Don't count your chickens until you've looked too closely at how sausages are horse-traded, or something like that

Eric Alterman, thinkologist for The Nation and prolific Reason pen pal, has been hinting darkly for a while now that the big Democratic midterm victory is going to end with Santa shaking his head sadly and saying "We'll just have to cancel Christmas Winter Holiday! And the children have been so good this year!" While Alterman's never quite willing to consider that the left is just selling a product few Americans want to buy, he's got a strong point about how districting has greased the ladder for the Democrats, linking to this interesting study:

Why are things so tough? Looking at the 2004 election, the Democrats won their victories with an average of 69% of the vote, while the Republicans averaged 65% in their contests, thus "wasting" fewer votes. The Republicans won 47 races with less than 60% of the vote; the Democrats only 28. Many Democrats are in districts where they win overwhelmingly, while many Republicans are winning the close races—with the benefit of incumbency and, in some cases, favorable redistricting.

I find gerrymandering issues heap-big confusing—isn't there a pretty low ceiling to how much you can redistrict your way to victory at the national level? (That is, I can see how it works to gerrymander a district at the local level, but when you're talking about all 535 seats in Congress, shouldn't it tend to even out?) But as Delaware Dave Weigel reminded me the other day, since the days of the Contract With America, the GOP has had plenty of time in control of a variety of state houses to set things up. It's going to take some doing for Powerhouse Pelosi to orchestrate a national turnover, and a general sense that "Americans are unhappy with the direction of Congress" is not enough to do it.

Studying how the Republicans have stacked the deck against Democrats (sort of like how the Democrats did the exact same thing to Republicans until the 1990s) takes some of the sheen off President Bush's performance as head of the Republicans, but not much. As I never tire of pouring icy water on political hopes, I'll point out again that Bush is still way ahead of the average presidential-coattails performance in off-year and midterm races. Even if the GOP lost both houses in November, Bush would still be ahead of the average. He's already an electoral success for his party. How such a small man had such a big effect is something future historians, with their smellevision and massive frontal and parietal lobes, will have to puzzle out.

Comments

[Patricia Mathews]

Party-wise, gerrymandering may even out. TRhe problem with it is that it's now an Incumbents' Lifelong Employment Act. They try very hard to eliminate swing districts of any sort.

[joe]

Gerrymandering doesn't even out nationally because the Republican gerrymandering strategy - packing as many black people into a landslide-majority black/Democrat district, thus leaving the four or five surrounding districts solid-majority Republican - has been assisted by black Democratic politicians, who've been supportive of the same strategy.

[TJIT]

Incumbent protections regulations, oops I mean campaign finance regulations are another reason it is difficult to unseat incumbent politicians.

They also markedly increase the power of PACs and lobbyists.

[Michael Stack]

Think about it this way. Let's say you played baseball for the Rs against the competing team, the Ds. Let's further assume that you played 50 games, and that over the course of those 50 games, the Ds scored 60 runs, and your team, the Rs, scored 50. Now you'd expect, given the run distribution, that the Rs would lose a few more games than you would win.

But - let's say I give you complete control over how all of those runs are apportioned per game. The optimal strategy would be to push all of the D's runs into one game, and none of the Rs. That'd give the D's an incredibly lop-sided victory, 60 - 0. However, you're then free to take one run for the Rs for each remaining game, winning the next 49 games, 1-0, giving you a record of 49-1, despite the fact that overall, you were outscored.

That's why the article talks about not "wasting" votes, because as long as you win by even one vote, it's inefficient to waste the remaining votes when you can push them into a different district via gerrymandering.

[PapayaSF]

Joe is correct, it's the result of an informal bipartisan agreement: Democrats wanted very safe black seats, and Republicans said "Sure, let's do that!"

[Jim,MtnViewCA,USA]

"How such a small man had such a big effect..."
Wise choice of political enemies?

[Seth]

"While Alterman's never quite willing to consider that the left is just selling a product few Americans want to buy..."

More to the point, does the left have a product many Americans can recognize? Aside from "not Republican", I mean.

[Paul]

What Joe and Papaya said. Much of the race-based gerrymandering was done at the behest of Democrats.

But at the same time, gerrymandering isn't everything. It's arguably contributed to republican victories, and may very much make the difference in razor thin margins-- a million votes here, a million votes there, eventually you're talking about real victories.

