A puzzle on US politics

Yes. You’re an Aussie, so it’s not clear to you, but in the US, it’s no puzzle.

The “safe” seats are created by redistricting so that the voters are overwhelmingly of one political party. That means in effect whoever wins the *party* nomination, wins the general election for the seat.

Party primary voters are well-known to be more partisan. The way to win the *party* primary is to appeal to those partisans.

Plus the Republicans have very explicitly focused on motivating the partisans as the key to victory (it’s not that they invented the strategy, but they’ve been working that angle with much greater success in the recent past). Making it work has been taken to be Karl Rove’s great achievement in practical political science.

There’s plenty of discussion of this in some US political punditry, it’s pretty standard analysis, though wonkish.

The triumph of incumbency is all about redistricting and pork. Polling is reliable here: people will give big disapproval ratings to Congress in general, but approval ratings to their own porker.

When you have districts gerrymandered so that incumbents require a big swing to lose, you get the creation of rotten boroughs — some of which are now hereditary fiefdoms — and partisanship spilling out of electoral politics.

When you are talking about the House of Representatives, the lack of competition is primarily the result of gerrymandering.

Gerrymandering has become very efficient within the last 10 years, with some pretty sophisticated software allowing the line-drawers to draw boundaries at the household level.

You don’t need to draw ultra-safe districts to maximize your party’s representatives, you need to create the maximum number of districts with a reasonably safe margin – say 10%. Barring any big swings in party affiliation, you maximize your party’s victories by diluting the other party’s voters.

At the level of the Senate, redistricting doesn’t matter. Money plays a much bigger factor in Senate races. Elections for Senate seats occur every 6 years (compared to every 2 years for the House), allowing more time for incumbents to fill their war chests. Senate seats are also more valuable per seat, so the national party has a lot more incentive to channel money into the seats that need it the most. As a result, it takes a lot of money to beat an incumbent.

The wildcard in US politics is ethnic voting patterns. As noted above, you only need to create 55-45 districts to make it very difficult for the other party to win. Blacks in the US are a major democratic voting group, with a partisan margin of 90+% for democratic candidates. It’s very easy for the GOP to create almost all-black districts (which is supported by black politicians) and then surround them with 55-45 GOP districts.

This year may actually make this policy backfire for the Republicans, because if the voter sentiment against them is strong enough all of those 55-45 districts could flip to the Democrats, and the advantages of incumbency will make them somewhat tougher to flip back.

John,

I think the logic runs from the safe seats(at the party level) to the partisanship.

In other words, in a majority of the races, winning the party primary of the dominant party means winning the seat. So, if nationally the parties are split 50-50 on some political scale, then instead of battling over the center you are battling over either the 25th or the 75th percentile (on a national scale).

Even in the senate races this can happen, for example it may happen with Lieberman. Too much identification with the opposite party may cost him in the primary.

The other thing that leads to partisanship in the USA is the amount of geographic mobility and self-sorting that take place in state of residence. There tends to be a self-reinforcing cycle of politcal differences leading to states attracting more of the same kind of people.

For example, the parties tend to be well differentiated on the number of children in the family. Less than 1.5 kids per household: very democratic, more than 2.0 kids per household: very republican.

On the other hnad, I don’t there is much data about a change in partisanship. From 1987 to 2004 the split between Democrat, Republican and unaligned hasn’t changed more than a percent or 2 either way, with the unaligned group being larger than either of the party groups (38,33,29 for UA, Dem, Rep.)

The thing that has changed in the last 15 years is the shift in control of the House of Representatives from the Democrats to the Republicans. The near lock on the house for the democrats before 1994 effectively meant Republicans had to be bipartisan. The question is “why were the democrats more bipartisan when they had control of the entire government?” The answer likely is that the South had conservative democrats which made for more ideological difference within the Democratic party.

Effectively you could say that what has happened in the last 15 years is that the South changed parties giving it better ability to enforce political control on its party.

