Democrats need to do better with white voters, part two


UPDATE: a commenter came up with another variable

Last December I wrote about, counter-intuitive as it might sound given what Republicans are going through, why Democrats need to win a bigger share of the white vote. This follow-up is getting into how to win more white votes, but given that a month has passed since the first post, here’s the short version of why:

  • Considering presidential elections, Obama won only narrowly, and improving on Obama’s 39% of the white vote by even a few percent provides a margin of error in case Republicans have some success reaching out to what we might lump together as Democratic-leaning demographic groups (DLDGs). Succeeding at such an improvement in the popular vote will probably have some effect on the electoral college too.
  • We’re losing the majority of congressional and legislative districts despite winning the popular majority (that’s the case in the US House, and though I don’t know about state legislative races collectively, it’s true in enough states that I feel safe saying that). Our huge lead among DLDGs isn’t translating into seats. If we can’t improve among whites, we can’t win the US House, nor most legislative chambers.

Why do I feel so sure we can do this, improving on Obama’s 39% a few points to, say, 43%? In 2008, Obama’s share of the white vote was — 43%. That would seem to be definitive proof we can do this.

Since figuring out how to win more of the white vote is an overambitious undertaking for a blog post, let’s just start with identifying the variables of our task. At least then we can start with some sense of the scope of the problem and where to start actually working on it, and the first variable that comes to mind is the exceptions.

Exceptions: OK, maybe it seems obvious there will be always be exceptions to any trend, problem, guideline, whatever, that we come up with, but sometimes we need to stop and look at exceptions to see if they really are unique, or if they tell us something. For example, in the first post, I included this graph showing how Obama won heavily in the most densely populated precincts, but Romney won where the population is sparse:

Source: Ezra Klein’s Wonkblog, “2012 The Year in Graphs”

One of the whitest and sparsest states is Vermont — where Democrats win easily.

Did Vermont Democrats get lucky, or do something pretty smart, or some combination? I don’t know nearly enough to speculate, but I do know Vermont proves that areas that are white and sparse can be won, and I know it’s worth looking into how it happens there.

For an example I know better, Northeast Minnesota is also white and sparse, and has long been a Democratic stronghold. The area isn’t as agricultural as other rural areas of the state, and the workforce is heavily unionized. Could that be the difference? We know union members vote about two-thirds Democratic. I’ve never seen union members broken out by demographics, so I’m going to make an educated guess that white union members vote more Democratic than white non-union members. Does that suggest anything to us? A couple things. One, white union members might be a place to start looking for white votes. Two, organizing more workers might lead to more of them voting Democratic.

Union membership: So there’s a variable to consider, union membership. I’m sure every good Democrat is thinking organizing more workers is screamingly obvious and not exactly an insight. Sometimes the obvious still needs to be checked in case there’s something we overlooked, like just how many white union members are voting Republican? Why are the one-third of union members who vote Republican voting for a party that wants them to stop existing? Organizing is tough and long term, but short-term, is there someplace where we’re leaving votes on the metaphorical table?

Density: Next, let’s get back to density and see if that graph above tells us anything. Asking whether density causes people to vote a certain way, or people who vote a certain way prefer the attributes of particular densities, or people are self-selecting somehow, got complicated. In fact, this section got so long as I was writing it that I cut it out and started a separate post, hopefully to be completed in the near future.

Even struggling for brevity, it seems worth pointing out that the graph matches the results we see in congressional and legislative elections. Democratic districts look like little blue dots on a red map almost right across the country. That, besides how many pixels it’s taking me to say what I have to say, is why I’m going to try to give density its own post. For now, I’ll just say that whether population density causes people to hold certain political views, whether choosing political preferences and choosing where to live happen to come from the same personality traits but don’t cause each other, or whether some sort of self-selection is going on, makes a huge differences in the next questions we ask and the approaches we take. I hate giving a conclusion with little about how I got there, but nonetheless, I’m inclined to think aspects of living in different densities lead people to certain political tendencies. Whites in urban areas should be the most winnable, but also most likely already won, and they help only to win statewide — not that winning statewide is pointless. Ask anyone living under a GOP ideologue governor elected in 2010. But we have to crack the problem of winning more sparsely populated areas to win majorities of seats, and that graph might suggest looking at districts around 800 people per square mile.