However, given the very good forum over at Cato Unbound over Democrats attracting that mysterious voting bloc called "libertarians", some interesting points in the bigger picture should be pondered.

The fact is, Democrats can win races if they attracted enough votes across a spectrum. They should learn this by studying Republican victories despite former Democratic gerrymandering.

I'm almost... almost ready to hope for Republican victories in the upcoming election in that it might send a message to Democrats to seriously differentiate themselves from Republicans.

I am not going to take Moulitsas advice and vote for Democrats and hope for a policy shift later. I've been burned before on that. So to speak: If you can't come up with the front money, you're not for real.

[Gene Berkman]

In 1991, the California legislature could not agree with the Governor on a redistricting plan, so Judges drew up the election districts. The result was that many districts were competitive for several elections.

In 2005, Governor Schwarzenegger suggested changing the law to put redistricting in the hands of a non-partisan commission of ex-Judges. The Democrats and several GOP incumbents spent massively to defeat it, so now California has no competitive districts.

This issue is bigger than who will win the election. The issue is whether voters determine who will be elected, or officials decide who can vote for them.

[We Are All Trots Now]

How such a small man had such a big effect

He hasn't. He has no coattails. Left and right are equally disgusted with him--conservatives more sensibly so, since he pretended, once, to be one of them, sort of. The left is fighting a devil of their own imagining. Their post-Kerry-loss attachment to "systemic" explanations for their haplessness is desperate and...hapless.

It's simple. The ground occupied by Bush (and Lieberman and Hillary) is the American "center"--boundless government, domestic and foreign. Slightly greater numbers of Republicans than Democrats are talking that talk lately, so slightly greater numbers of them are winning. It'll change when the talk changes.

It's why libertarians always lose--and, sadly, always will.

[Ari Tai]

Seems this is a case of poker and who's willing to take the greater risk. If the pubs are willing to spread their voters out more than the dems (and have a slightly better product) they will win big. If the dems are more interested in self-preservation (i.e. more selfish) and collaborate with the pubs to insure their own "secure seats," they set themselves up to lose to a more confident opponent.

It all works out in the end, what goes around, comes around, as long as the rules are uniform and slow-to-change.

Maguire has a good analysis of this affect at Just one Minute.

[Mr. Snitch!]

Reality TV + Smell-o-vision = Too much reality

[James Ard]

Historians likely will question why the "small man" was able to perform so well in elections despite his bold stands on social security, taxes and national security. But, I guess, historians that don't work for the Los Angeles Times might have had more time to appreciate history than a full time science fiction buff.

[Lemur]

When you add demographic trends in with gerrymandering, it becomes pretty clear that the current fiasco notwithstanding, we are in for a "permanent" Republican majority...

[Paul]

When you add demographic trends in with gerrymandering, it becomes pretty clear that the current fiasco notwithstanding, we are in for a "permanent" Republican majority...

I respectfully disagree. I think that if the Democrats can really appeal to a wide cross-section- especially a large percentage of so-called 'swing' voters, they can take majority of both houses. And forgive me for being Captain Obvious, but the white house is certainly up for grabs.

[madpad]

I wonder if gerrymandering to have such close races is all that wise.

In any given population, there are fence sitters and swing-voters. And in that population, over time, there will be people who change their minds, lose hope or get disgusted.

You can only game the system so long before human dynamics fails to work for you and your own strategy comes back to bight you in the ass. And in any complex system, attempting to control the variables can, in and of itself, create unpredictable results.

Now, if you've won by engineering close races and then:

1. fail - in a big way - to remember to follow the message that got you elected OR

2. your strategy is dependant on repeatedly trumping up the same hot button issues with no real attempt at any solution or resolution in order to "whip up your base." OR

3. Your party is so obviously corrupt, venal and obtuse...

...then you're pretty much setting yourself up for some ass-handing.

[madpad]

BTW, the Washington Stock Exchange's stock for the GOP retaining control of the House in '06 has pretty much nose-dived over the past 2 weeks and is down around it's historical low.

[Arkey]

There have been a number of "small" Presidents. Tyler, Lincoln, Roosevelts (both the shrimp and the cripple), Wilson and Reagan.
I think G.W. measures up!

