American Compass has a new survey out in which it finds, among other results, that “only 40 percent of workers have secure jobs.” This is the latest attempt by the outfit to portray the American economy as in dire need of “rebuilding.” The report summarizing the findings is titled, “Labor Market Not Yet Working for Workers.” Like American Compass’s previous attempts at empirical analysis, this one is far too doomerist.

American Compass looks at adults ages 18-65 working at least 20 hours per week. It creates a definition of a “secure” job based on five criteria. The first criterion is that a job pay at least $40,000 per year, according to the survey respondent’s estimation. This threshold seems completely arbitrary; move it up or down and the share “secure” changes accordingly. It also puts a floor on the share of workers deemed insecure. American Compass’s survey indicates that 69 percent of workers meet this criterion, so the share of American workers who are “insecure” must be no lower than 31 percent.

We can use the Current Population Survey’s Annual Social and Economic Supplement to assess the American Compass numbers. I find that 64 percent of adults ages 18-65 with earnings who typically worked at least 20 hours a week earned at least $40,000 in 2022. Not too different from American Compass.

Incidentally, the median female worker in this group earned $45,000. American Compass has selected an initial criterion for job security that almost ensures that over half of women workers will be deemed insecure once other criteria are considered.

The second criterion for a job being secure is that health insurance “is available to” a worker “through your job.” This is true of 67 percent of workers according to the American Compass survey. In the CPS, we have to get a little creative to try and replicate this figure. We know how many workers were covered by their employer in the previous year, and we know how many were not covered by their employer at the time they were interviewed despite coverage being available to them. I find that 70 percent of workers fall into this group.

American Compass’s third criterion for job security is that the job provide “Paid time off (PTO) or vacation” or “Paid sick time.” It finds that 61 percent are in jobs with these benefits. Unfortunately, the CPS lacks information on these benefits.

The fourth criterion used by American Compass is problematic. Respondents are asked, “How confident are you in your ability to estimate your earnings over the next twelve months?” For their job to be secure, they have to respond “very” or “somewhat” confident—answers that were given by 85 percent of workers. American Compass interprets such a response as meaning the job “allows for confidence in future earnings,” meaning it is a steady and stable position. But the question itself is about the “ability” to “estimate” future earnings. Many respondents were likely to interpret it as a self-assessment of their own predictive skills, not about attributes of the job. Even respondents who did answer the question thinking it was about job security, might have been thinking more about macroeconomic conditions than about what their job “allows for.” American Compass should have used that wording in the question if that’s what it wanted to elicit.

The final criterion for assessing whether a worker has a secure job is their indicating either that “I have a regular work schedule that is generally the same from week to week” or that “My work schedule varies from week to week, but I am satisfied with my control over it.” American Compass does not provide the alternative responses workers could give, but 96 percent of them gave one of these two answers. It is striking that nearly all workers (at least those working 20 or more hours a week) either have a stable work schedule or are OK with their varying schedule. In this light, the issue of “just-in-time scheduling”—where workers have little advanced notice about their work shifts—seems grossly overblown by many commentators. (Here’s a piece on the American Compass website advocating that the practice be “outlawed.”)

Take all of those requirements into account, and just 40 percent are in secure jobs. We are to believe that only 55 percent of four-year college graduates are in secure jobs. What should we make of these findings?

Nearly everyone is secure by the last of the five criteria, so we can ignore it safely. The fourth criterion should be dropped, as it is too ambiguous. I do not have information on paid time off in my CPS data. So let us look at job security according to the first two criteria used by American Compass.

In the CPS, 52 percent of employed adults ages 18-65 are “secure” in the sense that they earned at least $40,000 and either were covered by their employer’s health insurance last year or were not covered by their employer when interviewed even though they could have been. That is higher than American Compass’s estimate of 40 percent using all five criteria. My estimate of job security would be at least a bit lower If I could add the other three criteria to my definition. American Compass does not provide the share of the population that is “secure” by these first two criteria.

But we should not stop here. Some people make less than $40,000 for reasons that American Compass has no way of discerning but that are not worrisome. The same is true of not being covered by employer health insurance.

For instance, I may forego my employer’s health insurance if I am already covered under my spouse’s employer. That would increase my take-home pay. Turning again to the CPS, when I count as “secure” workers who make at least $40,000, are not covered by or eligible for their employer health plan, but are covered by some family member’s employer health plan, 56 percent are secure rather than 52 percent.

Similarly, I may take a job that pays less than $40,000 annually if my spouse has a high-paying job. Thanks to my spouse’s salary, I might work fewer hours, work only part of the year, or take a job with a lower hourly rate or salary. When I define as “secure” anyone who is covered by some employer (or eligible for their employer’s coverage) and who either makes $40,000 or is in a family with $80,000 in total earnings, then 64 percent of workers are secure. By this measure, 81 percent of those with a bachelor’s degree are (versus American Compass’s 55 percent).

By this two-criterion definition, then, 36 percent of workers lack job security. Comparing this to the 48 percent found when I try to mimic American Compass’s first two criteria, it looks like the group overstates job insecurity by about one-third.

Again, if I could add the paid time off criterion, my estimate of job security would decline from 64 percent. But the argument above applies here too: some people take jobs lacking these benefits for reasons that aren’t worrisome.

American Compass has produced another analysis that implicitly has a sole breadwinner family in mind as the norm. I’ve explored this view before, including with Jeremy Horpedahl. The group’s commitment to the sole breadwinner norm and the idea that all families without a sole breadwinner want one (or their indifference to what families want) continues to drive a policy agenda based on an overly declensionist analysis.

Don’t get me wrong: if 95 percent of workers had job security, I would still want to help the remaining 5 percent. But policy has to be driven by accurate empirical conclusions. American capitalism does not require “rebuilding” and with earnings at near-record highs and unemployment exceptionally low, the labor market is working pretty well. Our economic challenges are less pervasive than American Compass thinks and we need to focus like a laser on helping more people benefit from what remains a remarkable economy.