Region: By “region” I mean states, not urban, suburban, and rural differences. Obama’s 39% of the white vote was just 10% in some states, but he won the white vote in other states, and even the white male vote in a few. So the problem is not uniform across the country, and the course of action is not obvious. Does getting only 10% in some states mean that’s where we have to go for more votes? Are the states where we’re doing the best also the states where potential voters are most persuadable and that’s the opportunity? It strikes me when looking at the whole country that we need a balance of both where the opportunity is greatest, and where the strategic need is greatest. On the one hand, referring again to Vermont, Democrats already win districts, so maybe that’s the best place to find persuadable white voters, but the impact is low. On the other hand, there’s Texas.

Texas isn’t just the second biggest prize of electoral college votes and it isn’t just safely Republican; it’s a must-win for Republicans. Without it, they have no realistic path to the presidency (barring finding some way to muck around with the electoral college — not to give them ideas, but I’m amazed they haven’t suggested just letting the Republican state legislatures assign their states’ votes to the Republican candidate regardless of popular vote, which would be perfectly constitutional). They also get two safe US Senate seats without which their hopes of taking the Senate grow dim, and the gerrymandering of congressional districts is much of their House majority. The latter absolutely requires they hold the governor and state legislature. If Democrats could make Texas competitive, that would be a national gamechanger. Republicans would be forced to pour resources into Texas instead of using it as a piggy bank even if Democrats merely gave them a good scare. Forget about waiting for demographic changes to just happen, taking the risk a big enough percentage of DLDGs will vote Republican to keep the state red. Trying to win over enough current voters might make Texas purple an election or two sooner. It’s an uphill climb certainly, but strategically, it has the biggest payoff.

But should more winnable situations with much less payoff be ignored if we’re in an either/or situation? Well, good thing we’re supposed to be the ones who cope better with nuance.

Religion: The takeover of the Republican Party by Christian conservatives has set up something of a feedback loop. Being Republican became identified as part of being Christian, and being Christian is part of being Republican. Liberal Christians might not agree with the need to be Republican and conservative to be Christian, but conservative Christians and Republicans seem to think those things go together, and they get to decide who joins the GOP. Non-Christians and liberal Christians are among the DLDGs, even if by virtue of being pushed out of the Republican Party rather than choosing to join the Democrats. Well, so be it. The fact Christians are still a majority of all Americans, and a majority of whites, suggests that we have to win over more white Christians, even if they’re shrinking as a portion of the population. People with no religious affiliation vote Democratic already, but they’re the fastest growing group, so there might still be plenty of opportunity there. If it turns out they’re concentrated in heavily Democratic districts already, then we still have no alternative to trying to win more Christian votes, but one thing needs to be made explicitly clear: I do not mean to imply Democrats have to start making religious appeals for votes, but rather, trying to win the votes of voters who happen to be Christians. I wouldn’t for a moment suggest that non-Christians pretend to be Christians, but I am suggesting the non-Christian vote isn’t enough in most districts, and where it is enough, those are usually safely blue districts already.

Gender: Since white women are more likely to vote Democratic than white men, maybe the place to look for votes is obvious. Keep in mind though that the gender gap doesn’t just mean women lean Democratic. It means men lean Republican. Sometimes, the gender gap has helped Republicans. So do we go for women’s votes since that’s been a better place to look so far, or should we seek a way to close up the gap on the male side?

Occupations: People in some occupations are among the DLDGs even when they’re mostly white, like scientists, government workers, teachers, and artists, and I presume readers have already noticed something they all have in common: conservatives hate them. I suggest they’re Democrats more because Republicans pushed them away than because they suddenly turned liberal, though I suppose being hated by conservatives would lead you to consider whether liberals are really so bad. Blue collar whites used to be a strongly Democratic group, and the fact they aren’t is still hard for Democrats to wrap our heads around. If we could figure out how to win back blue collar whites, that’s a lot of the GOP base gone. Doing so is clearly a big topic, so for now I’ll just admit that what we’re doing isn’t working well enough.

