In many ways, 2016 has been the strongest year, economically, since the mid-2000s, or even the late 1990s. Unemployment was low. Wages were increasing. The markets hit new highs. Consumer confidence rose. Yet Donald Trump unseated the incumbent party and won the presidential election promoting economic pessimism and promising restoration: He swore to Make America Great Again, describing the country as falling behind, workers as jilted, growth as anemic, and the current period of expansion as a sham.

How could a message like Trump’s take hold in an economy like that? The answer is that it did not take hold with everyone—or, more to the point, everywhere. In ways that have been overlooked and under-appreciated for years, the strength of the post-recession recovery has varied dramatically not just from income bracket to income bracket, but from place to place as well. And it so happens that the places left behind are also places that wield outsized political power—enough power to vault the loser of the popular vote to the presidency and to make the economic story of 2016 more about stagnation, failure, and the dissolution of the American dream than it was about the headline economic numbers.

And those headline numbers were undeniably good. This year offered unremarkable but steady growth—about 2 percent, the same-same, just-okay annual pace that has characterized the Obama recovery. But over time, that same-same, just-okay annual pace has nudged the economy to a pretty good position relative to where it was eight years ago. The unemployment rate has fallen to 4.6 percent, its lowest level in nine years and “in the vicinity of maximum employment,” according to Janet Yellen, the chair of the Federal Reserve. Businesses added about 2 million jobs, and measures of underemployment dropped as well. “It’s a pretty good year bordering on a very good year in terms of jobs,” Mark Zandi, the chief economist at Moody’s Analytics, told me.

Competition for workers led businesses to boost wages, with earnings jumping faster this fall than at any other point during the recovery. “The job market was pretty solid for lower-skilled workers,” said Zandi. “Minimum-wage hikes helped, and you saw a lot of retailers and fast-food restaurants get ahead of those increases too.” Indeed, economic growth started to reach down to middle-class and even poor families towards the end of the Obama administration. According to Census data, median household income jumped 5.2 percent between 2014 and 2015, the first increase since 2007 (though it is still lower than it was in 1999, adjusting for inflation). Millions of people were lifted above the poverty line, with the drop in the poverty rate the biggest in a half century. Those trends are expected to continue next year.

2016 proved encouraging in financial markets as well. The total net worth of American households surged past $90 trillion, driven by a powerful buildup in the value of houses and stocks. At the national level, the housing market rose to new heights, though home prices remained below their bubble-era peak after accounting for inflation. The stock market climbed to heady new territory as well.

But those aggregate, big-picture numbers deceive as much as they elucidate. Income inequality became a dominant theme in the 2012 presidential election, given its power in explaining why steady growth and a booming stock market were leaving so many behind. Now, other sorts of disparities have taken on new importance. “In the recent past, the focus on inequality has all been about where you fall in the distribution, whether in the 10 percent, 1 percent, or top tenth of a percent,” said Jed Kolko, the chief economist at the job-search website Indeed. “Now that discussion is shifting to differences between groups, with those groups being geographic, and also different in terms of the types of jobs available and education levels.” It is these geographic inequalities that help to explain why the recovery left so many voters enraged, and the peculiar structure of the American electoral system explains why that rage proved so politically potent.

The recovery has been characterized by yawning gaps between the rich, the middle, and the poor. But, as Trump’s election made clear, it has also been characterized by yawning gaps between cities, the suburbs, and rural parts of the country. “Once you go beneath those big, national, largely encouraging statistics, you see a lot of variation,” said Mark Muro, a senior fellow at the Brookings Institution. “This is about big-county, cosmopolitan metropolitan areas pitted against small-population rural counties and some small metros. It is absolutely clear that the economic story has been much more difficult in that second group of non-big-metro counties.”

Indeed, the employment rate in rural areas was actually 2.9 percent lower in mid-2016 than it was in early 2007, just before the Great Recession started: There has been no jobs recovery in those spottily-populated swaths of the country. The employment rate in metropolitan areas, in contrast, is 4.8 percent higher than its 2007 level, and businesses are adding jobs twice as fast in urban areas as they are in rural ones. The far-flung counties that boomed during the Clinton expansion started to lag during the Bush expansion, and really fell behind during the Obama expansion. “The urban counties of large metros actually did better in the recession and recovery than they did” during the housing-bubble years of the 2000s, Kolko said.

And some metros are doing far better than others, with New York and the Bay Area in particular booming, minting millionaires and launching new businesses. (One stark data point: Wages are so high in San Jose, San Francisco, and New York that some economists think building more housing to let more people move there would boost national economic output by an astounding 10 percent.) The story is not just about New York beating rural Illinois, but New York beating St. Louis. Government policies that had made the country more equal across regions—among them antitrust protections—were forgotten or rolled back in the late 20th century. And fortunes diverged.

According to a report from the Economic Innovation Group, an entrepreneurship-oriented research and advocacy organization, the economy is growing more and more reliant on a smaller and smaller number of “super-performing” counties, some of them in or near Los Angeles, Miami, and New York City. Just 20 counties, out of more than 3,000 nationwide, accounted for half of the net increase in new businesses between 2010 and 2014. The number of counties seeing net job growth has dwindled, too. All in all, this adds up to, in the report’s words, “a massive and historically unprecedented imbalance” in geographic dynamism.

The kinds of jobs in places like Dallas and Las Vegas and San Francisco, versus eastern Kentucky or far-west Texas or the Nebraska-South Dakota border, might explain why the recovery feels more durable in denser areas, too. Workers in rural areas are far more likely to be involved in industries like farming, forestry, mining, and manufacturing, sectors in which job growth has been undercut and remains threatened by offshoring and automation. “These are thin economies. … Other than the hospital or local government, there isn’t a whole lot going on,” Zandi said. It’s just stagnation.” Workers in urban areas, on the other hand, are more likely to be involved in knowledge or service jobs, like corporate law or home health work, that are harder to offshore or pass off to a robot—which is true even for many jobs that are not particularly well-compensated.

Rural areas are lagging their urban counterparts in terms of home values—the single biggest component of most households’ net worth—too. The rise in home values has been sharpest of late in coastal cities like Seattle and Boston. And the market for downtown luxury condos has recovered in a way that the market for exurban ranch homes has not, with the spoils concentrated on the high end. Data compiled by the housing-data firm Weiss Analytics and reported by The Wall Street Journal shows that home prices in ZIP codes where the average house is worth $500,000 to $1 million are up 39 percent since the crash; in places where houses are generally worth $100,000 to $150,000, home prices are up just 16 percent. The long-term appreciation rate for houses in those cheaper, generally more rural ZIP codes is a quarter what it is for houses in those posher, generally more urban and coastal ZIP codes.

As rural areas have stagnated and urban areas have become more vibrant (as well as more unequal), optimism itself has become more unevenly distributed during the recovery. According to Gallup data, people living in big cities, urban suburbs, and college towns were the most sanguine about the overall economic outlook as of 2015, and they had grown much more positive since 2009—right when the recovery started to take off for them. People living in (in Gallup’s words) “evangelical hubs,” the “working-class country,” “aging farmlands,” and “rural middle America,” on the other hand, were dour about the economy’s prospects as of 2015, and had become only somewhat more optimistic since the darkest days of the crisis.

