Let’s say you are a Republican candidate for President of the United States. You have talking points and policies that seem out of touch with mainstream America and you say things that the typical American, which is to say the type of person advertisers want to reach, urban dwellers with cash between the ages of 18 and 54, would be appalled to hear. You run a campaign mainly in what many people believe to be the parts of America in steep decline; ‘rust belt’ cities and areas where industry and mining jobs have disappeared, rural areas in the South and Midwest, places with a high proportion of older white people who watch Fox and Friends, the very type of person advertisers don’t much care for, the parts of America where the incomes are stagnant and the population is declining. And then you win and become President of the United States of America. Surely this is an episode of Westworld or of Fantasy Island; any second now it will all be revealed as an elaborate charade.
And yet, it is reality. You don’t win the popular vote to be sure, but you win many more votes than might be expected in places that did not seem to have the votes to give. You win the Electoral College by eking out narrow victories in places like Pennsylvania and Michigan, previously seen as Democratic strongholds, although the presence of large numbers of Republican Congressmen and local legislators should have been a tip off to Democrats that perhaps all was not calm below the surface. You seemingly pull votes out of your ass and chalk up wins despite what the polls and pundits predicted in places like Pennsylvania.
So, how did Donald Trump win the 2016 Presidential election? By winning places like Westmoreland County, Pennsylvania. Trump took this county, 20 miles east of Pittsburgh, by 116,522 votes to 59,669 for Clinton, a margin of 56,853 votes. Trump won Pennsylvania by 44,292 votes; thus his margin in Westmoreland County, the 11th largest county in Pennsylvania by population, was 12,000 votes greater than his margin of victory in the state. By comparison, John Kerry, who beat George Bush in Pennsylvania in 2004 by over 140,000 votes, lost Westmoreland by only 22,283 votes. Trump picked up 34,570 votes in this one county over George Bush’s margin of victory. Incidentally, Kerry received 2,938,095 votes in the state in 2004 while Clinton received 2,926,095 votes in 2016, a difference of only 12,000 votes; that is to say, they both pulled in about the same number of votes statewide but Clinton lost Pennsylvania by 44,000 votes while Kerry won by 144,000 votes.
Trump won Westmoreland County with a record voter turnout in the county despite the fact that the population of the county has declined by more than 13,000 people in 12 years, by a little more than 1,000 people per year. In fact, no county in the state has lost more residents (7,213) in the last five years than Westmoreland County. Clinton ran up bigger margins than Kerry and even improved on Barack Obama’s 2008 and 2012 numbers in the five county Philadelphia area, an area that is fast growing and increasingly Democratic in orientation. By all rights she should have prevailed in Pennsylvania and frankly in the overall election results. Things seem suspicious. Perhaps there was voter fraud in these greying parts of the country, perhaps somehow the Russians got there mitts into the election machines in places like Westmoreland County. It seems almost plausible given the aforementioned data, but a deeper dive into the data shows that what happened can be entirely explained by demography.
The main conundrum that needs to be addressed is how a place with a steadily declining population produced the largest number of voters in the history of the county for the 2016 election? I have included below a typically busy chart with various pieces of data about population and voting results in Westmoreland County for the last four elections. I use the last four elections only because it was in 2004 that voter turnout numbers jumped significantly from the numbers in previous elections and so the elections are more easily compared. I know the chart wanders off the screen but I am too lazy/incompetent to bother fixing it up so please just scroll right to read all the data contained in the chart.
