Help Wanted

If the federal government can’t accurately forecast which jobs will grow, can anyone do it?

A makeup artist. Click image to expand
Has the BLS caused a boom in beauticians?

If the Bureau of Labor Statistics could predict the future, makeup artistry might look like a very good career right now. The government agency—perhaps best known for publishing the nation’s unemployment rate every month—released a forecast late last year predicting which jobs would grow fastest over the next decade. Once it crunched all the numbers, the bureau reported that makeup artistry would be the seventh fastest-growing occupation in America from 2006 to 2016, between personal financial advisers and medical assistants. Overall, the bureau said, the number of jobs in the field would grow by a whopping 39.8 percent over the decade.

Soon enough, makeup schools were advertising opportunities to get trained in one of the “hottest careers” in the country. Stories from the New York Times to noted that makeup artists made the BLS’ fastest-growing list.

The problem: There are only about 2,000 makeup artists in the country right now. Even if the field expands at the dramatic pace the BLS projects, that growth will only translate into a few hundred job openings over the entire decade. There may soon be a lot of people singing “Beauty School Dropout.”

Despite such potential for misinterpretation, the BLS occupational projections (PDF) are almost certainly the most influential economic statistics you’ve never heard of. They appear just about anywhere you might look for information about your next job: on Web sites like, at local job-search agencies, in high-school guidance offices and college career centers. States use the national data to produce their own estimates, helping job seekers from Alabama to Washington figure out where the hot jobs are in their communities.

The demand for these numbers is obvious: Everyone looking to start a career or switch jobs wants some confirmation that they aren’t entering a dying industry. And the projections—which are issued every two years—provide the kind of easy lists that are the bread and butter of your local newspaper’s “Jobs” section.

The only problem is that projecting exactly which jobs will grow and which will shrink is a nearly impossible task. And, sure enough, the BLS is often wrong.

Take the late 1980s projections for the year 2000. According to an analysis published by the bureau (PDF), among the job categories predicted to expand at a “much faster than average” clip, about one-quarter either fell below the average growth rate or even shrank. Most egregiously, the bureau projected that the number of travel agents would grow by 54 percent—failing to anticipate that the rise of Priceline and Expedia would actually cause the industry to shrink. Data-processing equipment repairers and medical secretaries had a much worse decade than the bureau predicted; welders and cable installers did much better.

Indeed, the bureau has long struggled to get the bigger picture right. “The BLS projected a strong market for college graduates in 1970, just prior to the bust of the 1970s,” Cornell economist John Bishop wrote in a 1996 analysis. “The BLS projected a weak market in 1980, just prior to the 1980s boom.” More recent projections are a little harder to evaluate because of changes to how the BLS defines different job categories, but its outlook for the year 2008 (PDF) appears to have been far too optimistic for desktop publishing specialists and semiconductor processors. As the bureau predicted, demand in these fields is growing—but new technology and offshoring have limited the number of new jobs.

It’s not the BLS’ fault. Economists can’t usually agree on where the overall economy will stand a year from now. Projecting 10 years into the future—and in 750 different occupational categories—is far more difficult. The methods used by the BLS are intuitive: Its economists first try to project the overall growth in the economy and the labor force and then take historical data and a few assumptions about the future to try to model how fast individual industries will grow. Using a set of equations that predicts the relationship between the size of an industry and the number of people it employs, the economists are able to forecast new jobs within each individual job category.

For the BLS to succeed, it has to get each of these individual steps right. It’s not enough to know that more people are going to need help when their PCs crash; the model also has to figure out whether that work will be done here or abroad and how much of it will be automated. James Franklin, a BLS economist who oversees part of the process, says that the forecasters try to anticipate these changes. But, in the end, they aren’t clairvoyant. An analysis by outside economists found that the 2000 projections didn’t do much better than a “naive” model that simply projected 1988 data forward.

The BLS is very open in pointing out that its projections are uncertain, and it deserves credit for producing a warts-and-all evaluation of whether its own predictions turned out to be accurate. But by the time the numbers make it into the hands of actual workers, all those caveats and warnings have often disappeared. The presence of a list—a government-sanctioned analysis that estimates growth rates down to one-tenth of a percentage point—makes the numbers seem exact, even when they’re not.

Maybe it’s better to have a hazy glimpse of the future than none at all, but there are dangers in suggesting that we know more about how the economy will change than we really do. Americans anxious about how to “recession-proof” their jobs or survive in a globalized economy are typically given a simple answer: Get more skills. It’s hard to challenge that prescription, but it also doesn’t mean that much when “higher education” includes everything from beauty school to Princeton. For a 19-year-old figuring out what to do after high school or a laid-off manufacturing worker choosing a new career, the question isn’t whether more education is necessary—it’s what kind of education they should get and where.

The problem is that education—like any investment—is risky. It costs both time and money to enroll in college or start a job-training program. And if you get stuck in a job that unexpectedly takes a nosedive, you may not receive the payoff you expected. You can hedge your bets in case your chosen field doesn’t work out—by, say, getting a more general education rather than a highly specialized course of training. But there are trade-offs to doing that, too; if you have good reason to believe that forensic science technicians are going to be in hot demand, it makes sense to spend all your time training to be a forensic science technician. Schools and colleges—who are measured much more by the students they bring in than what happens to them after they graduate—don’t always provide much help navigating these trade-offs. So job seekers are left trying to divine the future on their own—and projections like the BLS’ can make them overconfident about their ability to do so.

It’s unnerving to learn that there is no way to get an accurate picture of what the market will look like in 10 years, let alone 30. If there’s a silver lining, it’s that it may make sense to think more about other things—what you are good at or what employers are looking for right now—when deciding on a career. In other words, unless you’ve dreamed about applying mascara to the stars, don’t enroll in makeup school quite yet.