July 17, 2024
Leading media outlets like The New York Times, The Atlantic, and The New Yorker have about as much concern for intellectual consistency as TikTok videos. In very serious and somber tones they will warn the rest of us about a major problem and then in the next issue, or the next article, present a story that is 180 degrees at odds without ever realizing the contradiction.
My favorite example of this “Which way is up?” problem is the simultaneous concern expressed that AI will eliminate all the jobs and that declining birth rates will lead to worker shortages. We got a taste of the former in an NYT article about efforts by OpenAI’s CEO Sam Altman to promote guaranteed income programs.
According to the piece, Altman and other tech types who share his views, are concerned that AI will lead to mass unemployment and further increase the gap between rich and poor. While this concern is voiced frequently in elite publications it is 180 degrees at odds with almost everything else they write.
First and foremost, if AI is going to lead to massive job loss there is zero reason to worry that declining birth rates will lead to a shrinking workforce. If we don’t have any jobs, why would we need more workers?
This issue also comes up with immigration. We are regularly told that we need immigrants because we have a shortage of native-born workers. But if AI is going to eliminate all the jobs, we won’t need large-scale immigration. To be clear, there are other reasons we might want immigration, but we can’t believe both. AI will lead to a massive loss of jobs and we also need immigrants to do our work.
The idea that AI will lead to massive job loss also means that productivity growth, and therefore economic growth, will be far more rapid than current projections indicate. The latest projections from the Congressional Budget Office show productivity growth averaging a bit over 1.6 percent annually over the next decade.
It’s not clear how fast the AI-will-kill-all-the-jobs gang expect productivity to grow, but we had productivity growth averaging just under 3.0 percent annually from 1947 to 1973, and we still maintained low rates of unemployment through most of this period. Clearly, they must envision considerably faster productivity growth, perhaps 5 or 6 percent annually.
If we had productivity growth anywhere near this fast, many other widely hyped concerns would quickly disappear. For example, the widely touted national debt would look pretty inconsequential if we could sustain anywhere near 5.0 percent productivity growth over a decade. (That would be the case even with 3.0 percent productivity growth.)
Also, the projected shortfalls in Social Security would be eliminated if average wages rose anywhere near as fast as productivity, as has largely been the case in the past. (It’s true that the median wage has not kept pace with productivity growth, but that is mostly due to the upward redistribution of wealth to highly paid workers like CEOs, Wall Street types, and well-placed STEM workers, not a failure of average wages to keep pace with productivity.)
We also would have little reason to be concerned about inflation if we’re seeing the sort of extraordinary productivity gains anticipated by the AI proselytizers. As a first approximation, inflation is equal to wage growth minus productivity growth. If we have productivity growth of 5.0 percent annually, we can have 7.0 percent nominal wage growth and still be at the Fed’s 2.0 percent inflation target. That pace would be more rapid than the peaks we saw in 2021–2022 when we were bouncing back from the pandemic.
In short, if the worldview being hyped by the AI folks in this piece is anywhere close to accurate, many of the concerns expressed in public discussions are completely off the mark. Either the AI gang is wrong or many other people writing in elite publications are wasting their time.
Is There a Case for a Guaranteed Income?
Since people often ask me, I might as well throw in a few words. I find the basic question sort of silly since it rarely is posed in the real-world context of being a question of trade-offs. (This is assuming that the AI gang is wrong in thinking that AI will take all the jobs.) Like many people, I would say that abstractly a guaranteed income would be a good idea, but in a world where millions of parents can’t afford child care, and many have inadequate housing and health care, I would have a hard time arguing that a guaranteed income should be a top priority.
A guaranteed income makes more sense in countries that have more fully developed welfare states. But $1,000 a month, or whatever is picked as the income target, will not pay for people’s health care, food, housing, and their kids’ child care. I would place the priority on ensuring that everyone has access to these necessities.
I’ll also add my other standard line. The massive inequality we see in America today is not simply due to market outcomes, rather it is the result of how we have chosen to structure the market.
The most important issue here is government-granted patent and copyright monopolies. These redistribute over $1 trillion a year from the rest of us to people in a position to benefit from these government-granted monopolies. This is seen most clearly with people like Bill Gates, who might still be working for a living if he couldn’t count on the government arresting people who used Microsoft’s software without his permission. While it is possible to argue for the merits of patent and copyright laws as policies for promoting innovation and creative work, the point is that they are policies, not intrinsic features of the market.
I realize that people who write for elite outlets don’t like to acknowledge the fact that inequality is the result of how we have chosen to structure the market, not the result of the natural workings of the market. This allows them to pontificate about how the winners should best toss a few breadcrumbs to the losers, as the AI gang seems to be doing. It avoids the question of whether we should not have rigged the market in a way to produce such extreme inequality in the first place. (Yes, this is my book, Rigged [it’s free].)
Comments