March 16, 2004

the New Hampshire Anti-Incumbent Ratio

George W Bush, running virtually unopposed for re-election this year, only won 80% of the vote in the New Hampshire Republican primary. Some of the other 20% went to an assortment of obscure Republicans, but most were write-in votes for the Democratic candidates. In fact, John Kerry’s 4% earned him second place. This got me to thinking that the level of write-in support for candidates from the opposite party might give some indication of how endangered an incumbent president is, at least in New Hampshire. After fiddling around with the data, I decided to also consider the number of write-in votes for the incumbent in the opposite-party primary. For example, Bush got 257 write-in votes in the Democratic primary this year.

As a result of this not-nearly-exhaustive research, I present the NHAIR, the New Hampshire Anti-Incumbent Ratio. Simply, it gives the number of “anti-incumbent” votes (i.e. write-in votes for opposite-party candidates in the incumbent’s own primary) for each “pro-incumbent” vote (write-in votes for the incumbent in the opposite-party primary). Here are the results going back to 1972, covering all the primaries except for 1988 and 2000, when there was no incumbent running. (Data sources are here and here)

Incumbent/Year Votes for incumbent in opposite primary votes for opposite party candidates in incumbent primary Ratio (anti-incumbent votes per 1 pro-incumbent vote)
Nixon 72 854 1821 2.1
Ford 76 405 1465 3.6
Carter 80 788 3704 4.7
Reagan 84 5058 7681 1.5
Bush 92 1434 8845 6.2
Clinton 96 1972 8690 4.4
Bush 04 257 8083 31.5

So, for every Democratic primary voter who wrote in Bush this year, there were 31.5 Republican primary voters who wrote in one of the Democrats. That puts him slightly out of line with the trend since 1972. Here’s a graphic representation of the comparison (if the graph images below are not displaying, you can see them, in reverse order, here):

Now the question is, does this metric correspond in any meaningful way to the subsequent fortunes of the incumbent in November? Well, with such a small data set and so many unaccounted-for variables, its statistical and real-world significance are obviously suspect. The presence of serious intraparty opposition in some cases (Buchanan, Kennedy), the effect of third-party candidates in November (Perot, Anderson), solidity of each base, and possibly evolving mores about employing cross-party protest votes are but a few of the messy realities to consider. But that’s why you’re reading this in the blogosphere and not in a peer-reviewed journal. So with those caveats out of the way, take a look at this plot of the NHAIR with the incumbent’s percentage of the vote in NH in November:

Carter appears to be an outlier, but not in a way that gives much hope to Shrub’s fortunes. According to this trendline, Bush can expect to receive somewhere in the neighborhood of -150% of the vote in New Hampshire. And statistics don’t lie.