Race and ethnicity play into the geography of the recovery too: Black and Hispanic families are more likely to live in urban areas than white families, and their representation in the workforce has grown through the recovery. As of 2015, there were 66 million white prime-age workers with full-time jobs, down from 72 million in 2007. In contrast, according to Department of Labor data, the number of black prime-age workers with full-time jobs remained roughly the same; for Hispanic and Latino workers, it grew to 15 million from 13.6 million. More than half of the job gains since late 2007 have gone to Hispanics, who make up 14 percent of the labor force; 29 percent of the job gains went to Asians, who make up 5 percent of the labor force; and 25 percent of the job gains went to blacks, who make up 11 percent of the labor force, according to an analysis from the Economic Cycle Research Institute, a nonpartisan think tank. Demographic changes—specifically, the workforce becoming less white—explain much of the shift. But “part of the reason may be that these jobs, predominantly in [the service sector], were created in metropolitan areas, rather than in rural areas and small towns where factories were shuttered and the manufacturing jobs disappeared,” the report concluded.

This hardly means that black and Hispanic workers have come out ahead of white workers in the recovery. By all accounts, white Americans continue to enjoy far better economic conditions than do Americans of color. The unemployment rate for prime-age white workers is just 3.7 percent. For black workers, it is 7.8 percent, and for Hispanic or Latino workers, 5.3 percent. Moreover, the racial wealth gap sits at its worst level in a generation and, as Valerie Wilson of the left-leaning Economic Policy Institute wrote, the “rise in overall wage inequality [has] mechanically separated black and white workers’ wages” over the past 35 years, since black workers were more likely to be working low-wage jobs to begin with.

Finally, variance in education levels has overlapped with place and race to create divergent experiences of the recovery. People in rural areas are less likely to have a college degree, and the growth in the share of people completing college has been slower in rural areas too. But job gains during the recovery have been strongest among workers with more education, with the economy adding 8.4 million jobs for people with bachelor’s degrees. Those jobs are clustered in big metro areas, not in rural ones.

Taking these disparities into account, then, the state of the economy meshes well enough with the outcome of the election. On the whole, Trump voters were as likely to be employed as those who voted for Hillary Clinton. They had, on average, higher incomes. Many lived in places with an unemployment rate down in the neighborhood of 3 percent. Yet the trajectory of their communities felt far worse. Trump had extremely deep support in rural places that felt more stagnant, where the opportunities were dimmer, and where output and productivity were lower. To wit: Clinton’s counties account for two-thirds of the nation’s economic output, and Trump’s just a third.

For many voters, that sense of stagnation and being left behind mixed with the potent forces of racial animus, sexism, cultural disdain, and political polarization. Indeed, economic hardship alone may not explain Trump’s victory so much as a desire in some regions for a more general political shake-up. Still, in the end, place, race, and class became far more important to understanding how the economy influenced voter behavior than anything that was happening to GDP or median household income. And there are signs that that will continue to be true going forward. Democrats are sorting into higher-productivity, higher-growth, higher-inequality areas that are more racially diverse—and hold less weight in the Electoral College. Republicans are sorting into lower-productivity, lower-growth, lower-inequality areas that are whiter—and hold more weight.

What can be done to revitalize those counties that feel left behind, that the recovery in many ways has bypassed? That is the hard part. Policy is much better at redistributing money to individuals than it is at revitalizing regions. The federal government has every lever it needs to address income inequality, and very few to address the problems of living in an economy reliant on wheat prices, coal mines, or the operation of factories. “We have a well thought-out sense of density-based economies,” said Muro of the Brookings Institution, referring to economists’ understanding of cities. “Where we don’t have many ideas at all is in terms of economies that don’t have critical mass. One of the few ideas we have seems to be to invite people to move [to them].”

Economists do have a few other ideas. Infrastructure investment in struggling parts of the country—building rural prisons, say, or constructing a beautiful wall through a sparsely populated desert—would help, though such implementation can take time. Policies to help certain industries through tariffs or tax incentives or government stimulus—punishing foreign automakers, perhaps, or subsidizing American steel companies—might even things out as well, and Trump has shown a remarkable willingness to meddle in companies’ internal affairs and boost certain businesses over others. But the truth is that the forces holding rural America back and pushing urban America forward are profound ones. Just as income inequality has become a fixture in many Americans’ understanding of the economy, so too must geographic inequality. And not just in election years.

Related Video