Dean Baker’s screed, Bill Gates Is Clueless On The Economy, keeps getting recycled, from Beat the Press to Truthout to Real-World Economics Review to The Huffington Post. Dean waves aside the real problem with Gates’s suggestion, which is the difficulty of defining what a robot is, and focuses instead on what seems to him to be the knock-down argument:
Gates is worried that productivity growth is moving along too rapidly and that it will lead to large scale unemployment.
There are two problems with this story: First productivity growth has actually been very slow in recent years. The second problem is that if it were faster, there is no reason it should lead to mass unemployment.
There are two HUGE problem with Dean’s story. First, aggregate productivity growth is a “statistical flimflam,” according to Harry Magdoff, who pioneered productivity measurement in the 1930s. In the 1980s, Magdoff co-authored a Monthly Review article with Paul Sweezy, “The Uses and Abuses of Measuring Productivity,” detailing the methodological problems of aggregate productivity measurement. After discussing “phantom statistics” in the reporting of construction industry productivity, and the technical problems of aggregating productivity statistics, Magdoff and Sweezy explained why “there is no such thing… as a ‘true’ measure of productivity”:
One reason for including this somewhat technical discussion is to drive home the point that there is no such thing as straightforward or “true” measure of productivity. And if this is so even in the realm of commodities where a reasonable, limited, meaning can be given to the concept, what can said about the productivity of service workers? There are of course service jobs that consist of routine, repetitive operations — e.g., in typing pools — where productivity measures may have some meaning. But how would one go about measuring the productivity of a fireman, an undertaker, a teacher, a nurse, a cashier in a supermarket, a short-order cook, a waiter, a receptionist in a lawyer’s office? It is in the very nature of the case that in most services qualitative changes are intertwined with quantitative changes; hence there is no continuity in the “output” from one period to another with which changes in employment can be compared. Moreover, it is typical of many of the service areas that the “output” cannot be separated from the labor engaged in the performance of the service; for that reason too there is no sensible way of comparing changes in output and labor. In other words, the notion of a productivity measure for most service occupations is nonsensical and self-contradictory.
Unfortunately, such considerations of elementary logic have not prevented statisticians and economists from producing a whole array of productivity measures, applicable not only to the private economy (combining commodity-production and services) but in some cases to government as well, useful for ideological and policy-making purposes. And by dint of endless repetition and selective emphasis, these statistical phantoms (to use Business Week’s apt expression) have attained the status of indisputable facts and have entered into the realm of scientific discourse. What is in reality nothing but a crude fetish has thus become one of the most potent weapons in capital’s struggle against labor and in support of an increasingly irrational and destructive social system.
Fred Block and Gene Burns took up the critique of productivity statistics six years later in “Productivity as a Social Problem: The Uses and Misuses of Social Indicators.” Their analysis specifically addressed the second problem with Dean’s story, his contention that productivity growth is totally benign:
In short, there is no reason that productivity growth should ever be viewed as the enemy of workers. We just need the right set of policies to ensure that they share in the gains.
Leaving aside the benefits and risks of technological advances themselves, Block and Burns chronicled how the concept of productivity growth — and its faux measurement — has been used as a political weapon against workers, unions and collective bargaining. The use of productivity data had initially gained prestige for its role in providing a “rational and objective” basis for wage negotiations, but in the late 1970s, business and political leaders,
…seized on declining rates of productivity growth as proof of the need for national policies to restrain wages and limit the growth of state spending. The decline of productivity growth was attributed to inadequate levels of investment and it was argued that only measures that increased the flow of resources to business and the rich could possibly facilitate adequate levels of new investment. It was simultaneously argued that excessive government regulation was also responsible for the slowing of productivity growth leading to stronger demands for deregulation of the business community. The culmination of these efforts was the Reagan Administration’s dramatic reversals of long standing tax and regulatory policies which were justified as providing solutions to the productivity crisis.
While the productivity concept had initially been elaborated by the WPA’s National Research Project to provide justification for more generous wage settlements and government public works programs, by the late 1970s, it provided a critical justification for getting tough with labor and for dismantling key parts of the American welfare state. The process of institutionalization had resulted in a reversal of the political implications of this particular indicator.
In short, flimflam productivity measures were used by the enemies of workers to justify enacting a set of policies that ensured that workers wouldn’t share in the gains of technological advance.