“It’s mathematically impossible to win 40% in this field of candidates.” So Stuart Stevens, Mitt Romney’s chief political consultant, told me yesterday. This is part of the expectations game: Romney was running above 40% in New Hampshire polls in the days just after the Iowa caucuses but is now at 38% in the realclearpolitics.com average of recent polls, which includes the most iteration of each tracking poll and other polls taken in January. Stevens obviously wants to avoid the Maria Carrier effect, named after Edmund Muskie’s New Hampshire coordinator in the 1972 Democratic race, who said publicly that anything less than 50% for Muskie (who was from next-door Maine) would be a defeat. So when Muskie beat George McGovern 46%-37%, that was widely treated as a defeat—and McGovern went on to win the nomination.
How credible is Stevens’s claim that 40% is “mathematically impossible.” It’s certainly plausible. If you assume that Ron Paul gets 20% of the vote (which is right where he is in the RCP average) and that Jon Huntsman, Rick Santorum and Newt Gingrich average 13% each (they’re currently at 11.5%, 11.5% and 9.5%, or an average rounded-off of 11% in the RCP average) and that Rick Perry gets 1%, that leaves 40% of the vote for Romney and for the various other candidates on the ballot or written in (in 2008 such candidates together got 3% of the vote). The New York Times’s numbers guy Nate Silver is currently projecting Romney at 39% (I’ve rounded off his percentages), Paul at 19%, Huntsman at 16%, Santorum at 12%, Gingrich at 11% and Perry at 1%. That puts the average of Huntsman, Santorum and Gingrich at 13%.
My conclusion: Stevens taken literally is wrong. It’s “mathematically possible” for Romney to get 40%. But if you translate spin into English he's got a good point: 40% is a high target and would be a very impressive result in a race where five candidates have been actively competing.