Its not historical but it works to say that the South sold the northeast in exchange for the republican states sans the west coast.

What you should think when you hear people saying increasing partisan feeling, is that those people don’t run the show anymore.

Gerrymandering, surpisingly, has very little to do with the absence of competitive races. It might in theory, but in practice other factors trump it. In particular, incumbents enjoy a huge advantage in name-recognition and fundraising capacity. There’s even a feedback mechanism at work. Challengers have a hard time raising money because of the perception that incumbents are unbeatable–and each dollar matters more to a challenger than to an incumbent. That’s why, even in districts that are marginal by the numbers, incumbents are realatively safe.

Growing partisanship seems to me to be a separate issue entirely. To some extent, it’s a function of the two parties’ electoral strategies, which have focused on getting out the base vote. The need to cover your extreme flank in a primary may also play a role, although serious primary challenges are as rare as losing incumbents. Also, the House has been very closely divided for the last decade, and the ruling Republicans have dealt with that by enforcing strict party discipline. Granted, that wasn’t the only option, but having chosen it they ensured a period of fierce partisanship.

“If partisanship is strong, people in 55-45 districts who have become annoyed with the ruling party should switch, regardless of any liking for the incumbent candidate.”

Exactly. In Australia they likely will do so; in the US they don’t have to, as they have the option of not voting at all.

Let me propose a highly simplified model that describes a situation in which intense partisanship coexists with a very strong correlation between incumbency and electoral victory, but where gerrymandering plays no essential explanatory role.

Suppose there are two political parties, A and B, and that each voter is affiliated with one of those two parties.

Let’s measure ideological partisanship on a 10 point scale: 1 representing the most intense degree of A-partisanship and 10 representing the most intense degree of B-partisanship.

Let’s suppose a highly partisan electorate: the typical A-Party voter falls in the 2 range on the scale and the typical B-Party voter falls in the 8 range on the scale. Let’s also assume universal partisanship – the average voter rating is independent of district, and depends only on party affiliation. So 2 and 8 are not just the national averages, but the average in each district.

Suppose in 1/3 of the districts, call them “A-districts”, 2/3 of the voters are affiliated with and A-Party and 1/3 are affiliated with the B-Party. Suppose in an additional 1/3 of the districts, call them “B-districts”, 2/3 of the voters is associated with the B-Party and 1/3 with the A-Party.

Suppose that in the remaining 1/3 of districts, the contested districts, the voters are equally divided by party affiliation.

Now let’s posit the following relationships between candidates and voters:

Suppose that in both the A-districts and the B-districts the typical A-Party candidate has a 2 rating, and the typical B-Party candidate has an 8 rating, just like the voters in those districts. In other words, in districts where the outcome is not in doubt and the challenger lacks a realistic chance of winning by moderating, the candidates tend to line up with voter preferences.

Suppose in the contested districts, by contrast, the candidates move toward the middle to pursue victory. So, while the voters in those districts are highly partisan, with the typical A-voter a 2 and the typical B-voter an 8, the typical A-candidate is instead a 4.75 and the typical B-candidate a 5.25.

I think we could predict the following relationships between incumbency and outcomes:

In A-districts, A-candidates almost always win. Thus incumbency is a highly reliable predictor of electoral victory – not because the power of incumbency itself produces victory, but because the clear partisan affiliation of the candidate produces continued victories for that candidate, with incumbency as a causal consequence rather than a cause. Incumbency thus correlates strongly with victory.

The same strong correlation will hold between incumbency and victory in B-districts, for the same reason.

In the contested districts, the weak party affiliation of the candidates gives even the highly partisan voter little reason either to prefer or oppose those candidates, since from the perspective of the partisan voter, the candidates are close to indistinguishable. Thus in these close races, the power of incumbency plays a deciding role. If the candidates are indistinguishable ideologically, then some voters may decide the contest based the advantages of incumbency: for example, based on which candidate already possesses a senior committee position in the legislature, and the accompanying ability to bring government spending to the district. A partisan A-voter may vote for a 5.25 incumbent B-candidate over a 4.75 A-candidate, because the B-candidate has established influence, is the “devil you know”, etc.

