Is American Economic Growth Over? (Thoughts on Robert Gordon's New Book)

dollar growth diagram
dollar growth diagram

Robert J. Gordon, an economist at Northwestern University has recently published an important new book, The Rise and Fall of American Growth, which argues that the U.S. has entered a new age of stagnation. Paul Krugman has a good review of the book here.

While Gordon's argument is often characterized as being the opposite of the argument I have made in my two books about the impact of advancing automation technology on the job market (most recently, Rise of the Robots), there are many areas in which I think we would agree.

Gordon's main point is that we no longer have the kind of robust, broad-based innovation that powered economic growth and rising living standards between roughly 1870 and 1970. Electricity, cars, planes, indoor plumbing and all the rest completely changed the quality of our lives during this period, and we haven't seen anything comparable in the decades since. This is certainly true.  However, I think it's clear that innovation since then has continued (and even accelerated) but has focused largely in the information and communication technology arenas. This has not had the same impact on median incomes and living standards, and I think one important reason is that information technology is increasingly substituting for cognitive human labor.

There seems to be a fundamental assumption underlying Gordon's analysis: IF ONLY we could once again have that broad-based, robust innovation, then nearly everyone would be better off, and we would again see real incomes for average workers once again rising (just like pre-1970). I think most economists would probably buy into this assumption.

I think that assumption is wrong. It's wrong because information technology (and specifically artificial intelligence) is going to intertwine with any innovations that occur in the future and make everything less labor intensive. Unless we change our economic rules (perhaps with something like a guaranteed income), broad-based prosperity will remain elusive even if those robust innovations do eventually show up. The innovations may come, but the people at the top of the income distribution will continue to capture nearly all the gains.

The key insight is that those 1870-1970 innovations were all powerful job creators, and in the market economy jobs are the only mechanism that distributes income to the bulk of the population. (Think of the millions of solid middle class jobs created by the rise of the automotive industry--manufacturing, driving, repairing, fueling, insuring, and even washing, cars and trucks).  The innovations of the future--regardless of how broad and disruptive they may be--are very unlikely to create that number of jobs, and the jobs that are created may well require skills and education beyond the capability of the average worker.

The chart below shows how compensation for average workers and productivity have decoupled since the mid-1970s. Suppose we suddenly have broad, 1900 to 1950-like innovation. Would that cause these two lines to converge?  I doubt it.

I think that to some extent, Gordon may have missed the real story here. The most important insight is not that we no longer have broad-based innovation. It's that innovation no longer guarantees broad-based prosperity. To solve that problem, I think we will ultimately have to change our economic rules.