I was a guest on the Trickle-Down Socialism podcast last week. It’s a fantastic pod that focuses on politics and problem-solving from a progressive point of view, but avoids unnecessarily pedantic and didactic conversation that can make some Leftist discussions tedious. The host, Pat, is a civics teacher in Boston and has had some interesting recent guests, such as Suffolk County (Boston) District Attorney Rachael Rollins and progressive Congressional challengers from around the country. Pat and I discussed redistricting and apportionment. You can check out my interview here.
New Mexico Special Election
Today there is a special election in the 1st Congressional District in New Mexico to fill the seat vacated by Interior Secretary Deb Haaland. The Land of Enchantment has an unusual method for filling House vacancies in which the parties hold conventions to pick their own nominees instead of through a primary or caucus. That means today is the general election between state Rep. Melanie Stansbury (D) and state Sen. Mark Moores (R).
The Republican campaign is testing some messaging for the 2022 mid-terms in this race. (Biden won this district by 23 points and Stansbury has to be considered the favorite in this race.) There are at least three distinct messages that the GOP is trying out here for possible use in potentially competitive districts next year. The first is attacking Democrats for wanting to “Defund the Police.” This unfortunate slogan is easy to attack, especially when one is being disingenuous about what activists mean by this. Moores has been focusing a lot of attention on the crime rate in Albuquerque while launching attacks on Stansbury for supporting the BREATE Act, which calls for redistributing some police funds to social and health services.
The next line of messaging is that Republicans “have a lot of leaders.” This is the talking point response to questions about Trump being the de facto leader of the GOP. While tethering oneself to Trump might be a winning strategy in Republican primaries and other safe GOP elections, polling suggests that it might cost the party seats that would otherwise be competitive.
The final message, which will come up in connection with the “lot of leaders” talking point, is that “it’s time to move on.” This message serves a few purposes. First, it is designed to downplay the seriousness of the January 6th coup attempt and the role some Republican officials and Trump played in encouraging and supporting it. This also supports their argument for voting down the independent commission to investigate the insurrection. Second, it is intended to deflect questions about whether Biden actually won the election. Maintaining that the election was fraudulent despite zero evidence of it is not winning Republicans support in potentially competitive districts.
Moore’s advantage here is that he did not have to run in a Republican primary. If he had a primary he might have had to use more hostile and disingenuous messaging that would have made him a much worse general election candidate. There is little reason to suspect an upset in this race. However, we need to look at the margin and the exit polling (if there is any). Since Democratic House candidates ran about two to three points behind Biden in November, we’ll want to see if Moore can come within 20 points of Stansbury today. However, for a number of reasons it’s probably more reasonable to think of a 15-point advantage for the Democrat. (Haaland won reelection in 2018 by 22.8 points, but that margin fell to 16 points in 2020. The last time the race for this seat was close was 2010.) If Moore closes the margin by a little, that might suggest that local issues played a larger role in the race. If he closes the margin to single digits, then that might suggest problems for Democrats.
I will be following/discussing the results on Facebook and Twitter.
Ohio Special Election
Tuchin Research (B/C; sample size of 600; MOE of 4.0) has released a poll in the Democratic primary for the special election for the 11th Congressional District in Ohio. It finds state Sen. Nina Turner in first place with 50% of the vote. Cuyahoga County Councilor Shontel Brown is in second place with 15% followed by a half dozen other candidates in single digits. Twenty-one percent are undecided.
More white voters than Black voters are undecided, but overall the share of white voters supporting each candidate is similar to the share of Black voters. For instance, Turner is winning 52% of the Black vote and 48% of the white vote. Brown gets 16% and 14% respectively. The difference tracks down to the minor candidates too. This results not only in fewer undecideds among Black voters (17%) but quite a bit more among whites (28%).
Turner, who became well known as national co-chair of Bernie Sanders’ 2020 presidential campaign, is a favorite of progressives and from this poll seems to be the front-runner. There is still time for this race to change. The primary is on August 3rd. The 11th CD is one of only three Democratic districts in the Buckeye State. The district has a partisan voter index of D+30. Biden and Clinton both won just shy of 80% of the vote in the past two presidential elections. This is a safe Democratic seat.
