote for Sanders if he were nominated. …

The case that Bernie Sanders is just as electable as the more moderate candidates thus appears to rest on a leap of faith: that youth voter turnout would surge in the general election by double digits if and only if Bernie Sanders is nominated, compensating for the voters his nomination pushes to Trump among the rest of the electorate.

(BTW: The Sanders campaign also believes it will bring Trump-leaning non-college whites back to the Democrats. The Broockman/Kalla data does not support this claim.)

So you can try to be as data-driven as you like, but in the end you come back to a “leap of faith”. Will that youth-voting surge really show up? Young people who say they will only vote if Bernie is on the ballot — might they change their minds?

How I wind up thinking about electability. Some pundits go so far as to say there’s nothing to know here, so you should just forget about the whole notion. Unfortunately, I find that impossible. November is so important, it’s hard not to form opinions about it.

Whatever conclusions you come to, though, you should hold them lightly. Use your notions of electability as a tie-breaker between candidates you like, not as your only criterion. Few political experiences are worse than to give up on someone you believe in so that you can win, and then not to win. Cast a vote you can live with.

For what it’s worth, my hunches about electability — and they’re really just hunches — come down on the moderate side rather than the progressive side. My confidence in a 2020 Democratic victory comes from the 2018 victory: Democratic candidates got 53.4% of the vote in congressional elections in 2018. If all the people willing to vote for a Democrat for Congress decide to vote for the Democratic presidential nominee, it’s a landslide.

I see that win as mostly supporting the swing-voter model. Democrats flipped seats by running moderate candidates in suburban districts where educated professionals used to be reliably Republican.

Conversely, I have never seen the turnout model work. If some progressive candidate had won an unexpected victory in some red state senate or house race by using radical policy proposals to bring in vast numbers of new voters, then I could more easily imagine the same thing working on a national level. But I don’t know of any such example. The best-known progressives in Congress come from liberal bastions like Vermont (Sanders) and Queens (AOC). They don’t flip red states. Or at least they haven’t.

Post a comment or leave a trackback: Trackback URL.


  • David Scott Carlick  On March 2, 2020 at 12:11 pm

    This article makes interesting points.

  • Brian Douglas  On March 2, 2020 at 12:14 pm

    I agree with you that “electability” is a somewhat nebulous concept most often dependent on which ever partisan you talk to. I think generally the more centrist candidate has the advantage unless there is also a “likeability” issue. For example, Reagan was portrayed as the more “radical” against Carter. But Reagan was undoubtedly more likeable. Nixon was the more centrist versus McGovern even though Nixon was not exactly “warm and fuzzy”. In 2016 after 25 years of relentless attacks, enough folks disliked Clinton to flip three otherwise Democratic states. In my opinion, Sanders will not only be viewed as the more radical candidate, but also not particularly likeable. To me at least he projects as a stubborn, angry old man (and I’m 70). I have been supporting Klobuchar and unless she drops out before early voting starts in Florida, I’ll vote for her because I have felt that she has the most qualities I look for in a President. But like most moderates, she does not inspire much passion and has not been able to translate her qualities to votes. So at this point, provided he can do reasonably well tomorrow (Super Tuesday) I think Biden is the most electable. He has obvious issues which I am sure will be exploited by the Trump campaign, but in my view he has the best combination of likeability and moderation.

  • George Washington, Jr.  On March 2, 2020 at 12:37 pm

    “Electability” is nothing more than speculation on how other people will vote. Since that’s usually based on little more than stereotypes and assumptions, it’s meaningless.

    • Brian Douglas  On March 2, 2020 at 12:49 pm

      I think “informed” speculation is closer. In Florida, a candidate proposing to levy a state tax on social security benefits in my view would have a difficult time getting elected.

    • Guest  On March 2, 2020 at 2:41 pm

      This sounds like things the Romney and Clinton camps were telling themselves, George. Polls, including electability polls, aren’t magic crystal balls, but you ignore them at your own peril.

  • Guest  On March 2, 2020 at 12:46 pm

    You’ve often relied on the comparison of off-year congressional election outcomes to presidential elections, Doug, and I think it’s an apples and oranges mistake. State-level races have too many local quirks. If we look at presidential elections in the context of the last few presidential elections, I believe a pattern emerges. Reasonable people can disagree, but I see steadfast moderates who hold the party leadership line on the status quo losing (Kerry, McCain, Romney, Clinton) and those who can better credibly occupy the outsider agent of change space in a campaign winning (Obama, Trump). Notably for 2020, we’ve not seen an incumbent lose to a moderate party-leadership water holder in this period.

    But since you insist on the centrality of 2018, I’ll only challenge your assumption that the story of that election cycle was all about the moderates. I recall a lot of the excitement and turnout of that time was fueled by unabashedly lefty organizations and people (the Women’s Marches, the rise of the Squad, etc).

    It’s curious that you come back again and again to the concern of whether or not a Sanders could flip a red state, but always fail to consider that Clinton not only failed to flip red states, but lost key purple ones.