[joe]

"When you add demographic trends in with gerrymandering, it becomes pretty clear that the current fiasco notwithstanding, we are in for a "permanent" Republican majority..."

What?!?

In about 15 years, Texas will be a majority-minority state. Ditto New Mexico, and Arizona. A few years after that, Colorado.

If current demographic trends, and the demographic voting patterns of today continue, the Republicans will no longer be able to win a national election, or hold the House.

I suspect by 'demographic patterns,' Lemur is talking about religious identification and church attendance. It's a mistake to draw a simplistic conclusion from this, however, because the vast majority of weekly-church-attending black and Hispanic Americans are pulling the lever for the Democrats - a pattern which Katrina and the Immigration Wars are only exacerbating.

[thedifferentphil]

Since the 1994 GOP sweep, high powered gerrymandering software and computing has been developed. Thus, GOP state government majorities have had the power of this computing to help national majorities in the House. The flipover of many Governorships to the Dems will probably mean that the 2010 census will not lead to so many GOP districts. I have not seen any polling or predictions on this, but I have to assume that state houses and senates are probably poised for big Dem. gains, with the national GOP troubles causing big, but unpolled, losses in state elections. I personally believe that the country would be better served by a uniform apportioning algorithm, such as all East to West, or all North to South within states. This would divide some cities awkwardly, etc., but would over time better reflect the population and keep politicians focused on that, rather than how to rig the system to stay in power.

[madpad]

When you add demographic trends in with gerrymandering, it becomes pretty clear that the current fiasco notwithstanding, we are in for a "permanent" Republican majority...

You may be right via lucky guess (or insider information from Diebold), but absent the latter, it is far from clear.

I've already laid out my suspicion with current "trends in with gerrymandering," so such a provocative statement of surety and confidence might warrant a righteous smackdown...but I'd be interested in your defence of that position with some solid explanation.

[joe]

thedifferentphil,

In addition, a Democratic Congress in 2007-2009 will not hamstring Census 2010 the way the Republicans deliberately uncut efforts to reduce undercounts in Census 2000.

[kevrob]

There are several demographic trends at war with each other inside the U.S. population, and I'd hesitate to predict which one will dominate politics in the first half of Century 21. To wit:

1.) The huge number of immigrants. Only those who arrive through official channels are going to be eligible to vote, unless some sort of amnesty, however qualified, is enacted. But a plurality or even a majority of new naturalized citizens will be Democrats, and the citizen-kids of even the illegals are also likely to trend that way.

2.) The comparative fertility of conservatives of the church-going stripe v. their sisters who are more or less secularist, from the professional woman to the more..bohemian..of lifestyle. It's not that lefties and Career Women don't have children, but they have fewer than the Red State Moms do. Does the membership of NARAL think its going to outbreed the Baptist Ladies' Sodality?

3.) The power of internal migration. One result of the shift of population within the U.S. is that the old Northeastern states have been losing seats in the Congress and the Electoral College to our nation's warmer climes for some time. Florida, California and Texas are out-clouting New York, Michigan and Pennsylvania. You also get weird effects, such as the influx of non-white immigrants to CA being offset by an out-migration of middle-class whites to states like Nevada and Utah. Both groups may just be following employment opportunities, but some people who are "Red Staters" in their souls just got sick of living in a blue one.

It is possible that immigrant-heavy states may become more Democratic, while conservative Sun Belt states, which have been Republican strongholds for some time, will continue along that path. Extrapolating existing trends without limit is a good way to make stupid predictions, though. I imagine life will throw up enough complexity to make monkeys out of the prognosticators.

Ahhnold's attempt to replace political-map drawing by the Legislature with a state commisiion failed, but Iowa has had one for a while, and people there seem pretty happy with it. In many states, the decennial mapping trauma has been a wrangle among incumbents intent on nailing down even safer seats, party operatives trying to game the system to increase the number of seats they hold, and racial and ethnic interest groups trying to carve out districts some of their members can win. The latter are often backed by the federal Justice Dept., especially in the states under special scrutiny from the Voting Rights Act. The whole mess frequently winds up in the laps of judges anyway, which is why the Iowa Plan has such appeal. If you lose a redistricting lawsuit, and the judiciary chooses to impose someone else's map, you can get screwed but good. A commission may not allow you to steamroll your opposition, but you won't get run over, either.

