The Myth of Poor Democratic Electoral Performance
At the end of a somewhat rambling post, Ezra Klein observes:
Meanwhile — and this will be the last paragraph of this rambling post, I swear — Rick asked how they — the Republicans — keep winning elections. And the answer, basically, is that they don’t. Democrats have won the popular vote in four of the last five presidential elections. They’ve enjoyed congressional gains in five of the last seven elections. Since 1992, the Republican Party has had three unambiguous wins: 1994, 2002, and 2004. In the grand sweep of American history, that’s not such an impressive record. But it’s sometimes hard to think about the sweep of history when you’re in the thick of the present, and so Democrats grew more impressed with Republicans successes than was probably warranted. But it’s time to let that go.
I would actually grow stronger than that. Check out the two-party share of the vote in House of Representatives elections:
Here you see that even though the elections of 1996, 1998, 2000, and 2004 all returned GOP majorities, most people actually voted for the Democrats. The vagaries of how congressional districts are drawn produced Republican majorities in those years but that’s not a problem of Democrats being insufficiently persuasive. Indeed, in 2004 the Republicans actually picked up seats even while securing fewer than half the votes. A lot of political commentary in the 21st century has been focused around the idea that Democrats have some sort of “problem” or other persuading voters to support them but in fact now that we’re past the 2008 election we can say that Democrats have narrowly-but-consistently prevailed in terms of voter preferences over the past 20 years.
Now of course much as close only counts in horeshoes and hand grenades, things like the Republicans securing more than half the House seats with fewer than half the votes or securing the White House with a second-place finish in the popular vote have real consequences. But if you want to understand voter behavior you need to look at voter behavior. It’s just that understanding electoral outcomes requires you to look at the intersection of voter behavior and the electoral system. That’s a different issue.
53 Responses to “The Myth of Poor Democratic Electoral Performance”
Register
RSSLogin
Topic Cloud
ArchivesBlog RollAbout Matt YglesiasContact Me
Donate
November 12th, 2008 at 11:12 am
Links to the data??
November 12th, 2008 at 11:14 am
duBois, that’s because when Republicans control a legislature they redistrict to maximize the number of Republicans, but when Democrats control a legislature they redistrict to maximize the safety of Democratic incumbents, which is essentially the same as maximizing the number of Republicans. (See California and Illinois for beautiful examples of this.)
November 12th, 2008 at 11:16 am
Perception and social rank are all important. From 1932 to 1980, Democrats and liberalism were ascendent socially in Washington and among the press. Since Reagan’s day Republicans and conservatism have conquered the Beltway and the mass media socially and culturally.
No matter how well Democrats do by the numbers, their victories are percieved as a deviation from the social norm and downplayed.
When Republicans win, they are free to do as they please.
When Democrats win, it is an aberration and the pundits (both conservative and non-conservative) start debating how soon normalcy, by which they mean conservative rule, will be restored, and how it can be best managed. The winning Democrats are expected to compensate for their breach of social rules by adjusting their behavior and policies to act acting like Republicans.
November 12th, 2008 at 11:18 am
This is one reason why control of redistricting is so important. You can win a majority of the vote, but still lose control of Congress, if too many of your votes are tied up in ultra-safe districts.
According to the Wikipedia page on Cook PVI, the most Republican district in the country is UT-1 with a PVI of R+26. The Dems have thirty districts with PVIs of D+27 or greater. That’s one serious handicap.
Meanwhile, the most Democratic district represented by a Republican is Delaware’s lone CD, which has a PVI of D+7, and which will flip whenever Rep. Castle retires.
While Congresscritters in ’safe’ districts obviously want them to continue to be safe districts, maybe they should decide that, for redistricting purposes, D+15 is ’safe’ enough that they wouldn’t be in any danger of losing their seats, and should be willing to redraw their districts, in states where they control the legislatures, so that they trade some of their excess Dem voters for more GOP voters to bring the PVI of their districts down into the D+15 to D+20 range. They’d still be safe districts, but we’d be able to improve the PVI scores of a bunch of other districts, turning safe GOP districts into competitive districts, turning R-leaning districts into D-leaning districts, and so forth.
