Technology and the Crushing of Labor

I don’t have the technical knowledge to evaluate this claim but it certainly feels right that the declining share of income going to labor and the implementation of technology are closely connected and that AI will make this all the worse.

Given that diagnosis, the prescription might seem to write itself. Senators Bernie Sanders and Elizabeth Warren and others on the left argue for breaking up large firms, particularly in tech. So do Senator Josh Hawley, Vice President JD Vance and others on the right. The impulse to rein in corporate giants isn’t limited to the United States. Regulators in Europe, Japan and Britain are also supercharging their competition policy and stepping up enforcement.

That may be the wrong solution, because it’s not rising market power, but new technology — A.I. being the latest example — that is probably the most important force reducing the labor share. Breaking up big companies won’t help restore labor’s share and may not help workers at all. Designing policies for a change this significant requires understanding the real cause.

A little over a decade ago, Loukas Karabarbounis and I documented a global decline in the labor share. Labor shares stood well below their historical values in the United States and in most places we looked. That included advanced economies such as Germany and Japan where labor laws or unions were strong, and developing countries where they were not, such as China and India. Since the decline was generally largest where the price of computing and communications equipment fell most, we concluded that replacing workers with machines in production — substituting capital for labor — was bringing the labor share down.

…

If the labor share is falling because of a decline in competition, antimonopoly policy would indeed be the right medicine. But if, as I believe is likely, A.I. is the key force that continues to drive the labor share lower, our focus should be elsewhere. Stronger policies that cushion job losses and speed up retraining will be essential.

We will also need to reduce the bias toward capital income in the tax code. For example, firms that immediately write off the expense of machine purchases shouldn’t be allowed to fully deduct the interest on the loans they used to fund them. And capital gains on investments shouldn’t escape taxes simply because those investments are passed on to heirs. Oddities such as these lower the cost of capital, which further encourages the shift away from labor.

The overall scale of the social safety net could be indexed to trends in the labor share — expanding the net as labor share falls. Taxpayer-funded accounts, seeded at birth and invested in broad equity indexes, would also help. The new Trump Accounts move in this direction, but to narrow wealth gaps and limit strain on the budget, the deposits should flow to children in the neediest families.

I do have a lot of issues with the nonsense “solutions” offered here. Taxpayer funded accounts are extremely unlikely to make a difference unless taxation of wealthy strips the advantages in income they have taken in recent decades and that then gets into the offshoring of wealth and all that. Retraining is always complete nonsense–retraining for what? Learn to code baby! If AI is going to strip wealth, then it is going to come for whatever we are being retrained to do. And what I am, a 52 year old professor, going to be retrained to do exactly? Never mind someone who is a 32 year old computer programmer with zero transferrable skills since all they learned in college was a now completely irrelevant computer science degree or a 45 year old worker on the margins with little education or skills at all? Concrete examples are required!

Still, the connection between technology and wealth at least feels very real to me. And if this is true, it’s going to cause upheavals in our politics that make the Trump years look like chicken feed.