Kevin

[PapayaSF]

the Republicans deliberately uncut efforts to reduce undercounts in Census 2000.

You mean when they insisted that the Census actually count people, as written in the Constitution, as opposed to the Democratic plan of "estimating" more people than could be counted in Democratic areas? Those villainous Republicans!

[Wally the bird]

What realy annoys me is whan they use the terms BCE BEFORE COMMON ERA and CE COMMON ERA as if their afraid to iffend the athesit wackos at the most infamous ACLU and the rest of the leftists groups

[ellipsis]

Words don't convey how pathetic you appear, Wally.

You have the grammatical comprehension of a 12 year old.

[madpad]

What realy annoys me is whan they use the terms BCE BEFORE COMMON ERA...

I'm with ellipsis on this one, Wally. You're a boob.

Consider: Maybe the 65% of the rest of the world that's NOT Christian get's annoyed when we try to reframe their history around our religion and shove our beliefs down their throat.

Besides, B.C. & A.D. don't even start on the right year to correspond with Christ's birth so it's a pretty pathetic argument to begin with.

[joe]

"You mean when they insisted that the Census actually count people, as written in the Constitution, as opposed to the Democratic plan of "estimating" more people than could be counted in Democratic areas?"

No, I mean the plan by the professional demographers working for the Census bureau to adjust known-inaccurate numbers to make them more accurate.

Classic Democrat vs. Republican issue: the entirety of the professional statistical and demographer fields endorse a scheme to improve the accuracy of the Census; Republicans notice that the errors favor their poltical interests and kill the plan; Democrats complain; the press reports the story as "he said/she said."

The question of accuracy never even crossed your mind, did it PapayaSF?

[madpad]

Backin' up joe on this one, too. Republicans have zero cred when it comes to abiding by peer reviewed, scientifically-based reality.

Faith in God is one thing. But when given a choice of faith in established experts vs faith in Republican pandering, I'm gonna go with the experts.

This is the same crap behind the over-the-counter emergency contraception issue in the FDA. Experts said do it. Partisan wankers said no.

Or FEMA. Sensible people said get someone with experience. Bush said, "Hire Brownie."

[Laika's Last Woof]

Once the Census goes off the real count there's a strong possibility it will also go off the rails, as every political faction with an axe to grind will try to get its fingers into the now-up-for-grabs demographic pie.

Then of course there will be lawsuits out the yingyang, as various activist groups demand "justice" in the form of bigger census estimates for their aggrieved "under-represented" constituents and activist judges will start throwing out numbers and replacing them with new ones -- always bigger, of course.

Local and even state governments would also have an interest in inflating the numbers. They get more money, more representation, more clout, and more power with increasing population.

So many stand to gain at Federal taxpayer expense from inflated Census numbers that anything but a straight-up count leaves too much room for corruption to ever be considered.

All this is moot anyway, as the Constitution calls for a count, not an estimate.

[joe]

"Once the Census goes off the real count there's a strong possibility it will also go off the rails, as every political faction with an axe to grind will try to get its fingers into the now-up-for-grabs demographic pie."

That's not a reason to take an inaccurate census; it's a reason to have a reliable system in place to shield the professionals from political interference.

"Local and even state governments would also have an interest in inflating the numbers."

And other local and state governments, as well as national organizations who favor certain groups over others, have in interest in an inaccurate census that undercounts certain populations. See above comment about independence from political interference.

"All this is moot anyway, as the Constitution calls for a count, not an estimate." No one is proposing to end the count, or replace it with an estimate. The statisticians and other professionals who design the census want to use sampling and imputation to adjust the count, not replace it.

There is nothing in the Constitution that forbids improving on the actual enumeration that fails to actually enumerate everbody. The Constitution does not include the words "...and that's it" after "actual enumeration." The Repubicans simply object to having an accurtate census, because the inaccuracies benefit them.

Sadly, the legislature has been controlled for the past 12 years by people who consider partisan interest more important than effective government or scientific evidence. It shouldn't matter who would benefit from making the census more accurate.

[Al]

Umm-

There's always the Senate!

You can't gerrymander the Senate. I just had a thought- if the Senate is as corrupt as the House, does that mean that gerrymandering does not cause corruption?

[madpad]

if the Senate is as corrupt as the House, does that mean that gerrymandering does not cause corruption?

