The average loss of House seats for a 3rd election of a presidency is, I think I heard, 41. This means the Republicans could lose a fair chunk of seats and still, by historical standards, come out way ahead.

It's possible that the Democrats will do well in Nov (unlikely, but possible), but that's just going to be worse for them in the long run, as they'll no doubt convince themselves that it's the start of a new trend, rather than a slight deviation from the nigh unstoppable slide rightward.

Never in modern times has a president been able to add Senate seats in the dreaded sixth-year election.

[W]e've discussed at length the "sixth year itch" election in two-term, eight-year presidencies (see here). There have been six of these elections in the post-World War II era (1950, 1958, 1966, 1974, 1986, and 1998). The average loss for the White House in these sixth year elections has been six Senate seats--double the overall midterm average loss of three seats. Thus, if Republicans just barely hold onto the Senate in 2006--a sixth-year election--the Bush White House will try to declare a major historical victory! Of course, just try selling a 50-50 Senate with Vice President Cheney breaking the tie as a glorious outcome.

Instead, the White House will be hoping to match the remarkable achievements in four specific midterm elections in the last five decades. [...] But there's one fly in the ointment: All four of these triumphant White House midterm elections were in a President's first term. Never in modern times has a president been able to add Senate seats in the dreaded sixth-year election. Of course, look at George W. Bush's remarkable electoral record so far. Sooner or later, there's always a first time for every mark in the record book.

From The American Congress, 3rd ed. (PDF):

For House midterm elections, (but not for Senate midterm elections), the number of Congressional seats that the two parties win is predicted well by the state of the economy, and by the public's evaluation of the President's performance. - page 25

Between 1946 - 1996, the President's party suffered an average loss of 24 seats in midterm elections, with second-term midterm elections resulting in almost twice as many losses as during first-term midterm elections, on average.

However, in 1998 and 2002, the President's party gained seats - but in both cases, the President was very popular at the time. - page 26

So...

The economy's booming, but the President's approval ratings are only fair...
Bush and Republican fatigue vs. the increasing moonbattery and Leftward tilt of the Democratic establishment...

It seems reasonable to expect the GOP to lose 20 (+/- 5) House seats in '06.

Posted by: Michael Herdegen at January 13, 2006 5:10 PM

Yep, I don't remember any great Dem loss in '98, either. Wasn't that the election that directly lead to Gingrich's getting booted for leading the GOP off a cliff instead of exploiting the "six year itch?" If the GOP hold this year, what will that do to this infallible predictor being wrong twice in consecutive applications in less than a decade?

Where are they going to lose these 20 ±5 seats? I thought that thanks to gerrymandering most House seats were safely under the control of one party or the other.

All these numerological predictions should come with that disclaimer you hear at the end of investment advertising: "Past performance is no guarantee of future results." Too many of these "trendlines" are based on someone slapping a straightedge against two points and noticing that a third happens to come near the line. (And a misunderstanding of the difference between interpolation and extrapolation seems to play a part, too.)

The opposition Labor Party read the signs — and accused the United States of meddling.

"Aside from saying, 'Go vote Sharon,' they did everything else," charged Isaac Herzog, a Labor member of the Knesset parliament.

...as opposed, of course, to when James Carville was handling Ehud Barak's election campaign against Likud seven years ago. He was merely a non-aligned independent contractor in that one, with no connection whatsoever to the sentiments of the Clinton Administration.

This is a pdf, but must reading for the midterm:

http://www.psqonline.org/cgi-bin/99_article.cgi?byear=2003&bmonth=spring&a=01free&format=view

In the pdf that oj references, Gary Jacobson makes a good case for GOP control of the House being a "natural" result, due to sociological and demographic factors, combined with gerrymandering.

However, the GOP and the Bush admin are heading into the most unfavorable type of election for a ruling party, a non-first-term, midterm election.

IF the election were to be held next week, there would be NO POSSIBILITY WHATSOEVER of repeating '98, i.e., the ruling party picking up House seats in a second-term midterm.
However, there's also little chance of '94 being re-run, i.e., the out party picking up 50+ seats.

Jacobson identifies four factors that affect House election results: The national economy, Presidential popularity, the number of seats that each party has to defend, and voter turnout.

The economy strongly favors the GOP, this year.

However, Bush's approval numbers aren't going to be stellar, barring the unexpected, and even many Republicans find themselves growing weary of the Bush admin.
Abramoff may well tar both parties, but it's a sure-thing that most of his money went to GOP Reps, if for no other reason than that Republicans control Congress.
A drawdown of U.S. troops in Iraq removes some of the impact of being the "war party", but it also allows focus to shift to domestic issues, where the Dems are better-trusted by a majority of Americans.

The GOP will be defending 30 more seats than will be the Dems. Since there are dozens of House Republicans from districts that barely went for Bush in '04, or that went to Kerry, they especially are vulnerable to low Presidential approval numbers, and low GOP voter turnout.

Turnout will definitely favor the Dems.
Bush fatigue, Abramoff, two "stolen" Presidential elections, a back-burnered war, few "gay friendly" ballot initiatives...

The biggest asset regarding turnout that the GOP has going for it, as Jacobson points out, is that Dem voters tend to be highly concentrated in fewer districts, whereas House Republicans tend to be elected with smaller margins - but in many more districts.
The downside to such GOP voter-distribution efficiency is that it takes less movement in public sentiment to move districts from leans-GOP to leans-Dem.

While my guess is that the GOP will maintain a slim House majority after the elections, I should point out that I expected bush to win the '00 election going away, and in '04 I thought that the GOP would pick up 7 more House seats than they actually did.

Therefore, a slim Dem majority in the 110th Congress wouldn't surprise me.

Posted by: Michael Herdegen at January 14, 2006 4:11 AM