Although the primary election is still months away, I’ll go out on a limb now to make a prediction: Ray Ellis will win the District 1 City Council seat in June, giving Republicans their first Council majority in more than two decades.
Discerning meaning from election results is like taking a Rorschach test – everyone sees what they want to see. Some will no doubt interpret the outcome as a referendum on Mayor Kevin Faulconer’s leadership; others will blame ex-Mayor Bob Filner for sullying the Democratic brand in San Diego. Perhaps others will point to Ellis’s impressive (or Barbara Bry’s disappointing) campaign performance.
Don’t pay much attention to these explanations. The outcome of this year’s election was largely predetermined five years ago: The credit – or blame, depending on your point of view – goes to the San Diego Redistricting Commission, the seven-member citizen group that’s formed every 10 years to set new boundaries for City Council districts. The maps the commission drew in 2011 made it very hard for Democrats to translate their large voter majorities into a sufficient number of City Council victories.
Although Democrats have a huge voter registration advantage in San Diego – and President Barack Obama carried the city by overwhelming margins in 2008 and 2012 – Democratic voters are concentrated in a few Council districts in the southern part of the city. Indeed, David Alvarez’s and Marti Emerald’s districts are so blue that the San Diego County Republican Party has stopped bothering to field credible candidates there.
The downside to concentrating so many Democratic voters in a few safe seats is that it leaves too few Democrats for the remaining Council districts. Thus, although Democrats still have modest voter registration pluralities in Districts 1, 2, 6 and 7, these districts can easily be won by Republicans because Republican turnout tends to be higher (especially in non-presidential election years) and Democrats tend to drop off at higher rates. Indeed, Republicans currently control three of these four districts, and will add District 1 to their column in June. I hate to say “I told you so,” but I made this prediction five years ago, after the Redistricting Commission adopted its final maps.
In other words, a great number of the Democratic votes in San Diego City Council elections end up being wasted because Democratic voters are packed into a small number of uncompetitive Council seats. Normally, this would be a telltale sign of partisan gerrymandering – an effort by one party to intentionally screw the electoral prospects of its opponents. In San Diego, however, the political pain is entirely self-induced.
We Stand Up for You. Will You Stand Up for Us?
I’ve been following this argument, but listening to the podcast prompts me to comment:
1)Jason Roe and John Nienstedt are right. I agree with their points on the meaning of the elections.
2)The premise of the article and the podcast discussion seems to be that the purpose of redistricting is to maximize the number of elected Democrats.But nowhere mentioned in the article or on the podcast are the by-laws of the Commission, nor the federal voting rights laws, and the many court cases that interpret them.
The Commission was required to respect communities of interest, in particular “racial, ethnic and language minorities” and to draw districts that are “geographically compact - populous contiguous territory will not be bypassed to reach distant populous areas.”
Though the election results may disappoint Professor Kogan, the Commission was required to follow its own bylaws and the relevant law.
@John Hoy The purpose of redistricting is to produce a redistricting plan that accurately aggregates voter preferences into policy outcomes.
Federal voting rights laws were not at play because: (1) as far as I know, there is no history of racially polarized voting between Latino and black voters in San Diego, so it is not necessary to separate them into different districts; (2) the black community is too small to make up a majority of a council district (as a percent of citizen voting age population), which means that the VRA does not require drawing a black district; and (3) LGBT community is not covered by the VRA.
You are right that the city charter does require respecting communities of interest, but there are so many overlapping communities of interest that one could justify almost any set of maps in these terms. For example, in 2001 Donna Frye argued (and the redistricting commission agreed) that the "Mission Bay watershed" was a community of interest.
One key point that seems to be missing in this discussion is that DTS voters are far more likely to be flexible in the way they vote than registered D or R voters. The idea that DTS voters are closeted Red or Blue voters is just wrong. If DTS voters loved one party they would be in it, at least that is my opinion.
The biggest swing in coming elections will be those pesky independent DTS voters. And this type of voter is growing faster than either of the old parties. Established parties can raise more money but can they raise more voters?
@Glenn Younger Glenn, as an empirical matter, that is incorrect. A little dated but still very much accurate: https://books.google.com/books?id=V72ZMHZktZEC&lpg=PR9&ots=2iSy5l5Ak5&dq=independent voter myth&lr&pg=PP1#v=onepage&q&f=false
@Vlad Kogan @Glenn Younger In my case, and in the case of others I know who have become independent in the last several years, I would say that the book's data is dated and no longer accurate. At least not accurate for me as an independent voter.
