Arnold Kling  

Lectures on Macroeconomics, No. 3

Lectures in Macro, No. 2... Kling on Financial Markets...

In this lecture, I get to the punch line and offer my explanation of unemployment. Markets are constantly in adjustment, with the number of people in different occupations changing. Usually, the task of adjustment and adaptation goes remarkably smoothly. Occasionally, however, markets are overwhelmed by the amount of adjustment required within a relatively short period of time. When the adjustment mechanism is overwhelmed, we have a recession.

Thank you for your time
Oh, you've been so much more than kind
You can keep the dime
--Jim Croce, Operator (That's Not the Way it Feels), 1972

Jim Croce's classic song conveyed the pathos of coming to terms with a broken relationship. Forty years later, it also conveys the pathos of the end of a bygone era.

The protagonist in the song is evidently in a phone booth, and the lyrics are the words that he speaks to the telephone operator. There are still telephone booths, but when was the last time that you used one? There may still be human operators who can connect calls, but when was the last time that you spoke with one?

It is safe to conjecture that, compared with forty years ago, we need fewer people to install and maintain telephone booths and fewer people to act as telephone operators. Instead of phone booths, we need cell phones and cell phone towers. Instead of operators, we have computerized systems for connecting calls and automated voice-response systems for looking up phone numbers. (To be sure, even by 1972 most calls were automatically connected without an operator.)

A dynamic economy requires constant adjustments in the labor force. New types of jobs emerge--consider that none of the job titles associated with web site development could have existed prior to 1994. Other jobs disappear.

Usually, these adjustments take place remarkably smoothly. In 2003, I wrote an essay called Manufacturing a Crisis, in which I presented the following data:

YearTotal Nonfarm EmploymentManufacturing Production WorkersPercent
195248.9 million12.8 million26 %
196255.7 million12.0 million22 %
197273.8 million13.5 million18 %
198289.7 million12.3 million14 %
1992108.7 million12.0 million11 %
2002130.4 million10.8 million8 %

For 2007, the figures are 137.6 million for total employment and 10.0 million for manufacturing production workers, respectively.

It should be noted that over this long history of falling employment, total U.S. manufacturing output was rising. Productivity increases have been larger then the declines in employment.

The fact that there are 3.5 million fewer manufacturing production workers today than in 1972 does not mean that there are 3.5 million former manufacturing production workers now standing around idle or in need of government retraining. It seems safe to presume that more than 3.5 million manufacturing production workers in 1972 have reached retirement age since then. Many of the retirees have not been replaced. However, there are many people, perhaps millions, working as manufacturing production workers who were not doing so in 1972.

Economists who notice job displacement tend to leap to the conclusion that we need government retraining programs. They forget that the natural process of retirement and new entry into the labor force tends to take care of the marginal adjustments in occupational choice. No, not every manufacturing production worker can retire at once, but they do not all have to. Many of them have to change firms or change industries, but the overall process of adjustment among occupations is reasonably gradual.

When workers are unemployed because of job displacement, economists call this structural unemployment. This is one of three types of unemployment described in traditional macro textbooks.

Another type is frictional unemployment, presumably named for the natural "friction" that exists as people leave jobs to seek new jobs. Someone might quit because they are fed up or want to move to a new city. Or someone might be laid off by one firm in a market where jobs are readily available at other firms, but it takes time to find a suitable opening and make a choice.

Finally, there is cyclical unemployment. For example, when the U.S. economy was more heavily dominated by manufacturing, the big auto companies and steel companies would occasionally find themselves with excess inventory. They would put workers on temporary layoff until supply and demand are in better balance. Because these industries were so large relative to the overall economy, the layoffs of their workers caused demand to fall in other sectors, so that employment also fell elsewhere. Eventually, however, the inventory correction would be over, manufacturers would recall their workers, and the economic decline would reverse. Alan Blinder, among others, has observed that most of the recessions in the U.S. economy between the end of World War II and 1980 could be described as "inventory recessions."

