Monday, January 9, 2012

Past Performance in Presidential Primaries as a Benchmark in Future Contests

Fair or not (and FHQ leans to the latter), Mitt Romney's vote total/percentage last week in Iowa being so close to the former Massachusetts governor's total/percentage from the caucuses in 2008 has set off at least some expectations-setting chatter about the now-versus-then in other states. FHQ has followed the presidential primary process for a long time and I don't know that I have ever seen this particular metric pop up in the past. The reality is that we just simply don't have that many viable but ultimately failing candidates from one cycle coming back to be viable candidates in the immediately subsequent cycle (emphasis on viable and immediately).

The best corollaries we have in the post-reform era on the Republican side are Ronald Reagan (1976 and 1980), George H. W. Bush (1980 and 1988), Bob Dole (1988 and 1996) and John McCain (2000 and 2008).1  And right off the bat, one gets into all the "yeah, buts". In Reagan's case, the race was a two person battle between the would-be 40th president and the sitting, but unelected president, Gerald Ford. That two person contest is tough to equate with the multi-candidate race in 1980. For the remaining examples, there is a lag that encompasses three total presidential election cycles instead of two back-to-back. In other words, the comparisons are being made with eight years, not four, in between the two points of observation.

To FHQ, these sorts of benchmarks are very tricky because of how many caveats can be involved. Are they more trouble than they are worth? Let's have a look at the early states with similar positions on the calendar from one comparison point to another for Republican candidates, 1976-2012:

Past Primary Performance by State (Early) in Republican Races (1976-2012)
% of vote (point #1)% of vote (point #2)Won State
(point #1)
Won State
(point #2)
(point #1)
(point #2)
Dem. Race?
(point #1)
Dem. Race?
(point #2)
Open Primary?
New Hampshire4850----
New Hampshire2338--
South Carolina1549--
New Hampshire2926------
South Dakota5545----
South Carolina2145----
New Hampshire4837
South Carolina4233--
New Hampshire32??

South Carolina15??



Now, everyone will probably get the most out of the first few columns with the vote percentages and win rates. But I'm hard-pressed not to include other relevant considerations here like whether or not the race was a multi-candidate race, which party had nomination races and whether or not the state had an open/semi-open/semi-closed primary allowing particularly independents to vote. Again, if there are multiple candidates, then there are more candidates vying for a piece of the pie. The result is typically, but not always, a smaller share of the vote for the winner (and other participants). That Reagan's percentage of the vote in New Hampshire went up between 1976 and 1980 despite the race having gone from two candidates in the former to multiple candidates in the latter is noteworthy, for example. That the former president went down in Iowa under similar circumstances is more understandable.

Additionally, the combination of if the Democratic Party simultaneously held a contested nomination race and if any given state allowed independents to vote in the primary matters. If, in one cycle Democrats were involved and peeled votes away from the Republican contest in an open state, but were not in the next contest of comparison, it could have an impact across races/cycles. There are not all that many examples of this occurring as the Democratic Party has had so many contested nomination races in an otherwise Republican era of presidencies. 1996 offers our best hope, and Dole's   numbers went up in South Carolina from a Democratic-contested year like 1988 to one where the Republicans were the only game in town in 1996. FHQ is skeptical just how much of an impact the presence of Democrats in 1988 and their absence in 1996 had on that. [Overall, this is a factor that is likely to play a larger role -- hypothetically -- in New Hampshire if it was to play a role at all. There's no evidence of that above.]

As for whether a candidate improved or declined from cycle to the next in the early states, it is a coin flip in the 14 state cases above. In seven cases, the candidate improved and in the remaining seven cases the candidate lost ground from one cycle to the next. And not even a tiebreaker works if one wants to throw Romney's Iowa performances into the mix. Romney essentially tied his vote percentage from 2008 last week in Iowa. Five of the cases saw an improvement from one cycle to the next lead to a victory for a candidate. But in four other cases, candidates lost ground and either won or still won across the sequential comparison points.

But the question remains: Is this a good metric for setting expectations much less examining performance for a repeat candidate across cycles? The record is mixed, but if ever there was a candidate/cycle combination where it would work, it would be Romney in 2012. But FHQ just isn't sold on whether even that would be effective. So, while we will hear at least some discussion about Mitt Romney's share of the vote -- particularly in New Hampshire tomorrow -- in 2008 (32%) as a means of setting the expectations for the former Massachusetts governor in 2012, it may be flawed simply because independents will flock to the Republican contest while the Democratic primary remains idle.

Fair or not, when you are the frontrunner and basically playing a home game all comparison points will be examined.

...and FHQ is fine with that so long as caveats are added.

1 Incidentally, along with Romney (2008 and 2012) that is all of the competitive Republican nomination races in the time since the McGovern-Fraser reforms.

Are you following FHQ on TwitterGoogle+ and Facebook? Click on the links to join in.

No comments: