A Landslide Election in 2004?
In last month's Washington Monthly, Chuck Todd argued that "the next election won't be close
," because "elections that feature a sitting president tend to be referendums on the incumbent--and in recent elections, the incumbent has either won or lost by large electoral margins."
As the table at the end of this post shows, incumbents usually win or lose by landslides. In 7 of the 9 post-war elections, the winner received at least 370 electoral votes, 100 more than the 270 needed to win. The winner has averaged almost 57% of the 2-party vote. Post-war elections without an incumbent have generally been closer, with 3 of 5 decided by a whisker, with the winner getting very close to 50% of the 2-party vote.
Todd doesn't really offer an explanation, beyond the sentence I quoted above. The obvious beginning of a theory is that voters have more information about an incumbent, and they're better able to evaluate his competence and character. The extra information tends to push a good chunk of voters one way or another.
One problem with this theory is that not all elections with an incumbent are blow-outs: Truman/Dewey and Carter/Ford are the exceptions. I would be tempted to write these elections off as aberrations that aren't important enough to discard such a nice and simple theory. But both of these elections have something in common: Vice Presidents who ascended to the Presidency without first being elected on their own.
So why does this kind of partial incumbency make elections closer? Truman and Ford were President for several years, which I'd think would be plenty of time for voters to form an opinion one way or another. I can't think of a neat explanation. Any theories?
Votes for the Winner in Post War Presidential Elections
Electoral Share of
Votes 2-Party Vote Winner/Loser
--------- ------------ ---------------------
Elections with an Incumbent
1948 303 52.4 Truman/Dewey
1956 457 57.8 Eisenhower/Stevenson
1964 486 61.3 Johnson/Goldwater
1972 520 61.8 Nixon/McGovern
1976 297 51.1 Carter/Ford
1980 489 55.3 Reagan/Carter
1984 525 59.2 Reagan/Mondale
1992 370 53.5 Clinton/Bush
1996 379 54.7 Clinton/Dole
Average 440 56.8
Elections without an Incumbent
1952 442 55.4 Eisenhower/Stevenson
1960 303 50.1 Kennedy/Nixon
1968 301 50.4 Nixon/Humphrey
1988 426 53.9 Bush/Dukakis
2000 271 49.7 Bush/Gore
Average 349 51.9
Source: Statistical Abstract of the United States.
In comments, posters emphasize the unique features of the 1948 and 1976 elections, and suggest that strong conclusions can't be drawn from a sample of two. I certainly agree that these two elections can't prove
any general theories. But I think they can generate useful hypotheses that could in principle be tested with better data (say data on gubernatorial elections). Why is it that not having faced the electorate before seems to befuddle the voters so, despite the candidates' several years in office?