Arnold Kling

The PSST Just-So Story

PRINT
Comments on Eurozone Solutions... Greenspan: Let Them In...

I will put this below the fold. It is from an essay that I am drafting.

PSST emphasizes the effect of technology on employment. In any industry where productivity rises faster than demand, employment will decline. This will be the case when the elasticity of demand is not sufficiently high to absorb the increase in productivity, so that labor resources are released from the sector where productivity is rising. In the long run, this should be offset by increases in employment in other industries where demand is rising faster than productivity. However, entrepreneurs have to discover which industries to enter, they have to design jobs within those industries, and workers have to adapt to fill those jobs. These adjustments can take many years.

The PSST story might be based on the pattern of diffusion of what economic historian Paul David calls general purpose technologies. The dynamo (small electric motor), the computer, and the Internet would be examples of such technogical innovations.

The first phase of diffusion might be dubbed the Solow phase, after Robert Solow's famous 1987 quip that "We see computers everywhere but in the productivity statistics." During this phase, firms experiment with new technology in order to learn. However, because the technology is in a relatively primitive state and its most effective uses have not yet been discovered, early adopters show little gain in productivity from the new technology and laggards see little threat. Thus, we get gains in employment in firms that are bullish on the new technology, with little or no displacement of existing workers. This might describe the 1980s with personal computers and the Internet boom of the late 1990s.

In the second phase of diffusion, which we might call the job-loss phase, the productivity gains kick in. Firms that have adopted the new technology gain market share without having to add workers (think of Amazon, the online book seller). Firms that lag in employing the new technology are driven out of business (think of Borders, the large bookstore chain that went bankrupt in 2011). This is the phase in which many workers suddenly become superfluous. Expanding firms are less labor-intensive than failing firms. Overall, the economy experiences a surge in unemployment.

In the third phase of diffusion, which we might call the renewal phase, entrepreneurs develop new firms and new industries to employ the surplus labor released by the productivity gains. These new industries, along with the adaptation of the labor force, foster a return to full employment.

The foregoing is something of a "just-so" story for the Great Depression and the most recent slump. For the Great Depression, the 1920s would be the Solow phase of the internal combustion engine and the dynamo, the 1930s were the job-loss phase, and the decade after the second World War the renewal phase. In the more recent cycle driven by the Internet, the 1990s would be the Solow phase, the current period would be the job-loss phase, and we have yet to experience the renewal phase.

In both of these episodes, the job-loss phase coincided with financial distress. In the AS-AD story, the financial distress is the causal factor. In the PSST story, the financial distress would be somewhat of a byproduct or symptom that accompanies the real adjustment. Financial euphoria in the 1920s and 1990s might have served to amplify employment growth and perceived wealth during the Solow phase of these cycles. The financial euphoria that accompanied the housing bubble might have helped to delay the onset of the job-loss phase of the current cycle.

Indeed, the phenomenon of financial euphoria followed by financial crisis is important to address in either AS-AD or PSST. Economic historian Charles Kindleberger argued that euphorias tend to take place following what he calls "displacement," such as major wars. In our terms, displacement creates a disruption in patterns of specialization and trade. New trading opportunities, like new technologies, can trigger a cycle. Perhaps if we look at these episodes, we can find evidence of the three phases, with the initial trading forays having little effect on productivity ("we see voyages to the New World everywhere but in the productivity statistics," so to speak) but eventually having disruptive effects.


Comments and Sharing


CATEGORIES:



COMMENTS (10 to date)
fundamentalist writes:
In any industry where productivity rises faster than demand, employment will decline.

Shouldn't prices fall instead of employment? Rising productivity should reduce prices and spur demand, thus preventing employment from falling.

Mark V Anderson writes:

fundy -- When productivity increases, prices will fall. That will of course increase the demand, but the quantity increase will vary. I think in most cases the increase in demand won't offset the increase in productivity, causing employment to fall in the industry of increasing productivity. I know my reaction to lowered prices is usually to buy a few more of the items that have decreased in price, but also to redirect part of my savings to buy some different things.

I think what Arnold is saying here is that this redirection of consumers buying different things in reaction to price savings takes some time to take effect. Thus the fall in employment of one industry is not immediately offset by increasing employment in other ones. Thus, an unemployment rate of 9%. It sounds to me that PSST isn't so much a separate way of looking at unemployment from AD-AS as an explanation as to why aggregate demand has decreased. It sounds like a pretty good explanation to me, although I'd like to see actual numbers to prove some of his suppositions, as well as actual examples so it is more clear which industries to which this is happening.

