This is a guest post by Catholic University political scientist Matthew Green.
When Eric Cantor became the first House majority leader in history to lose his primary, it initially seemed there might be a spirited race among the House Republicans to fill his leadership post. But within 48 hours, one candidate — the House GOP whip, Kevin McCarthy of California — had been all but anointed majority leader. How did McCarthy lock up the race so quickly?
McCarthy’s opponent, Rep. Pete Sessions of Texas, explained his decision to withdraw from consideration as wanting to avoid “painful division” in the GOP. The more likely reason is that Sessions was going to lose.
How do we know this? Certainly some news stories written before Sessions bowed out suggested that McCarthy’s victory was all but inevitable. But such reports are often based on information leaked by the candidates themselves, who hope to create an aura of inevitability around their campaigns and bring undecided lawmakers to their side. When Sessions dropped his bid Thursday evening, only 15 Republicans had announced whom they would support, hardly enough to draw conclusions about the strength of McCarthy’s candidacy.
One could surmise that Sessions was toast based on some broad advantages that McCarthy had — not least that he was already whip and had considerable experience in counting votes. But dark horse candidates have won leadership races before, and history has shown that being party whip is no guarantee of a promotion.
- Ideology (in this case, conservatives should prefer Sessions to the more mainstream McCarthy)
- Campaign contributions (from each candidate’s leadership PAC to his colleagues)
- Seniority (with more senior legislators leaning toward more senior candidates like Sessions)
- Being in leadership (which in this case is expected to predict support for McCarthy)
- Serving as a committee chair (likely predicting support for fellow-chair Sessions)
- Whether a lawmaker is from the same state as the candidate (in this case, California or Texas)
Based on how these factors explained preferences in the 2006 GOP majority leader race, I found that only 41 Republicans are estimated to have had more than a 50 percent chance of supporting Sessions — less than one-fifth of the GOP conference in the House. (That includes Sessions himself, for what it’s worth.) One reason for this lopsided result is that, while more Republicans hail from Sessions’s home state, McCarthy gave far more money to a lot more of his colleagues: over $380,000 to 68 Republicans, versus Sessions’ $41,000 to 17 Republicans.
This sort of back-of-the-envelope statistical exercise should be taken with many grains of salt (though it did successfully predict seven of the eight lawmakers who initially endorsed Sessions). Had Sessions continued to campaign, for instance, he might have found ways to win more votes in the actual election.
Moreover, statistical models of leadership races fail to account for a good deal of variation in legislators’ vote choice — variation caused by things like personal relationships, which matter a lot in these races but which are very difficult to measure. Yet the results do add credence to the claim that McCarthy had a big advantage over Sessions from the get-go.