How to build a winning democratic coalition

This comes from a mailing list of social-science informed advice on effective political action. You can find past posts online here, or subscribe to the list using this link.

Building a Winning Coalition: Summary

Debate over the future of the Democratic party is generally framed as the party having to choose between two strategies:

1. Win New Supporters: Advocate for policy positions that will win over more supporters. Generally this means “moving to the ideological middle,” where most independent voters are located.
2. Embrace the Base: Rather than trying to win new supporters, Democrats can focus on increasing the probability existing supporters will turnout by embracing more liberal policies that excite the base.

In this post, I review the social science research on this question, and conclude:

• if the Democratic party wants to be successful in achieving their policy goals, they have to reach out to Trump voters, especially those living in rural America. There’s some possibility that if they do not, they can still form a coalition of urban voters that will be successful in electing a president, but due to the spatial distribution of voters in the US (more on that below), the cost would be losing more and more races in Congress and State Legislatures. And that’s a recipe for perpetual gridlock, not progression policy.
• Contrary to this traditional “base or moderates” frame, I’m not convinced that this has to come at the cost of massive policy concessions. The data suggests that many of the voters who have left the party — voters who are predominantly rural, white, and lack a college education — did so because perceived Democrats to be biased towards urban populations in their rhetoric and policy. Voters felt ignored or, worse, mocked by the party.

For these reasons, I contend that the Democratic party cannot (and should not) give up on rural America. With changes to both tone and attitude, as well as some changes in policy priorities, Democrats can appeal to rural voters without making concessions that challenge its core principles.

Making this case requires diving into a lot of social science research, so I’m going to divide this analysis into four sections:

1. The centrality of “Rural Consciousness” in 2016
2. The political geography of America
3. What this means for the Democratic party
4. What this means for you

In the interest of not crushing anyone’s inbox, todays email will include only Sections 1 and 2, and I will followup soon with Sections 3 and 4. However, if you get to the end of this email and are dying to keep reading, you can find the last two sections online now at ScienceofPolitics.org

Section 1: The Centrality of “Rural Consciousness” in the 2016 Election

The story Americans are told about the US electorate is that there are red (middle America) states and blue (coastal) states. But that “fact” only emerges because we’re aggregating vote totals up to the level of states. If instead we look inside states, we find instead that it’s much more accurate to say we are a country of blue towns and red rural communities, and this division has been growing for decades.

(Side note: the phenomenon in which one gets a different view of the world when one looks at different levels of geographic aggregation is known as the Modifiable Areal Unit Problem (MAUP). Use that phrase at a cocktail party!)

This simple observation is critically important for understanding where the modern Republican party gets its support. Pundits point to Trump’s successes in states facing post-industrial decline like Ohio, Michigan, and Pennsylvania and argue, simply, that Trump supporters are laid off factory workers. But while it is true that many states with post-industrial decline voted Republican, towns and cities (the places where those factories used to be) are still voting Democratic.

Don’t believe me? Go take a look for yourself — you can find a great, interactive polling-station level map of vote shares here. I promise — it’ll change the way you think of the US. Go wander Pennsylvania, Ohio, Michigan. Everywhere you go, blue cities, blue towns, red rural America.

OK, so why do towns vote Democrat and rural communities vote Republican?

I will not pretend there is a single definitive answer to this, but the explanation I’ve found most compelling is that rural Americans identify strongly with their rural identity. Katherine Cramer has an outstanding book on the subject. Before anyone was talking about Donald Trump, Cramer spent five years in rural Wisconsin sitting in on coffee klatches and sewing circles doing exhaustive ethnographic research on the nature of the rural urban divide. Based on these years of interviews, she concluded that rural voters were driven by a sense of “rural consciousness,” which she defines as

an identity as a rural person that includes much more than an attachment to place. It includes a sense that decision makers routinely ignore rural places and fail to give rural communities their fair share of resources, as well as a sense that rural folks are fundamentally different from urbanites in terms of lifestyles, values, and work ethic. Rural consciousness signals an identification with rural people and rural places and denotes a multifaceted resentment against cities.

According to this perspective, this division is not about concrete policy differences. Instead Cramer argues that this division stems from a sense that (a) rural Americans are being ignored by policymakers and the media, (b) that the hardships of urban America are being prioritized over the hardships of rural America, and finally, to make matters worse, © urban elites are condescending, smug, and paternalistic towards rural Americans. (You can find an interview with Cramer here that summarizes her views if you don’t have time for her book, and someone making a similar point though more anecdotally here.

Sociologist Arlie Hochschild makes a similar argument based on 5 years interviewing people in rural Louisiana. She does not use the term rural consciousness, but she argues:

They feel their cultural beliefs are denigrated by the culture at large. They feel that they’re seen as rednecks, that they live in a region that’s being discredited. Many of them are deeply devout, but they see the culture at large becoming more secular. And then they see economically that this trapdoor that used to only affect black people and people one class below them is now opening and gobbling up them and their children too.

Note that Hochschild puts more emphasis on the idea that the people “jumping the line” are minorities and not just the people in cities — a point I’ll return to shortly. But in general, she also supports this idea that rural communities are angry in significant part because they feel like attention is being focused on urban communities, and that among the elite, they’re viewed as a joke.