I find it interesting that the % of DTS voters is often the % that a candidate will outperform their party registration. See Kevin Faulconer in last election. But I could be wrong.
It may be time for somone to do a newer study on why someone would want to be a DTS or Independent voter, and how they vote. Again, I could be wrong.
Faulconer received 52.9 of the vote. Since my stated prediction based on our tracking was Faulconer 52.5, the polling was accurate. You keep saying it was a low turnout election but 44% is not low turnout. Our turnout model is accurate. Strength of partisanship may be correlated with turnout, but what's far more correlated is turnout history, so we use that.
I think the overall point is that we each use the tools we have at our disposal to try to understand election outcomes. My tools are polls which are based on the individual voter and include lots of demographic and attitudinal variables. Your tool is the EI model which is based on the precinct and uses 2 variables, partisanship and vote share w/in precinct. Your tool causes an over-emphasis on partisanship as a predictor. My tool inherently acknowledges that some people's minds change from time to time and over the course of a campaign and they don't necessarily vote like their neighbors or in rigidly along the lines of how they registered (if they even know what party the candidate belongs to... in January '14, 41% were uncertain Faulconer was a Republican). And, after all, the candidates in this race did not have Ds or Rs after their names. Those situations -- with the right candidate(s) and campaign -- can produce a lot more crossover than strictly partisan races. Because Faulconer was running so well with Dems, he still would have won with turnout in the mid-50s.
@john nienstedt I hate to beat a dead horse, but if you are right, there is an observable implication. Suppose we calculate a new precinct-level measure: Faulconer Overperformance=Faulconer Vote Share - Republican Share of Voters (just those who voted). If you are right, and Faulconer got some meaningful percentage of D cross over votes, then Faulconer Overperformance should be higher in heavily D precincts (the more D voters, the more cross overs, as percentage of total voters). If I'm right, and Faulconer won based on R-leaning NPPs, then Faulconer Overperformance should be higher in R precincts (where there are more R-leaning NPPs). Take a look at the charts below. The raw precinct-level data is here, if you are curious:
https://osu.box.com/mayor (need to copy and paste link)
There appears to be no room, in the era of “identity” politics, for non-partisan elected positions, even at the municipal level. Sad, really, as San Diego elections have morphed, largely thanks to the efforts of local unions and Democratic activists, into red vs blue events. And identity as black, brown, female, gay and other “categories” is playing an increasing role in party stereotyping. If you’re a small business owner, you must be a Republican, a union leader a Democrat, a farmer a Republican, a schoolteacher a Democrat and so it goes as expectations harden. Black Republican? Gimme a break!
I remember, e.g., how Carl DeMaio was treated, even in a relatively centrist publication like VOSD, in the recent congressional race that narrowly returned Scott Peters. It’s just not OK to be a “gay” Republican. Late in the campaign, one of DeMaio’s former staffers made allegations about Carl’s conduct toward him, and instead of saying “prove it” to the accuser, who had been fired before the allegations surfaced, DeMaio was challenged to prove a negative. Unfortunately, the proof came after the election, courtesy of the FBI, not the local media.
Prof. Kogan engages in this stereotyping in his article, explaining that Democratic voters can’t be expected to vote to protect their own interests, except in “presidential” election years. He then further reveals his biases by predicting that Ray Ellis will win District 1 because of redistricting five years ago. I don’t get it. Democrats have a higher number of voters in District 1 than Republicans, and education or poverty is hardly a factor in the most affluent area of the city. Can it be that candidates who happen to be less ideological are more attractive sometimes? This certainly happened in the District 2 election a few years ago, as the “Democratic” candidate outlined her big plans to help Bob Fillner remake the city government into a Democratic bastian. According to Kogan, it was a matter of turnout, not expressed ideology, that did the trick. I’d like to think he might be wrong.
@Bill Bradshaw, I think much of your comment is complete confirmation bias.
@Gregory Hay @Bill Bradshaw Out of curiosity I scrolled down to see if you had made any other comments on Prof. Kogan’s term paper. Voila! I was the second person you had accused of “confirmation bias”. Maybe you could, instead of calling people names, give us your own reaction to Kogan’s opinion piece.
The truth is that even following the 2011 redistricting, the 1st District provides an advantage to Barbara Bry, rather than Ray Ellis. In 2012, President Obama won 57% in the current version of the 1st District, just two points off his 59% vote in 2008 in the previous iteration of the 1st. So the district underwent only a minimal partisan shift, and still gave a strong majority to the president in 2012, post-redistricting. Moreover, polling data shows that even though Republicans have small party registration advantage among the likely primary electorate, more voters in this electorate today consider themselves Democrats (40 percent) than Republicans (35%). The totality of the data clearly shows that District 1 leans Democratic. The idea that the partisan composition of the electorate works against Bry is simply not evidenced by the data.
