'Democracy moneyball': How to beat Mitch McConnell at his own rigged game
There's no question that democracy itself is on the ballot in 2020, as Barack Obama argued in his Democratic convention speech. But it's not just a matter of getting rid of Donald Trump or appealing to voters "to embrace your own responsibility as citizens – to make sure that the basic tenets of our democracy endure." Those basic tenets are themselves inadequate. Trump did not come out of nowhere. He was the result of prolonged democratic dysfunction. If that underlying dysfunction isn't dealt with, an even more destructive Trump-like figure js virtually inevitable in the near future.
One glaring facet of that dysfunction is the centuries-old practice of gerrymandering, which was brought to new heights by Republicans in the 2010 cycle, as former Salon editor David Daley described in "Ratf**ked: Why Your Vote Doesn't Count" (Salon author interview here). Indeed, the Election Integrity Project's report on the 2018 elections stated that its panel of 574 experts judged issues involving district boundaries as "the most problematic issues of electoral integrity in America."
One leader in combating that dysfunction is the Princeton Gerrymandering Project, which is "a gerrymandering team the way that cancer researchers are against cancer," in the words of founder Sam Wang, a Princeton neuroscientist who also founded the Princeton Election Consortium. Wang was speaking at a recent town hall called "Redistricting Moneyball 2020."
As with the original "moneyball" concept, made famous by Michael Lewis' book, the idea is to use smart statistics to identify undervalued prospects as a way of leveraging the power of small donors.
"The original Moneyball concept can be easily translated to elections," PGP team-member Connor Moffatt explained. "Each baseball team has a limited budget just as each citizen has a fixed amount of time and resources they are willing to invest in politics. Baseball teams are trying to win games whereas in elections political parties are trying to gain governing power."
Next comes an intriguing parallel: "Baseball has superstars that are overvalued and politics has races that are high-profile and overvalued." The most obvious example is the Senate race in Kentucky between Majority Leader Mitch McConnell and Democrat Amy McGrath, which as Moffatt says, is attracting more financial resources "than the leverage would warrant."
What does he mean by that? Kentucky is a solid red state and most polls have McGrath well behind. Democrats are pouring money into the race because they dislike McConnell so intensely — but that's not a smart use of resources. "If McGrath wins," says Moffatt, "it's likely that the national save for Democrats is so large they'll have already flipped the Senate. So the best way for either party to influence whether Mitch McConnell is Senate Majority Leader in 2021 is to invest in seats closer to the majority-control tipping point."
Finally, and most importantly, is the question of "undervalued gems," which in baseball means players whose true value is not obvious. Those exist in politics too, Moffatt said. "There are Texas State House races that have the potential to influence multiple congressional districts over the next decade."
What's more, those races overlap considerably with the "Texas Nine" congressional districts that political scientist Rachel Bitecofer has highlighted as prime targets for flipping in her 2020 forecast. That's not even considering the psychological and strategic impact of accelerating the shift of Texas from solid Republican territory to a purple or even a blue state. Texas, in short, is loaded with undervalued gems in this election cycle.
Of course, the McConnell/McGrath example brings something else to the fore — the passionate, irrational side of politics that can never be eliminated, but only complicated or offset. Visceral hatred for a figure like McConnell — who literally stole a Supreme Court seat for conservatives and has repeatedly boasted of his anti-democratic intransigence — is always likely to motivate some people. All the more reason, then, for those with cooler heads to seek a more rational approach. The moneyball approach also transforms poll-tracking from a passive and potentially demobilizing activity into its exact opposite: An active way to engage in the electoral process, drawing on a key metric of electoral moneyball: voter power.
"Voter power is defined as the ability of one or a small group of voters to move the probability of one of three important questions," Wang explained in an early August PEC post. These questions are:
- President: providing tipping-point support to get Biden or Trump above 270 EV.
- U.S. Senate: providing tipping-point support to get 50 Democratic-plus-independent seats.
- State legislatures: providing tipping-point support to gain control of one chamber of the state legislature, in states where this would lead to bipartisan control over redistricting.
"In all cases," Wang wrote, "voter power tells you where to put your efforts as a citizen. It's what we are calling the 'Moneyball 2020' approach."
For example, at the time I'm writing this, the U.S. Senate race in Montana, between incumbent Republican Steve Daines and Democratic Gov. Steve Bullock, has maximum voter power: It's a close race in a small state, and if Bullock wins that's likely to mean Democrats win control of the Senate. By convention, the race with the most voter power is set to 100, with all others calibrated against it. You can see the full list of Senate races, margins and voter power here.