(BTW, if your interested in the idea people vote on the basis of identity and not because they think one party is more likely to pass policies that will be personally beneficial, you may be interested in a new book by Achen and Bartels — very respected American politics scholars — that argues that most US voting is motivated by identity.)

In other words, this argument is about us versus them identity politics. But the identity division is largely between urban-sophisticates versus rural true Americans, not just white versus black.

Wait, you’re saying race wasn’t part of the 2016 election?

No. The most reasonable read of the data, in my view, is that racial politics played a substantial role in the 2016. In this, I’m particularly influenced by work by John Sides, Michael Tesler, and Lynn Vavreck (who just released a book on race in the 2016 race which I’m embarrassed to say I haven’t read yet, though I’ve seen many of its results presented). In short, they argue that prior to the Obama presidency, white voters without a college degree didn’t really see Democratic or Republicans as more inclined to offer support to blacks. But during the Obama presidency, this changed, and white voters without college degrees started seeing Democrats as being more supportive of pro-black policies (see the data here). They argue that this shift in perspective didn’t matter in 2012 because Romney never appealed to racial politics, but when 2016 came along, Trump spoke to these voters (data here).

But while race played an important role, I don’t think it was the only factor. I’m sure some voters voted for Trump because of racial animus, and I’m sure it was a factor for others. And yes, there’s concern that in some cases resentment towards urban elites is coded language for racial animus.

But I think it’s important to not conflate the racial animus and rural consciousness. There is substantial evidence, going back decades, that rural consciousness has been rising in America.

I find ethnographic studies like Cramer’s — work that largely finds that people in rural communities talk about politics in terms of their rural identity and not ethnic politics — compelling. Scholars like Cramer spend time earning the trust of the communities that they work with. Subjects in these studies do talk about racial politics with researchers — they are not merely coding their racism in rural solidarity. Instead, they talk about each as distinct phenomena. Moreover, it’s worth noting that, if current trends were only about racial animus, white towns should also be fleeing the Democratic party, and also that Obama (who, as you might recall, identifies as Black) did better among rural voters than Hillary.

It’s also worth noting that at least some of the evidence that racial animus drove some previously-democratic white voters to Trump is likely being driven by the increased salience of rural identity. Because people with less progressive views on racial policies also tend to be people who live in rural areas, increased rural-identity-motivated voting is likely contributing to an apparent uptick in racially-motivated voting.

So yes, in my view rural consciousness and racial animus are probably correlated, but they are not the same. And treating these as distinct phenomena matters because, if people are voting for reasons of rural identity, then changes in tone and attitude towards rural America may lead to changes in voting behavior.

To be clear, this is not the “consensus view” of political scientists. No such consensus view exists, and is unlikely to emerge. Voting is multi-factoral, meaning lots of things influence why people vote for a given candidate, and it’s very hard to pin down the exact size of the contribution of each factor because we don’t get to see people vote over and over in races that are almost the same but differ only by one specific factor. Instead, elections are relatively infrequent and rarely that similar to one another, both because the actual candidates running for office change and also because the context is always changing. We do try to answer these questions using lab experiments in which people are asked to pick between two hypothetical candidates, where the policy positions of the candidates are randomly varied. At the end of the day, however, subjects will always know they’re being observed, and these “paper candidates” are fundamentally different from real candidates in many ways. Because of this, we have to look at the evidence we have and try and deduce what more general pattern underlies the specific data we collect and can analyze. This is my interpretation of that data, but other reasoned interpretations exist.

Section 2: Political Geography

If you accept that diagnosis, then the choices are relatively clear: try to win back rural voters through a shift in attention, tone and attitude, or alternatively embrace the “coalition of the ascendent” and really emphasize progressive policies — don’t try to lessen the inter-party rift, run with it to drive turnout.

In choosing between these options, there’s one other very important consideration: electoral geography. And as it turns out, the electoral geography of the US dictates that the more Democratic support is concentrated in cities, the fewer seats Democrats will win in legislatures at both the state and national level. Not because of gerrymandering, just because of the nature of electoral geography.

Most elected officials in the United States (like members of the House of Representatives and most state legislatures) are elected by winning a plurality of votes from a geographically-defined district. To facilitate this, we cut up our states into electoral districts, and elect one representative from each district.

How these district boundaries are drawn can have significant consequences for political representation. Consider a state with 50 Democratic voters, 50 Republican voters, and 4 districts, each of which must contain the same number of voters. Given the overall distribution of voters, we’d probably expect 2 seats to go to Democrats and 2 to go to Republicans. But if one can draw the lines any way one wants, one can potentially create one district with 25 Democrats, and two districts with 17 Republicans and 8 Democrats and one district with 16 Republicans and 9 Democrats. In that configuration, Republicans get 3 seats, and Democrats get 1.

When this is done strategically, it’s referred to a gerrymandering. But here’s the thing: even if a computer were creating district boundaries with sole aim of making reasonable looking districts (i.e. make them look like compact squares or triangles, not ugly things with tentacles reaching out everywhere) of equal size, Democrats in the US will end up with fewer seats then you’d expect.

