Wednesday, March 30, 2011

Where Gerry meets Dummymander

States are currently doing redistricting for congressional seats. In states where one party has control, they’re trying to gerrymander the districts to get as many seats for their party as they can. A new term is floating about, the Dummymander. A dummymander is when a party creates a district that they’ll win and then they lose it within the next election or two. Thus, they’ve created a district for the other party, not their own.

After redistricting in 2002 most of the country’s congressional districts had a similar number of people. Now, ten years later, 116 of the 150 most populated congressional districts are represented by Republicans in congress. Ninety-one of the 135 least populated congressional districts are represented by Democrats. This is remarkable and couldn’t happen by accident.

Democrats dominate urban city districts, while Republicans dominate more rural districts. In the last ten years there’s been a mass exodus from northern cities. The 50 districts with the smallest number of people includes four in Chicago, five in Northern Ohio (Cleveland), two each in Detroit, Northern New York (Buffalo-Rochester), Minneapolis-St. Paul, St. Louis, Pittsburgh, and Providence. There are also districts in Kansas City, Cincinnati, New York City, Newark, Philadelphia, and Flint. These districts aren’t just represented by Democrats. Seventeen of them are represented by African-Americans.

So, there’s been an exodus of Democrats, especially African-Americans from mostly northern mostly urban areas. Where have they gone? Thirty-nine of the 50 most populated congressional districts are in 9 states. Twenty-one of those are in Florida and Texas. All 3 districts in both Nevada and Utah are among them.

Besides growth from northern Democrats, some of it is from Hispanics. They tend to vote Democratic. It’s likely that these 50 districts have a lot more Democrats than they used to be. That means that while the district may be represented by a Republican now, the addition of Democrats may change that. In Utah and Texas this is probably not a big deal. An R+16 can become more Democratic and still be safe Republican.

The biggest potential trouble for Republicans is in California, Florida, and Nevada. The GOP doesn’t control redistricting in California and Nevada, but they do in Florida. Fair Districts Florida may restrict how Republicans can draw the districts, but they’d be wise to make sure that they draw a nice compact district with lots of Democrats in Central Florida, because if they dummymander they’ll lost more seats than the one they might save now.

The flip side of all those Democrats moving into Republican areas is that there are now less Democrats in Democratic controlled districts. Many of those districts are heavily Democratic, but there are ten districts that would be opportunities. Because these districts have lost so much population, six of them are in states that’ll lose seats. While that might limit the opportunities, there’ll be districts that cover these areas. There may be opportunities there.

Of course the population shifts will impact Senate and Presidential campaigns. Thus, Nevada, Florida, North Carolina, Georgia, and Arizona are likely going to become tougher for Republicans to win. Texas would be a big linchpin, but I think it’s too Republican now to move over to the Democratic side.

While those states maybe tougher to win, Missouri, Michigan, Ohio, Pennsylvania, Minnesota, and Iowa should become easier. The two groups have roughly the same number of electoral votes. The big question will be then, which states will flip soonest.

Tuesday, March 29, 2011

Obama's Strategy

When people look at the 2008 election they assume that Barack Obama won because he was an exciting new candidate and the Republicans were ridiculously unpopular both with everyone, including Republicans. This was't the whole story. A state by state analysis shows that Obama's performance was better than the 53% of the vote he got. Here's how the 2008 vote looked.

What if the two candidates split the vote in 2008? In order to do that we'd need to deduct 3.3 points from Obama and give them to McCain. So we'll take 3.3% of the votes from Obama in each state and give them to McCain.

Florida, Indiana, Ohio, Virginia, and North Carolina swing over to McCain. Obama still wins, although it's close.This shows that Obama's campaign was effective. They weren't trying to win more votes overall, but trying to win votes in swing states. And they were successful. McCain could rack up all the votes he wanted in Texas, but Obama wasn't concerned with how many he could get in New York. In fact, Obama could've lost the popular vote and still won the election. And not by the half a point Bush lost the election in 2000.

McCain could've won by 1.7 points and still lost. Obama knows how to run an effective campaign, efficiently using his resources. That makes him tougher to beat.

