Who should be the Democratic nominee to oppose Donald Trump’s re-election?  That question boils down to two: who among the Democratic aspirants would make the best president, and who would be the strongest candidate against Trump? The answers to these two questions could very well be different.

My own answer to the first question is clear: Elizabeth Warren.  That choice of course reflects my personal policy preferences.  I believe that the current gilded age requires significant structural change in American capitalism, and only Warren and Bernie Sanders among the Dems are really serious about pushing for such change.  I like Bernie but I don’t think a 78 year-old man with heart trouble who is also an avowed socialist is a credible candidate for president.  (Others disagree, apparently—Sanders is now ahead of Warren in the polls among Democratic voters.) I think of Sanders as a cause—an admirable cause—rather than as a candidate.  So, my candidate would be Warren.

Would be.  But, because the Trump presidency is such an ugly disaster, the question of relative electability looms very large.  Beating Trump is the overriding necessity in 2020.  Every single one of the Democratic candidates is infinitely preferable to Trump.  If I thought one of my least favorite among them–say, Biden or Bloomberg–had a significantly better chance of winning than Warren, I would go for him, unless it seemed that even Warren was a heavy favorite to beat Trump.

At this point in the campaign, I don’t think anyone but a blind partisan would make a confident prediction about the outcome next November.  The best we can do is look at the current polls.  The website Real Clear Politics (RCP) does us the great service of compiling all the most reputable polls related to the 2020 horserace and, very helpfully, averaging the results. The average of several polls on any single question or election can be assumed to be more reliable than any one poll.

If we are to believe that the current polls are a decent indicator of the probable election results 11 months from now, then Joe Biden is the strongest Democratic candidate: the RCP polling average shows him beating Trump 51%-43%. Sanders and Warren also beat Trump, but while Sanders does very nearly as well as Biden, Warren’s margin over Trump is much less comfortable.  The only other Democratic candidates for whom RCP has compiled polling data are Buttigieg and Bloomberg, both of whom are more centrist than Warren/Sanders and both of whom do less well than they do against Trump. The fact that the leftiest candidate—Sanders—does nearly as well against Trump as Biden seems to deflate the argument that the Democrats need to stay near the center to win.  But, overall, the current polling evidence is a mixed bag as to whether the Dems need to stay centered or go left.

It gets murkier still when we consider the electoral college.  All the polls cited above deal with the national popular vote.  I would be very surprised if the Democratic nominee didn’t win the popular vote, but that’s not how we elect the president.  The electoral college result and thus the election will be decided in the six big swing states that Trump carried in 2016: Michigan, Pennsylvania, Wisconsin, Ohio, Florida and North Carolina. The Democrats need to take three of those states in order to win.  They have very good chances of doing so, but it’s no slam dunk—the election is likely to be close in all of these states. Biden polls more than a little better in these states than do Warren and Sanders.  So, I’m still on the fence. If a centrist has a clearly better chance in the swing states, I couldn’t justify staying the course with Warren.

Of course, a lot can change between now and next November.  Most Americans aren’t yet paying nearly as much attention to the 2020 race as you and I are.  As the campaign heats up, the Sanders/Warren economic populist message could catch fire with working class voters in the swing states.  Or it could crash and burn with moderate suburbanites. I just don’t know.  Neither does anybody else.