|Total Registered Voters
|Total Votes Cast
|Registered R Voters
|Registered D Voters
|Registered Voters as % of Total PA Voters
|Votes as % of Total PA Votes cast
|Population Westmoreland County
|Population as % of PA population
|Voting Age Population Westmoreland County
|VAP as % Pennsylvania Voting Age Population
|Votes cast as % VAP
There are some obvious points to make right off the bat. First of all, the number of Democratic votes has been dropping steadily for years as the number of Republican votes increases. This fits the pattern that many pundits have observed of the Democrats losing white voters in places such as this county which is usually ascribed to the party being ‘out of touch’ with voters and their needs. To which I say, bullshit! There is a much easier and much more concise way to describe what is going on here and places like this in the rest of the country: White voters in places like Westmoreland County are aging, are more than likely not to have a college degree, and are very likely to watch a steady diet of Fox News. They also live primarily in communities that are more than 95% white that are very slowly becoming less white as the number of working age people declines and younger Black and Hispanic workers come in to do the jobs that need to be done, like nursing or custodial work. They are told on Fox (‘News’ seems inaccurate) that the cities of America are burning down and that ‘the minorities and the illegals’ are coming to your town soon, with their drugs and their crime and their desire to impregnate poor, unsuspecting white girls. No wonder the voters rush (waddle, more likely) to vote for their ‘savior’, a man who speaks their language, the language of Fox.
Surely I exaggerate? The data says otherwise. Let’s take a look at the ratings of Fox News Channel, the most viewed channel on cable television last year. The average number of viewers of Fox News was 2.48 million in 2016, half a million viewers higher on average than the second-ranked ESPN at 1.91 million viewers. However, If one looks at viewers between the ages of 18-49, Fox News is only the nineteenth-ranked network on cable, with 354,000 viewers, behind first place ESPN’s 816,000 viewers, behind even CNN. Assuming few children watch Fox News (a pretty good assumption I think, although there are likely a few Alex P. Keaton’s out there), more than 85% of the viewers of Fox News are over the age of 50. In fact, half of Fox viewers are over the age of 68, according to an article in Atlantic Monthly. Also, the above data show that young people clearly do not watch as much TV as older people. In the same article in the Atlantic, The typical Fox News viewer is described as over 70, white, less likely to have a college degree. There is also evidence that people tend to become more conservative as they age ; it is certainly clear that older people in recent years have voted in larger numbers for the more conservative candidate in elections world wide (although it is difficult to discern whether these patterns are age effects or just cohort effects, that it is merely Baby Boomers, for example, that are becoming more conservative and they happen to be aging).
So what does this have to do with Westmoreland County? If you were looking for ground zero of Fox News potential viewers you would find it in this county in western Pennsylvania. According to the most recent census data (2015 estimates) Westmoreland County is 94% White (Not Hispanic), 26.4% of the residents have college degrees, compared to 49.1% of the population of Chester County, a fast-growing, Democratic-leaning county near Philadelphia, 28.6% of the residents of Pennsylvania, and 29.8% of the population of the United States, only 1.5% of the population is foreign-born compared to 6.3% of the population of Pennsylvania and 13.2% of the national population. Most importantly, more than 21% of the population of Westmoreland County is over the age of 65, compared to 15% of Chester County, 17% of the population of Pennsylvania and under 15% of the population of the United States. Is it any surprise that the voters steadily drift into the arms of the Republican Party? Even the voter registration numbers indicate the conservative nature of the electorate of Westmoreland County: registered Democrats continue to exceed the number of registered Republicans (see chart above), despite the fact that the Democratic candidate has failed to win the presidential vote in Westmoreland County since 1996. People clearly do not change their voter registration at the same rate that they change their voting preferences.
Incidentally, in 1996, the year Bill Clinton defeated Bob Dole in Westmoreland County by a narrow margin (44.4% to 43.3%) Ross Perot pulled in over 11% of the vote. In 1992, Clinton defeated George Bush (45.2% to 30.6%) with Perot collecting an astonishing 24.0% of the vote! prior to 1992 the Democratic candidate routinely carried this county and most of the other counties in the southwestern corner of Pennsylvania, even Michael Dukakis versus George W. Bush and Walter Mondale against Ronald Reagan. Since the ‘Perot’ elections the Democrats have failed to win the county, losing by increasingly large margins at each election. I believe that Ross Perot voters and the continual rise of Republican votes culminating in the surge of voters for Donald Trump are the same people, just older.
