Barack Obama is ahead of Hillary Clinton in the (regular) delegate count, and it looks almost certain that he will remain so when all of the primaries and caucuses are completed. Still, the gap is not large, and it is possible that Clinton will end up with the most popular votes. Rightly or wrongly, it appears that some — perhaps many — of the as-yet-uncommitted Democratic superdelegates intend to take electability into account in deciding which candidate to support (see here, here, and here).

What information should they consider in making a decision?

Partisans and pundits have suggested a number of reasons why one might stand a better chance than the other of defeating John McCain in the general election. Clinton is more strongly despised by conservatives and thus may generate larger Republican turnout. On the other hand, she appears to have stronger support among working-class voters, women, and Latinos, whereas Obama is stronger among professionals and African Americans. If the latter groups are more likely than the former to vote Democratic regardless of who is the nominee, this is an advantage for Clinton. Obama seems more likely to inspire independents, but Clinton may be better prepared to effectively confront Republican attacks. I’m not convinced that these considerations clearly favor one or the other.

National polls pitting Obama or Clinton vs. McCain are another potential source of information. They often show little difference between the two, but it’s too early for them to be of much use.

If I were a superdelegate trying to assess electability, I’d be inclined to focus on which candidate is most likely to win states that are not solidly “blue” or “red.”

The following chart shows the 24 states in which the popular vote in the 2000 and/or 2004 presidential election was within 10 percentage points. The numbers in parentheses indicate the 2000 and 2004 vote results. A plus sign means the Democratic candidate (Gore, Kerry) won the state; a minus sign means Bush won it. The states are ordered by the number of electoral votes they’ll have in the 2008 general election. The markers (“Ob” and “Cl”) indicate the winner of the state’s primary or caucus.

Consider first the 20 states at the bottom. Together they have 159 electoral votes. All but one have had their primary or caucus already. Of them, Obama has won states that have total of 98 electoral votes, and Clinton has won states with a total of 56 electoral votes. (West Virginia, with 5, holds its primary in May.) These results favor Obama.

One can argue that perhaps some of these states aren’t truly in play. New Jersey seems likely to go Democratic regardless of who is the nominee, and so too do Hawaii, Maine, and New Hampshire. Similarly, neither Obama nor Clinton is likely to win Tennessee or West Virginia. As best I can tell, though, subtracting these types of states would not change the picture much.

What could change it significantly is if all of the “big four” states — Florida, Pennsylvania, Ohio, and Michigan — go for Clinton. These states have 85 electoral votes between them. They decided the 2000 and 2004 elections, and might well do so again this time. If Clinton were to win these four states, she’d have won in-play states with a total of 141 electoral votes, versus 98 for Obama. Though primary results don’t automatically translate into general election performance (see this and this), this would give her a case for claiming greater electability. Otherwise, it seems to me that electability either is a draw or favors Obama.

Clinton won Ohio. Pennsylvania holds its primary April 22nd. If I were a superdelegate concerned about electability, I would want a true primary result from Florida and Michigan.

Addendum:  Jeff Weintraub has some sensible thoughts on what the Democrats should do about Florida and Michigan.

7 thoughts on “Electability

  1. To me, this seems an odd way of figuring out electability. A survey of democratic voters (basically what the primary results can be interpreted as) wouldn’t seem to have that much relation to the voting patterns of the entire (voting) population (which is what would matter in a general election.

    At the very least, it would seem that one would have to figure out to what extent the democratic voters who participated in the primaries represent the voter population in general.

    My guess would be not very much.

  2. Interesting, but–absent an Obama meltdown,

    It’s over.

    Hillary needs circa 60% of the remaining delegates/votes to claim any kind of lead.



    Again, absent an Obama meltdown, that just isn’t going to happen.

    The superdelegates have never overridden the popular mandate, and they never will. It would cause a nuclear meltdown in the Democratic party.