The reason is that Democrats are much more likely to live near other Democrats, and so when one draws boundaries without thinking about politics, one tends to get a few urban districts with lots of Democrats (far more than the 50% the Democrats would need to win the district). And when you have more of your supporters in a district than you need to win that district (i.e. more than 50%), those extra supporters are, in effect, wasting their votes.

Republicans, by contrast, are much more evenly distributed, and as a result, tend to be in more districts where they represent closer to 51% of voters. You can show this through simulations, and (in some of my own work with Jonathan Rodden) you can show it just by comparing the difference in the spatial density of Democrats and republicans with representation in legislatures.

As a result, if Democrats decide to forsake rural constituencies and concentrate in cities, they may win the presidency, and other “state-wide” offices like governorships and Senate seats. But they will fall father and father behind in the US House of Representatives and in most state legislatures. And that, at best, is a recipe for gridlock.

And that’s the last piece of why I think the Democratic party has find ways to reach out to rural voters. First, because I think they’re probably right that urban communities tend to ignore and condescend to much of rural America. Second because I think that there are inroads to be made by changing tone and attitude without making massive policy concessions (though to be clear, some will also likely be necessary). And third because if we don’t, I think we’re facing down the barrel of ever more gridlock and disfunction as a country.

Section 3: What the means for the Democratic party

Up till now, one could fairly argue I’ve presented a relatively rosy picture of the choice faced by Democrats — all the party has to do is change its tone and campaign style, not its policies (I think changing how one things and talks about a different group is anything but easy, but fair enough). So what do we do when there are real policy differences?

I think one inherent tension will always be policies that disproportionately help racial minorities. As noted by Sociologist Arlie Hochschild, rural Republicans often feel like urban minority communities are jumping ahead in line ahead of Americans. Here, Democratic support for groups like Black Lives Matter are likely to always have a cost in terms of support in rural communities. So how can we address this?

I absolutely do not feel Democrats should fall back from supporting marginalized communities and causes like Black Lives Matter — to do so would mean abandoning what the party stands for. But what Democrats can do is try and pair their support for these causes with attention to causes that matter to rural Americans. Rural communities aren’t mad these communities receive attention per se, they’re mad because they feel that this attention comes at their expense, and that is something policymakers can address.

Like it or not, Democrats largely have abandoned rural America, at least in spirit — reporting strongly suggests Hillary Clinton’s campaign, for example, made a deliberate decision to not spend energy trying to win over rural voters and to focus instead on urban Democrats (you can find a lot on that here, and smaller snippets here and here). Oddly, democratic policies actually are pretty beneficial to rural America. But when those actions aren’t matched by rhetoric and actual time spent campaigning in rural communities, the end result is that rural voters think they’ve been forgotten by the Democratic party, even if they’re being helped by their policies in many ways (see, among many others, here, here, here).

But are there issues we may just have to de-prioritize? Yes, probably. Gun control is probably the clearest example of a position that’s not particularly tenable. I’m personally in favor of gun control, but the reality is that (a) after 30 years of pushing for gun control, Democrats have made little to no progress, (b) the Supreme Court recently discovered an individual right to gun ownership, meaning any legislative progress may be futile int he face of a Supreme Court that’s likely to be pretty conservative for some time, and © it’s something the rural constituency cares about passionately. So if letting go of the issues opens the door for other more likely liberal policies — a carbon tax, sentencing reform, progressive taxation, improvements of Obamacare, immigration reform, etc. — I think that’s something we may have to accept.

Section 4: What this means for you

If you’re not currently running for office (and maybe you should be?!), then you may now be saying “Ok… but what does this mean for me?”. Here are my take-aways for average citizens:

• Stand on principle and avoid ad hominem attacks: It’s easy to slide into attacking Trump and the Republican party on petty grounds. It feels good to call people on the other side idiots, to compare Trump to Hitler, or to insult his hair color. But avoid the cheap jokes. Make your case on the basis of the principles at stakes — protection of marginalized groups; equality; equal rights for all groups; universal health insurance — not by attacking the people on the other side. This serves two purposes:
• If you insult the other group, it just reinforces group distinctions and makes it more socially difficult to shift ones support.
• If you want to connect across “group” lines, the way to do it is find common ground, and that common ground will often be found on common values and principles like caring for your community, or respect.
• Be generous, especially to Trump voters: Try to give Trump supporters the benefit of the doubt. Don’t jump to conclusions about why they support Trump. Some Trump supporters are white supremacists, and they should be called out as such; but most Trump supporters are not. Until you know why the person you just met voted for Trump, give them the benefit of the doubt. And next time you have a chance, ask a Trump supporter why they voted how they did and listen, rather than debate.
• Avoid questioning the legitimacy of Trump’s win: I am very displeased Trump won, but he did. I think chants like “Not my president” at rallies suggest that you don’t accept the legitimacy of the votes of your fellow citizens, which is just insulting. Similarly, leaning too hard on the idea that Trump supporters were “duped” (by Russia, by Comey, by anything) just alienates people we want to bring into the party.