  • Guest  On March 2, 2020 at 3:25 pm

    I applaud you for going out on a limb, Doug, but you have to laugh. You hold Klobuchar up at the best counter to Trump and she immediately drops out of the running. If Warren loses MA tomorrow, the Weekly Sift kiss of death is confirmed 😉

  • bradtem  On March 2, 2020 at 4:44 pm

    Very worthwhile article. One flaw, you cite a national poll at one point. National polls are highly misleading and as we know, winning a big national poll and $5 can get Hillary a latte at Starbucks. The only polls that matter are swing state polls and I would hope that, in your analysis of polls (flawed as they are) you would focus on those.

    They do measure “opinion today” but they can offer some basis of comparison. They can and do reveal that Bloomberg does much better with the large block of Independents than the other Democrats, for example, and that is less likely to change with time than certain other preferences.

    There really needs to be much, much more study of low-probability voters. They are usually not queried in polls and not well understood. The real answer is of course a mix of swing voter theory and turnout theory and the questions lie around what sort of mix it is. In addition, turnout analysis has two components — people on your side inspired to the polls by you, and people on the other side not scared to the polls by you. It is the latter group which are the target of most political advertising, which is negative advertising aimed at scaring weak supporters that the enemy is so scary they must vote. We’ll see a lot of that in this election, though the faults of Trump are widely known, and the GOP will work hard to show the faults of the Democratic nominee when the time comes.

    The two most important groups to examine are indeed, first these “Bernie-or-bust” voters. How many are there, how committed are they do that approach, and are they so committed they would even lie to pollsters? The second is the “not-fond-of-Trump” conservative voters, to get real quantification of how many will stay away from the polls under different Democratic offerings.

    Then we might be able to decide who is most electable.

    • Guest  On March 3, 2020 at 9:38 am

      The low-probability voter angle is a great point, Brad, as they’re a huge bloc in the country and tend to be younger, poorer, less educated, in a word, the disadvantaged in our system. Clinton essentially ceded this bloc to Trump. As in 2016, I only see Bernie being competitive with Trump there. Is there any rationale for Biden doing better? I honestly haven’t seen any.

      On the Bernie-or-Bust voters, I think a floor of 10-15% will jump ship if Biden gets more votes, based on past experience of any candidates. 12% of Bernie supporters voted for Trump in 2016; for comparison, 25% of Clinton supporters voted for McCain against Obama. I think those numbers will explode upward if super-delegates conspire to steal the nomination from Bernie provided that he does win more votes/delegates. On top of that, if that nightmare comes true, it’s easy to see a lot of the low-probability voters who feel disempowered and disengaged getting turned off. Double-whammy.

      On the not-fond-of-Trump/disillusioned Trump voters, my hunch is that they voted for Trump because they are pissed off at their own party establishment, and that their contempt of the other party’s establishment is equal or greater to their in-house contempt. Because honestly they’ve been screwed over by both. Bernie’s outsider status, and the very contempt he earns from the comfortable Democratic establishment, plays to his advantage here. Will it be enough for Sanders to beat Trump? I’d like to think so, but the flip side of that coin is I don’t see any other candidate putting up half the fight. I see Biden following Clinton’s path, they being so close on issues and policies, and having an even juicer scandal than her damn emails with the Burisma nepotism/corruption story.

      • bradtem  On March 3, 2020 at 2:06 pm

        The answer is that a big effort should be taken to determine the size of these voting blocks, particularly the bernie or bust block. In theory, we can start with the lists of people who vote today, which parties get access to. See how many people are voting who are new to the process. Sample them to find out how many are for Sanders. Sample the other things too.

        As you can guess, in a contested convention, “steal” is a very loaded word. After all, the whole _point_ of the contested convention rule is to give the party the chance to select somebody other than the plurality winner. If it’s wrong to give it to other than the plurality winner, you would not have the contested convention rule, you would just make the rule say something like, “the nominee is the winner of the first vote who gets more than 40%” instead of “more than 50%” as it says. The rule exists for exactly the situation where there are factions within the party, and one minority faction consolidated and the other one didn’t, and now the other one has the chance to consolidate into a majority. (The superdelegates rule is a different one and harder to justify as purely democratic, because it is not.)

        But if It’s Sanders 40%, Biden 35% and Bloomberg 25%, and most Bloomberg supporters prefer Biden, isn’t it the correct Democratic thing to give it to Biden? Why do you call it stolen? Problem is that vocabulary will end up destroying the party and help Trump.

        But anyway, measure, measure, measure. Let’s stop having people describe their intuitions on how many of each class of voter there are, including Sanders voters who will stay-home/vote-Trump. Let’s measure them. The cost of that is tiny compared to the other money being spent.

      • Guest  On March 3, 2020 at 4:02 pm

        Thanks, Brad. I’m all for measurement and paying attention to the data, no disagreement here. Wouldn’t be surprised if some of the campaigns have already been tracking this stuff internally, although I’m not sure the DNC would care either way – they certainly didn’t mind the big red flag polling data in 2016. In any case, absent the hard data you are calling for, we can only hope to make reasonable, evidence-backed assumptions. Anything from a 10% to 25% defection rate is a reasonable starting point given recent history. Considering the support Sanders has amassed to far, it looks like it will again be a significant handicap for the centrist. I haven’t seen anybody making the case that the DNC invalidating the votes of Bernie supporters would endear them to the candidate selected by the very leadership in-group that invalidated their votes to begin with. You are of course free to dismiss it as intuition, but I find it naive or utopian to think that wouldn’t exacerbate the defection rate. So yes, let’s measure, but Super Tuesday voters for example don’t have the luxury of waiting for your (very reasonable) polling request to be granted. We have to make due with the data we have.