One way to think about cyclical unemployment is that it represents a system overload in the adjustment mechanism in labor markets. There is too much adjustment required in too little time, resulting in excess unemployment. This excess unemployment tends to amplify because of multiplier effects--workers out of a job are going to spend less, and this will reduce labor demand elsewhere.

Ordinarily, the need for adjustment is sufficiently gradual that lateral movements (workers changing jobs but not changing careers), retirements, and new entry into the labor force can meet the needs of a dynamic economy and maintain full employment. A recession takes place when the adjustments required are large and the economy is sluggish about making them.

With an inventory recession, the market is saying that we need less heavy manufacturing output and more of something else. However, workers laid off by manufacturers to not go to work producing something else, because they instead wait to be recalled by their old jobs. In a typical inventory cycle, most workers in fact do get recalled.

Sometimes, however, major shifts in employment are called for. For example, the development of automobiles, trucks, and paved roads in the 1920's brought about a restructuring of production. Farm produce could travel farther, which took away the economic advantages of dense urban areas surrounded by farms. Eventually, we would see in the 1950's the configuration of suburban housing, shopping malls, fast-food restaurants, major grocery stores stocked by out-of-town produce, and other familiar features of the mobility-driven economy.

However, in the 1930's, the adjustment process got stuck. The jobs that were added in the 1920's in boom industries such as electrification were cut back when the stock market crashed. Bank failures, monetary contraction, and deflation made matters worse, in ways that will be explained in a subsequent lecture. The financial turmoil caused further job cuts. In addition, the turmoil meant that industries that otherwise would have been expanding were instead struggling. The workers displaced by the transportation revolution and by layoffs in the boom-bust segment of the economy could not find work elsewhere. They reduced their spending, and multiplier effects kicked in.

Between 1925 and 1955, the U.S. economy achieved a massive transition. Suburbs replaced close-in rural communities. Farm labor contracted and service employment expanded. Unfortunately, this transition was not smooth. The veterans of World War II built a new economy, because the one that their parents had known in the 1920's stopped functioning during the Great Depression.

The nature of labor has changed since the 1930's. A much larger proportion of our labor force has specialized skills, and a smaller proportion are general laborers. The increased levels of specialization and training will have mixed effects. On the plus side, not everyone is tied to key manufacturing industries. An inventory correction in automobiles does not affect nearly as large a percentage of the work force as it did fifty years ago.

On the other hand, major adjustments may be more painful for highly trained individuals. If we need more health care management personnel and fewer mortgage securities traders, that adjustment process may not be as simple as moving a farm laborer into a factory.

I suspect that the modern labor force poses different problems for policy during a recession. With an inventory recession, a generic fiscal stimulus might serve to shorten the duration of the downturn. However, when the economy is in the process of expanding some sectors and contracting others, and when the expansions and contractions involve workers with different skill sets, generic stimulus may not be able to affect the dynamics of the process very much. I intend to talk more about policy issues in the next lecture.

Further reading: Much of the transition from landline to cellular telephones took place between 1990 and 2004, as recounted by Christopher C. Carbone in a n article published by the Bureau of Labor Statistics. A few data points from the article:

--from January 1990 to August 1993, total employment in telecommunications fell by 43,000, even though wireless communications firms added 25,000 workers over that period.

--From the beginning of 1996 to March of 2001, total employment in telecom increased by 352,000 jobs, as landline telecom services built capacity to market Internet services and to handle the increased data load, while cell phone services continued to expand.

--From March of 2001 through the end of 2004, total employment in telecom fell by 300,000. It turned out that the capacity build-out had gone too far, too quickly. Some of this was the Internet bubble, but much of it reflected the failure to take into account reduced demand for landline phone calls with the growth of cellular.