Sunday, March 27, 2011

Why Hispanics don't matter as much as you think

The media has discovered Hispanics. And they're in love. Add Chris Cillizza, Christian Science Monitor, Dave Wasserman, MSNBC, Others to those alerting us to the Republican Hispanic problem.

Having worked in Hispanic advertising for many years, I do find it funny that the MSM is acting as if 20-30 million Hispanics suddenly showed up in the U.S. There has been a steady increase for 20 years. Hispanics, however, aren't the force the MSM seems to think they are.

Because Hispanics don't vote. Some of the Hispanics aren't currently citizens and are unlikely to become citizens any time soon. So that isn't a vote to worry about. Yet, even setting them aside Hispanics just don't show up at the polls.

This chart shows how Hispanics, despite being 15.7% of the population, were roughly 9% of the electorate in 2008. In 2010 it was 8%. They were 8% in 2004 and 2006 and 7% in 2000. Despite being the driving force for U.S. population growth Hispanics aren't voting much more. Only 23% of Hispanics vote, while 42% of Blacks, and 45% of Whites do. Because many Hispanics are under 18, the numbers are a little better as a percentage of the voting age population. White people, on the other hand, show up to vote. The Democrats got roughly half the White vote in the 1990s, but they have dropped sharply in the last ten years.This isn't reported because getting Hispanics is cool and any party that goes after White people is racist.

The proclamation that the sky will be falling on California Republicans in 2012 ignores the light Hispanic turnout. Instead the pundits are proclaiming high Hispanic populations will doom the GOP.

If turnout in our sample district is similar to what it was in 2008, a high watermark for Hispanics, a 50% Hispanic population district is actually a district with 35% of the electorate being Hispanic. Such an electorate is winnable for a Republican getting around 35% of the Hispanic vote. That might not happen, but a 35% Hispanic electorate isn't one that's 50%.

The low turnout shows up in the vote totals. If you look at the 10 lowest voter turnout districts in California you'll see that 8 of them are in the top 10 most Hispanic districts.

Republicans undoubtably have some problems with the Hispanics right now. Will they continue to do so in the future? Predicting how the electorate will vote in 12-20 years when Hispanics might actually be a force is difficult. Voting patterns change a great deal. The Democrats' major appeal to Hispanics is "the racist Republicans hate you." That may work right now, but what happens if the immigration issue goes the way of "Don't Ask. Don't Tell?"

If Obama had done as badly as the Democrats did with Whites in 2010 he would've lost in 2008. Democrats having been turning off White more and more each year. That's a far bigger factor in 2012 than the Hispanic vote.

Wednesday, March 23, 2011

Is the House in play in 2012?

Chris Cillizza and Aaron Blake at the Washington Post are asking that question. Democratic pollster Democracy Corps argues that it is. Glen Bolger at Republican pollster POS disagrees.

I've looked at this before. I've been pessimistic on this.
Below are Presidential re-election and how many seats the Democrats will get if they perform like each President. This chart is derived from determining how much above the expected average number of seats each President won when re-elected.

The good news for Democrats is that in almost every case they'll gain seats. It'd take a Jimmy Carter size loss for them not to. If they follow Reagan, Eisenhower, or the two Bushes they'll fall a little short. If they perform like Johnson, Truman, and Nixon, they'll do it. Johnson's circumstances were unusual. He benefitted from sympathy as a result of John Kennedy's death and a poor Republican candidate. Johnson also got the highest percentage of the popular vote of any President.

Truman is a better example, as his foe wasn't that bad and there wasn't much sympathy from Franklin Delano Roosevelt's death. Still, Truman had never been elected before, making it different from Obama's situation.

Nixon had a poor opponent and was also one of the three Presidents to top 60% of the vote. Reagan also won in a landslide and the expected 217 total is within a margin of error. So Truman is the only non-landslide President. If Obama wins by 18 or more points, as Reagan, Nixon, and Johnson did the Democrats have a good shot to flip the House. Otherwise, it's a long shot.

If we look at the 11 times the House has flipped since 1894 we see that the incumbent President's party was the one doing the flipping twice.