        Superdelegates are unfortunately a key aspect of our brokered convention process as is. And being as they definitionally non-democratic as you point out, I think “steal” while loaded is justified in this case. I realize that isn’t polite language, but there’s nothing nice about stabbing the guy with the most votes in the back a la Estes Kefauver in 1956 and telling the plurality of Sanders voters “thanks for the input, but WE the party insiders know better and we’re choosing the guy who will serve us.” Whether or not that’s the point of the DNC rules is irrelevant to me on the level of principle, because I value democracy far higher than whatever self-serving, non-democratic rules the DNC put into place without accountability.

        If you’ll allow a little intuition, Nomiki Konst, who has channels into high level DNC committees, has said that a brokered convention is something that most everyone there wants to avoid and is therefor not likely in her view. Grain of salt and all that, but it makes sense. Going against the plurality was disastrous for the Democrats in ’56, while going with the plurality paid huge dividends for Republicans in 2016. My guess is that anything over 40% support for Sanders in the first round and the DNC will be embarrassed into accepting their fate. To your hypothetical scenario, no, I don’t think it’s a given that handing the nomination to Biden in that case is necessarily the democratic move. Based on history, it seems that it would be the strategically imprudent move. A more democratic thing to do would be to hold a runoff election between the top tier candidates among actual voters. Since we do not have such democratic structures in place, go with the will of the voters, follow the most votes.

        As a personal aside, I’m a big fan of EFF, thank you for your service! Everyone go download Privacy Badger and HTTPS Everywhere!

  • Bill  On March 2, 2020 at 5:42 pm

    If I’m not mistaken, voting data from the first four states reveals no groundswell of new voter participation as Bernie always touts. I believe the participation numbers are no better or less than 2016 and 2018. If that’s an accurate reflection of what has happened so far, and accurately foretells what’s likely to happen in Nov., then Bernie as the nominee for the Dems. will be a spectacular catastrophe….. not only for him, but the House and Senate as well. Bernies “leadership” skills could be best used elsewhere.

  • The Serapion Brotherhood  On March 2, 2020 at 10:14 pm

    So we’re talking about the moderate swing voters in the Philadelphia suburbs who went for Clinton at twice the rate Blue collar voters defected from her, just like Shumer said, is that correct? The people who voted for Trump and would vote for Sanders are the ones who believe that government as it is is doing nothing to help them. Since Trump turns out to have nothing for them they would be all the keener to support Sanders this time. One imagines that the pick up of black voters in Flint, Mich., who did not bother to vote for Hilary but who might reasonably believe Sanders would do something to help them, would be enough to swing that state.They would not vote for Bloomberg, and if the nominee isn’t Sanders, it surely won’t be Warren, but Bloomberg, who would scarcely be an improvement on Trump.

  • frankackerman0617  On March 3, 2020 at 7:28 pm

    As a retired computer scientist I can intellectually understand how Trumpsters come to be fixed on their erroneous understanding that Trump in leading America to a better place, but emotionally I fantasize that this cult-like following can be turned aside. In trying to get a better grip on current reality this article: was helpful.

    The article is billed as “The reckonings of one of the South’s white suburban women whose loyalty is key to whether Trump is reelected” but for me the article strongly shows that there is very little chance that there are enough white suburban southern women who will break free from tradition to make an iota of difference in 2020.

    What was more interesting to me was the insights it gave on the subject’s Trumpster husband and the Trumpian culture in which he is embedded.

    My take-away: No imaginable social, political, economic or environmental event is going to keep Trump from winning Augusta Georgia’s congressional district. And I suspect that although some of the issues will be different in other parts of Trumpland, the same conclusion holds.

    Putting aside the nuances of an electoral college win, I believe we can reasonably assume that 40% of the presidential voters will certainly vote for Trump and 40% will vote for his opponent. Thus the continuation of the reign of Trump depends on at most 20% of the voters.

    I seems to me that what Doug points out here is that at this point it is quite impossible to predict with any degree of certainly how this (at most) 20% will vote, but at the moment it looks to me, David Scott Carlick’s post none withstanding, that odds are that the majority of these votes will go to Trump.


  • By I’m Voting for Warren | The Weekly Sift on March 2, 2020 at 12:09 pm

    […] Vote for the candidate most likely to lead your party to victory. This vote requires that you identify who the most electable candidate is, which is not as easy as a lot of people make it sound. […]

  • By Best Courses | The Weekly Sift on March 2, 2020 at 12:56 pm

    […] week’s featured posts are “The Coronavirus Genie Escapes Its Bottle“, “Does Anybody Know Who’s Electable?“, and “I’m Voting for […]

Leave a Reply to Brian Douglas Cancel reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: