It’s not rocket science to figure this out. Clinton got more votes out of the cities and more votes out of the suburbs (where turnout was actually up) and she still turned a 300,000 cushion into a nearly 70,000 vote defeat. What happened is that where Obama lost the smaller more rural counties in the 70-30 range, Clinton lost them in closer to an 80-20 range. And if you spread that out across the commonwealth’s 67 counties, it swamped the Democrats and cost them not only the state’s 20 Electoral College votes but what should have been a pickup of a U.S. Senate seat.
For Clinton, and most observers, it wasn’t thought possible that she could lose if she not only netted her votes out of the cities but did better than Obama in the suburbs. It made perfect sense for her to stay focused on urban turnout and persuasion in the suburbs and she undoubtedly met all her benchmarks.
Now, Pennsylvania is unique in some ways and it can’t be perfectly compared to any other state, but this general phenomenon can be observed in Florida and in Ohio, Michigan and Wisconsin, too.
What happened here wasn’t a general failure to persuade white professionals in the suburbs or to mobilize the targeted cities in key swing states. It was a collapse of Democratic support in areas that were already voting heavily against the Democrats by unhealthy percentages.
People will want to go over these results with a fine-toothed comb, and I await that analysis, but the collapse of the Democrats in these rural counties isn’t tolerable if the left ever wants to control the House of Representatives or the state legislatures in most of the country. I know people are angry that these voters were willing to overlook or even actively support Trump’s racist idiotic campaign, but the Democrats will need to do better than winning 20% in these areas. It’s likely that the suburbs will continue to move away from the Trumpian Republican Party and that may solve the Electoral College problem. It won’t solve the other problems, though.Two issues here:
First, why didn't this show up in the polling? Going into Election Day, 538's polling averages had Clinton up 3.7%, a pretty substantial lead. I'm not a statistician, but this would be pushing the boundaries of the margin of error. I'm trying to look at the crosstabs for the final CNN poll, but there's no geographic breakdown. Did the polls generally undersample rural areas? It's going to be tough for us to know how to properly campaign and GOTV if we're not polling correctly.
Second, was this the case in other states where Dems underperformed the polls, such as NC, FL, OH, WI, and MI? I haven't had a chance to take a look yet; I'm wondering if anyone else has. If this was the case, Longman is probably right. In the David Remnick New Yorker article I talked about yesterday, he mentioned that one of the first signs that Obama and his political director saw something being wrong was that there were surprising results coming out of rural areas in Florida.
I don't think we're anywhere near making conclusions, and won't be for a while, but I'll have another post up in the next 24 hours with a little more about Longman's post, particularly about why the Dems would be having problems in rural areas and what they might do to change that.