Democratic Primary for NYC Mayor: New Polling and Challenges with Ranked Choice Voting
Last Tuesday, the Boston-based Emerson College Polling released a survey finding Ms. Garcia leading the field, followed by Mr. Adams and Mr. Yang. That survey came just the day before a poll by Core Decision Analytics finding Ms. Garcia in the top three, trailing Mr. Yang and Mr. Adams.
In the past week, there have been two new polls that show increasing strength for the candidacy of Kathryn Garcia who until recently was barely registering in the polling. This is reminiscent of the 2013 election when current mayor Bill de Blasio was polling in single digits just weeks before the primary. Many credit a major union endorsement for de Blasio for changing that race. This time, it might be a major newspaper endorsement that changes the race.
A poll from Emerson (A-) showed a big gain for Kathryn Garcia, with her finishing in first place. The former head of two city agencies is in first with 21%, followed by Brooklyn Borough President Eric Adams at 20%, Andrew Yang at 16%, and Manhattan Borough President Scott Stringer at 12%. All other candidates are in single digits. Garcia recently was endorsed by the New York Times. It’s unclear what kind of impact that had on Garcia’s candidacy, but it sure does not look like it was a negative one. The last poll had her support at 8%.
The next day Core Decision Analytics (no rating) released a poll finding Garcia in third place behind Adams and Yang. These polls tell us a few things. First, Garcia is a serious candidate to win the primary. Second, so is Adams. Third, despite losing some support Yang is still in the running. Finally, undecideds and voters’ second and third choices look increasingly likely to make the difference in the outcome of the June 22nd Democratic primary.
New York is using ranked choice voting (RCV) for the Democratic primary this year. Considering the number of candidates, we can start to see the wisdom in the Alaska model in which the top four finishers regardless of party designation advance from the primary to the general election. Alaska’s primary is modeled on the jungle primary that California (and others) use, except that four rather than two candidates are put forward to the general election. Alaska is not using RCV for the primary, but it will be using it to determine the winner in the general election. This way, the electorate gets a clear choice of four candidates.
New York is using a more common version of RCV in which voters get to rank their top three choices from all candidates. Once all the voters are in, election officials count ballots in rounds in which bottom finishers are eliminated and their second and third choices are redistributed to remaining candidates until one person hits 50%. Unless there is a dramatic change in the polling (including a lot more polling), we are unlikely to have any idea who is going to win this primary until the results are announced. We can narrow it down to at least four, possibly six, candidates. But even that is a bit of guesswork.
It’s difficult for voters to make informed choices with so many candidates on the ballot and the resultant lack of clarity about which of them have realistic chances to win. This is my biggest concern about RCV – the lack of operational intelligence voters have about the relative strengths of the candidates. Proponents of RCV have often responded to this concern by claiming it is actually a feature rather than a bug, but that’s only true in the abstract. The argument is that because any candidate can get support from 100% of the electorate throughout voters ranking their choices, the incentive to play nice benefits the electorate and everyone can feel like their preferences mattered when the winning candidate got a majority of voters supporting her through multiple rounds of counting. Sounds nice, but it’s flawed in reality.
In a first-past-the-post system like ours, run-offs make some sense. In 1998, Somerville (MA) Mayor Mike Capuano won the Congressional seat held by the retiring Joe Kennedy II against more than a dozen other candidates. Capuano won the Democratic nomination with just 23% of the vote, largely on account of a big turnout in his home town. Being a safe Democratic district, Capuano’s nomination was tantamount to winning the election (which he did with 81% of the vote). Almost any of the other Democrats would have won the seat by a similar margin in the general election (which is no slight to Capuano, but in addition to the district being safe Democrat many of the candidates in the primary were well-known Democratic politicians).
