It was a classic campaign promise, overly ambitious and cleverly vague. What exactly did “winning” mean? Certainly, many reporters believed voters perceived the promise as an economic one. So let’s measure the promise’s success that way. How have Trump voters fared economically, compared with Hillary Clinton voters?
Not noticeably better, according to the data. By most measures, my latest research shows, Trump counties — and especially counties with higher proportions of Trump voters — continue to fall farther behind the rest of the country economically. The story of our economy, like the story of our politics, continues to be a story of division and divergence.
Trump’s America vs. Clinton’s America
It is no secret that the country is as geographically fractured as it is economically unequal. In fact, the two trends are intertwined. In separate studies, economists Rebecca Diamond and Peter Ganong and Daniel Shoag revealed a widening gap in incomes, skills and wages between low-income and high-income regions, beginning around 1980. After decades of converging, in other words, our cities and states have been growing apart. The wider the income gap grows between the regions, they show, the harder it becomes for those in service and even blue-collar jobs to afford to live in high-income, high-rent places with high-quality amenities such as clean air, good schools, low crime, strong job markets, transportation infrastructure and retail stores.
Driven out of thriving communities by those rents, people who were just getting by are surrounded by others who were also struggling, in areas that the better-off had fled. That leaves a skimpy tax base, shrunken opportunities and economic segregation.
Thus, we increasingly live in two Americas, and we vote accordingly.
Consider the stark differences in basic measures of local economic performance — employment and housing prices — between counties where the majority of votes were cast for Donald Trump and counties where the majority voted for Hillary Clinton. The average Clinton county employs seven to eight times as many workers as the average Trump county, with nearly double the market value per single-family home. In part, this difference reflects the higher population density of the urban areas, which voted disproportionately for Clinton. But as my analysis shows, it has been growing over time, as the Clinton counties outperform their Trump counterparts.
Post-election, the more things change, the more they stay the same?
After November 2016, many Trump supporters told reporters that they expected this gap to narrow. In essence, they were hoping to see faster job growth — and income growth, which would drive up housing prices — to catch up to the rest of the country.
Looking at 13 months of data since the election, we can see that that hasn’t happened. The average Trump county added 1.13 percent more jobs, while the average Clinton county added 0.49 percent. These increases are quite small, especially considering that significantly fewer jobs existed in Trump counties to begin with.
Housing prices tell a similar story, with even more data stretching into 2018. Regardless of how I compare the counties, Clinton supporters consistently come out on top. Even though their housing prices started significantly higher than their counterparts in Trump counties, their value increases even faster after November 2016.
The major shortcoming of this comparison is that it fails to account for pre-election trends. Maybe the Trump counties aren’t growing faster than the Clinton counties, but at least they are improving relative to their previous performance? Maybe they’re bending their trend closer to the trend of the Clinton counties, even if they’re not overtaking them?
Not at all, it turns out. Using a standard statistical technique called “difference-in-differences,” I estimate the difference between Trump and Clinton counties before and after the election and show whether the difference … differs. In other words, I look at whether the economic performance gap narrows. The answer: No. Statistically, there appears to be no significant improvement in job growth. The gap in housing price growth actually widens. In fact, the larger the Trump electorate and the larger the degree of Trump support, the worse the county’s economic performance.
Finally, you might wonder whether it’s fair to lump all Trump voters together. Many of them would have voted Republican regardless of the nominee. Perhaps it makes more sense to focus on the voters who switched to Donald Trump after voting for Barack Obama in 2012. Arguably, these voters were wooed specifically by the candidate’s promise to “start winning again.”
Even among these counties, however, there does not appear to be any improvement. Their performance is statistically indistinguishable from the performance of their fellow Obama counties that stuck with the Democrats in 2016.
What can we expect for the next two years?
The two Americas remain as economically divided after the midterm election as they did after the presidential election two years ago. We do not know, however, whether these different economic directions will now begin to converge. Two years is a very short time in which to reverse macroeconomic trends. It is still possible that Trump counties will be rewarded in the long run. For that reason, I will be updating these estimates throughout the coming years.
Let me acknowledge that it’s possible that these comparisons don’t account for differences between voters within counties — or what statisticians call the “ecological fallacy.” Perhaps Trump voters are experiencing stronger economic growth than their neighbors, even if their counties are underperforming the nation. Perhaps future data will allow us to make this person-to-person comparison within counties.
Until then, these preliminary findings reveal the economic state of our nation remains much as it did before Trump’s election: divided as much by economics as it is by politics.
Anthony W. Orlando (@AnthonyWOrlando) is assistant professor of finance, real estate and law at California State Polytechnic University, Pomona, and faculty affiliate of the Bedrosian Center on Governance and the Public Enterprise at the University of Southern California.