The most interesting piece of data in the chart above helps us to resolve the apparent paradox of the declining population and the record number of votes cast. Although the population of Westmoreland County has declined steadily in recent years, the absolute number of voting-age residents has remained remarkably stable. In fact the number of voting-age residents in 2015 (291,462, the most recent year for which there is data) is higher than the number in 2004 (290,793). The population is aging, and the absolute number of residents under the age of 18 has declined precipitously in recent years, from 77,867 in 2004 to 66,494 in 2015, a decline of over 11,000 children. Even the number of younger voting-age residents has declined: in just five years, from 2010 to 2015, the number of people between the ages of 18 and 54 in Westmoreland County has declined from 169,690 to 157,578, a drop of more than 12,000 residents.
On the other hand, the number of people 55 and older has increased from 122,868 to 133,884 residents, a jump of over 11,000 people in only five years. In a very short span of time (2004-2015), the population of Westmoreland County has transformed from more than 21% under 18 (77,867) and 18% over 65 (66,997) to 21% over 65 (75,735) and 18% under 18 (66,494), a complete reversal in a little over ten years. Astonishingly, almost half of the potential electorate (133,884 out of 291,462 or 46% as of 2015) is over the age of 55. Older people tend to vote more reliably as well. In all likelihood, therefore, the typical Westmoreland County voter is your average Fox News viewer cohort.
So, it is NOT an imaginary stage set for old Republican fantasists: the absolute number of what we can call ‘Fox Voters’ has actually increased in places like Westmoreland County, confounding the data that shows that Pennsylvania should have remained in the Democratic column. It is relatively slow-growing but is still a growing state, especially in Philadelphia and environs, where the population makes up an ever larger percentage of the total population of Pennsylvania, a population that increasingly skews younger, more highly educated, and less white . The percentage of people in Pennsylvania under the age of 18 is 20.9%; in Philadelphia it is even higher at 22.9% and even reaches 24.9% in nearby Chester County, a county increasingly trending Democratic. Thus, there is a higher proportion of voting-age population in Westmoreland and this population is much older than the average potential voter in the state or in these ‘blue’ areas. Hence the seemingly unlikely increase in the vote in places like Westmoreland County compared to counties like Chester County, where the vote was expected to increase (and did) as a result of overall population growth. Westmoreland shrank in overall population but the voting-age population was stable and even slightly larger than it was in 2004, but was markedly older as well.
It should be immediately apparent that this anomalous oddity cannot continue. For one thing, older people die. Despite the fact that the population is aging, as each year passes the number of people in a given age cohort slowly evolves until a dramatic change will occur as the population bulge commencing with those born in the 1940s and continuing through to about 1960 starts to reach the years of higher mortality. When will this happen? I have extrapolated population data from Westmoreland County through to 2020 to look at the potential electorate. The population should continue on its downward trajectory, reaching 353,000 in 2020. The number of people turning 18 each year is is currently about 4,000 and is slowly declining. On the other hand about 4,500 people die each year, and this number is going up slowly as more people reach their later years, so the natural voting age population is slowly declining by about 500 people per year, and is likely to become greater each year for the foreseeable future.
Additionally, there is net out-migration from Westmoreland County. This is hard to calculate precisely, but it stands to reason that if, in 2005, there were 25,000 people between the ages of 75 and 84 and about 5% of that cohort typically dies each year, there would be 15,000 people from that cohort alive in 2015, to go along with the people remaining who were already 85+ years old ten years previously. Assuming about 15% of 85+ year old people die each year, the approximately 9,000 people 85+ in 2005 would be reduced to about 1,700 people 95 years and older by 2015. Thus there should be about 16,700 people 85+ years old alive in Westmoreland County in 2015. However, there are only 11,513 people matching that description; thus the missing 5,000 people have presumably left the county for Florida, Arizona, a rest home in another county or state or the home of a son or daughter, for example. Conversely, some young families from Pittsburgh have surely decided the price is right in Westmoreland and have bought a house and started a family, slightly evening out the balance. Overall the number seems to be slightly negative, judging from the difference in births and deaths (about 3200 births and 4400 deaths, or net -1,200 in recent years) and the overall population change in recent years, which has been greater than -1,500 per year in recent years.