Age:The Democratic lead among younger voters is partly due to younger voters being less white than older voters. Among younger whites, we’re still losing, just by less than among older whites. Then again, we need just a bigger minority of whites, not a majority, so that still works for us. On the other hand, do we need to give up on older whites? For example, the photo ID for voting constitutional amendment in Minnesota failed as badly among senior citizens as other age groups. This was an amendment rather than candidate, but still, among older white voters, the Democratic position won (our older voters are so heavily white that I feel safe saying that). So it can be done.

Income: I haven’t seen a cross-tabs of different income groups and race or ethnicity. As I started thinking about differences in income, I realized I’ve got nothing but guesses and presumptions. It seems worth asking though if rich whites vote differently than middle class or poor whites. Does each income group vote the same way regardless of where they live, and what other groups they fall into? If I had to guess, I would guess income isn’t predictive of much of anything. I don’t like guessing though, so best to look.

Multiple variables: A white male is actually a member of two demographic groups. A suburban white male is in three groups. A young suburban white male … it’s probably obvious where I’m going, and probably the point is obvious. Just remember that when you’re trying to step back and rethink what you’re doing, you have to check the obvious too just in case you’ve actually been forgetting it. So multiple variables is a complication without a magic formula. The message or policy that we hope wins over the young suburban white male might have a strong appeal to suburbanites, but then young people hate it. I have no solution to offer except to be sensible that this is what we’re dealing with. In fact, rather than assuming undecided voters are oblivious people who somehow missed an election campaign that dominated the news for year, they could be undecided because the suburban part of their identity likes what we have to say, but the young part prefers the Republican position. In other words, an individual could be experiencing conflicting values or conflicting interests, even while being informed, and thus gets to election day unable to decide. That might explain why the Obama campaign broke down their targets for TV ads to such discrete units.

Where you live: How is where each of us lives a variable in winning the white vote? While we’re discussing the macro view, and we can try to encourage those in a position to actually make strategic decisions to do one thing or the other, at a local level, we can do more than just share a link on Facebook or recommend a diary on Kos. We can become the person making decisions. Sometimes the trick is jut showing up. Even when you’re looking just at your own time, we have our biggest impact at home. By time I mean we can travel to volunteer elsewhere, but you can only help with a campaign that way. You can’t stay and build something like you can at home. So I’m hardly going to blame someone who devotes their time to the safe district that helps not at all to win swing districts, if the safe district is where they live. When we try to influence the people who run statewide campaigns or statewide parties, we’re thinking at a large level. When you have three hours free next Thursday, you help at home. And that’s fine. Nobody knocking doors in their own district should be made to feel like they’re wasting their time, and you’re certainly not wasting time building a local party or advocacy group where you live and therefore can keep at it. At the same time, our overall topic is winning more white voters so we can win more districts in a situation where Democrats are packed and maybe even gerrymandered. So to the extent each person has flexibility, just consider that. Maybe you live in a safe district but just over the line from a swing district. That might allow you to not just show up once or twice, but to be a volunteer who can be around for years.

It also shows the importance of local party building. If local parties are built in safe districts only, that doesn’t help. It means in every other district, each campaign requires reinventing the wheel. That’s short term thinking. If you do happen to be living in a red or purple district, I encourage you to focus on building the local party, hopefully building up an experienced group and broadening the base of volunteers. Be the 50-state strategy. So I guess as well as saying think local, I’m also saying think long term.


Thanks to a commenter on the cross-post on Daily Kos pointing out another variable, which is media. Maybe it’s actually an aspect of density, but media availability might actually be a variable by itself. Urban and suburban areas usually have a bunch of local TV stations, full radio dials, ready access to broadband, and large dailies and alternative weekly newspapers, while some parts of the country still have no broadband, and dial-up works fro not much of the modern internet. Many cable systems don’t carry MSNBC let alone Current/Al Jazeera, but they all carry Fox News. There are few local TV channels in many places, and even few radio stations, and nothing liberal. Even some big cities have no liberal radio, not the terrestrial radio is as important as it used to be, but it’s far from nothing … and everybody gets conservative radio. Good luck getting alternative weeklies much outside the urban core. So in short, people can vary widely in terms of the available media.