The situation with "redistricting moneyball" is quite a bit more complicated. In this case, as Wang told me afterwards, "It depends on the closeness of chamber control, how many voters there are per district, and how close the tipping-point districts are." But before we explain that any further, it's helpful to recall why this is even necessary in the first place.
Wang came to focus on gerrymandering by accident — through his own mistake, actually. A neuroscientist by day, he'd been analyzing polls since 2004, blogging at the Princeton Election Consortium, but in 2012 he messed up. "I made a prediction error," he explained in the town hall. "I correctly said that the House was in play. I correctly said that Democrats were likely to win more votes at a national level and then I made an incorrect prediction … that therefore Democrats were likely to take control of the House."
His blog readers "are really smart," Wang noted — "string theorists, economists, financial traders, social scientists" — and they pushed back."They said, 'No, you're crazy. You haven't paid attention to the redistricting that happened in 2010.' They were right and I was wrong." That led Wang to write a New York Times op-ed, "The Great Gerrymander of 2012," and from that, the project was born.
"At the time I thought, 'This is terrible! What are we going to do about it?'" But, he recalled, an early member of the gerrymandering group named Brian Remlinger told him, "Sam it's really time for us to start working on individual states. It's time for us to discover our inner federalist," which led to a third New York Times op-ed, "If the Supreme Court Won't Prevent Gerrymandering, Who Will?" in which Wang discussed a variety of available approaches. ("You can tell that I'm a good academic because I can turn these bad moments into publications," he joked.) Summing up, he concluded, "Putting all federalist routes together — courts, voter initiatives, laws and elections — I estimate that reform is actually possible in the vast majority of states, even without the Supreme Court's help."
And so the most recent evolution of the project was born. Wang described a number of different paths, represented in a color-coded map of the states — some with independent citizens' redistricting commissions, some with advisory commissions, some (like Virginia) where redistricting reform is on the ballot, and others where court challenges are possible via racial gerrymandering arguments.
All of this Wang referred to as "wearing our white hats … our nerd outfits." But in other cases, especially where there is divided government, he has called for "a brute force raw political power approach to getting fairness" by getting "as many people as possible to put both political parties at the table" and let them fight it out, whether by donating to races or by volunteering to get out the vote, and so on. Wang's team identified six states, where this "raw power" approach might work: Texas, Florida, North Carolina, Minnesota, Connecticut and Kansas.
So this is where "redistricting moneyball" comes in. It's inherently more complicated than U.S. Senate moneyball for at least three reasons. First, there are a lot more seats that could at least possibly be in play, and second, there's a lot less statistical information available about the much smaller electorates involved. In the town hall, Moffat and another PGP staffer, Jacob Wachspress, explained how a model was constructed to compensate for those problems.
In that first op-ed, Wang offered two proposals to fix the gerrymandering problem. First, "nonpartisan redistricting commissions in all 50 states," and second, adopting "a statistically robust judicial standard" for identifying partisan gerrymandering. At the time, Wang noted, Justice Anthony Kennedy was on record as being open to that idea, "if a clear standard could be established."
A little over two years later, Wang published a law review article, "Three Tests for Practical Evaluation of Partisan Gerrymandering," and a second op-ed, "Let Math Save Our Democracy," explaining how one such standard worked, and arguing that statistical tests in general could be used in a number of different ways to weed out meritless cases or to supplement other mandates such as geographical compactness, or to help balance them with other requirements, such as compliance with the Voting Rights Act.
Others advanced other test methods as well, and Wang co-authored an amicus brief discussing and analyzing some of them in the North Carolina gerrymandering case that eventually came to the Supreme Court, only to be rejected in June of 2019.
"This was clearly a bitter disappointment," Wang said in the town hall.
They began by creating a model to predict the margins of victory in state legislative races — ranging from toss-up (0 points) to safe (>20 points) and uncontested. That data was derived from five factors: statewide election results in each district, incumbent popularity, quality of challenger, campaign finance reports and demographic trends.
They supplemented that model with an "internal foundations model" using four components to predict winning margins: presidential election results within the district, the incumbent's past performance, 2016 statewide presidential election results and 2020 statewide presidential election forecasts. This internal model has improved the accuracy of the estimates, but modeling individual races — even thousands of them — is only the first step.
Next, Wachspress explained, they modeled the uncertainty around those margins "to find the chance that both parties will have a say in the redistricting process." There were three sources of uncertainty: the individual district race level, statewide shifts between parties and density-specific shifts. For example, "there's significantly more uncertainty in the ratings based on whether districts are in rural, suburban or urban areas," he noted. That model produced, for example, this "heat map" for North Carolina, a state where the governor plays no role in redistricting.