House gains big enough to flip the House are usually due to a President's unpopularity, not due to his popularity.

Of course the big 800 pound gorilla is redistricting. Right now there are 234 Republican leaning districts, 192 Democratic leaning, and 9 even. After redistricting, there'll be at least 240 Republican leaning districts, if not 250.

Monday, March 21, 2011


I was asked to volunteer for the Huey campaign today by someone I worked with before.

Me: I don't know enough about the candidates to choose one.
Her: Huey's the only one with money. With the tight timetable the other candidates won't have the time and the money to mount a good campaign.
Me: What's RPLAC's position?
Her: I think they're trying to thin out the field.

Saturday, March 19, 2011

Saturday at the convention

I talked to someone I know who is keyed into what's going on behind the scenes. He actually works in the 36th and goes back to the Deukmejian administration. He told me that the candidates have been talked to by party officials and he's spoken with a couple of them. Efforts to get candidates to withdraw have failed. No one is dropping out. I mentioned that maybe they all wanted to see how they'd do and use it as a springboard for 2012 when the district becomes more Republican. He told me that people are talking about the possibility that Palos Verdes might not be put in with the beach communities, but might end up with Long Beach in a Hispanic district. I shook my head. I don't think a Republican will make the run-off.

I spoke with a friend who isn't as well connected but has worked with party leadership and asked what she heard about the senate. She mentioned Chuck Devore. When I mentioned that I heard he was running for Orange County Supervisor she mentioned that people close to Michael Reagan had been floating his name out there.

Friday, March 18, 2011

California Republican Convention 1

I met a gentleman named Craig Huey. He owns the Direct Marketing Center, an ad business. A Google search shows he's an activist and an Evangelical. He has nearly 5,000 followers on Twitter. (!/reality_alert) He's announcing next week that he's running for the 36th district congressional seat. Three other Republicans are running. I mentioned this to him and he said that he'd have money behind his run and thinks those that don't have money might drop out. He hadn't been in touch with any other candidates, nor did he have plans to meet with them.

I met Forest Baker. He was edged out by Pete Stark by 44 points and plans to go back for more in 2012. I asked him why he ran against Stark, a certain losing proposition. He answered that he got a lot of Democrats to vote for him. Since he didn't get enough I don't think much of that answer.

As I was leaving the registration desk, I got a call from an enthusiastic volunteer from the California Republican Party. He went through his script, hitting on a lot of talking points. Then he asked me to join the party. Which I'd already done. 20 minutes before.

Wednesday, March 16, 2011

Flooding The Zone

It's early in the cycle. Few pollsters are doing polls. So one pollster, PPP, seems to be doing 70% of the state polls. PPP is a Democratic pollster. Without any independent pollster their numbers should be taken with that in consideration. Yet Politico, Swing State Project,, Talking Points Memo, and report the polls as fact. Even Republican sites like Hot Air, Fox News, and Washington Examiner accept the number without question.

No one poll should be taken as gospel, especially not when it's a partisan pollster. What makes it worse is that in at least half their polls, they have a higher percentage of Democrats than voted for them, even in 2006 and 2008 Democratic landslide elections. This is true in New Mexico, which has 55% Democrats compared to 44% in 2008, Colorado, Rhode Island, Virginia, and Ohio. PPP says that this can happen because they don't adjust for party. Since party is the best indicator of how a Republican or Democrat will vote, that would seem to be the first thing you adjust for. PPP did the same thing in 2008, only adjusting for party two months before the election when they switched to likely voter.

I don't mind a partisan pollster pushing their candidates. I do mind everyone on the Internet taking their numbers without even looking at the data.

Sunday, March 13, 2011

Congressional v. Presidential Voting

Last week I looked at how 2008 Presidential voting compared to how candidates normally did statewide. Republicans do much better in California, Florida, Iowa, Indiana, and Michigan, while Democrats do much better in Arizona. Some states weren't included because they didn't have enough 2006 and 2010 races to provide good data.

