Donald Trump Thinks the Unemployment Rate Is “Fiction.” Fine. Here’s What He Should Track Instead.

Leader of the free world, right here.

Chip Somodevilla/Getty Images

Donald Trump does not believe in the official unemployment rate. The president has called the stat a “fiction,” and his staff has apparently gotten the memo. Asked by a reporter on Monday to simply state the current jobless figure, which is released every month by the Bureau of Labor Statistics and reported widely, Press Secretary Sean Spicer spent about three minutes evading the question. The figure is 4.7 percent, for the record. This is what passes for a sensitive subject in 2017.

Trump’s skepticism toward the unemployment rate is clearly rooted in political self-interest—if he treated the number as legitimate, he would have to acknowledge that the economy improved significantly under President Obama. By undermining public trust in government data, he also makes it harder for anybody to hold his administration accountable for the economy’s performance.

But even if it’s for all the wrong reasons, our president is also, in his own awful, howling bloodhoundlike way, sort of onto something. In recent years, the headline unemployment rate has become a less useful yardstick for measuring the overall health of the labor market. Like all economic indicators, it has shortcomings, and those weaknesses have grown more problematic since the Great Recession. If an administration wanted to benchmark its economic performance by focusing on a single headline statistic, this might be a good time to pick a new one.

The BLS even produces a data point that Trump himself might like: The employment-to-population ratio for adults between the ages of 25 and 54—or “prime-age EPOP.” It’s great. Really terrific. Some even say it’s the best labor market indicator out there. Believe me.

Here is something Trump gets right: These days, the traditional unemployment rate makes the economy look misleadingly rosy, and the problem is baked into how the statistic is calculated. Every month, BLS divides the number of jobless Americans who are looking for work by the total labor force. If you aren’t actively seeking work, you aren’t counted among the unemployed. There’s perfectly sound logic to this approach, since it keeps retirees, stay-at-home moms, full-time students, and other Americans who aren’t really interested in working from inflating the unemployment rate. But it also has a perverse side effect: When out-of-work Americans lose hope and quit sending in job applications, it pushes the unemployment rate down.

Trump sort of gets this. “If you look for a job for six months and then you give up … they consider you statistically employed,” he told a crowd in December. That’s not quite right, but it at least gestures toward the basic problem that the government stops counting discouraged labor-force dropouts as unemployed.

And that flaw in the unemployment rate has become more critical since the recession. Today, labor force participation—the percentage of Americans working or looking for work—is stuck near lows last seen in the 1970s. That’s partly because aging baby boomers have started retiring. But participation is down among younger Americans as well. We may have a 4.7 percent unemployment rate, but 4.7 percent doesn’t mean what it used to.

The government does cook up other, more expansive measures of “labor-market underutilization” that address this problem. The broadest, known as U-6, stands at 9.2 percent, and includes part-time workers who’d like a full-time job, as well as people who say they want work but have given up looking for it. That stat gives us a more detailed picture of the labor market’s woes, but it might still obscure some important issues. For instance, working-age American men are much less likely to be in the labor force now than they were 10, 20, or 30 years ago. Nobody has been able to pinpoint a single reason behind the decline; the disappearance of blue-collar factory jobs, mass incarceration, chronic pain, the growing use of Social Security Disability Insurance, and even video games have been identified as some of the possible culprits (don’t get me started on that last one). But the fact is that there’s a long-term trend away from work among U.S. males, and many of them don’t even claim to want a job. A measure like U-6 won’t reflect that group. You need something wider.

Trump himself seems inclined to go as big as possible—he at least once suggested that the real unemployment rate could be as high as 42 percent, because there were more than 93 million adults without a job (it’s now up to 95 million). Implicitly, he’s arguing that we should use the overall employment-to-population ratio as our labor market barometer. That doesn’t make a lot of sense. The raw employment rate is bound to decline as more Americans age into retirement, and there’s really no point in treating grandmothers living out their golden years in Boca as if they’re among the unemployed.

That’s why prime-age EPOP might be a good alternative. It gives us a raw look at the employment rate, without any fancy caveats about who is and isn’t part of the labor force. And because it only tracks workers 25 to 54, it isn’t really distorted by the wave of retiring boomers or growing college attendance. It’s a simple snapshot of the portion of the population we most need to worry about. And if you want to get really academic about it, using prime-age EPOP might even save the Phillips curve—the relationship between unemployment and wages (or inflation) that seemed to have broken down in recent years.

Best of all, from Trump’s perspective at least, prime-age EPOP has plenty of room for improvement. Currently, the rate sits at 78.2 percent, below its pre-recession high of 80.2 percent, and far beneath its tech-boom peak of 81.9 percent. If Trump wants to argue that Obama left him an economy that was still hurting, this is one stat that will easily help make the case .

Prime-age EPOP has its own downsides, as far as econ stats go. While it isn’t affected by our growing elderly population, it is influenced by aging within the 25-to-54 cohort; as younger workers move toward late middle age, they become less likely to be employed. That’s why former Treasury economist Ernie Tedeschi likes to produce a “demographically chained” employment rate, which tracks EPOP as if the nation’s age ratios were totally constant. It also doesn’t tell us much about wages, or whether people are stuck working part time when they’d rather have 40 hours a week.

But that’s just to say that prime-age EPOP isn’t a perfect summary statistic for the whole economy—and there’s really no such thing that exists. Economists tend to look at a whole array of stats to tell them where labor markets and growth stand. Fed Chair Janet Yellen has a dashboard of nine different numbers she prefers. And in his own, ineloquent way, Spicer on Monday tried to suggest the Trump team would take a similar approach.

So why harp on the issue of which economic indicator Trump should favor? First off, while maintaining an intricate dashboard of stats is a perfectly sound way for a technocrat to monitor the economy, it’s pretty much rubbish for communicating with the public. Presidential administrations owe it to voters to explain how they think the country is faring economically, and offering a few, easy-to-grasp numbers helps do that. More importantly, after spending a campaign attempting to erode faith among his supporters in basic economic data, Trump ought to rebuild some of that trust by saying which stats he does believe in. And prime-age EPOP, which he could just call the “working-age employment rate” for simplicity’s sake, would be a good place to start.

Unless, of course, his goal really is just to keep sowing confusion and distrust. In that case, he can just keep up the good work.