Most importantly, Barbara Bry is a successful high-tech entrepreneur with a moderate, pragmatic stance on the issues, which gives her an appeal that transcends partisan politics. When elected, she will be the first high-tech entrepreneur to serve on the San Diego City Council. This is significant because studies show that high-tech and life science industries are driving the economic future of our region, and most of these companies are based in District 1. Someone with Barbara’s decades of experience helping to grow companies in these sectors sets her apart from the other candidate and makes her the best choice for driving District 1 and San Diego forward. She is committed to expanding the innovation economy to create more good-paying jobs so that we can build the tax base needed to fix our streets and sidewalks; ensure law enforcement has the resources to keep San Diegans safe; and protect our beaches, bays, and parks. Barbara Bry understands that keeping the 1st District and San Diego safe, clean and prosperous is a matter of principle, integrity, and thoughtful leadership, not party affiliation.
@Hilary Nemchik Good luck getting the UCSD students who voted for Obama in 2012 and 2008 out to the polls in June...
UCSD students have actually been a very reliable part of our grassroots, volunteer driven campaign - check out the recent Citybeat article discussing our broad base of UCSD volunteers. Barbara worked on the campus for over a decade when she was Associate Director of Connect (once based there). She also taught a UCSD course on entrepreneurship and mentors students. Her longtime connection to the university and her message and track record of job creation resonates there.
@Hilary Nemchik BTW, Although Obama won 57%, Marti Emerald won less than 55%. This is in large part because about 10% of the Obama voters didn't mark any vote for the city council race, as I documented in VOSD before. Take out Marti's advantage as an incumbent due to name recognition, etc. (~6%) and the fact the election will be decided in June, and I would venture a guess that Bry loses 45-55.
@Hilary Nemchik In that case, I will be happy to be proven wrong. If Barbara wins, I will send both her and you a very nice bottle of wine each!
Prof. Kogan's thesis is largely dependent on the assumption of lower democrat turn-out on election day. Therfore, if democrats fielded stronger candidates, if Labor put its money behind strong candidates (re: not Alvarez for mayor) and if the council election was held at the time of the presidential election - the outcome would easily be different.
@bgetzel You are right to a certain extent. But lower turnout among Democrats is not simply driven by the campaign dynamics. Poor, less educated voters are less likely to vote everywhere in every election, and this happens to be precisely the kinds of voters that support Democrats. Even in 2008 and 2012, Democratic turnout was lower than Republican. And even in these high-turnout elections, as I've showed in VOSD before, Democratic voters are also far more likely to "roll off" -- vote for the president, but then skip the down-ballot nonpartisan races.
I wish the premise were true...but only 1 of the 9 districts in the city has a GOP registration advantage. Republicans represent 3 Democratic plurality districts (D7 +4%, D2 +4%, D6 +2%). The difference isn't the lines, its the policies...one side is committed to reform and fixing the problems at City Hall, the other side isn't.
@Jason Roe It's not the "registration" advantage that matters, it's the "actual voters who vote" advantage. The Democratic plurality in these districts is illusory, because Democratic turnout is lower.
One side committed to reform? Hahahaha! You're funny, Roe. But not. One side is committed to lining the pockets of Chamber and Lincoln Club members. And screwing people like my fellow Barrio Logan residents. Which side is that? Yours.
@Vlad Kogan @Jason Roe So Democrats have a majority or a plurality in 8/9 districts and Republicans have the advantage? The city has a 26% GOP registration and Faulconer pulled 53% of the vote - 27% ahead of registration and he was outspent by $1 million. Apparently the maps should've been drawn so Republicans are irrelevant. I'd argue that the maps, which give Democrats 4 completely safe seats that a Republican will never win, and the Republicans one safe seat that a Democrat will never win, and 4 seats that lean Democrat, is a pretty good deal for Democrats. Let us also remember that the Independents lean left and Republicans have been winning them too. It's not the lines, its the agendas.
@Desde la Logan I'm sure you don't want any facts to get in the way of your opinion but "my side" has put more investment in communities historically underserved than at any time in years. Also interesting that "my side" spends a lot more time trying to build bridges to "your side" to find ways to move the city forward, while "your side" seems intent on making sure we stay divided. Well done.