Al, there are many forms and causes of corruption but in the end, the 'cause' of corruption is individuals choosing corrupt methods to achieve dubious ends.

Gerrymandering, per Wikipedia, is a controversial form of redistricting in which electoral district or constituency boundaries are manipulated for an electoral advantage. So it is essentially, by definition, a corrupt practice.

The real difference 'tween the House and the Senate, as regards gerrymandering, may be the level of partisanship it has led to. Some have asserted that the Senate is comparitively less partisan than the House, though I don't really have a solid opinion one way or another on that one.

[PapayaSF]

Laika's Last Woof: Exactly.

There is nothing in the Constitution that forbids improving on the actual enumeration that fails to actually enumerate everbody.

Joe, that's the kind of "let's improve the Constitution without actually amending it" thinking that's produced the weakened, half-ignored document we have today. Sure, accuracy is nice, and if you have ways of improving "enumeration," fine. But I don't think estimates are "enumeration."

I'll tell you what: I'll give you a census with transparent, carefully-done scientific "adjustments" for accuracy, and you give me photo ID for all voting and increased penalties for voting fraud. Deal?

[joe]

PapayaSF,

I see authorization in the Constitution to raise a Navy. I see authorization to raise an Army. I don't see squat about raising an Air Force. Do you know why that is?

Because an Air Force, while carrying out the same role as, and improving on the efficacy of, an Army and Navy, is based on technologies that did not exist at the time the Constitution was framed. You know, sort of like modern statistical analysis.

Should the fact that the Constitution only specifically referenced an Army and Navy be read as forbidding the creation of an Air Force to improve our military capabilities? Should the fact that the Constitution only specifically referenced an actual enumeration be read as forbidding the use of statisical analysis to improve our census-taking?

[PapayaSF]

"Army" and "Navy" clearly mean "armed forces," and an air force is a totally non-controversial extension. (Really: has anyone, even the most rigid Constitutionalist, ever argued that having a separate Air Force is unconstitutional?) The Constitution doesn't mention "Marines" either, but nobody ever said their existence contradicts the Constitution, even though we've had Marines since the days of the Founders.

But "enumeration" has always meant "counting," not estimating. Is it perfectly accurate as a means for a census? Nope, but it's a lot harder to twist an actual count for political purposes than any form of estimating, as LLW wrote above.

[Knemon]

"Sadly, the legislature has been controlled for the past 12 years by people who consider partisan interest more important than effective government or scientific evidence."

12?

Short memories ...

[madpad]

PapayaSF,

You're argument, though rigidly constitutional, is becoming somewhat disingenuous.

It's perfectly reasonable to assume that when the constitution was written, to most folks back then, counting heads was considered the most logical, reasonable way of arriving at an accurate census.

Lo some 200 years later, we find ourselves realizing that the scope of a census for 300 million people is beyond the capacity for civil servants to merely do a head count.

It's obvious that the framers were looking for an accurate count to equitably ensure representation. These were, after all, men of the enlightenment.

I'm figuring that they would probably find sticking with an inaccurate headcount method over a more accurate approach simply to accord to the letter of the 'as written' document to be pretty stupid.

[joe]

PapayaSF,

"and an air force is a totally non-controversial extension"

Among census takers, demographers, anthropolotigists, and staticitians, incorporating statistical analysis to improve the accuracy of the count is as totally non-controversial as using air power is among the professional military. The only controversy that exists over incorporating modern statistical methods into the census is that whipped up by partisans with an interest in inaccurate, biased census numbers. I do not for a second believe that this phony "controversy" - on with as much grounding in scientific evidence as the link between marijuana smoking and spree killing - has any relevance whatsoever to the proposal's constitutionality.

Let me run a not-so-hypothetical by you: a census taker gets forms back from every house on Elm Street except one. He knows it's there, he knows that people live in it. Like a good census taker, he follows up, leaving messages and knocking on the door, but it keeps getting slammed in his face. Perhaps the people who lived there are the remains of a family murdered by the Khmer Rouge, and they don't like to deal with the government. Or perhaps they just travel a lot. Neighbors report that there are a few people living there, but they don't know how many. They report seeing kids. They think the house has been divided up among more than one household, but they're not really sure.

