Missing Productivity

 

What interests me lately about productivity is the seeming puzzle of its supposed increase and its lack of effect. The November trade deficit rose to a record $60.3 billion, reflecting the lack of American exports. While unemployment is around 5.4%, there are more uncounted unemployed people than ever (because they get no benefits). Jobs are being "outsourced" to Asia in record numbers, but consumer prices are no longer falling. All these things (and more) should be otherwise, if U.S. productivity is rising. Moreover, if productivity is high and rising, paying the costs of children, retired people and other non-workers should be easier because there should be more to go around.

The importance of productivity may be seen in a single number: 1.7%. That is the average rate of U.S. productivity increase for the next 40 years assumed by the Great White Bandit in calculating a huge future shortfall in Social Security. The same past average since World War II is, however, around 2.5-2.7%. One extra percent makes all the difference for Social Security, since, if economic growth scales as the past average, there won't be any tax shortfall.

What is Productivity?

Productivity is a key concept and measure that connects a lot of different elements in our thinking about how society works, and how it should work. The simplest definition of productivity is the amount of human labor required to produce a certain quantity of product in a set time. It is a measure of the efficiency of production, not the quantity of production.

The efficiency of machine production isn't directly included in "productivity," because the term, as used in economics, has a hidden human orientation. The point of the measure is how much human effort is required to get a result, not how much energy or intelligence or machinery is needed. The reason for that restriction is that economics is all about human needs and wants, not about what robots want. Thus, the productive work performed by machines and nature is considered ancillary to human effort, either under human control or subject to human harvesting. Any productive result obtained in the non-human world is attributed to some human. However obtained, productivity simply measures the amount of product a given human accrues per unit of time. Example: Humans who live in a banana belt can be very productive in producing bananas by watching them grow and occasionally picking them, whereas humans living in Alaska will have to make major investments in buildings, energy and time to grow a single banana.

In principle, productivity has nothing to do with comparative market prices of different products, but should be a dominant factor in the competitive pricing of the same or similar products. Thus, Alaskan bananas are likely to be very expensive whereas Costa Rican bananas are likely to be cheap, but the price of bananas has little to do with the price of oil or steel (except to the extent bananas are an input in the production of those commodities). Since productivity measures can be used in comparisons of similar products, it is possible to aggregate productivity over several different productive facilities or methods. The further implication is that productivity can be generalized to economic sectors, and maybe even whole economies. However, as with every generalization, the further it gets from its roots, the more unstable it becomes. Banana production may give us a hint about the production of other agricultural products, but is not likely to have a lot to do with semiconductor fabs. So, there are a lot of cavils when we start extending "productivity" beyond particular production facilities or industries. Nonetheless, the government and several private agencies regularly make estimates of productivity, reporting results for industrial sectors and the entire economy as a single number. (I think those reports would be fairer, if the estimates were presented as a 3-D scattergram showing the result in a graded series of sectors over time.)

According to those reports, productivity among U.S. firms has been increasing for many years. Productivity increases were more rapid 5 years ago than now, but that follows the classical economic recovery curve. Productivity does not rise as fast when hiring goes up, because new workers take time to "get up to speed." On the other hand, during recessions productivity typically rises as companies lay off people and squeeze more work out of those remaining on the job. Net productivity rises over the business cycle, if the gains during a squeeze are not lost during the following upswing.

Productivity also increases as a result of improved production methods or product substitution. Although it is still disputed, automation is widely considered to have improved productivity because machines (robots) replaced people The robots work faster, longer and more accurately than humans, so typically require less energy and resources per unit of production. The additional capital required to install a robot is justified because, over the robot's useful life, the total cost of operation is lower per unit than human wages and benefits.

Related to automation is product substitution; e.g., replacing humans by robots is a form of substitution. More commonly, new materials are used instead of old, as in the replacement of aircraft structural aluminum by carbon fiber. While new materials may be more expensive by the pound, less of them are often required. Or, the new materials may be stronger or more resilient or have other properties that result in savings over their useful life. Sometimes new materials are required to implement a new or automated procedure, as happened in the development of integrated circuits and robot-made circuit boards.

Finally, there is a controversial increase of productivity due to "outsourcing." In neo-classical economics, trade based on the differentiation (division or specialization) of labor increases productivity. This is based on the metaphor that a skilled blacksmith will shoe more horses in a day than a jack-of-all trades. It is also recommended by the fact that certain tasks, such as surgery, are best done by well trained individuals. The assumption is that specialized training costs less than the value added over a working life. The generalization of labor specialization within a community is the differentiation of labor on an international scale. Thus, the neo-classical doctrine of comparative advantage rests on the specialization of labor (which supposedly increases productivity).