@Jason Roe You're ignoring the large DTS block, which is made up 90% of closet partisans that break down roughly along the same lines as the Dem-Rep registration. As for Faulconer, he won in a super low-turnout special election when Republican turnout was way higher than Democratic. As we showed here (http://u.osu.edu/kogan.18/files/2015/02/sdmayor-2b9uf0b.pdf) he got almost no cross-over Democratic votes. See, in particular, Table 12.1 and 12.3.
the only side I'm on is that of my fellow Barrio Logan residents. Who always seem to get screwed by your side. I'm neither Dem nor Rep. Never have never will. But I do know which side ain't on mine. It's the side of the mayor that hasn't done a thing for Barrio Logan except use us as photo ops at press conferences.
@Vlad Kogan @Jason Roe Sorry, that's not accurate. We tracked for the final 30 days and routinely pulled 17-20% of the Democratic vote with our last track giving us 19%. And looking ahead, Faulconer's approval rating is well above 50% with Democrats and Latinos with the most recent poll showing a 70% approval rating city-wide. He's governed in a non-partisan fashion, just as the other Republicans on the Council largely have. I'm not a statistician but I know enough to say quite confidently that there is no model you can dream up that Faulconer ran 27% ahead of registration without significant cross-over.
@Jason Roe Are you talking about self-identified Democrats (including registered DTS) or registered democrats?
@Jason Roe Here is a quick sanity-check on our ecological regressions. If you look only at consolidated precincts where registered Democrats made up at least 70% of all voters who cast a ballot in the race (365900; 365400;360000;365800;365600;373200;350300;360700;371000), Alvarez's vote-share exceeded the Democratic share of registration (again, only among voters who actually voted) in every case. How could that be consistent with a large amount of cross-over voting? I trust our EI models (which examine election data and party registration data among those who actually voted) more than tracking polls that guess about who is actually going to vote. The die-hard partisans were much more likely to vote in a low-turnout election like this than the voters who were potentially persuadable.
@Vlad Kogan @Jason Roe Turnout was not super low in 2/14. It was 44%. In the previous special -- when Faulconer also came in 1st -- it was 35%, also not super low for a special. Super low would be the 27% turnout in the 6/14 primary.
More importantly, the partisan breakdown among VOTERS in 2/14 was +7% Democrat. Even if Faulconer had received every single one of the Republican votes (which he didn't), he would have had to receive 74% of the NPP/3rd party votes in order to have gotten zero Dem crossover votes and come out with 53% overall. There is no way that happened. The most likely breakdown of Faulconer's 2/14 votes was Dem 19%, Rep 90%, Other 58%.
@Vlad Kogan If your point above re DTS (now called NPP) breaking the same way as the partisans is correct, and I generally think it is, then the Democrats can count on MORE votes not fewer, and the party is more powerful, not less, than their registration numbers suggest.
@john nienstedt It’s funny you say that. Our EI model (Table 12.3 in the above-posted link) estimates Faulconer winning 68% of DTS and 74% of “Others.” I think this is explained by two things. First, I think Faulconer did decently well among D-leaning NPP voters—winning, say, 15-30% of them. But, second, the main story is turnout. Your overall turnout figure of 44% is deceptive, since it masks huge differences between parties. It was 57% for Rs, 45% for Ds, and an abysmal 32% for NPP! The gap between Ds and Rs was big, but I would bet it was even bigger among D-leaning and R-leaning NPPs. This is consistent with what we know about D-leaning DPP voters: They tend to be younger and less white, precisely the kinds of voters who are least likely to vote in an off-cycle special election. This helps explain why Faulconer did so well among NPP (the NPP who turned out).
@john nienstedt I don’t get your second point about Democrats counting on more votes when NPP is taken into account. If NPP breaks the same as partisans, then D power will be exactly the same as its share of the two-party registration. In other words, San Diego is a 60% D, 40% R city (not 26% R, as @Jason Roeslyly suggests). With turnout among Rs and R-leaning voters ~15% higher in the special election, Faulconer’s winning 53% looks far less impressive.
@john nienstedt BTW, I've been waiting eagerly for the next installment of the super bowl image study. I use it in class and students absolutely love it.
@Vlad Kogan @john nienstedt If it's 60/40 Dem, then no Rep should ever win. And yet Faulconer did. Because he picked up a large % of Democrat votes. The turnout point you're making is immaterial. The turnout was in fact D+7. That means Faulconer must have won a large % of Dem votes. To answer your question about where Dem-leaning NPP/Other voters went, Faulconer won about 40% of them. Your EI model vastly underestimates the number of conservative Democrats and/or wrongly categorizes too many of them as Alvarez voters.