He can "actually enumerate" the houseing unit, so obviously he reports that. What about the people - since he doesn't have the opportunity to "actually enumerate" them, should he write the number of people he "actually enumerated" there: zero? What about the number of households?

Should that census taker provide population, household, and ethnic-group figures for that block that he knows to be untrue? Or should he go beyond reporting what he's actually been able to count first hand, in order to provide the most accurate information he can?

You know that right definition for an "actual enumeration" is? One that actually enumerates everyone, as closely as the Census bureau can get.

[PapayaSF]

Madpad, if the Constitution is outdated and needs to be changed, it includes a mechanism for making such changes.

Joe, see above. Also: your hypothetical seems nitpicking to me. Frankly, I doubt that this happens often enough to be statistically significant, just as I doubt your claim that there are somehow vast numbers of legal voters who are somehow unable to get photo IDs.

But if your example is all we were arguing about, I wouldn't object too much. However, that's not what the argument was circa 2000. IIRC, it was Democrats and liberal social scientists who wanted to add in estimates of people they couldn't count. It wasn't "We know 4 people live at 123 Elm St. even though they won't answer the door, because that's what the people at 121 and 125 say." It was "Our statistical estimate says there are X thousands of homeless in this city, so let's add them to the census count." Once again, that's not an enumeration, that's a guess.

[juris imprudent]

It's obvious that the framers were looking for an accurate count to equitably ensure representation. These were, after all, men of the enlightenment.

Eh, these were the same fellows that counted "other persons" as 3/5ths of a person for census purposes?

I'm all for making the census a better COUNT, but not a statistical free-for-all.

[joe]

PapayaSF,

"Frankly, I doubt that this happens often enough to be statistically significant..."

Frankly, you are wrong about that.

"IIRC, it was Democrats and liberal social scientists who wanted to add in estimates of people they couldn't count."

Yes, of course, the statisticians and anthropologists at the Census Bureau, and the American Statistical Association, are liberal Democrats. How do we know this? Why, they made a proposal that would not have been good for the Republicans. So they must be biased, and wrong.

'It was "Our statistical estimate says there are X thousands of homeless in this city, so let's add them to the census count."' Yes, it was statisticians and anthropologists who knew how to idenfity what the undercounts were. Why are you assuming that they don't know what they're talking about, or that they're cheating? Because they're telling you something you'd rather not hear? Because the undercounts they've identified are concentrated among groups of people you don't like? Because the outcome of a better census would benefit the Democratic Party?

You've only read the spin about this, and you've concluded that you know who is making a legitimate argument, and who is a partisan cheat. Well that's just great. It's this blinkered partisanship, this refusal to acknowledge inconvenient, scientific truths, that's gotten this country into so many messes over the past six years, and you still refuse to learn your less.

You can't ignore the truth just because you don't want to admit it's the truth!

There's a good book titled "Who Counts?" about the debate, if you'd like to learn more about this issue. If you'd rather wallow in your comfortable, self-serving ignorance, fine. But there is absolutely no question that the methods the Census Bureau recommended, and that the Democrats in the Congress endorsed, would have made the census more accurate. None. The benefits of statistical sampling in the census have been proven beyond a reasonable doubt, and it is a serious disservice to this country, and an injustice to the people who live here, to sabotage the operations of the government for partisan gain.

[Laika's Last Woof]

"Lo some 200 years later, we find ourselves realizing that the scope of a census for 300 million people is beyond the capacity for civil servants to merely do a head count."

And yet the IRS seems quite good at finding us.

"The only controversy that exists over incorporating modern statistical methods into the census is that whipped up by partisans with an interest in inaccurate, biased census numbers."

This is a straw man, but as it happens also undermines your argument for estimating: If partisans are motivated to bias the census figures then using estimates leaves more room for those motivated to bias the figures to do so, not less. As with the direct counting of votes in elections, a direct count of people for the census enforces the strictest measure of accountability.

Allowing estimates in such a political climate opens the door for even more inaccuracy, particularly if activist lawyers and biased judges start throwing out estimates they don't agree with and replacing them with their favored estimates.

If people are as motivated to produce fraudulent results as you claim, maintaining the accountability of strict enumeration becomes critically important. We must under no circumstances open ourselves to the kind of fraudulent "counts" that would result from adding in "estimates" whose formulae would, by your own admission, be subject to political pressure.