Gerrymandering need not be a factor here. Even if the shapes of legislative districts are constitutionally fixed, and the current incumbents possess no power to redraw them, there can be start imbalances in voter affiliation among the districts. We can assume that the imbalances are due to historical and geographical factors, for example.

Of course the US system differs from this idealized model in many ways. For one thing, in the US system there are large numbers of independent voters, affiliated with neither party. But the two major parties have an absolute lock on electoral politics. Since Democrats and Republicans are highly partisan, and the Representatives and Senators they elect are also highly partisan, these independent voters are often faced with a choice between a partisan Republican and a partisan Democrat. They may find themselves evenly split between the candidates, not because they find them indistinguishable, but because they find them equally unattractive. In such a situation, incumbency may again prove a deciding factor.

I think that it’s true that Republican gerrymandering has cut it a little too close for comfort. Even Delay’s seat supposedly isn’t safe any more.

Of course, with strong partisan loyalties you expect a fair number of safe seats for either party, but the discussion of incumbency is mostly about the strength of individual incumbents.

Correct me if I’m wrong, but isn’t it true that the parties simply don’t allow primary challenges in safe districts? To run in a primary against an incumbent you’ll need at least some party support and if it’s not there the incumbent will get in without a hitch.

If the incumbent is sufficiently disciplined in toeing the party line, the party will try to make sure there is no primary challenge – and that’s a typical scenario; if he/she doesn’t want to play ball – then the party will promote and finance a challenger, but that’s a rare scenario: they’ll play ball, they know what’s good for them.

I think this is pretty much the answer to your question.

Just about everything in the American political system was designed to be sluggish, inefficient, and slow to turn with the winds of political change.

It has taken 40+ years for the country to turn so far to the right; it’ll take just as long for it to turn to the left. Yes, we are becoming more partisan. Yes, thos 55-45 districts will turn around. Yes, soon after that happens, the country will be less partisan, things will correct, we’ll go back to ignoring politics and focussing instead on making mone and watching T.V. and eating junk food. Then, when the country falls to hell again, we’ll stick our heads up and start paying attention to politics again, start being more partisan, and those 55-45 seats will be in danger of flipping again.

Ask this question 20 years from now; you’ll have your answer.

No one seems to have mentioned the movement conservatives yet. They really took off with Gingrich, though their roots were way back at the Goldwater defeat.

The movement conservatives are determined to destroy liberalism and completely restructure the US. They do not recognize the legitimacy of liberals and Democrats, or even of conservative and moderate Republicans who are willing to work with liberals and Democrats. Along with a highly partisan strategy they’ve adopted uncompromising, highly partisan tactics, which include smears, character assassination, and a preference for electing extreme candidates and passing extreme laws in close votes (rather than trying for consensus.) For them a candidate who gets 65% of the vote isn’t conservative enough, and a bill in Congress which gets 65% of the votes isn’t either.

It’s only just starting to be symmetrical. I’m a Democrat (or anti-Republican) who’s as partisan as the Republicans are, but until recently many or most Democrats haven’t believed that victory is possible, and many of them still long for the old days of comity and mutual respect.

I still think gerrymandering is the largest contributing factor, and if you look at the whole process, the relationship between partisanship and incumbency bacomes clearer. (This is a VA state-level explanation.)
Gerrymandering is done:
1) by parties
2) to protect their “good” incumbents
3) from both internal and external challenge

The key point is “good” incumbents. In Virginia, as in most states, it is fairly clear who the reasonable challengers for legislative seats are. Quite often, district line-drawing is is done to keep possible challengers out of incumbents’ districts–BUT that is only done for “good” (e.g. partisan) incumbents. If you are Republican and buck the leadership too much, they will redistrict a strong challenger into your district. (The Democrats did the same when they controlled the GA). Thus, partisan legislators are rewarded by partisan districts in their favor–the parties use gerrymandering to increase their number of seats, but also, crucially, to keep their own members aligned with the party.

