David Autor and David Dorn had an op-ed in the NYT this weekend, outlining their case that technology has been the primary headwind for middle-class living standards over the past generation. We tend to have a different view; we think that the cumulative effect of economic policy decisions made over this time-span has been the real barrier to decent growth in middle-class living standards.

The first question lots of people have upon hearing this dispute is “does it actually matter?” Autor and Dorn accept (and document) just how rough a go it has been for middle-wage/middle-income workers and clearly accept that the rise in inequality that has seen rewards flow away from the middle-class is a problem that needs to be rectified. Who cares why they think it happened if they’re on-board with attempts to fix it?

The answer to this is simple and important: the diagnosis matters because it implies a prescription. In Autor and Dorn’s NYT piece, the only real policy solution they nod to is boosting the share of workers with higher education, noting that “the payoff for college and professional degrees has soared.” And if one believes that technology is boosting demand for high-skill workers and that’s why middle-wage jobs are suffering, this policy solution does make sense. But in fact, the payoff to a college degree has actually been pretty flat for more than a decade now. Wages for those with an advanced degree have done better, but this group is just over 10 percent of the workforce—which is a key reason why Autor and Dorn note that boosting the share of workers with college or advanced degrees can’t be a “comprehensive solution to our labor market problems.”

Conversely, we would argue that institutional factors like the eroding value of the minimum wage as a labor standard and eroding protections for willing workers that want to form a union are more important drivers of inequality and disappointing middle-class living standards over the past generation. This implies a different policy priority—raising minimum wages and changing the legal landscape to protect workers that want to form a union and bargain collectively.

The problem with analyses (like Autor and Dorn’s) that greatly de-emphasize these solutions is not just that they lead us to different solutions that will be less effective in combating inequality, but they lead us to much different and much more expensive solutions that will be less effective. As I’ve written before:

…boosting the share of workers with a 4-year college degree is indeed a heavy economic lift—it takes real resources (books, labs, classrooms and most expensively teachers) to provide the skills and education needed to qualify for a college degree. But boosting the share of workers with union representation really doesn’t cost much at all—there is really no serious research linking economy-wide productivity declines to increased unionization. Instead, boosting the share of union workers in the U.S. would redistribute money, but would not cost the U.S. economy anything in the aggregate. And given that so much of the decline in unionization seems to be policy-driven (PDF), the real lever to make this increase happen is essentially the costless act of changing the policy stance towards unionization (there is no CBO score, for example, for the Employee Free Choice Act because it doesn’t cost anything)…

And yet you still find many policymakers and DC think-tank types willing to salute the need to boost the share of U.S. workers with a college degree than willing to publicly say we need to make it easier for American workers to unionize—even though the college commitment is far more expensive in economics terms. Further, as Janelle Jones and John Schmitt have shown in an excellent paper, the expected gain in job quality that would stem from boosting the share of U.S. workers represented by a union far exceeds the gain that would stem from boosting the share of U.S. workers with a 4-year college degree.

Given how much cheaper (in actual economic costs) it is to change policy to boost workers’ bargaining power, it seems odd at first how little enthusiasm there is to make this the first priority for those concerned about middle-class troubles: after all, shouldn’t the low-hanging fruit at least be picked first?

But of course it’s not that odd – because boosting workers’ bargaining power would redistribute income away from those being served very well by today’s status quo, even if it’s cheap in overall economic terms, it’s a very heavy political lift to try. And as often happens in today’s United States, politics trumps economics.