In order to take another view of congressional races, I'm comparing how a party's statewide congressional total (indexed to 50%) compares to the presidential total (indexed to 54% Democratic 46% Republican). I wasn't able to include all the states again, however. Some states have such a small congressional delegation that the congressional vote is heavily/entirely influenced by one strong candidate. Earl Pomeroy has done much better than Obama did in 2008, but that may be all Pomeroy. Looking at his numbers wouldn't be able to tell us how another Democrat might do. I once again averaged 2006 and 2010, with the belief that the two years that bookended the 2008 election would produce a good average since one was a Republican year, while the other was a Democratic one.

I also excluded some legacy states. Mississippi, Arkansas, and West Virginia have had very strong Democratic congressional totals, but increasingly weak presidential ones.

While I was able to fill in an expected number of votes to give an accurate read when a candidate ran unopposed Oklahoma, New York and Texas had too many unopposed candidates to the point that I'd be using too many estimated votes.

Florida, Michigan, and Iowa again come up very strong, showing that Republican statewide elections translate to congressional races. Illinois isn't much of a surprise, as President Obama likely skewed that state too Democratic.

Tennessee may qualify as a Democratic legacy state. So I'm not sure their strong congressional showing, when compared to the Presidential race, is that telling. The rest of the list doesn't contain a lot of surprises. Arizona was overblown Republican due to John McCain, while we've seen strong Democratic congressional wins in Connecticut, Maryland, North Carolina, and Washington.

Overall, the numbers skew slightly Republican. This may mean that 2010 was a better Republican year than 2006 was a Democratic one or that Republicans are very good on a congressional level or John McCain was very bad. You could deduct about 1 point from each if you think that's the case That'd still leave a significant number as congressional overachievers. Republicans shouldn't have that much to fear from Fair Districts Florida. They also should be able to gerrymander their gains in Wisconsin and Michigan, and do better than expected with the contraction in Illinois and Iowa.

Notable on the list are the states which are fairly neutral. Republicans may have a tough time gerrymandering their gains in Ohio and Pennsylvania. as mediocre McCain numbers may translate into mediocre congressional numbers.
When looking at the

Tuesday, March 8, 2011

Senate Retirements

The retirements are coming one after another, with Jeff Bingaman, Daniel Akaka, and now John Ensign retiring. I've talked about the likelihood of Republicans taking over the Senate in 2012 and almost every retirement has helped the GOP.

Without the retirements Republicans had virtually no chance in New Mexico or Hawaii, while the retirements in Nevada, North Dakota, and Virginia have made Republican pick-ups more likely. I don't think Republicans are favorites New Mexico or Hawaii, but I think they now have a 45% chance to win one or the other. It was 10% before the retirements. Retention is far more likely in Nevada now that the scandal plagued Ensign won't be running for re-election. Heather Wilson is considered the strongest GOP candidate in New Mexico and she's in. Republicans are hoping for Linda Lingle in Hawaii and if Dean Heller runs in Nevada, he's not a lock, but he will be a heavy favorite.

I'm putting the senate at 52.6-47.4 Republican right now and giving Republicans around a a 2/3 chance of winning the senate.

Sunday, March 6, 2011

Where else does McCain-2008 give the wrong story?

Earlier this week I looked at California and how, on average, Republican congressional candidates did a point or two better in Majority-minority districts, 2-3 in Democratic districts, and 12-20 in Republican districts. Even though Barack Obama won 6 of those districts in 2008 and lost 2 others by percentage points, Democrats have only been within 10 points of a Republican congressman in 6 of 57 elections in the last 3 cycles.

So I decided to look at statewide elections in 2006 and 2010, a strong Democratic year and a strong Republican one, both of which were mid-term elections. I took all the elected offices that had a Republican and Democrat running both years, averaged them, and then took the average of the two years. The sample size is small and the actual neutral average may not be at the midway point between the two. Yet this should tell us whether Obama's numbers are good to use when determining how Republican or Democratic a district is.

Obama outperformed the average statewide Democrat by 3.5 points in 2008. Thus, a district won by a few points probably doesn't lean Democratic.

Arizona and Iowa produce different, results. One would guess McCain, when running against a non-Arizona, would overachieve two Arizonans in race. Iowa, on the other hand, leans Republican on the state level despite leaning a little towards Obama in 2008. This should bode well for Republicans in 2012 when five congressional districts become 4.