Previous entries in this series:
1. Introduction
2. Misconceptions about labor markets, both among the lay public and among economists

Comments and Sharing

COMMENTS (12 to date)
fundamentalist writes:

The recognition of differing labor markets is a big improvement, but the rest was a little disappointing. It’s not much more than Schumpeter’s technology shocks and “stuff happens.” While what Dr. Kling writes is true, and necessary, is it sufficient? It doesn’t explain the regularity and frequency of business cycles, nor the fact that the greatest variances occur in the manufacturing and commodities markets. And it leaves money out completely as if money has no effect on the real economy whatsoever. Any theory that ignores money seems deficient. I think if Dr. Kling would consider disaggregating the capital market as he has done with the labor market and include the effects of monetary stimuli on the different capital markets as well as the labor market he would have a good theory of business cycles.

bgc writes:

Very lucid - thanks.

Could you explain in a similarly lucid way what happened to the trillions (?) of dollars which were leveraged on mortgages. Where did the money go? What will be the overall effect of its disappearance? (If it has disappeared).

burger flipper writes:

Very helpful. Keep 'em coming.

Jesse Blocher writes:

Dr. Kling -
I think that a potential amplifier of recessions is the tacit regulation of human resources in corporations via threat of litigation. My experience in Corporate America was that it is very difficult to fire someone for performance reasons once a company is large enough - the resources are now there for a lawsuit based on wrongful termination.
The best way around this is to skip performance-based terminations and rather focus on economy-based layoffs. That way, firms can slowly increase their labor force during good times, and then if/when things become a bit untenable, declare a 'layoff' in which, say, 10% of the workforce is cut, and you get around (or perhaps minimize) wrongful termination suits.

Thus, we have bigger downswings in unemployment during down economic times. Perhaps this is a secondary effect, but my (albeit anecdotal) experience seems to confirm it.

Pietro Poggi-Corradini writes:

I wonder if a theory of regulation should be included as some sort of "macro" red-lights. It seems to me that there can sometimes be too much coordination that emerges either spontaneously within certain sectors or that is being imposed from the outside. The Basel accords would be an example, the Fed, the New Deal, etc...The problem with these traffic-lights is that they encourage uniformity of behavior and this in turn is more susceptible to sudden phase-changes. The analogy in physics is with gasses that when polarized (so that all the spins are aligned) then start exhibiting quantized behavior.

Matt writes:


How can you be so sure? All we have is an intuition which hasn't been formalized sufficiently to provide precise predictions.

To that end, I think a computational turn will turn up more insights. Talking about equilibrium with 25 different yet related labor markets is one thing. However, what seems to be interesting here is the processes by which the market returns to equilibrium (or at least heads towards). Agent based models seem to be the most promising avenue of capturing this process.

Adding things like money to the model may add more insights later, but it's best to simplify first to gain basic understanding before complicating the model with all of the features of reality.

Ajay writes:

fundamentalist, do you blog somewhere? You're the one commenter here I'd like to read on a more regular basis. If you don't blog, why not start?

phineas writes:

I found this 3-part lecture worth my while to read, and I essentially agree with it. It has the failing that it does not explain why the labor market adjustment process got stuck in the 1930s.

During the 1930s there was major improvement in labor productivity, and also major progress and modernization in the structure of the private-sector economy. There was also pretty major growth in gross economic output: GNP grew by 34% between the start of 1933 and 1936. By 1941 GNP was 58% higher than in 1933, despite undergoing a setback of 13% during the 1937-38 recession. Despite such growth, unemployment was still very high throughout the 1930s. You can't place most blame for 1930s unemployment on adverse non-labor-market economic multiplier effects exacerbating adverse labor market dynamics, because the economy was actually growing pretty healthily (although some of the growth from 1933 on was "relatively easy catch-up" from the contraction pre-1933, as GNP in 1932 was 33% below 1929).

I've taken the above figures from Michael Gordon and for more about the 1930s labor market dynamics see commenter Michael Gordon at

Jacob Oost writes:

I too have always been confused as to why economists specifically recommend *government*-paid training for laid-off workers, but then never explain why they believe vocational training is best done by the government. Friedman says that as there are no truly external benefits to vocational education, it is a private good and best handled by the market, rather than being socialised.

