President Donald Trump is a phony populist. Everyone in the intertwined worlds of politics, the mainstream media, and show business understands this. Then-president Barack Obama made the charge during a self-described “rant” in mid-2016, just before appearing at a campaign rally for the decidedly non-populist, and ultimately defeated, Democratic presidential nominee Hillary Clinton.

The only Americans supposedly holding any illusions to the contrary are his core supporters. Those at the lower reaches of the national income scale are supposed to be among the most pathetic victims of this epic con.

Except an obscure set of official national statistics has undermined this thesis. Overlooked as usual when it was released in May, the latest version of the Labor Department’s quarterly County Employment and Wages reports contains abundant evidence that the Trump victory in 2016 has paid off handsomely for the everyday Americans who gave him his Electoral College margin.

That’s because the Labor Department data make possible examining the incomes of workers in the counties that voted for Obama twice and then flipped to Trump. These counties number 206, and statistics are available for 203.

No fewer than 62 are located in the key Midwestern battleground states of Michigan, Iowa (the individual leader with 27), and Wisconsin. Nineteen more are found in Minnesota (which Trump lost narrowly and is targeting for 2020). In other words, they’re the quintessential jurisdictions whose voters were understandably attracted to the Democratic message in 2008 in the wake of the financial crisis and Great Recession and again in 2012 as the economy recovered, but turned to Trump when growth and its benefits ultimately disappointed. (Oddly, Pennsylvania, a major state Trump turned red, contained only three such counties.)

And the county data show that, so far, this bet has paid off for most. The most important statistics are those that compare how these counties have fared economically during the last two years of the Obama administration and the first two years of Trump’s presidency. For these two periods are not only the same length, they also came close together during the current expansionary phase of the business cycle. (Examining economic performance during different phases of business cycles—e.g. expansions versus recessions—produces often misleading apples-to-orange comparisons.)

The leading sign: in 130 of the 203 Trump flip counties in question, average annual salaries rose faster during the two years Trump has been in office than during the last two Obama years. That’s 64 percent of them. The rest enjoyed faster annual salary growth during those Obama years.

In addition, just over 23 percent of those counties (47) suffered absolute declines in average annual salaries during one of the two final Obama years, and four saw those salaries shrink in both years. During the first two Trump years, fewer than 10 percent (20) of the flip counties suffered absolute salary declines during one, and none experienced such decreases during both.

The only grounds for economic complaint that Trump flip county residents might legitimately have concerns the gap between their well-being and the rest of the country’s. During the last two Obama years, the number of flip counties whose annual salary growth rates topped the U.S. average rose from 89 to 116. During the first two Trump years, their numbers dropped to 84 and then 83. (Between 2015 and 2016, the salary growth in one of these counties matched the national average.)

But some context is needed here—and quite possibly it hasn’t been missed by flip county voters. Average annual U.S. salaries rose considerably faster during the first two Trump years (3.41 percent and 3.36 percent, respectively) than during the Obama years (3.08 percent and 1.21 percent, respectively). So since the national salary bar for the flip counties has been significantly higher during the Trump years, the fall-off looks pretty moderate, and in absolute terms, the annual increases remain an improvement.

The situation in the key Midwestern states doesn’t jibe well with the idea of Trump voters as economic dupes either. In Michigan, six Trump flip counties saw faster annual salary growth under this president while six saw salary growth slow, and in Iowa, the Obama years in question bested the Trump years in 16 flip counties as opposed to the opposite pattern in 11. But in Minnesota and Wisconsin, the Trump years prevailed by margins of 12 to 7, and 15 to 8 in the flip counties, respectively, while all three Pennsylvania flip counties have seen faster salary growth under Trump.

Four data years don’t make a slam dunk case. And with the current U.S. recovery setting new longevity records each month, neither a slowdown nor even a recession can be entirely ruled out between now and the 2020 presidential election (even though the latest data indicate that the economy is moving past a soft patch experienced earlier this year). Moreover, presidents rarely deserve all of the credit for good economic times, or all of the blame for slumps.

But so far, it’s the Trump critics who have mainly been wrong-footed over how working- and middle-class voters have fared during his administration. And barring big changes in the economy’s direction, unless they start looking at the data rather than trafficking in their own biases, another Trump shock could be in store come next November.

Alan Tonelson is the founder of RealityChek, a public policy blog focusing on economics and national security, and the author of The Race to the Bottom.