Thus, it seems that we may have reached the high-water mark already in terms of overall voting-age population(VAP) in Westmoreland County given current trend lines. However, the overall number is still likely to remain in the neighborhood of 290,000 potential voters in 2020. And these voters will be even older in 2020! Extrapolating to 2020, the number of ‘Fox Voters’ (65+ years old) in Westmoreland will likely increase from the 76,000 in 2016 to about 84,000 in 2020. The number of voters 55-64 will also increase slightly from 58,000 to about 59,000, making the total number of voters older than the 18-54 demographic reach 143,000 about half the total number of potential voters. It seems likely that Westmoreland will become even more Republican in 2020. The same battle between aging voters in places like Westmoreland and younger, more educated, less white voters in Philadelphia will play out, only the vote totals in Philadelphia will rise even higher while the vote totals in Westmoreland are likely to remain static or begin to decline. Thus the balance again will revolve around turnout. Here the Democratic party will likely have an advantage as the Trump surge will probably look like the ‘Perot surge’ and many of the infrequent voters who showed up in 2016 to cast a vote will likely stay home in 2020. The turnout in Westmoreland County, which had an exceptionally high 74.6% turnout rate in the 2016 election, higher than any previous recent election, will be difficult to match and extremely unlikely to exceed the 2016 rate.
Looking beyond 2020 to 2024, which theoretically will be an open election, what can we say about Westmoreland County? The number of ‘Fox Voters’ will continue to rise, to about 92,000, but the overall voting-age population will drop to about 280,000. Thus the voting trend towards the Republican party might continue, although it is far from certain that today’s 55 year old will be a Fox News viewer in 2024. Again, the issue of whether it is a cohort effect or an age effect is unclear but the upshot is that overall vote totals will diminish further as the VAP declines. Meanwhile, hundreds of thousands of young Hispanic, Black, Asian and other ‘Non-White’ residents of the Philadelphia area will have joined the VAP and replaced the close to one million older white voters who will have died since 2016.
In conclusion, the results of the 2016 Presidential Election reflected a unique concatenation of circumstances which resulted in a ‘surge’ of older, white, less educated voters from the ‘rust belt’ and the rural areas of the South and the Midwest, which can be entirely explained by demographic analysis of the voting age population of places like Westmoreland County, Pennsylvania. Older ‘Fox Voters’ made up an usually high proportion of the voting age population and these voters showed up in unusually high numbers to affect races in normally reliably ‘blue’ states like Pennsylvania, Michigan, Wisconsin and, to a lesser extent, Iowa, Ohio, Maine, and Florida. On the other hand, a ‘surge’ of non-White, increasingly well-educated voters is on the horizon in many states, including Pennsylvania, but also in Florida, Texas, North Carolina, and Arizona which, combined with the slow decline of the very voters who turned out for Trump, will likely result in a reversal of fortune in the respective parties.
To take the example above as a model: Westmoreland, the epitome of a ‘Fox Voter’ county, has peaked in terms of possible voter turnout based on an analysis of short and long term demographic trends. Although the voters who do turn out will likely continue to move in the direction of the Republican Party for a few more years, the net demographic changes in Philadelphia, Montgomery, Chester, and even Lancaster and other counties in southeastern Pennsylvania will more than match the vote totals in places like Westmoreland and the difference will begin to accelerate after 2020. Unless the Republican Party changes its positions markedly in the next few years (which, based on current evidence seems extremely unlikely; if anything they are doubling down on ‘Fox Voters’–See Sessions on Marijuana, Trump on race, religion, immigration, and virtually any Republican and LGBTQ rights, and so on) the future seems ‘blue’. And that is not a fantasy.