Fiorina, Abrams, and Pope: The 2000 US Presidential election

Fiorina, Abrams, and Pope. 2003. The 2000 US Presidential election: Can retrospective voting be saved?. British Journal of Political Science 48:723-41.

Overview

All the retrospective voting models predicted that Gore would win in 2000 by a comfortable margin; after all, the economy was strong and things were going well. But he lost. Apparently, Gore's efforts to distance himself from Clinton worked--and as a result, retrospective facts (i.e. Clinton's policy successes) had less influence on voters in 2000 than in other years. Using some sophisticated time-series statistical analysis of ANES data, the authors estimate how many percentage points Gore lost as a result of several factors. (Note that the effects are not additive.) The only hypothesis not confirmed by the data involves Gore's personality; despite his personality flaws, his personality does not seem to have influenced the vote. However, certain campaign variables did appear to influence the vote, as did retrospective variables.

Four Hypotheses as to Why Gore Earned Several Percentage Points Fewer than Predicted

Retrospective variables had a weaker influence in 2000 than usual

The (ANES) data support this hypothesis. If retrospective variables had the same influence in 2000 as in 1988 (a similar election, in which a vice president ran for president largely on the success of the preceding president), then Gore would have had 8 percentage points more. But why did retrospective variables matter more? Four possible explanations heard from observers:

1. Vice presidents don't get credit. Probably not true, since Bush (1988) cashed in on Reagan's success.
2. 'What have you done for me lately?' The Clinton years had such prolonged success that people no longer saw it as a Clinton-Gore product; they hadn't done much lately. (The authors claim to have disproved this possibility, but they don't really test for it. A better test would substitute incremental variables (e.g. yearly change in economic indicators) for the standard retrospective variables and see whether they mattered in 2000.)
3. Entrepreneurs and Alan Greenspan got the credit. Possibly true. Most survey respondents credited these two groups, not Clinton, for the success of the 1990s (as of a 1999 survey).
4. Al Gore did not try to take credit. Definitely true that Al Gore distanced himself from Clinton, at the expense of claiming credit for Clinton-Gore policy successes. Hard to know how much this contributed to the overall outcome.

People didn't like Al Gore personally, but Bush was likable

Although people talk about Gore's bland personality a lot, perceptions of Gore's personality (relative to Bush's) don't seem to significantly predict vote choice when other things are controlled.

Al Gore was Too Liberal

Al Gore chose to run to the left of Clinton's more moderate position. This may have cost him; 43% of voters thought Gore was too liberal (in an exit poll), but only 34% thought Bush was too conservative. Using ANES data, the authors estimate that Gore would have had 4 percentage points more if he had run as a centrist like Clinton instead of as a liberal.

2000 was an anomaly caused by Clinton fatigue

People were tired of Clinton. And although he had high job approval ratings, he had dismally low personal approval ratings--a factor that led Gore to distance himself from Clinton during the campaign. Perhaps retrospective voting models simply need to control for personal approval ratings to be corrected. Clinton fatigue probably cost Gore 3 to 4 percentage points.