I have a lot of respect for Ro Khanna. He is an articulate and energetic progressive from the belly of the beast, Silicon Valley. I also appreciate the sentiments behind his latest op-ed [2] in the NYT on how we should seek to ensure that the benefits from AI are broadly shared. However, he badly errors in his framing of the issue and description of recent economic policy.
Khanna describes our policy on trade in the 1990s:
“Today the Democratic Party is at a crossroads, as it was in the 1990s, when the dominant wing in the party argued for prioritizing private-sector growth and letting the chips fall where they may. The criticism of this approach offered around that time by Senator Paul Wellstone, Senator Russ Feingold and Representative Bernie Sanders (as he was then) — that the offshoring globalization debacle was not helping the working class and was, in fact, hurting it — was largely ignored.”
It is wrong to say that our trade policy was “letting the chips fall where they may.” It was in fact far worse.
The policy was to deliberately subject manufacturing workers to competition with the low paid workers in the developing world. This had the predicted and actual effect of eliminating millions of manufacturing jobs and lowering the wages of manufacturing workers in the jobs that remained. Furthermore, since manufacturing had historically been a source of high-paid employment for workers without college degrees, this put downward pressure on the pay of less-educated workers more generally.
At the same time that we were using trade policy to depress the wages of less-educated workers, we continued to leave protections in place for more highly educated workers like doctors and dentists. We also increased protections in the form of patent and copyright monopolies, which we made stronger and longer both in domestic policy and internationally through trade agreements. This raised the profits and the wages of those in a position to benefit from these monopolies.
It is a gross misrepresentation to say that the government policy in this period was just leaving things to the market, it was a deliberate policy of upward redistribution. It is also important to recognize that most of the upward redistribution was within the wage distribution, not to profits. The profit share of national income changed little from 1980 to 2000, at which point most of the upward redistribution had already taken place.
The big gainers from the upward redistribution were high-end professionals, well-placed STEM workers, Wall Street types, and high-level corporate executives. We need to recognize this fact if we want to structure the benefits from AI in ways that lead to broadly based gains.
AI does offer enormous potential in this way. For example, AI could allow a trained healthcare professional, like a nurse practitioner or a physician’s assistant, to make diagnoses as well or better [3] than doctors. AI could also replace many of the higher paid workers in various STEM professions, like drug development. If we don’t allow research findings to be bottled up with government-granted patent rights or other protections, AI could benefit the bulk of the population by giving us lower cost health care and legal services, in addition to savings in many other areas.
The key point here is to recognize that policy has always consciously directed trade and technology. It is to the benefit of the winners to pretend that the upward redistribution of the last four decades was “letting the chips fall where they may,” but that is not the reality. And, they won’t let you say that in the New York Times.