Bryan Willman writes:

It's "Solow" versus "diffusion" and above all changing from "doing task A better" to "don't do A at all, do task C instead."

This is a big deal. Yes, there's some advantage to having an order taker on the phone typing in a terminal or PC that helps get the order right while they are still on the phone. But it is swamped by having the buyer enter the order directly into the order system. So rather than reducing clerk staff from 25 to 23 with a much better error rate, we reduce it from 25 to 0.

There is inertia against this sort of change, and lots of complicated details to work out (the moral equivalents of 3 phase motors, standard voltages and hertz rates, fuses, breakers, disconnects, mundane stuff you need to have electric power really work well.)

A recent story I commented on elsewhere in this blog, talks about a company that went down to 1 shift running lights out for 2 more, making more product with radically fewer people. USING MACHINERY IT ALREADY HAD. Probably they had to get a fair amount of support machinery and work out details, but they had the key elements in hand already.

Electric motors did not appear in the 1920s, nor did networks appear in 1990s. But the details of operations, and some social constructs, came to a reasonable level of maturity then. (Including social changes like having a long relationship with a bank in which you go years at a time without seeing any bank employee. Would anyone have felt comfortable doing banking that way in the 1970s? The 1980s?)

Is this diffusion? Maturation? In any case, changing to the right task and getting the details right are required for the productivity gains and job losses to occur.

Ed Bosanquet writes:

I like this story. I have a couple of thoughts on it.

This PSST would work for any change not just technological. For instance climate change could create a desert out of current farm land but the Canadian and Siberian tundra would become arable. Overall the change is positive but we would need to adjust to the new growing patterns. Until we adjust, we suffer.

Since change (technological or otherwise) creates a set of temporary winners and losers, this would seem to indicate a transfer from the winners to the losers while the losers are reallocated. That would slow the motive for change but would seem more cosmically fair. You don't seem to advocate compensation to the losers but your theory seems to point to it.

Steve Roth writes:

"In the AS-AD story, the financial distress is the causal factor. In the PSST story, the financial distress would be somewhat of a byproduct or symptom that accompanies the real adjustment."

How about instead, an "either-or" story.

Couldn't both be true, with separate waves that are to some extent synchronized? Drop two pebbles in a pool....

Kevin writes:

It is tough for entrepreneurs to make the adjustment when states are hostile to commerce and success.

Dave Schuler writes:

There are at least two unspoken issues in this story. First, what enables businesses to introduce non-productive innovations, e.g. the personal computer in the 1980s over a substantial period of time?

Second, do those factors continue to be in place for some technology that is yet to become productive? And what is that technology? Or are we in some consolidation phase during which non-productive investments are no longer made?

Bryan Willman writes:

For computers (I worked in that industry) we were always befuddled by a dictomy.

In our own work and our own operations computers in general and PCs in particular were huge winners. (As smart phones are now and tables are or are becoming.)

Yet productivity studies often didn't show it.

I think *part* of the answer is that things like word processors and page layout software *could* (and often did) improve the quality of the output a person made. But for a salaried person doing a fixed number of letters, say, that quality wouldn't be monetized (so GDP doesn't go up), their compensation is fixed by salary, so labor costs don't go down. And for many documents, having nice fonts and spell checking doesn't actually matter.

For many tasks (taking notes at a meeting say) using a tablet is easier and faster, and the output may be more complete. But probably none of that shows in costs or is visible in output.

So *part* of the answer to @Dave Schuler is around things that "increase wealth" without increasing GDP numbers or decreasing labor costs.
(That might be peculiar to computers/phones.)

Rolf Andreassen writes:

You use a lot of "would be" where "is" would serve the same purpose and waste fewer electrons.

c141nav writes:
"Firms that have adopted the new technology gain market share without having to add workers (think of Amazon, the online book seller). Firms that lag in employing the new technology are driven out of business (think of Borders, the large bookstore chain that went bankrupt in 2011). This is the phase in which many workers suddenly become superfluous. Expanding firms are less labor-intensive than failing firms. Overall, the economy experiences a surge in unemployment."

Andy Kessler's book Eat People directly addresses this point.

Comments for this entry have been closed
Return to top