Coding boot camps are all the rage right now. These intensive, accelerated learning programs are like trade schools for the digital age, holding the promise of teaching previously untrained students the latest computer programming technologies and providing fast-track access to high-paying jobs in the computer industry.

Under the gun to respond to and feed innovations in skills training, the Department of Education is looking to experiment by extending federal aid eligibility to these boot camps (as well as other “partnerships between accredited colleges and alternative providers”), with the intent of promoting program growth, increasing access for low-income and working-class students, and confronting the computer industry’s appalling lack of diversity.

As great as this initiative sounds, the Department of Education should proceed with extreme caution, as history has shown that, whenever you open the federal financing spigot in the education sector, you risk being flooded with low-quality programs run by charlatans.

The Current Financing Pipeline

Students seeking to attend a boot camp have a lot to choose from, including what they have to pay. Some boot camps charge nothing upfront, but demand a portion of the participant’s first-year salary after graduation. Others charge traditional tuition that can be as high as $1,500 or more per week for day-and-night immersion programs that last for eight to sixteen weeks.

So far, none of these programs have taken the steps necessary to make their participants eligible for federal Pell Grants and student loans, so low-income students in particular have their program choices limited.

The conventional wisdom in Silicon Valley and in Washington, D.C., is that “innovative” training like these coding academies are excluded from federal aid by an accreditation “cartel”—the word used by senator and presidential candidate Marco Rubio—dominated by old-fashioned sector insiders that place a premium on campuses with grand libraries, sprawling lawns, and tweedy professors with the letters Ph.D. after their names.

Washington’s Program Assessment Challenge

Responding to the call for equal opportunity, the U.S. Department of Education’s undersecretary, Ted Mitchell, is seeking input on how to access the “vibrant marketplace for learning” that is emerging “outside of the traditional colleges and universities.” The primary idea? Shortcuts to federal aid that could be implemented on a trial basis.

Now, first off, it turns out that the conventional wisdom about the college cartel is wrong: there are, in fact, hundreds of short-term and long-term nontraditional programs in the federal system that are aimed at specific skills and careers. So this emphasis is not really all that new.

The real problem is that, like the conventional system of job training programs, Mitchell’s draft approach does little to ensure that the sector isn’t flooded with low-quality programs. The current excellence of some coding academies is no guarantee that they or their imitators will be worth the time and tuition after the sector expands.

Undersecretary Mitchell’s tentative plan is to insist that program outcomes are affirmed by assessments that are “reliable and valid.” Danger ahead! While it may sound reasonable to ask for evidence of learning that is statistically defensible, that is exactly the thinking that has led, in K–12 curricula, to coma-inducing standardized tests.

In coding academies, students prove their skills through unique, creative projects, usually as part of a team. To meet the test of statistical validity and reliability, coding boot camps will have to change their curricula in ways that will produce programmers who lack exactly what tech companies need: creativity. Indeed, any test that can be scored by a machine is, by definition, judging a skill that can be done by a computer. To get the high-paying jobs of the future, the students need to be doing what only a human can do (and cross their fingers that some future programmer won’t be creative enough to later make their jobs obsolete).

In conversation with Mitchell and his staff, I suggested that they offer an alternative. Rather than insisting on standardized tests, the Department of Education should have programs that want to use the shortcut to federal money make the actual projects completed by their graduates available for inspection. This approach would allow the programs to continue catering to changing industry needs without getting stuck with static, quickly outdated assessments. Crowdsourcing accountability would also let employers—and potential students—see the coding skills and creativity that a boot camp is capable of instilling in its participants, and would help to ensure the ongoing integrity of the program’s own quality control processes.

Beware the Profiteers

Opening up federal aid to coding academies creates another hazard because, since so many people are salivating for programming jobs, unscrupulous people might look at that aid simply as a bounty on the head of future students.

Anyone who wanted to take unfair advantage could follow a simple script: (1) Recruit in low-income neighborhoods where adults would qualify for $15,000 in grants and loans (“don’t worry about the money—we’re approved by the federal government because this program can lead to great jobs with companies like Google and Twitter”). (2) Spend only a couple thousand dollars per enrollee by using lesson plans available for free online, and scouring Craigslist for cheap space and temporary instructors. (3) Pocket a cool million in profit free and clear on every hundred students enrolled, even if the program’s outcomes are dismal.

To prevent that from happening, I recommended to Mitchell one additional accountability mechanism: require that programs enroll students whose tuition is being paid by employers or by the students themselves, not just those being financed by taxpayers. (This would be similar to the 90-10 rule currently applied to for-profit colleges.)

Customers who are digging into their own bank accounts have a built-in incentive to analyze whether the price is reasonable, given the alternatives. When everyone’s tuition is being covered by the GI Bill, student loans, and other government aid, it is too easy for schools to charge far more than appropriate. To ensure a market-determined price and continued quality while allowing growth, programs should be limited to no more aided students than the number who had paid without aid in the prior boot camp. This mechanism, forcing programs to continually demonstrate market viability, makes it difficult for them to grow crappy, overpriced—yet highly profitable—versions of a coding academy.

Over the past decade, through the GI Bill or student loan programs, hundreds of thousands of students have been recruited into high-cost, low-reward programs for which they are now paying the price. It is worth remembering that the for-profit college scandal, which is still in the process of being cleaned up, began as a noble effort to allow companies to gain access to federal funds only if they ran innovative training programs that led to good jobs. We must be careful that, in opening federal aid to coding boot camps, we do not let that happen again.