Rockefeller Center is lit up to reflect the results of the U.S. electoral college votes in New York. Mark Kauzlarich/Reuters

Large metros voted for Clinton. Everywhere else went for Trump.

America’s contentious Electoral College may be organized around states, but its economy is organized around metropolitan areas. U.S. metro areas house more than 85 percent of the nation’s population and generate 90 percent of its economic output. Just the top ten metros account for more than a quarter of the U.S. population and more than a third of national economic output.

How did those metro areas vote in the 2016 election? Election returns are not typically collated by metro area, but county-level results are available. My Martin Prosperity Institute (MPI) colleagues arrayed the county-level data from Dave Leip’s 2016 Presidential General Election Results to America’s 350-plus metros. (Note that while there are still some vote tallies yet to be certified, the overall shares of the vote at the metro level are unlikely to change much at this point. And because of the large number of individual metros, any unforeseen shifts in final vote tallies will not significantly change the analysis here.)

All told, Donald Trump won many more metros, 259 to 122, than Hillary Clinton. But Clinton captured a greater share of the metro vote, 51 percent, compared to 44 percent for Trump. Metro areas accounted for 85 percent of total votes, 110 million of roughly 130 million total.

Metro Size

Clinton Share

Trump Share

Share of National Vote

Share of Population

Share of Economic Output

1 million or more people






500,000 to 1 million people






250,000 to 500,000 people






Less than 250,000






Clinton captured the largest metros. She bested Trump with 55 percent compared to 40 percent of the vote in metros with more than one million people, and won eight of the ten largest metros. These metros accounted for more than half the vote and generate two-thirds of America’s economic output.

Trump took the rest. He won metros with between 500,000 and a million people by 48 percent, compared to 46 percent for Clinton; those with 250,000 to 500,000 people by 52 percent, versus 43 percent for Clinton; and those with under 250,000 people by 57 percent, versus 38 percent for Clinton.

The average Clinton metro is home to almost 1.4 million people, more than three times the size of the average Trump metro, which is about 420,000. And outside of metropolitan areas, Trump beat Clinton 61 percent to 33 percent in micropolitan areas, and by 67 percent compared to 29 percent in rural areas, as Jed Kolko has calculated.


What economic and demographic factors are behind these results?

To get at this, my colleague Charlotta Mellander ran a basic correlation analysis between the share of metro votes for Trump and Clinton and metro characteristics such as size and density; education; income; and other indicators of socioeconomic class. We then compared the results of the 2016 election to those of 2012. As usual, I note that correlations point to associations between variables and by no means imply causality.

The results are in line with my previous analysis of the election results at the state level. Rather than being a significant break with the past, the 2016 election reinforces America’s deepest divides: between the country’s larger, denser, more affluent, more highly educated, more knowledge-based and more diverse metro areas; and its smaller, less advantaged, less educated, and less diverse metro areas. The 2016 election appears to have hardened these long-standing divides.

The first thing that jumps out, again in line with the state level pattern, is that despite the very different outcome—Trump’s victory compared to Obama’s win in 2012—the basic voting patterns are closely aligned. Across metros, Clinton votes were extremely closely correlated with Obama’s in 2012 (.94), and Trump’s votes were similarly closely correlated with Romney’s votes in 2012 (.90).


Class remains a persistent feature of America’s great metro divide.

The first dimension of class is income. Clinton support was concentrated in metros with higher incomes and wages, while Trump support was concentrated metros with lower incomes and wages. These correlations were up slightly from 2012.

Clinton Trump Obama Romney
Income (per capita) 0.40 -0.42 0.38 -0.38
Average Wage 0.58 -0.60 0.50 -0.51

Education is a second dimension of social class. Clinton support came from metros with higher shares of college grads, while Trump support came from metros with smaller shares of college grads. These correlations are substantially higher than in 2012.

Clinton Trump Obama Romney
College Grads 0.53 -0.60 0.42 -0.44
Creative Class 0.49 -0.54 0.40 -0.41
Working Class -0.51 0.53 -0.45 0.46
Service Class 0.04 -0.02 0.11 -0.11

*Figures in italics are not statistically significant

The kind of work we do is a third marker of socio-economic class. Clinton support was concentrated in metros where knowledge, professional and creative class workers make up a larger share of the workforce, while Trump support was negatively associated with this. Conversely, Trump support was much greater in metros with a larger share of the working class, with Clinton support negatively associated with it.

Most analyses of America’s divide juxtapose these two classes, the new knowledge class and the older working class. But few look at the largest class—the service class—which is made up of nearly 70 million American workers, 45 percent of the workforce who persist in low paid, often precarious jobs in retail shops, office and clerical work, and food service.

The share of the workforce in these low-paid service jobs was not significantly associated with support for Clinton or Trump this election (it was very weakly positively associated with Obama support and weakly negatively associated with Romney support across metros in 2012.) Instead of trying to re-cultivate the vote of the blue-collar working class, which has move solidly into the GOP, it would seem to make more sense for Democrats to aim to make inroads on this much larger, more multi-ethnic and more poorly paid service class.

Clinton support was also much greater in metros with larger concentrations of startups, venture capital investment, and high-tech industry, while Trump votes were negatively correlated with each. These correlations are all stronger for 2016 than they were in 2012.

Clinton Trump Obama Romney
Startups 0.49 -0.51 0.40 -0.40
Venture Capital 0.41 -0.43 0.28 -0.29
High Tech 0.46 -0.49 0.35 -0.35

Size and Density

Size and density are key features of America’s political divide.

