I was correct in predicting (along with Nate Silver) that most of the post-mortems on the special election in Georgia’s 6th District would be “dumb.” Specifically, I said that the fact that election was going to be close was the most important thing we needed to know and that discovering who actually won wouldn’t add much value.

Now, I’ve been arguing strenuously for some time that the Democrats are at risk of making a big miscalculation if they put all their hopes in consolidating their gains in affluent well-educated suburbs and do not address their weaknesses in rural and small town areas. So, on the surface you might think that I’d shake my head in agreement with this tweet:

Our own David Atkins made the same observation a bit more expansively in a piece last night in which he argued that “There is no Democrat so seemingly non-partisan that Romney Republicans will be tempted to cross the aisle in enough numbers to make a difference.”

But this takes my argument too far.

When I did my analysis of the presidential election results in Pennsylvania, I noted that Hillary Clinton had done much better than Barack Obama had in the Philadelphia suburbs. In my own Chester County, which is an affluent, highly-educated, traditionally Republican area, Clinton won by more than 25,000 votes. By contrast, Obama won Chester County against McCain and lost it to Romney, both times by less than a thousand votes.

Most of those 25,000 votes came from “Romney Republicans.” Something similar happened in Georgia’s 6th District.

Year Office Results
2000 President George W. Bush 68% – Al Gore 32%
2004 President George W. Bush 70% – John Kerry 29%
2008 President John McCain 62% – Barack Obama 37%
2012 President Mitt Romney 61% – Barack Obama 38%
2016 President Donald Trump 48% – Hillary Clinton 47%

Of course, we can get bogged down in defining what we mean by our terms. What exactly distinguishes a “Romney Republican” from a Trump voter? I think people are trying to create a distinction between working class Democrats who defected to Trump and more affluent and professional Republicans who defected from Romney. But that’s not a dichotomy that quite works in this case because the theory is that Romney Republicans won’t defect.

Let’s try to be clear about what we mean. Hillary Clinton won a lot of votes in the suburbs from people who had voted for John McCain and Mitt Romney. She lost even more votes from folks in small towns and rural areas who had voted for Barack Obama.

So, if I understand what Jeet Heer and David Atkins are saying, it’s basically that the Democrats can’t make much more progress in the suburbs than they’ve already made and that the easier task is to win back Democrats that they’ve recently lost. Either that, or they’re just wrong about how likely Romney Republicans are/were to defect.

I don’t have a strong opinion on which would be the easier task. But I do know that so far this trade has not favored the Democrats. The left’s votes are already too concentrated and I can make this point clear fairly easily.

When suburban Chester County was voting 50-50 in the presidential elections of 2008 and 2012, it was possible for the Democrats to also win down ballot seats. And the Democrats have succeeded in electing representatives from Chester County to the state legislature. Gaining 25,000 votes at the top of the ticket helps, but the area is still competitive. But in many other counties in Pennsylvania, the Democrats went from winning 50 percent or 40 percent to winning only 30 percent or 20 percent. The result is that many more legislative seats became so lopsidedly red that downticket Democrats no longer have a fighting chance.

In this sense, not all votes are equal. It’s more valuable for the Democrats to add a voter in a rural area than one in a competitive suburb, and rural votes are definitely of more use than added votes in seats where Democrats are already winning by comfortable margins.

What distinguishes Georgia 6th District is actually that so many Romney Republicans have already defected that it’s become a competitive seat. It’s definitely worth contesting for exactly that reason. Perhaps we’ve reached the limit on how many defections there can be, but that’s uncertain.

My concern is that even if the Democrats start winning at least some seats in places like Georgia’s 6th District, this is part of an overall shift in the electorate that doesn’t seem to favor the left. It would work somewhat well if the tradeoff were 50-50, with every lost rural vote being matched by an added suburban vote. So far, this hasn’t even been close to being the case, but it would at least allow the Democrats to win blue states like Pennsylvania and Michigan in presidential contests again. Yet, that would still be an even trade that favors the GOP because it would make their lock on state legislatures all the stronger and reduce how many seats they have to worry about defending.

So, I disagree strongly with the idea that the Democrats haven’t and cannot win over Romney voters in affluent well-educated suburbs. I see little reason to believe they’ve tapped that source out, either. But it runs the risk of solidifying and even accelerating a realignment of the electorate that empowers the right.

The Democrats should not leave competitive seats on the table. The answer is not to abandon efforts to makes gains where they’ve already been making them. The answer is to have a dual strategy than can do two things at once, which is to work on solidifying gains while also working even harder to reverse recent losses.