November 12th, 2008 at 11:21 am
How much is this due to the facts on the ground and how much to strategic districting? How much does districting matter given vote shares and where voters live? Say if Democrats controlled all state legislatures, how many House seats would they gain? If Democrats controlled no state legislatures, how many House seats would they loose? I understand that there are constraints on what state legislatures do, so I’m assuming those constraints are respected.
How does this relate to how I should think about redistricting reform/litigation?
November 12th, 2008 at 11:21 am
I would actually grow stronger than that.
Ezra is a little frail.
November 12th, 2008 at 11:24 am
To follow up, one reason that Democrats have “thirty districts with PVIs of D+27 or greater” is that urban areas are very democratic and have lots of population, often enough to fill a House District. Getting more seats for Democrats would require breaking up some of these safe urban seats. And that gets us in to the big majority-minority districting fight. Any movement there?
November 12th, 2008 at 11:31 am
This also relates to the issue of how a vote for President in North Dakota carries more weight than one cast in California. There are simply more people per congress-critter in CA, NY, etc. than in the red states.
November 12th, 2008 at 11:31 am
Yeah, I think the reason you have so much more safely Democratic districts is because of these urban districts that are very small geographically. To make these less “safe” for the Democrats, you’d basically have to make them really, really weird-shaped. And people don’t like crazy, weird-shaped districts. It might even be nearly impossible in the biggest cities, due to the way Democrats cluster.
November 12th, 2008 at 11:35 am
The Dems have thirty districts with PVIs of D+27 or greater.
That’s a natural function of the Democratic death-grip on urban areas. I don’t know how you’d carve a bunch of competitive districts House districts in New York, for example. Do you draw a line from Bronx up until you hit some Republican farms in upstate?
November 12th, 2008 at 11:36 am
More people per congress person, but fewer voters per congress person. Democratic Congressional districts on average have fewer eligible voters since they have populations with proportionally less eligible voters (more non-citizens, ineligible more felons). Democratic voters are also less likely to vote (younger, more minority, poorer, less married).
November 12th, 2008 at 11:40 am
I should also add that
a) House districts are drawn by population, not by eligible voter numbers.
b) though Democratic voter demographics vote less than Republican voter demographics, it seems that the non-voting Democratic demographics potential voters tend towards ‘independent’ voter behavior and are not a huge help to Democrats, at least the white non-voters.
November 12th, 2008 at 11:40 am
This strikes me as a bit like the way the BCS compares teams in part based on whether they won 55-0 over a horrible team or 49-7 over a merely very bad team. The answer to that question tells you virtually nothing about the strength of the two teams, yet it’s for some reason included in the calculation.
Most House votes are case in races that are either de facto or de jure uncontested. It would be much more interesting to see a more sophisticated analysis that either excluded those races or handled them separately.
After all, whether fifty thousand or ninety thousand people showed up to vote for Steve Incumbent (D-NY) tells you basically nothing about anything.
November 12th, 2008 at 11:50 am
Getting more seats for Democrats would require breaking up some of these safe urban seats. And that gets us in to the big majority-minority districting fight. Any movement there?
Gerrymandering a few “safe” districts into many Democratic leaning districts works in the short term, but once the political tides shift just a little bit, it ends up collapsing with great drama. If you create a bunch of 50+1 Democratic districts out of a few 60%+ Democratic districts, you leave yourself exposed, as the Republicans learned in 2006.
November 12th, 2008 at 11:56 am
The reason there are so many super-majority Democratic districts is because the amendments to the Civil Rights Act mandates that you maximize minority representation. Those 80% Democratic districts are really 80% black/latino districts. And as a practical matter, there’s no way you’re going to get away with breaking them up.
November 12th, 2008 at 12:04 pm
actually it isn’t the ‘intersection’ of voter behavior and electoral system, it is the two-time (cross) correlation function between electoral system and voter behavior, or maybe the ‘memory function’, or possibly the commutator between their respective actions, if such exist.
November 12th, 2008 at 12:08 pm
The reason there are so many super-majority Democratic districts is because the amendments to the Civil Rights Act mandates that you maximize minority representation. Those 80% Democratic districts are really 80% black/latino districts. And as a practical matter, there’s no way you’re going to get away with breaking them up.
