What is education actually for? What is its value? These may seem like silly questions with obvious answers. But the country is in the midst of a crisis that calls the "obvious" answers into question. We've saddled Americans with an intolerable $1.5 trillion student debt load, but it hasn't delivered the promised income benefits, and Democratic presidential candidates are seriously debating overhauling the entire system.

The debate over education often boils down to a duel between two theories: the "human capital" theory of education versus the "signaling" theory.

The "human capital" theory remains the dominant one in American policymaking and politics: Education gives people knowledge, skills, and social habits that make them more valuable to employers, thus they make higher wages. It's the key that unlocks upward mobility. Hence the desire by much of the center-right and center-left to move as many Americans as possible through the higher education system.

The "signaling" theory says something quite different: that what education does is make employers more likely to hire someone and pay them more than another person with less education. If you imagine job applicants as a line with the most coveted people at the front, the more education you have the closer to the front you are.

There's a lot of evidence that the signaling explanation is more true than not. Graduating with a college degree gets you a lot more pay than finishing three years of undergrad with no degree, for instance. This suggests it's the credential employers are looking for, as opposed to the real underlying abilities which you would have mostly gained in those first three years. Job postings that used to not require a college degree now often do, even though it's hard to imagine being a secretary requires more "human capital" now than it did 40 years ago. The wage premium — the additional money people with college degrees get over those without — basically hasn't grown at all since 2000, which is the opposite of what you'd expect in an economy supposedly hungry for more skilled workers. In fact, for the bottom two-thirds of earners with college degrees, wages are actually lower now than they were in 2000.

Now, these two explanations aren't necessarily mutually exclusive, and the reality is probably a mix of the two. But if the signaling theory is the dominant explanation, that matters. The human capital theory is positive-sum: the more education everyone gets, the more income everyone makes. The signaling theory is zero-sum: Everyone can't be first in line, so giving everybody more education doesn't actually increase real economic incomes.

Plenty of this evidence in favor of the signaling story comes from left-wing analysts and think tanks. But the big champions of the signaling story tend to be conservatives. As opposed to the left, which wants to make college basically free for everyone, people on the right recommend that the government pull back even more from investing in higher education — stop giving grants to universities, stop supporting student borrowing. They argue that all these policies do is encourage a never-ending arms race over education credentials and pump ever more money into inflating the cost of a college degree. And like I said, a lot of the evidence is prima facie on their side.

But there's another confounding variable to contend with here, and that's the relative power workers and employers have in the labor market.

If that market is a sellers' market — if there are constantly more jobs on offer than there are unemployed Americans looking to get hired — then workers have the leverage. Employers basically have to hire all comers and pay them well regardless of their education level, because businesses are constantly desperate for the labor they need. If the people they hire don't have the desired skills or know-how, the employer will eat the cost and inconvenience of training them or allowing them to learn on the job.

Conversely, if the labor market is a buyers' market, then employers can be picky and use educational degrees as a convenient way to sort applicants — leaving Americans to claw over each other to get ever more credentials, so they'll stand out in the scrabble for an insufficient number of good-paying jobs. In fact, research actually shows employers' demands for education going up in the immediate aftermath of the Great Recession, then falling again as the economy recovers. It's also pretty clear that the U.S. jobs market has very much been a buyers' market for a long time, well before 2008. Indeed, most of the evidence for the dominance of the signaling story is also the evidence that employers hold all the cards — their power is why signaling matters as much as it does.

But this actually undermines supporters of the "human capital" story and the "signaling" story alike.

For the former, it means that, on an individual level, how much human capital a worker has simply can't be disentangled from how much bargaining power they have — both things are measured the same way, by how much an employer is willing to pay you. If "human capital" has any use as a concept, it's only on the national level, in how much know-how we all have together as a cooperative society.

For the signaling story, it means that neither making higher education free, as several progressive Democratic presidential candidates have proposed, nor removing existing government aid for higher education, as conservatives would prefer, will likely end the arms race. What ultimately drives the arms race is the imbalance of power between employers and workers — and that remains unchanged. (And withdrawing aid entirely will likely just make the arms race even more financially punishing for American households.)

Ironically, the ultimate fix for the education crisis may not lie in education policy at all, but in macroeconomic and labor market policy: reforming fiscal and monetary policy to prioritize full employment and rebuilding unions; shortening the work week, strengthening over-time laws, providing more paid leave, and expanding the social safety net, so that more Americans can comfortably stay out of the labor force, and thus shrink the supply of workers relative to the demand for them. All that will return power to workers and make the labor market a sellers' market again. That will force employers to de-prioritize credentials in their hiring, and start taking all comers.

That does still leave the question though: Should we make college a free, universally available public good, the same way K-12 already is? Sure! But not because it increases anyone's "human capital" or "prepares them for the modern workforce."

We should do it because education is a good thing that can make one's life richer. It is an end in itself.