Charlie Cook has been more sober-minded about Republican gains in 2010 than, say, Michael Barone, so when Cook says that the Republicans could win a majority in the House his words have to be taken more seriously. Nonetheless, talk of a GOP takeover of the House still seems very unrealistic. Andrew Kohut is correct when he says:

While there is every reason to believe that the party is in trouble and will lose seats this year, there is no solid data that would justify a view [bold mine-DL] shared by many here in Washington that the Democrats are destined to lose control of the House. This certainly could happen, but it is really too early to jump to that conclusion.

Kohut notes that this year neither party has strong favorability ratings, which differs from 2006 when Democrats were viewed more favorably and differs from 1994 when both parties were viewed quite favorably. This difference in favorability is something Nate Silver has observed before. Kohut also draws attention to the difference in presidential approval ratings between Obama and Clinton. While it is true that Clinton had managed to climb back to 46% by the time of the midterms, his approval during 1994 had sunk as low as 39% and had been as low as 37% in his first year. In Gallup’s tracking and in both the RCP and Pollster.com averages, Obama has never fallen as low as 45% since he took office. Obama suffered a significant drop in approval over the course of 2009, but since about September his approval rating has been fairly stable and has not shown the volatility of Clinton’s during the same period. During the first two years Clinton’s rating went from 58% at inauguration down to 37% in June 1993 back up to 58% around the time of his first State of the Union address, and then down to 39% in late summer, and then back up seven points before the election. If Obama seems to have nearly the same approval rating that Clinton did before the ’94 blowout, that obscures how much worse Clinton had been doing earlier in his first two years.

Something keeps bothering me about the midterms this year. Unemployment is nearly 10%, and Washington conventional wisdom has determined that the new administration has tried to do too much too quickly. The same conventional wisdom has unaccountably concluded that left-wingers are dominating the scene and decided that the Democrats have therefore offended the basic center-right alignment of the public. The economic conditions and the extremely convenient political narrative suggest an electoral rout of major proportions, but the generic ballot and presidential approval numbers show nothing of the kind. Reagan’s average Gallup approval rating for 1982 was 43%, and that is the sort of approval rating we should already be seeing if the Democrats were going to suffer losses similar to the last midterm election that followed a very severe recession, much less the epic defeats and loss of majorities that some are now predicting. We hear a lot about an intensity gap, but at least two surveys in the last month have found more respondents stating that they are going to the polls this fall to express support for the administration rather than opposition. That guarantees nothing, but it is consistent with previous surveys from 1998 and 2002 that showed more Clinton and Bush supporters than opponents.

Kohut cites one of these, and I have mentioned the NBC/WSJ poll from late January that showed pro-Obama midterm voters outnumbering opponents 37-27%. As I said, taken at face value, these figures are consistent with other midterms where the presidential party gained seats in defiance of the traditional pattern. I’m not so contrarian and crazy as to claim that anything like that will happen, but it is an odd detail that does not fit the convenient story we are hearing these days.

I noticed something else when I was looking over that Pew survey on the Millennials. Looking at Pew’s generic ballot numbers from 1994, 2002, 2006 and 2010, we see that the two midterms that are most alike in their numbers are 2002 and 2010. According to Pew, Democrats lead 45-43 for this year, and they led 46-42 in 2002. As it turned out, Republicans made small gains in ’02 (8 House seats, 2 Senate seats) despite being the presidential party, and it seems reasonable to expect that they will do slightly better than that this year. They do have the advantage of being the party out of power in poor economic times, and first midterms are typically good for the out-party in any case, so the argument I have been making for a 1978-like midterm election with opposition gains of 15 to 20 House seats and 3 to 5 in the Senate remains quite reasonable. Obviously, much will depend on individual races and circumstances in each district and state and on results of the competition between the parties for fundraising. As they usually are, events during the rest of the year will be very important. That said, the evidence we have right now suggests modest, slightly below-average to average gains for the GOP.