Ohio went slightly more Republican in the Presidential race, but the difference is negligible. McCain numbers are fairly accurate.

Pennsylvania had the smallest sample size, as most statewide offices weren't up in 2010. It leans slightly more Democratic than Obama-McCain did. Considering Republicans have done well in Pennsylvania toss-up districts, this is a bit of a surprise. Florida is no surprise, however, as Republican candidates have done well there.

Nevada had a slightly stronger Democratic lean in the Presidential race.

Michigan produced the biggest disparity. Despite Obama's strong showing Republicans have done very well here. So Obama-McCain numbers aren't very indicative. Minnesota's slight Democratic lean is confirmed by the statewide vote.

The surprise isn't that Republicans won every statewide election in both years, but how well Obama did there in 2008. It seems unlikely he'll do so well again. Missouri's Obama-McCain numbers were very indicative.

Illinois Democrats do about as well as Obama did. You'd expect a stronger lean for Obama, given that he wasn't running against someone else from Illinois. Oregon's Democratic lean is fairly consistent.

Georgia has a remarakable number of statewide office holders and the numbers indicate that.

There are several other states that have sufficient 2006 and 2010 elections that I should be able to add later, but this data shows that Republicans do at least two points better than Obama in several states, but in only one, Arizona, do the Democrats do the same thing..

Tuesday, March 1, 2011

How McCain-2008 Lies

I've been perplexed for a while how Republicans do so poorly on California redistricting plans that people draw up using Dave's App. After all, Republicans hold only 36% of the congressional seats now. If you only include Republican and Democratic vote totals, Republicans have never been beaten 64%-36% in a statewide California election.

All of these calculations use Obama-McCain numbers to judge how competitive a congressional district is. Those numbers don't tell us the truth.

Barack Obama won California 62%-38%. If you look at the 16 statewide elections in 2006 and 2010, Obama beats the Democrat in 15 of them. Even at their best Obama outpaced California Democratic candidates by 2-12 points. Unfortunately California held no statewide elections in 2008, so we can't compare apples to apples. With 16 elections before and after the Obama election we have a good sample how candidates will do statewide. "Aha," you say. "But Barack Obama was an exceptional candidate." Exactly. And he won't be running for California statewide or congressional office. Yet we look at his numbers when judging how congressional candidates will do. The Democrats only have one Barack Obama. Other candidates don't generate that excitement.

So let's look at how congressmen did.

This group of Democratic congressmen are in majority-minority districts. These are overwhelmingly Democratic. Due to the Voting Rights Act rules, there likely will be a similar amount of these districts in 2012. Since they'll have the same proportion of Blacks and Hispanics as these districts do, we'll likely be able to set aside those voters for these districts. They won't impact the remaining 33 districts very much.

Here are 14 White Majority Democratic districts. Republicans did a lot better in House, Senate, and Gubernatorial races in 2010 compared to Barack Obama in 2008. Despite having strong incumbents and often having weak challengers these congressmen didn't do much better in 2008 and 2006 than Obama did.

Republican congressmen outperformed John McCain in 50 of the 52 contested congressional elections from 2006-2010. This includes a Republican year and two Democratic years, Republicans with strong challengers and weak ones, and 3 open seats. (highlighted in pink) At their worst, Republican congressmen outperformed John McCain by 12 points. At their best they were 21 points better. If you average the seats that'll be contested in 2012, the range was 4-12 points more Republican.

If the McCain numbers were reflective Republicans would win 16 seats in an average year. If you add 4 points to each Republican, you'll get Republicans ending up with 23-24 congressmen in an average year. Based on Republican share of the vote in California that's about what you'd expect. If you add 2 points to each Republican, they'd get 18-19 congressmen in an average year. About what they have now.

It's possible that an open seat will be close to McCain but incumbent Republican congressmen will outperform him, in most cases by 12 or more points. If Barack Obama wins a district 51-49, an incumbent Republican will likely do at least 55%-45%. And incumbent Republicans will be running in 2012.

Thus, any evaluation you read based on John McCain's vote total will be biased to the Democrats.