Bell Curve The Law Talking Guy Raised by Republicans U.S. West
Well, he's kind of had it in for me ever since I accidentally ran over his dog. Actually, replace "accidentally" with "repeatedly," and replace "dog" with "son."

Monday, July 05, 2010


Here's an amazing fact. In the H of R, 202 of 435 were elected before Jan 1, 2001 (47%). In the Senate, just 47 of 100 (47%). I picked this date more or less at random. (Other dates are similar. 65 Senators were seated before Jan 1, 2005, while only 142 House Reps. That's 32% of the House and 35% of the Senate in the last five years). The first time I ran these figures I got different results, since I can't count. Given the two-year vs. six year terms, this is an astonishing result. The number of freshmen House members (those who took office on or after Jan 1 2009) is 64, or 15%. The other 85% have been reelected, and nearly half (47%) more than four times. 35% of Senators have served fewer than six years.


Raised By Republicans said...

So you are saying that the incumbency advantage is greater in the House of Representatives because of gerrymandering, right?

That's interesting. It really drives home the "All politics is local" maxim. Since district boundaries are determined (in most states) by the majority party in the state legislature, whichever party happens to have control of the most state legislatures in the year following a census will have a distinct advantage in getting a majority and keeping it in the House of Representatives.

The Law Talking Guy said...

Yes, that's what I'm saying. It's all the more astonishing because you would think intuitively that a Senator with the higher profile and longer exposure time would have a greater advantage from incumbency.

It really shows you why there are so many people groping for a solution to the gerrymandering problem that has made the H of R so noncompetitive. It has removed, in many ways, the potency of the main institutinoal check on the H of R - the frequency of elections.

Raised By Republicans said...

I wonder though how much of that advantage is due to the particular shapes of the boundaries and how much is due to the small size (and therefore less diverse electorates) of those districts.

My guess is that even in states where districts are delineated by balanced panels that try to minimize gerrymandering (like Iowa for example), there is a strong incumbency advantage.

I'm not up on the research on US Congressional electoral rules unfortunately. I bet someone has studied this.

The Law Talking Guy said...

Why would small size create less diverse electorates? I think that's a fallacy. Also, the size of districts - approx. 650,000 is pretty big. If smaller size created less diverse electorates and mroe staying power, you would expect almost all statehouses to be more fixed. In fact, the NH lower house, with its 500 members (1 for 4000 people) is famously prone to major shifts every election.

We actually know that in many states, such as CA, the gerrymandering is deliberately designed to protect incumbents. They say so. The 2000 redistricting in CA was a known compromise, the Dems kept a big advantage but tried to make sure existing GOP elected officials wouldn't lose their jobs. (note that in CA, a 2/3 majority is needed not just for the budget, but for other Big Things, because legislators can just hold the budget hostage...).

Raised By Republicans said...

LTG, to start, I'd like to clarify that I don't think gerrymandering is meaningless. I just think there is evidence of other factors at work - such as district size.

To respond: You are raising three objections to my point about district size. I'll try to deal with them one at a time.

1) A district with 650,000 people is bound to be less diverse than an entire state with several million - unless we're talking about that handful of really small states with only one or two districts. Please note the RELATIVE part of my assertion. I'm not saying that house district voters are an undifferentiated bloc. Rather I'm saying that a district that includes one or two towns is likely to be dependent on a smaller set of interests than an entire state that includes big cities, suburbs, small towns and rural areas.

Imagine a district that contains a large factory. If that factory closes it would be devastating to that district but not necessarily as devastating overall to the state at large.

2) I agree that gerrymandering is intended to ensure incumbent advantage. That doesn't mean it's the only thing that ensures this or even the most important thing. The fact that states with less gerrymandering (states with more balanced redistricting panels), also have strong incumbency advantages would suggest that there are other ways to get to the same result in addition to gerrymandering.

3) State legislatures are second order (or third order) elections. I'm not surprised that they are more volatile. Also, state reps in many state are known to be "less professional" that is they are known to be either part time (even unpaid) positions that have a high turnover because candidates either seek higher office or drop out of politics after a shorter career. Indeed, the example you pointed to, the New Hampshire General Court is one of the least professionalized legislatures out there ( ).

Bob said...

I am intrigued, and then confused, by your post.

First: this is the kind of interesting data that would be worth tracking and demanding comparisons to when people talk about whether X was a "throw the bums out" election and suchlike.

Second: there are words missing, or something. As written, the second sentence says 202 Congresspersons were elected before 2001. The first parenthetical seems to say 142 Congresspersons were elected before 2005. Which can't be right.

Third: this might be fixed by whatever's going on with my second point above, but the data seems to show that the time of incumbency is roughly equal in the House and in the Senate. Is that about right? If anything, it strikes me that I didn't realize how many career congressmen there are. I always assumed that it was more often a steppingstone job to the Senate (or maybe Governor). Then again, there's only so many seats available, so maybe there's lots of congressmen frustrated in their ambitions.

The Law Talking Guy said...

Bob is right. I meant to say that 35 Senators and 142 House Reps were seated AFTER 1/1/2005, 65 and 293 beforehand. Every time I do the numbers I run into some brainfart problem with the seniority rank and the # of persons. 35 senators all rank at 65 and above, you see, in terms of seniority. Thanks.

The Law Talking Guy said...

"The data seems to show that the time of incumbency is roughly equal in the House and in the Senate. Is that about right?"

Yes, it seems that way. Which is shocking if you realize how much more frequently house members have to stand for reelection. It's like dog years.

When compared election by election, you see that the House has a smaller % of turnover than the Senate by quite a bit..

The last year (2008) which was a big house year had 64 new reps, about 15% of 435 who stood for election. In the Senate, 16 new members came in since 2008. That is almost half of the 34 seats up for election.

Dr. Strangelove said...

As far as I can tell, there are two main strategies for gerrymandering: (1) to protect incumbents, and (2) to get more seats for your party. For example, California chose the first strategy for its congressional delegation in 2000, while Texas chose the second. Either way, supporters of the minority party will get concentrated in fewer districts. Supporters of the majority party however may be concentrated or spread out, depending on the strategy chosen. So I don't think it's accurate to say that gerrymandering always favors incumbents. It may be designed to unseat incumbents as well.

Statistically speaking, RbR is of course correct that a smaller sample of voters drawn at random from a larger population is more likely to show a net bias on the political scale than the whole. And since districts are not random but correlated with geography to some extent, the expectation of some sort of net bias can only be higher. However the same sort of statistical argument also suggests that smaller districts should be more volatile, so I believe the example of the NH statehouse if anything cuts the other way.

Finally, it occurs to me that the fact that both houses of Congress show similar incumbency rates might reflect an efficient use of campaign funds on the part of the national party organizations to protect their incumbents across the board.