Babinich writes:

Jacob Oost asks:
"I too have always been confused as to why economists specifically recommend *government*-paid training for laid-off workers, but then never explain why they believe vocational training is best done by the government."

I am with you an this too. How is this not similar to the current bailout? With the current situation you have opened Pandora's Box and the result is that a number of different industries appear to be lining up for a bailout.

With government training of the unemployed I have two basic questions in need of answers:

1) Which disciplines are eligible for training when jobs in those disciplines disappear?

2) What disciplines are offered as training for those who've lost their jobs?

MattYoung writes:

I am stuck at the basics.

I get the following causation:

1) Minimize scarcity results in 2) Inventory management over the environment results in 3) Economies of scale results in 4) Specialization and wealth.

The idea that the original village was designed to store goods over time to get a better optimization of resources. Managing inventory requires an economy of scale to minimize gathering and storage effort, and that economy of scale yields larger wealth.

Wealth would force specialization, though the causality between wealth and specialization is uncertain in my mind. However, economies of scale do determine a hierarchical flow of wealth because the wealthier persons would be estimating a larger fraction of inventory requirements of larger time periods. This subdivision of the problem by hierarchy minimizes the work effort to manage inventory.

fundamentalist writes:

Matt: “How can you be so sure? All we have is an intuition which hasn't been formalized sufficiently to provide precise predictions.”

I earned a master’s in econ back in 1991 where I learned mainstream econ. The macro class compared Keynesian, monetarist and neo-classical econ. At the time it seemed that the monetarists and neo-classical economists had the data on their side and that Keynesian econ was wishful-thinking. I was particularly offended by Keynes’s “paradox of thrift.” It just didn’t make sense to me. And studying the stagflation of the 70’s made me realize just how little economists knew about the economy. Finally, a class on third world economic development contradicted much of what I had learned in macro, though micro was still obviously very sound. Over the years, I followed economics fairly closely and realized that no one could predict recessions, which seemed to me to be a very important point. Then I learned how bad the macro models were and that they couldn’t predict anything at all and I lost interest.

Eventually I got interested in economic history as I tried to learn the origins of capitalism. That led me to the New Institutional School of Douglass North. And I began reading work by cultural economists. I put the two together and realized how important culture is to creating the institutions necessary for capitalism to work. I also read a lot of transition economics that studied how well former communist nations made the transition to free markets. Again, the keys were culture, institutions and property rights. Mainstream macro ignores all three. Austrian econ emphasizes all three, so when I stumbled upon it was like a home coming.

I happen to agree with Austrian epistemology that says you must approach data with sound theory. I had already been prejudiced against the idea of trying to build theory from data by statistics classes I had taken in which I was taught that you can find anything you’re looking for in data if you know how to torture it long enough. Statisticians should always approach data with a theory in mind and try to find support in the data. Using data to formulate theory is called data mining and has led to a lot of bad theory. Real statisticians hate that.

As for business cycles, the idea that business failures sometimes cluster is an intriguing idea, but I think it’s somewhat of an excuse to quit thinking. The first guy to come of with the “Austrian” business cycle theory was Richard Cantillon in the early 18th century. He discovered it by observing how economies worked at the time. Over the years other people refined it by more observation and critical reasoning. The thing about depressions that stood out was that they were concentrated in the capital goods industries. The theories they came up with were mainly to explain that phenomena. Even today, while economists deny it, financial people will tell you that the capital goods industry is very volatile. That’s why retail and utilities are “defensive” investments when you think a recession is approaching.

The explanation for the wild swings in capital goods industry and milder ones in retail goods was price differentials. Austrians also have a very good grasp of how money works. I think if you read them you’ll agree. Money is not neutral. Changes in its value not only cause all prices to change, but the prices of capital goods relative to consumer goods, which change the way people invest.

Ajay, thanks for the kind words, but I really don’t have time for my own blog. Besides, I only repeat what I have learned at

Comments for this entry have been closed
Return to top