Ahead of the presidential election, publications like the New York Times, the Huffington Post and 538 gave Hillary Clinton somewhere between a 65%-99% chance of winning the election. You know what happened next. As November 8 ticked into November 9 and it became less and less likely that Clinton would win, people began wondering how forecasters got it so wrong.
Now that a month has gone by and pollsters have had a chance to look into the data, some partial explanations have emerged: pre-election polls underrepresented Trump voters, two candidates with historically low approval ratings caused voters to stay home and white working class voters in the Rust Belt came out strong for Trump. As personal finance is SmartAsset’s specialty, we decided to look at voter statistics through that lens.
Check out our mortgage calculator.
We examined data on homeownership rates, median home values, credit card debt, costs of living, the percentage of the population with health insurance and the percent change of median income from 2010-2014 in 2,127 counties. We compared that data to the percentage of people in each county who voted for Donald Trump, the winning candidate. By doing this we can see which personal finance metrics correlated with voting for Trump.
Even in cases where the charts show a strong correlation readers should not assume the relationship is causal. There are a variety of reasons people pick their preferred candidate, some economic, some cultural and so on, making it hard to isolate any one cause.
- Rural vs. city divide – Much has been made of the fact that rural voters helped clinch the White House for President-elect Donald Trump. Hillary Clinton on the other hand won most big cities. Our data largely backs that up. According to our data, Clinton voters tend to live in counties with high costs of living and lower homeownership rates: both factors which are associated with large metropolitan areas. Our data also shows that Trump supporters tend to live in areas with lower costs of living and lower home values – the hallmarks of many rural areas.
- Income change was a toss-up – One interesting finding that emerged from our data is that Trump won the vast majority of counties that saw big income changes in both directions. Of the 50 counties which saw the biggest income decreases from 2010 to 2014, Trump won a convincing 46 counties. And of the 50 counties which saw the biggest income gains from 2010 to 2014, the president-elect won 44 counties. However, perhaps the more interesting discovery to emerge from our study is that income change did not play as big a role in the election as some may expect. That is (as the flat trend line on the graph in the corresponding section below shows), there was not a particularly strong relationship between how counties voted and the median income changes in the area.
Homeownership Rates and Home Values
In the 192 counties in our study where the homeownership rate is greater than 80%, voters were twice as likely to vote for Trump. The county with the highest rate of homeownership in our study is Elbert County, Colorado. More than 88% of people in Elbert County own homes and the county voted for Trump at a rate of 73%.
The three counties with the lowest rates of home ownership are three counties in New York City: Kings County, New York County and Bronx County. Of the three, Trump got the highest percentage of votes in Bronx County, where just 17% of voters voted for him.
Median Home Values
In the 45 counties in our study where the median home value exceeds $400,000, voters were more than twice as likely to vote for Clinton. Trump support dwindled in the most expensive places. Trump won only two counties where the median home values are over $400,000: Morris County, New Jersey and Richmond County, New York.
The county with the highest median home value is New York County, New York (the bottom right on the graph). Less than 10% of people living in New York County voted for Trump. The area with the lowest proportion of Trump voters was the District of Columbia. The median home value in D.C. is $454,500 and just 4.12% of voters there cast a ballot for Trump.
At the other end of the spectrum is McDowell County, West Virginia. McDowell County has the lowest median home value ($38,100) and Trump won the county by 49 points.
Find out: Should you rent or buy?
Cost of Living
The average cost of living across all counties we looked at is $17,600 (for one adult). In the 52 counties in our study where living costs exceed $24,000 per year, Trump lost by 33 points on average. Once again the District of Columbia is the county which voted for Clinton the most and has the highest cost of living in the country at about $28,000 per year. Trump lost D.C. by a whopping 88 points.
Trump received more support as a place became less expensive to live in. In Converse County, Wyoming, which has the third-lowest cost of living in the country at $14,400, Trump received 84% of votes.
The county with the highest cost of living that voted for Trump was Fauquier County, Virginia which has an annual cost of living of $27,507. Almost 60% of the population in Fauquier County voted for the president-elect.
Related Article: Here’s How the Trump Tax Plan Could Affect You
Health Insurance Rates
Trump spent a lot of time during this election season hammering home his opinion that Obamacare is a disaster for the country. One of his big campaign promises was that he would repeal President Obama’s signature health care law. Did the success or failure of Obamacare determine how people voted? The data suggests not really.
The trend shows that people who voted for Trump tended to be slightly less likely to have health insurance but the relationship was not very strong.
The least-insured county in our study is LaGrange, Indiana. Only 55% of the population has health insurance and 73% of votes there went to Trump. On the other end of the spectrum 97.5% of people living in Norfolk, Massachusetts have health insurance and only 33% voted for Trump.
Credit Card Debt and Income Change
Credit Card Debt
Trump tended to outperform Clinton in counties with less credit card debt. The average credit card debt by county is $2,311. The 755 counties in our study with credit card debt levels averaging less than $2,000 preferred Trump by about 30 points.
On average residents of Jackson County, Kentucky have just $816.66 in credit card debt. Trump won by the county by 80 points.
In general our data shows that counties with higher costs of living tend to have higher amounts of credit card debt. One explanation for this may be that people who live in places with higher costs of living end up having to take on more credit card debt.
The county with the highest average credit card debt is Pitkin County, Colorado with an average credit card debt of $6,727. Pitkin County had a strong preference for Clinton, with only 24% of Pitkin residents voting Trump.
Median Income Change
Another key part of Trump’s election messaging was that so-called elites had betrayed American workers by outsourcing their jobs and by extensions their livelihoods. Despite this rhetoric, our data suggests that whether or not a counties’ median income grew or shrunk over the past five years was not a major factor in the election.
You can find evidence for places which saw incomes shrink voting for Clinton, like Hertford County, North Carolina where median incomes shrunk 10% on average from 2010 to 2014, and 70% of the voters voted for Clinton. You can also find places like Knott County, Kentucky where incomes shrunk 14% on average over the same timeframe, and Trump won by 50 points. The same is true for counties in which median incomes grew as well.
Interestingly, in areas with large changes in income in either direction, Trump came out on top. Of the 10 counties in our study with the most income growth, only one county – Iberville, Louisiana – went to Clinton. Similarly, of the 10 counties in our study where incomes shrunk the most, only the aforementioned Hertford, North Carolina went to Clinton.
Of course income alone does not reflect the entire economic reality of a county. Specifics like job quality and job type are also potentially important indications.
In order to create these charts, we gathered data on six metrics and compared them to the election results in 2,127 counties. Election results data comes from the American Press Bureau. The data does not include Alaskan counties, nor does it include counties for which we did not have data.
- Homeownership rates by county. Data comes from the U.S. Census Bureau’s 2014 5-Year American Community survey.
- Median home value by county. Data comes from the U.S. Census Bureau’s 2014 5-Year American Community Survey.
- Cost of living by county. This is the estimated annual cost of living for one adult. Data comes from the MIT Living Wage database.
- Percentage of a county’s population who have health insurance. Data comes from the U.S. Census Bureau’s 2014 5-Year American Community Survey.
- Average credit card debt by county. Data comes from SmartAsset estimates using Federal Reserve data.
- Median income change from 2010-2014 by county. This is the difference between the median income in a county in 2010 and in 2014. Data comes from the U.S. Census Bureau’s 2014 5-Year American Community Survey and the U.S. Census Bureau’s 2010 5-Year American Community Survey.
Questions about our study? Contact us at firstname.lastname@example.org.
Photo credit: ©iStock.com/adamkaz