@john nienstedt If it's 60/40 Dem, turnout is low, and Republican turnout is 15% higher then Dem, of course Republicans will win. For example, if D turnout is 30% and R turnout is 45%, the voters who vote will be exactly 50/50, even with a 60/40 registration advantage.
Turnout was not D+7 -- because you keep ignoring the huge turnout difference in D- and R-leaning NPPs.
The EI model doesn't estimate anything about the ideology of voters. The only thing it looks at is: (1) Alvarez/Faulconer vote-share in each precinct; (2) the proportion of voters who actually vote in that precinct who are D, R, NPP, and Other. It backs out the proportion of each group that votes for each candidate based on how the vote share varies across precinct voter composition.
@Vlad Kogan @john nienstedt "Turnout was not D+7" is factually incorrect. ROV has the actual #s, as I'm sure you know. On the other hand, "The EI model doesn't estimate anything about the ideology of voters" is correct. But that's the problem. Because it doesn't see ideology the model -- and, in the words of Monty Python "it's only a model" to begin with -- underestimates ideology's moderating effect on Democrats. Instead of attributing Faulconer's success to D-leaning NPPs, you should be attributing it to right-leaning Ds. But you can't because your model is constrained only to partisanship and not other variables like ideology (and race and age and length of residency and type of residence, etc.) which are all things an actual poll can include in its analysis.
@john nienstedt One thing a poll can't tell you is who among the people you poll will actually vote.
@john nienstedt Let me rephrase, "Turnout was D+7" is true only if you count people who register as Ds, and exclude people who register as NPPs but are functionally Ds and Rs. We can't tell anything about ideology, but we can tell what proportion of Ds, Rs, NPPs, and Others are voting for what candidate. Almost no Ds voted for Faulconer, and many NPPs did. Among NPPs, you're right we have no way of figuring out how much support was coming from D-leaners, how much from R-leaners, and how much from pure independents.
@john nienstedt BTW, I have nothing against polls in general. But I don't think they are that accurate for predicting low-turnout elections like this one, because strength of partisanship is going be correlated both with who actually votes and which candidates they support. E.g., strong Democrats are more likely to vote, but also less likely to crossover to support Faulconer. Unless your turnout model is really accurate, you're going to overestimate the crossover support.
Since you know who you polled and can see whether they ultimately voted, I'd like to know how accurate your turnout model was. And, in particular, how the accuracy varied between subgroups. Did you overestimate turnout more among Faulconer-supporting Ds and Alvarez-supporting NPPs? That would be my prediction.
@john nienstedt One last thing: I think you'll be really interested in this paper: http://www.stat.columbia.edu/~gelman/research/published/swingers.pdf
@Jason Roe, you keep saying things like "It's not the lines, its the agendas" but without offering any proof. Barring that, you're just offering an *opinion* without any validation. (Confirmation Bias.)
I am shocked that this hasn't received more comments. It's really very interesting. I know at least a couple of those commissioners and really can't wait to hear their thoughts. Also, this is a little picky, but shouldn't you wait until *after* the election to bust out "not to say I told you so?" Seriously, this is a fascinating point about politics because it sort of supports what many people - including more than a few of my apolitical friends - have to say about the system. If Kogan is correct, this might be an overwhelmingly sad affirmation of the popular view that one's vote doesn't matter. I actually think Ray Ellis and Barbara Bry are both incredibly sharp and they are obviously both very accomplished. And perhaps most importantly, both are very personally compassionate human beings who give of themselves in service to many, many less fortunate. I digress. Mr. Kogan, please tell me you have some antidote that is better than Mr. Blackson's bike to ward off the apathy such grim news about voting creates.
Mr. Kogan's assertion that I drew up a City Council redistricting plan in 2011 is absolutely false. Had Mr. Kogan interviewed me about my role in the process, I would have gladly explained it, and I could have put him in touch with the leaders of San Diego's African-American, Latino, LGBT, and Asian/Pacific Islander communities who actually did draw maps to maximize their communities’ opportunities for political empowerment. This is precisely the community-driven process the voters of San Diego were seeking when they created our independent redistricting commission. Also, the process of trying to weave these communities' separate priorities into a unified citywide proposal was led by Empower San Diego (http://goo.gl/fAj0Rg), and while I lent technical assistance to all of these groups, I never created a citywide map, never testified at the Redistricting Commission, and never asked anyone else to do so for me. Mr. Kogan has done a great disservice to dozens of volunteer activists from these diverse communities by ignoring the hundreds of hours of work they put into the redistricting process.