The most likely outcome is Republican control of both chambers (lower left quadrant), and thus another egregious partisan gerrymander. (It was the previous North Carolina GOP gerrymander that the Supreme Court refused to consider.) The most likely remedy is Democratic control of the House (lower right quadrant). So that's where the project's attention is focused for North Carolina. Here, raw voter power is the amount that adding one more Democratic vote in a given House race would impact the probability of bipartisan control — and thereby a more balanced redistricting process.
In addition to all the other variables, Wang told me that voter power "depends most of all on how close a chamber is to the edge of control." In that respect, he sees state legislatures in Texas, Minnesota and Kansas as being "right on the edge."
That shows up in the voter power scores themselves. Texas has the most congressional districts, so it's not surprising that it's got the highest voter power ratings. It has 11 State House districts with voting power of 70 or more, compared with just two that are 50-plus in Minnesota, and three in Kansas. There's a lot more at stake in Texas, simply because of how much more mischief can be done in redistricting the Lone Star State to favor Republicans at a time when its electorate is clearly shifting. The most valuable legislative seat in the country, in terms of voter power is the 112th Texas House district, just northeast of Dallas, with a voter power rating of 100, followed by two more suburban districts, the 26th (outside Houston) and the 66th (north of Dallas) with ratings of 93 and 91.
The Democratic candidate for the 112th district, Brandy Chambers, was a participant in the PGP town hall. So, naturally I was interested in her district and her race. She came within two points of winning in 2018, and is running again to unseat five-term incumbent Angie Chen Button. Chambers' website accused the GOP legislature of being "focused on fringe issues," and I asked her to elaborate.
"When Texas was suffering a CPS [Child Protective Services] crisis and public school funding was drowning, the leadership in Austin prioritized bathroom bills and laws intimidating immigrants," she replied. "I've talked to people all over our district and not one person said, 'Please more tax breaks for yacht owners and no to lifesaving and billion dollar saving Medicaid expansion.' No voters asked for more than a dozen pieces of anti-LGBTQ legislation."
"My opponent votes against local control, neighborhood schools, property tax reform, public safety, wage-theft prevention, and private family planning," Chambers said. "She works for special interest donors. I will work for the people."
While some specifics may vary ("wage theft prevention" is important, but probably made Chambers' list because she's an employment law attorney), those broad themes surely echo in thousands of other races nationwide, along with other issues her website highlights, including climate change and criminal justice reform. She also includes fair redistricting maps ("Politicians shouldn't pick their voters"), which as an issue remains more muddled.
As Wachspress noted in his presentation, both parties have "announced campaigns to target close state legislative chambers in key redistricting states" and claim that fair redistricting was their concern. But their "choices of targets plainly show that their goals are not so noble": Democratic targets include Pennsylvania, for example, while Republican targets include Texas. In each case, their own party's governor — who is not up for election in 2020 — holds veto power over the process, meaning that only one party is in position to potentially draw unfair maps.
"If partisans are the only people investing in these races, then there's just as much organization on the pro-gerrymandering front as on the anti-gerrymandering front," Wachspress said. "In order to tip the scales toward fair redistricting, people have to be willing to organize for both parties depending on the political landscape of a given state."
That may sound overly noble in these deeply divided times — and it's not strictly true. PGP, is explicitly nonpartisan and devoted to fighting gerrymandering by either party. That makes sense: A majority of all voters, across party lines, oppose partisan gerrymandering. So PGP enables the ordinary citizen to take a proactive role, but doesn't tell you what to do — it simply makes the potential impacts clearer and more readily comparable. But since Republicans have been far more organized in their partisan gerrymandering, there are more available targets, with higher stakes, for Democratic activists to focus on.
There are other groups out there offering similar guidance, such as the Future Now Fund, which provides a more broad-brush approach, multi-issue framework for working to flip or defend state legislatures. Or Swing Left with its Super State strategy (including a focus on redistricting), which I wrote about as part of my 2020 preview in early 2019. There are also state-level efforts, such as Flip the Texas House. You can use PGP's "redistricting moneyball, in combination with any or all of these other guides, or all on its own.
The point is, you have options and you can find information. You can decide what battles matter most to you, and prioritize them on your terms, based on much better evidence and better data than hand-waving promises from politicians and misleading media narratives. And we can all count on the Princeton Gerrymandering Project to keep on developing new tools to help citizens fight for fair representation in the next big wave of redistricting battles that's just ahead.