No, no, most of them are 51-65% black/latino. You can actually fall afoul of the Equal Protection Clause if you overpack. And it’s the Voting Rights Act, not the Civil. And it doesn’t really mandate it; it’s just kinda been interpreted that way.
November 12th, 2008 at 12:19 pm
The graph would look different if districts were drawn differently, because voter behavior is also beholden to the likelihood that your vote will matter. So in landslide districts probably less people will vote which skews the overall popular vote total. Basically you’re counting something that doesn’t actually matter and trying to assign worth to it – it’s obviously an indicator but it will be a very limited one.
November 12th, 2008 at 12:38 pm
Not surprising that it took 19 comments to reach Lee Atwater’s most pernicious legacy — the Voting Rights Act that crams minority voters into Bantustans and allows the Repubs to be disproportionately represented everywhere else. I recall George Will pointing out that Dick Gephardt would have been Speaker of the House during several Congresses if we believed a color blind society should be fostered by our institutions.
November 12th, 2008 at 12:45 pm
Not surprising that it took 19 comments to reach Lee Atwater’s most pernicious legacy — the Voting Rights Act that crams minority voters into Bantustans and allows the Repubs to be disproportionately represented everywhere else.
The VRA was passed in 1965. The Court narrowly construed it in 1981 in Mobile v. Bolden, holding that the VRA only applied to intentional minority vote dilution. In reaction, the Congress, led by folks like Ted Kennedy, not Lee Atwater, amended the VRA to include a results test. I don’t know where you got this idea.
November 12th, 2008 at 1:14 pm
Asher,
Au contraire,
Atwater was instrumental [with an unholy alliance with the Congressional Black Caucus] in shepparding thru the VRA and vigorously pointed out to anyone who would listen that this would ensure Republican gains for exactly the reasons we are discussing today.
November 12th, 2008 at 1:37 pm
Atwater was instrumental [with an unholy alliance with the Congressional Black Caucus] in shepparding thru the VRA and vigorously pointed out to anyone who would listen that this would ensure Republican gains for exactly the reasons we are discussing today.
Wayne, barely anybody voted against the 1982 Amendments to the VRA. He wasn’t needed to “sheppard” it through. Show me a link.
November 12th, 2008 at 1:55 pm
The Republicans won the House popular vote in 2004. Matt is correct about 1996 and 1998.
Also, in presidential elections since the Civil War the Democratic candidate has gotten a majority of the popular vote just seven times: in 1876 (which the Democrats wound up losing), FDR’s four electoral wins, 1964, 1976, and now 2008. Democrats have won thirteen presidential elections since the Civil War, but seven of those wins were plurality wins. By contrast, the Republican presidential candidate got a majority of the vote in sixteen post-Civil War elections, plus two more plurality wins, plus another three elections where the Republican candidates got fewer votes than their opponent but still won.
At the presidential level, its fair to say that the Democrats have had a branding problem, and those are the election most people pay attention to.
November 12th, 2008 at 2:10 pm
The demographic groups that vote overwhelmingly for Democrats are growing relative to the overall population. As the number of blacks and Hispanics grows, the number of districts where the Repubicans are not competative will grow.
In 2008 there were something like 100 districts where the Democratic candidate did not face a challenger but only 25 districts where the Repubican candidate did not face a challenger. In the future there will be more districts where the Republican have no chance of winning and the U.S. will soon reach a point where more than half of the seats in Congress will be safely Democratic and without challengers in the general election.
So, in the long run, what will happen the U.S. House and SEnate have not affected by general elections?
November 12th, 2008 at 2:22 pm
Asher
I admit my memory was faulty – Atwater’s participation was not during the renewal of the VRA in 1982, it was during the redistricting after the 1990 census drawing the new majority-minority districts required by amendments to the act.
November 12th, 2008 at 2:56 pm
wayne -
Atwater might have been involved in redrawing districts in the early 90’s, but the redrawing was not required by the ‘82 amendment, which focused on Section 2 of the Act (which allows a private right of action) rather than Section 5 (which required preclearance by the DOJ of changes in election law, including electoral district boundaries). Almost all of the minority-majority gerrymanders were created because of Section 5. Suits under Section 2 sometimes lead to maps being redrawn, but it is not nearly as common.
