“The number of jobs lost to more efficient machines is only part of the problem…automation may prevent the economy from creating enough new jobs… But automation is beginning to move in and eliminate office jobs too… In the past, new industries hired far more people than those they put out of business. But this is not true of many of today’s new industries… Today’s new industries have comparatively few jobs.”
This is a pretty standard example of today’s worries, particularly the part about today’s new industries being different than in the past (think of the tech sector). Problem is that quote is from Time Magazine in 1961. It’s the same exact “this time is different” argument, except that we have the hindsight to know it was wrong. I vaguely remembered that Keynes wrote on this same subject so I went back and looked. Sure enough, in 1930 Keynes wrote (in an essay titled The Economic Possibilities for our Grandchildren):
“[I]n our own lifetimes…we may be able to perform all the operations of agriculture, mining, and manufacture with a quarter of the human effort to which we have been accustomed… the very rapidity of these changes is…bringing difficult problems to solve…namely…unemployment due to our discovery of means of economizing the use of labor outrunning the pace at which we can find new uses.”
It’s the same argument! Only by 1961 we had the hindsight to know it was wrong. Keynes correctly identified it as “only a temporary phase of maladjustment.” Since the 1960s (or 1930s for that matter), against the backdrop of higher productivity and women entering the labor force, employment has continued to grow. The employment-population ratio rose over the rest of the 20th century (it is now back to where it was in the mid-1980s due the recessions of this century and demographic change).
To address such concerns, the Johnson Administration formed a Commission on Technology, Automation, and Economic Progress. They hit the nail on the head:
“Thus technological change…is an important determinant of the precise places, industries, and people affected by unemployment. But the general level of demand for goods and services is by far the most important factor determining how many are affected…and how hard it is…to find jobs. The basic fact is that technology eliminates jobs, not work”.
Technology eliminates some jobs, but compliments others, increasing demand for such labor. And the typical job contains a mix of tasks, some of which are easily automated and others not. More generally, technology increases productivity, which means lower labor cost per unit of output. This will translate into a combination of increased production (because of lower costs), higher real incomes (because of lower real prices), and higher consumption (because of higher real incomes). All of which increase demand for labor. This is why the decline in agricultural employment in the US, from 41% of the labor force in 1900 to 2% by 2000 didn’t result in mass unemployment[3].
A frequent response to this simple, comforting story is that the costs and gains do not accrue evenly. Indeed technological innovation does cause inequality, particularly because cutting edge technologies are most expensive when new. In a recent paper, David Autor presents evidence that such worries are at least over-simplifications. While automation eliminates some low skilled jobs, robots cannot do everything a human can even at the bottom end of the skill distribution. As a result, automation increases the productivity of jobs at all skill levels. The rising incomes that result will cause people to consume more low skilled services.
Autor finds that over time employment increases across the skill spectrum typically have a U shape, though not always. That is, gains have been disproportionately at the ends of the spectrum[4]. However, because lower skilled jobs have lower barriers to entry (such as educational or licensing requirements), an increase in wages due to higher demand causes an increase in the quantity of labor supplied, subduing wage gains vs jobs with higher barriers to entry. Autor theorizes that middle skilled jobs have not seen as rapid growth because technological improvements are allowing automation to creep further up the skill ladder.
So the inevitable march of technology will not hit everyone equally, but it isn’t eliminating low skilled jobs opportunities generally. Some will lose while a majority benefit on net. The Luddites advocated destroying and banning technologies that threatened their livelihoods. Or, put another way, reducing the economy’s productive capacity, thus impoverishing wider society, in order to protect one class of labor from adjustment and competition. Surely there are better ways of helping the minority who lose out due to automation.
As Herbert Simon, a great economist and Nobel laureate, put it in the 1960s, “the world’s problems in this generation and the next are problems of scarcity, not of intolerable abundance. The bogeyman of automation consumes worrying capacity that should be saved for real problems.”
Part II
1. I think I used that correctly but someone let me know if I didn’t. If anyone reads this that is.
2. Classic economist “on the other hand” hedge.
Another example of this tendency is the “Secular Stagnation Hypothesis” credited to Alvin Hansen, who postulated it in 1938 (although Keynes’s General Theory of 1936 has similar ideas, specifically that the economy may lack a self-correcting mechanism to return to its previous growth path). Basically, the theory is that, unlike in any previous recovery, savings will continue to be in excess of investment for some mysterious "secular" reason (as in they can’t come up with a real theory to explain it), the result is that the economy will chronically fail to reach its potential / full employment (this is a gross generalization stemming from my distain for the theory in case you didn’t notice). Hansen’s theory looks spectacularly wrong in hindsight. It even looked pretty wrong in 1938: by 1937 Industrial Production (we didn’t have GDP back then) had risen by an average of 14% a year from the Great Depression’s low in 1932. To be fair, in 1938 it crashed 21% before returning to strong growth in 1939. But the Great Depression was over by 1940. Hansen’s theory was almost immediately proven wrong. Larry Summers resurrected this theory a few years ago, towards the end of the recovery from the Great Recession, and looks just as foolish for it now. Again, great way to get attention in the moment and look like an idiot later.
3. Same with the decline in manufacturing jobs, same with the decline in coal mining jobs, same with the decline in blockbuster video jobs, etc. etc.
4. Though some amount of this finding is due to other factors such as increased off-shoring, which tends to disproportionately impact middle-skilled jobs.
Admittedly this is a pretty easy position for me to take. I'd like to see a robot do economics (not really, but I don't think they could is the point). In seriousness, I'd like to see a robot do all the math of economics and I can just do the abstract theorizing.
No comments:
Post a Comment