I want post this now, while all of my readers are watching college football. That way, if I’m right I’ll get to say, “I predicted it!” And if I’m wrong, all y’all can point at me and laugh at the data nerd.
I just finished reading a fascinating analysis of the 2016 election, authored by a couple of data geeks. Sean Trende and David Byler are analysts at Real Clear Politics, a must-read-daily website full of charts, projections and analysis. Their analysis following the 2016 election looks at trends in the electorate over time with a geographic filter, breaking voters into six distinct groups: rural, small town, large town, small city, big city and mega city.
Here’s a tell-all chart presented in the conclusions section of the report:
This shows that over the last three decades, Democrats have dramatically increased their share of voters in the largest cities and dramatically lost their share in areas that are large town or smaller. Their worst losses are in the smallest population areas. Democrats may be getting more votes nationwide, but concentrating those votes in cities that they already dominated is like shooting a dead man. It’s a wasted bullet because dead is still… well… dead.
Getting an extra four or five million votes across the cities of Chicago, Los Angeles, New York and Philadelphia is not enough to offset votes lost by the collective tens of millions from the thousands of Anytown, USA locales. There are at least 38 states that do not have a mega city in them or nearby. There are at least 25 that don’t even have a large city in them. Yet these “fly over” states have well over half of the eligible voters.
Democrats have created a very effective coalition of voters, but they have concentrated them in too few places, blunting their effectiveness and allowing Republicans to win elections without winning the majority of the popular vote.
This is important: You don’t need to win a national majority to win a national election.
California has a population of about 40 million. It has 53 members of the U.S. House of Representatives, or about one member for each 755,000 people. Alabama has about 4.9 million people and seven House members, or one member for every 700,000 people. Fewer people in Alabama have the same influence as a larger number of Californians. It’s not hard to put together a coalition of rural-ish states with fewer people wielding greater influence than states with high concentrations of urban dwellers.
This effect plays out in electoral politics but it also plays out in popular culture.
Pop culture is driven by urban dwellers. Urban areas are where the studios, producers, financiers, technicians and pop stars live because urban areas tend to have the physical infrastructure to support mass media production.
Urban areas are also more diverse demographically. They have higher concentrations of people of African, Hispanic and Asian descent. They also have higher concentrations of those identifying as having a non-traditional sexual orientation. As you watch your favorite commercial television programming over the next season, pay attention to the ads for alcohol, automobiles, toothpaste, mortgage lending and healthcare services. The vast majority of those will show the kinds and mixes of people you are likely to encounter in everyday life in the big city. Why? Because the people who make commercials use urban-ish people.
This may puzzle the Anytown citizen and for reasons he or she doesn’t realize. The small town banker may not run into a biracial or homosexual couple seeking a mortgage. The rural grocer may not see a large south Asian family shopping together in her store. The large town teacher rarely has the child adopted by the gay couple in his class.
This cultural divide causes issues when the votes are counted on (and sometimes for weeks after) election day. The urban dweller can’t believe that their House or Senate candidate lost because all of the people just like them voted the same way. The rural-ish folks don’t see what the fuss is about, because all of the people just like them voted the same way.
The Ruralites may not (yet) understand the power that they wield, especially in midterm elections when turnout is typically low. But the Urbanites really don’t understand because everything they see on TV, hear on radio or dig on Spotify says everyone is diverse like them and diversity is just another word for Democrat.
They don’t understand that despite the diversity surrounding them in their urban enclaves, non-Hispanic whites still make up at least two-thirds of the electorate and most of that electorate lives in areas that are large town and smaller areas.
As a result, Urbanites are likely to conclude that the Ruralites are racist homophobes.
They would be wrong. Ruralites probably get the Melting Pot concept much better than Urbanites are willing to give them credit for. Most rural folks can trace their lineage back to immigrants, especially those in the rural areas east of the Mississippi and in the Mississippi River Watershed. Latin cultural influences are felt from rural Michigan to Louisiana and from south Florida to Wisconsin. The difference between those roots and the ones forming today is that in the ‘old days,’ immigrants assimilated and became Americans, contributing their culture to the greater American tapestry. This is less so with today’s Hispanic, Asian and growing Middle Eastern communities. Assimilation is now slower to occur, especially given the speed at which we get news today. Also, we have not helped assimilation by printing instructions for Life in America in languages not named English.
This frustrates Ruralites, who grew up speaking only English and maybe had bilingual grandparents. It’s not racist for immigrant-born Ruralites to expect today’s immigrants to do what everyone before them had to do. This is what made America great, they think, and when someone comes along with a promise to Make America Great Again, Ruralites are more willing to give the guy a shot.
The pollsters and election gurus missed the Ruralite movement in 2016, and may be missing it again in 2018.
Most of the 2018 polling, projections and analysis still look to the so-called ‘generic ballot’ as the most influential predictor in the outcome of congressional elections. This is a useful indicator of national sentiment, but it is still just an indicator of how the popular vote may go. It does not predict how people will vote. Even if it did, it ignores the first fact from above—you don’t have to win the popular vote to win a national election.
Two models: If both are right, Republicans will hold both the House and Senate
Ray Fair is a Yale Economics professor. He uses a numerical model driven by economic conditions to predict election outcomes and has a pretty good record—his was the only model that favored the Republicans in both the 2014 midterms and the 2016 election. The Fair Model predicts that Democrats will win 50.70% share of the vote.
Nate Silver is the architect of the Five Thirty Eight prediction model. Silver currently gives the Democrats an 85.9% chance of taking the House (and an almost nonexistent chance of taking the Senate). However, this prediction is based on Democrats winning by at least 5.9 percentage points.
Fair's model says Democrats only get 0.70 points, just barely a majority. As we know from Trende and Byler’s analysis, and as we can see from Silver’s model, this means that Republicans are poised to surprise the talking heads in the media on Tuesday night.
I believe they will.