Another factor in increasing partisanship is that there is a change in the political power balance; the people who used to win everything (liberals) are now losing sometimes, and are extremely unhappy about it. They are in the same position as conservatives in the 1930’s (another era of intense partisan rancour).

People who want less partisanship should be careful what they ask for.

The most concensus election in terms of electoral vote was Nixon in 1972 520 to 17 and in terms of percent of the popular vote Reagan in 1984 at 58.8%.

The increased partisanship from this point of view is California changing the type of governors it will support. The country has become less republican in presidential elections leading to more partisanship.

John Quiggan- there is a rule about American politics. Learn it and you will never go wrong. Here is the rule: Whenever you don’t understand something, the answer is race.

The important fact is that until the 1970’s the Republicans and Democrats were not genuinely national parties. They were coalitions of regional parties.

The Democrats were a coalition of Northern urban immigrants (Irish, Italians, Jews), western farmers, and southern white racists of all economic and social classes. Right up to 1972, white Southerners would not vote Republican under any circumstances because the Republicans were the party of Lincoln. But after the New Deal, white Southern Democrats were far more conservative than Northern Democrats on every issue.

Republicans were also a coalition party: Northern industrialists and professionals; mid-western country club, main street businessmen; western resource owners (oil, mining, ranching); Southern blacks.

Because party organization was on the state level, not the national level, each state Democratic or Republican party got along quite well on its own, coming together only once every four years in a loose coalition that held just long enough to run a presidential election campaign.

So once in Congress, legislators found that their party affiliations had little to do with their political inclinations. Region and class made more of a difference. So “bipartisanship” flourished.

But then came Lyndon Johnson and the Civil Rights Act and the Voting Rights Act. The Democrats unequivocally declared that to be a Democrat, you had to accept that Black people are human beings. White Southerners wouldn’t accept that, so they left the Democratic party, voting first for George Wallace in 1968 and then for Nixon in 1972.

It took 30 years to purge every last conservative Democrat out of the South, but it’s happened- the last to go was Zell Miller, who famously supported GW Bush at the 2000 Republican convention.

The Republican embrace of the racist, fundamentalist South has had the effect of driving the former liberal and moderate Republicans of the northeast into the Democratic party.

So the parties have realigned along ideological lines, and partisanship is now the natural order of things.

It took 30 years to purge every last conservative Democrat out of the South, but it’s happened- the last to go was Zell Miller, who famously supported GW Bush at the 2000 Republican convention.

Alas, we haven’t been rid of him that long: he supported Bush–very flamboyantly and very obnoxiously–at the 04 convention.

Two other aspects, so far unmentioned, have to do with personalities rather than political theory:
1. For 50+ years, with one short hiatus, Democrats were the majority in both houses. Over the years, Democrats learned, or acquired the habit, of treating Republicans with a certain noblesse oblige; Republican leadership would be consulted respectfully. And Republicans learned or acquired the habit of surrendering themselves to a minority position. Newt Gingrich changed that. The Republicans in majority were not used to being in the majority and the Democrats were not used to being in the minority. There is no doubt in my mind that this created a heightened state of personal rancor among the elected representatives on Capitol Hill.

2. The Clarence Thomas hearings elevated the personal over the political in a new way. This has led to a still-ongoing cycle of “gotcha politics.” Republicans, having been forced to admit that a person who did what Thomas was accused of should be disqualified from high office, insisted on applying that same rule to Clinton, “exposing” the fact that Democrats didn’t care about the principle, but only about opposing Thomas. Democrats, having listened to months of rhetoric about whether lying is a grounds for impeachment, want to “expose” the hypocrisy of Republicans who will oppose impeachment of Bush.