What the traditional ideas do not account for is the ability of just about anyone to perform many or most of the tasks required in a modern economy, when assisted by automated equipment. In this case, very few people can design and implement machinery that allow everyone else to do a job. The specialized few designers clearly add huge value to the economy, but what of the other workers? If a would-be doctor is reduced to answering the telephone in a call center, we might think there has been a loss of productivity. Of course, perhaps society already has too many doctors; we just don't know. If foreign workers are displaced from other tasks to serve the demands of international trade, is there comparative advantage? The answer again depends on what are the alternative employments of those workers. Generally, this question amounts to inquiring about the cost of specialization relative to the eventual social utility.

While productivity seems simply defined as the amount or value of production per unit labor, it is actually very difficult to pin down in the modern world. This is the result of the complexity of modern economies, in which many resources and skills interact.

In Utopia

A lot of time and effort has gone into developing new ideas and products not just to make (more) money, but to make life easier and more enjoyable. Consider a utopian society in which:

Automated machines make almost all of our clothes and furnishings. They make most of modules that are put together to build houses. They make most of the gadgets we use everyday, such as telephones, television, microwave ovens, refrigerators, cars, bicycles, etc. They grow most of the food we eat from seed to harvest, prepare and package it, and direct its shipment to storage and distribution locations. They direct the transportation system, loading and unloading freight cars, ships and trucks, and driving the carrier vehicles most of the time.

When I was a teenager, people scoffed about the possibility of such a society. Now, some 50 years later, First Worlders are very close to living in such a society. Simply put, the machines do most of "the work;" people do most of the consuming (enjoying the products). In such a world, how do we measure productivity?

The traditional measure of productivity involves output per unit of [human] labor. The silent word before labor is "human," not just any labor. This definition can be fixed by removing the human reference. Of course, that requires developing new standards by which machine productivity is measured. How much energy does the machine use? How many material resources are required? How long does it last? What does it cost to build? By taking careful measure of the inputs - costs - of machinery, it should be possible to find a basis for measuring its productivity. In such an analysis, the currently assessed "capital cost" of equipment can only be taken as a gross guideline, a first estimate, of the machine's value in production. When a refined estimate of productivity is made without any reference to human labor, the important results are the rate and costs of production, so the measure should be product per unit cost per unit time. Obviously, "per unit cost" has replaced "labor cost" in this formulation.

The economy does not fall apart or disappear in utopia. It is just measured in a different way. We can use the utopian case as a productivity limit; i.e., the value of productivity as the human labor factor approaches zero. This suggests the possibility of estimating the "ultimate" productivity of a system by plotting measured productivity against the amount of labor, and extrapolating the curve to 0 labor. (This plot should probably be logarithmic.)

Where's the beef?

If our society is already close to the Utopian case, and I think it is, we must ask why there is a problem in supporting the poor, the sick, the disabled and the aged?

An economy with very high and increasing productivity inevitably produces a surplus of products. This is not a new fact. It has been recognized at least since World War II as a major problem. The solution to this problem in a capitalist society is "managed demand." The productive system is not allowed to produce more than a certain amount, which the system's managers calculate on behalf of the capitalist owners. How much demand will be fulfilled is calculated to maximize profit, not to meet all needs. Facilities are not put in place to meet possible demand, but only to meet pre-arranged demand. Demand is, in fact, pre-arranged by the advertising industry. What happens is something like this:

1. An estimate is made of how much product can be sold (market research)
2. An estimate is made of the production costs
3. An estimate is made of the product life-cycle
4. An estimate is made of likely profits, investment costs, etc

When those four things are put together, it is possible to estimate how much stuff to make at a given price and at what profitability. The capitalist system does not concern itself with production for use; i.e., producing the things people need. Rather, what is produced is what will make a profit. All of us might be better off with a universal health care system, but it is more profitable to produce tobacco products.

The United States should have more than enough goods and services to support every resident without undue sacrifice by anyone. If so, why are there so many needy people? Where did the production disappear? The answers are given everyday in the Wall St Journal and other financial media. One of those answers, which is widely acknowledged, is that a disproportionate share of corporate profits are being distributed to certain shareholders and managers at the expense of ordinary workers. In other words, line workers are not receiving the share of profits justified by productivity.

Since 1980, the wealthiest Americans - the top 10% - have seen their incomes double and triple and more, while the (inflation adjusted) take-home pay of the lowest 50% of workers declined. The top 10% has appropriated all of the increased productivity to itself. If the rewards of our highly productive industries were distributed fairly, there would be no problem in paying for Social Security and many other benefits.

I conclude there's plenty of productivity to go around, but it just isn't distributed fairly. That was also the problem of the Great Depression, solved by the New Deal. It's time to raise that problem again, and enforce the solution.