    1. They know better than anyone which side their political bread is buttered on.

    2. They care more than anyone about the health and unity (and victory) of the Democratic party.

    Obama’s the presumptive nominee. It’s time for him to start acting like it. (Politely) ignore Hillary and her Ferrarogates, and turn all attention to McCain.

    Hillary can certainly keep hoping for the meltdown, and her best personal strategy is to campaign to cause it. But Obama should act as if it’s not even a possibility–and not raise to her bait.


    The best indicator of electability we have, btw, is the recent Survey USA state-by-state poll, which is of registered, not *likely* voters. Obama leads clearly.

    Now on the Hillary/McCain comparison, multiply the McCain number by the number of your choice to account for Hillary turning out Republicans.

    Do the opposite for the Obama/McCain comparison.

    You don’t need very big multipliers to turn it into a comparison between a landslide and a rout.

    As The Economist said, “If what should be a cakewalk in November turns into a rout, the Democrats will know who to blame.”


    Hillary: “a one-woman answer to the Republican’s problems.”

  3. Although you’re reasonably cautious, I think trying to infer the general outcome from the primary data is a crap shoot. You have to take in to account whether the primaries are closed or open. If you don’t have an open primary, it doesn’t tell you enough about general election support. Florida has a closed primary, so far as I know, Michigan’s was open, but I suspect fewer independents/republicans would vote in the democratic primary if they knew the vote wouldn’t count. Ohio was semi-open, so that’s perhaps more telling. Pennsylvania will be a closed primary (also important is that I think Rendell is right that Pennsylvania is likely to go democratic for either nominee). You’d also want to know whether there were simultaneous (competitive) repubblican primaries going on at the same time

    Anyway, I think a better way to gauge electability is to look at polling for the general, and/or demographic appeal. As you noted, there’s not much to go on there either.

  4. The real issue is not how well Clinton or Obama might do in battleground states, but that we shouldn’t have battleground states in the first place.

    We should have a national popular vote for President in which the candidate who gets the most votes in all 50 states should win. There would be no red, blue, or battleground states. Then, everybody’s vote would be equally important throughout the country.

    When the National Popular Vote is enacted by states possessing a majority of the electoral votes, these electoral votes (enough to elect a President) would be awarded to the presidential candidate who receives the most popular votes in all 50 states (and DC). This change to a national popular vote can be accomplished state-by-state, and Maryland and New Jersey have already passed this legislation.

  5. One of the messages from the primaries/caucuses that should disturb is this: Most of the working class white vote in the most populous (and therefore elector-rich) states will probably be influenced by race in a McCain-Obama fight. I will likely goto a white Republican rather than to a black Democrat, wherever the voters economic interests lie. I fear this will be one of the messages from the coming PA primary as it was from the Ohio primary (although that message was tainted by the news of Obama’s reported disclaimer of the promise he made to renegotiate NAFTA)..

  6. I have a better idea. I don’t know if there’s much indication that there are many voters who love one candidate and hate the other and will make a point of staying home or vote McCain if their primary candidate loses. The most probable effect is that some enthusiasm and therefore voter turnout will be lost among those whose candidate loses, so your superdelegate should be counting electoral votes based on who won in states where a small drop in voter turnout could end up losing the race. In other words, they should just stick with whoever won in the “swing states”.

  7. I agree with Alex, and would like to expand it:

    I don’t think we can completely discard initial polling in those states, because while there is significant movement between GOP-Dem that is likely, the initial results give us a bit of a coloring of the Obama-Clinton matchups in battleground states. Clinton does disastrously bad in the West, and Obama does significantly worse than Clinton in the Northeast. I don’t see the tenor of that trend changing much, even if the independents swing one way or another in the general.

    There is also this (qua Alex): Obama’s ability to charge the youth vote, while blunting the anti-Hillary conservatives, is a greater advantage I should think than Clinton’s advantages among blue collar whites.

Leave a Reply