Rep. Joe Walsh is one of many Republican incumbents in a difficult re-election race.
At the Rothenberg Political Report, we have counted 200 safe Republican seats and another 22 seats where the GOP has a clear advantage. Democrats have 161 safe seats and another 17 seats that lean their way. Even if Democrats win all 35 tossup races, they would fall five seats short of a majority.
Of course, Kristol's suggestion that Democrats could retake the House isn't based on an assessment of individual races. It's based on the generic ballot number and his assertion that a virtually even popular vote "tends to translate into pretty even results in seats split between the two parties."
But, contrary to Kristol's assumption, a generic vote that shows the parties even probably would not translate into a roughly even number of seats for the parties. Experienced Democratic observers point out that their voters are not evenly distributed throughout the country. Instead, Democrats tend to be packed into urban areas, which mean that they are likely to underperform a very close popular vote, at least slightly.
The recent round of redistricting only added to Democrats' problems, as Republicans protected their incumbents and took districts off the table.
More than a year ago, New York Times contributor Nate Silver wrote about a statistical analysis he performed. "If the House popular vote were to be a tie next year, Democrats would pick up a few seats - but very probably not enough to win the majority from Republicans."
Second, history suggests that it will be difficult for Democrats to regain the House next month.
Given that the chamber didn't flip for more than 40 years, regardless of the kind of election, it may not be meaningful to note that the House hasn't changed party in a presidential year since 1952 (or since 1948 when a sitting president beat back a challenger), as Paul Kane of the Washington Post has observed.
Still, it's hard to ignore the fact that in four of the last five times a sitting president was re-elected (2004, 1996, 1972 and 1956), his party gained or lost a dozen seats or fewer. In the other case, in 1984, the president's party gained 16 seats. In other words, election years when a sitting president is seeking another term generally haven't produced dramatic swings in the House.
What would explain the relatively small changes in the House in presidential re-election years, especially in light of the huge swings in midterms?
I have a working hypothesis: For most voters, federal elections are primarily about the president - either the performance of the sitting president or the choice for the presidency.
In midterms, the only way voters have to express support or dissatisfaction with a sitting president is to vote for or against his party. So, when a president is unpopular, there are large swings in the House.
But in presidential years, voters can make a statement about the president by voting for him or against him. Their vote for the House is a separate vote, and they are more likely than not to support the incumbent if he or she comes from the same party of the majority of voters in the district or has avoided major problems.