Clinton support was positively associated with both the size of metros and even more so with their density, while Trump support was negatively associated with both. These correlations are again slightly higher than they were for Obama and Romney in 2012.

Clinton Trump Obama Romney
Population 0.44 -0.42 0.34 -0.33
Density 0.59 -0.63 0.51 -0.51
Drive Alone Share -0.48 0.55 -0.38 0.40
Public Transportation 0.48 -0.49 0.44 -0.44

On the flip side, Trump support was positively associated with the share of people who drive to work alone, a proxy indicator for sprawl, while Clinton support was negatively associated with it. Clinton support was also higher in metros where a greater share of the workforce uses public transit, while Trump support was negatively associated with public transit use.

The kind of housing we live in and how much it costs is another dimension of America’s political divide. Trump support was positively associated with the share of residents who own their own homes, while Clinton support was negatively associated with it. These correlations are up substantially from 2012 and even more so from 2008. Housing prices also play a role. Clinton support was higher in metros with more expensive housing, while Trump’s was negative. This correlation was also a bit higher than in 2012.

Race and Diversity

Much has been made of the role of race in the election. The table below shows the correlation for race, ethnicity, and other measures of diversity.

Clinton Trump Obama Romney
White Share -0.39 0.30 -0.22 0.21
Hispanic/Latino Share 0.31 -0.29 0.17 -0.17
Asian Share 0.46 -0.46 0.37 -0.38
Black Share 0.14 -0.02 0.07 -0.04
Foreign Born 0.54 -0.52 0.37 -0.37
LGBT Index 0.65 -0.64 0.51 -0.53

As many commentators have pointed out, Trump support was highly concentrated among whites. Metros with higher shares of whites went for Trump, and even more so against Clinton. Clinton support was higher in metros with greater shares of Hispanic and Latino residents, while Trump support was negatively correlated with the Hispanic and Latino share of the population. All of these correlations are up compared to the Obama and Romney race in 2012.

While black voters remain core components of the Democratic coalition and voted in large margins for both Clinton and Obama, the metro share of the black population was only weakly associated with Clinton support and not significantly associated with Trump support.

Immigrants played a much bigger role. Clinton support was even more closely correlated with the share of metro residents who are foreign born, while Trump’s support was even more negatively correlated with the foreign-born population. This too was up substantially from 2012.

The share of the population that identifies as LGBT is another indicator of openness, diversity, and tolerance. It was a key factor in Clinton support, and very much negatively correlated with Trump support. These correlations are among the very highest in our analysis and up from 2012.

Inequality and Segregation

A common assumption is that inequality drove the vote toward Trump, but the data place more unequal metros firmly in the Clinton camp. Clinton support was positively associated with income inequality and even more so with wage inequality, while Trump support was negatively associated with each.

Clinton Trump Obama Romney
Income Inequality 0.24 -0.17 0.14 -0.14
Wage Inequality 0.42 -0.42 0.29 -0.29
Economic Segregation 0.47 -0.44 0.32 -0.31

America has not only become more unequal, it has become increasingly sorted and segregated by socio-economic class. Clinton not only did better in more unequal metros, she did better in more economically segregated ones as well. These correlations are also up from 2012. It is important to remember that both inequality and segregation are features of larger, denser, more affluent metros.


The great metro divide that I identified back in 2012 when Obama won has only hardened and deepened.

America is divided between cities of knowledge and skill and the rest. The residents of these knowledge cities not only do better economically, they are better-traveled, better-connected to the global economy, and more open to diversity. Perhaps because the work of the knowledge-based metros centers turns on knowledge, creativity, and abstract thinking, their residents tend to be more open to the notion that government can help improve the economy, better the environment, provide essential services (like healthcare), and protect the fundamental rights of disadvantaged or discriminated-against groups...

Those who live outside these places see knowledge-based centers as elitist and coddled by government. They are well aware of the growing gap between the metro haves and have-nots, and know they are losing ground. They'd like to somehow stop the forces of change that are leaving them behind and bring back the good old days when they, and their more traditional vision of, America was on top.

With the election of Trump and Republican control of the House and Senate, these divides are poised to deepen further still. So far, the incoming Trump administration appears to be uniquely uninterested in addressing such mounting crises as affordable housing and rising inequality. If Trump makes good on his promises to build a border wall and halt the flow of immigrants, he may also staunch the openness and innovation of the dense, high-tech metros that are currently powering the U.S. economy.

About the Author

Most Popular

  1. photo: A cyclist rides past a closed Victoria Park in East London.

    The Power of Parks in a Pandemic

    For city residents, equitable access to local green space is more than a coronavirus-era amenity. It’s critical for physical, emotional, and mental health.

  2. Coronavirus

    The Post-Pandemic Urban Future Is Already Here

    The coronavirus crisis stands to dramatically reshape cities around the world. But the biggest revolutions in urban space may have begun before the pandemic.

  3. photo: South Korean soldiers attempt to disinfect the sidewalks of Seoul's Gagnam district in response to the spread of COVID-19.

    Pandemics Are Also an Urban Planning Problem

    Will COVID-19 change how cities are designed? Michele Acuto of the Connected Cities Lab talks about density, urbanization and pandemic preparation.  

  4. Perspective

    In a Pandemic, We're All 'Transit Dependent'

    Now more than ever, public transportation is not just about ridership. Buses, trains, and subways make urban civilization possible.

  5. Illustration: two roommates share a couch with a Covid-19 virus.

    For Roommates Under Coronavirus Lockdown, There Are a Lot of New Rules

    Renters in apartments and houses share more than just germs with their roommates: Life under coronavirus lockdown means negotiating new social rules.