RCV helps us avoid the case where someone who get less than a quarter of the primary vote advancing to the general election. But it has its own problems. In 2010, Oakland (CA) City Councilor Jean Quan beat several other candidates, including front-runner state Senator Don Perata, and was elected mayor. Three of the candidates, including Quan, ran a campaign designed in part to stop Perata from getting 50%. But the race was confusing, and it was never clear who realistically had a chance to win other than Perata. What ended up happening was a race between Perata and Not Perata. But there was no Not Perata on the ballot. Quan won after several rounds of elimination during the RCV counting process even though Perata won the most first place votes. And that’s okay, because the idea of RCV – or run-offs in general – is to ensure that the winner is supported to some degree by more than 50% of the electorate. However, the race was never between Perata and Quan and we don’t know that more people would have picked her over him were the choices made clear. In fact, while a lot of people were excited by Quan’s election (I was one of them)* many were not and felt like the result was strange since Quan was the first choice of less than a quarter of the voters and in the final tally she just barely beat Perata.
Is this a bad thing? I think it is and this is why. RCV has a tendency to obscure rather than promote voter choice. Sure, voters get choices, but the information they get is far more likely to confuse than to illuminate - especially when it comes to operationalizing their choices. Yes, it is true that in the traditional election model voters do not know what the outcome will be until all the votes are counted, but polling is better able to capture voter preference and communicate that to the electorate. Polling may have error in it – and it can miss projecting the actual winner – but it is really good at showing the relative strengths of the candidates. Pollsters are still trying to figure out how to survey RCV races. It failed spectacularly in the Maine Senate race last year. Perhaps pollsters will figure it out, but it still won’t solve another problem with RCV – which is the way campaigns communicate to voters.
The thinking with RCV is that since voters can choose a number of candidates and someone who cannot muster a majority or even a quarter of first place votes can still win the election, that campaigns have to appeal to all voters rather than just one party’s or candidate’s base. This should make elections more civil and give the electorate the satisfaction that the winners represent a majority. There was some (anecdotal) indications when San Francisco first implemented RCV that the campaigns were more civil, but it’s not clear that campaigns now are any different in terms of tone than previous ones were. In Oakland, the campaign turned into an anti-Perata campaign which was not only not “civil” (in the way I assume proponents of RCV mean that term) but it made the race one big negative campaign. After all, politics makes strange bedfellows.
In Quan’s case, the vote ended up being 50.9% to Perata’s 49.1% - a race that is so close that it would have helped voters for them to have known what their actual choices were. This is something RCV hides from voters when there are many candidates. That doesn’t mean Quan would not have beaten Perata in a traditional head-to-head run-off, but at least voters would have known what their choices were. By knowing those choices clearly, voters can make informed decisions about which candidate they prefer rather than guessing at how everyone else’s second and third choices will add up.
Generic Ballot
We are just about 18 months from next year’s mid-term general election. The generic ballot – which asks respondents if they would support a generic Democrat or generic Republican for Congress – has shown some reliability in the past at predicting the total national popular vote share for Democratic and Republican House candidates. It was less predictive in 2020.
In November, Democratic candidates for House races won more votes than Republicans by a margin of 3.1 points. Nevertheless, the GOP netted 12 seats. The Democrats maintained their majority in the House. It was close, but the margin of three points suggests a course correction for House Democrats regardless of how many seats might be gerrymandered to support Republican candidates.
On November 4, 2020, the final USC Dornsife (B/C) tracking poll found Democrats winning the generic ballot by eight points. The most recent generic ballot poll from Quinnipiac (A-) finds Dems +9. There has been a number of generic ballot polls in recent weeks, which have been all over the place (from R+1 to D+10). Quinnipiac is the best pollster that has recently put a generic ballot survey in the field. However, it was the poorer pollsters who came closest to matching the actual House popular vote margin in the weeks before the November election (some good pollsters had it as close as D+5).
It’s too early to tell what any of this means, but think of it as a baseline as we start to enter the 2022 campaign season. The baseline – when taking into account the various generic ballot polling of the past month – tells us that Democrats are probably ahead by about four or five points. If this develops into a trend and turnout echoes last November’s, we would expect the Democrats to hold the House. But even so, it will be close. I will try to update on the generic ballot every month or so until we get into the heart of the mid-term campaign next summer.
————
* Quan’s term was largely disappointing, except for her support for the arts. In some ways, her administration was a cautionary tale about electing activists with no prior administrative experience to an executive position like mayor.