Regardless, Al at #29 is right – recent academic studies have shown the impact of racial gerrymanders to be minimal. They were blamed for a while for shifting control of Congress in ‘94, but almost all of the new maps were actually in place in ‘92. The ‘94 loss had more to do with the complete abandonment of the Democratic Party by Southern Democrats than with racial gerrymanders. I think the data showed results being different in like one or two seats in ‘94 based on different maps (it’s been a couple of years since I read the most complete study, so apologies for the faulty memory). So that myth is pretty much busted.
November 12th, 2008 at 3:03 pm
the redrawing was not required by the ‘82 amendment, which focused on Section 2 of the Act (which allows a private right of action) rather than Section 5 (which required preclearance by the DOJ of changes in election law, including electoral district boundaries). Almost all of the minority-majority gerrymanders were created because of Section 5.
Yeah but the Bush and Clinton DOJ’s thought, reasonably enough, that they had to upheld the Section 2 standard when they did their Section 5 preclearance. The Court ruled in Reno v. Bossier Parish School Board (2000) that that wasn’t the case (that all that mattered was retrogression, as per US v. Beer), but until Bossier Parish, that wasn’t at all clear.
November 12th, 2008 at 3:19 pm
Oh, and also, don’t forget that Congress overruled Bossier II at least when they reauthorized the VRA in ‘06. My last post was wrong though – it was Bossier I that delinked Sections 2 and 5. Bossier II just goes to discriminatory purpose.
November 12th, 2008 at 4:11 pm
The question, all this time (and I mean since Nixon won his first term), should have been: “why are they better at politics than us, and what can we do about it?”
Unfortunately, we spent years and years and years getting our collective rear ends beaten at the politics game, having our brand and our image ravaged. America fundamentally agrees with the vast majority of our positions over those of republicans. But, much of america has been convinced otherwise.
And its because the other team wins at the politics thing.
Whether its the rhetoric used, the political tactics used, gaming the system, controlling media and message, whatever, the Republicans have dominated for a long time.
We’re finally turning it around (hopefully), but we need to stay focused on turning things around. We also need to stay focused on reminding people why Republican governance is so bad.
That said, it is a testament to the strength of our overall message and weakness of their message that the Republicans spent over 30 years (from Nixon’s first term to W’s first term) tearing us down before they were finally able to wrest total control of the levers of power away from us, and then lost almost all of them in the space of one Presidential term. All they have left is the Supremes, and the Supremes are still narrowly Socially moderate (though very, very conservative on almost all other counts).
Think about how powerful it could be if Obama had tuned into the oratorical skills of FDR. Obama sounds, frankly, timid in comparison.
November 12th, 2008 at 6:17 pm
Matt writes:
“The vagaries of how congressional districts are drawn produced Republican majorities in those years…”
What Matt is euphemistically referring to as “vagaries” is the 1986 Voting Rights Act which required the gerrymandering of Congressional districts to produce “majority minority” districts that would elect black and Hispanic Representatives. The GOP thought this was a great idea because in order to create districts liberal enough to elect blacks to Congress, you have to cram supermajorities of Democrats into a few districts, leaving Republicans with small majorities in most districts. This worked great for the GOP in 1996-2004, giving them a majority in the House without a clear majority of the total vote, but the big shift toward Democrats in 2006 and 2008 swept the marginal districts to the Democrats. Still, if the GOP gets its act together and/or if Obama can’t actually deliver on his promises to give everything to everybody, then the Voting Rights Act will give the Republicans an advantage in the House again.
March 11th, 2009 at 4:24 am
Very interesting site. Hope it will always be alive!
March 14th, 2009 at 2:38 pm
I usually don
March 17th, 2009 at 2:22 am
If you have to do it, you might as well do it right
tramadol
March 22nd, 2009 at 6:00 am
tramadol
It is the coolest site,keep so!
April 2nd, 2009 at 5:02 am
Great site. Good info
buy cheap viagra
April 3rd, 2009 at 4:02 am
Very interesting site. Hope it will always be alive!
cheap brand pfizer viagra
April 12th, 2009 at 11:57 pm
I like this theme you are using… what is it?