Labor Day in the United States, and EP has taken the day off. Herewith a re-issue of EP’s edition of July 18, 2004.
Frank Levy, of the Massachusetts Institute of Technology, and Richard Murnane, of Harvard University, had just published The New Division of Labor: How Computers Are Creating the Next Job Market (Princeton, 2004). It turned out to be one of the most prescient books of the decade.
Murnane has continued to teach. His econometrics horn book for graduate students, Methods Matter: Improving Causal Inference in Educational and Social Science Research, with John Willett, appeared in 2010 (Oxford). Levy, emeritus now and a lecturer at Harvard Medical School, has continued the reconnaissance. He organized, with the late Seth Teller, a roboticist, a joint seminar of computer scientists and economists at MIT.
A World Without Shims?
A shim is a piece of stone, wood or metal, usually tapered and used as a filler or leveler between components that otherwise wouldn’t fit together smoothly. You see shims in the stone walls of old cathedrals, between pieces of furniture and the floor, and, if you know where to look, in the Boeing 727.
Such was the complexity of manufacturing that great icon of the jet age — it first rolled out in 1962, and nearly 1300 of the 1800 planes delivered are still flying — that one manager who worked on the project estimates that a 727 weighing 44 tons typically contained a half-ton of shims.
The development of the airliner took nearly seven years. Five thousand engineers worked on it. Nobody could be certain that the blueprints were consistent, so the first step was to build a full-scale model.
Only then could blueprint specs be translated into settings for machine tools. But naturally the parts fit together imperfectly. So assembly workers adjusted them by hand with metal shims to insure the plane was tight
This story, from The New Division of Labor: How Computers are Changing the Job Market, by Frank Levy and Richard J. Murnane, is designed to underscore the significance of what happened next.
For when the 777 model came along in 1994, it required only barely five years to build, even though it was much bigger and more complex. No mock-up model was required. No more paper blueprints. The 777 was the first airplane to be completely designed with computers, which insured the internal consistency of its parts.
Using computer-assisted design and manufacturing software purchased from the French engineering company Dassault, Boeing engineers were able to take their numerically-controlled machine tool settings directly from their plans. The plane fit together smoothly as a result.
The manufacturer boasted, “The first 777 was just .023 of an inch — about the thickness of a playing card — within perfect alignment while most airplane parts line up within half an inch of each other.”
Among the results: there no more jobs for highly-skilled operators of turret lathes, far fewer blue collar workers on the assembly lines. Component-manufacturing plants can be located anywhere around the world, partly in response to political pressure from customers, since Boeing knew that the parts would fit together seamlessly.
Boeing headquarters relocated from parochial Seattle to cosmopolitan Chicago. Competition with the European Airbus syndicate stiffened. Investors’ returns from airframe companies gyrated with each twist of what author John Newhouse long ago called “the sporty game.” And, not least, better planes were offered for a much broader market for air travel.
The New Division of Labor is a fascinating book. Not since the mathematical economist Truman Bewley interviewed 300 business executives and labor leaders for Why Wages Don’t Fall during a Recession have sophisticated economists waded so deeply into the real-world circumstances of the important problem they are seeking to understand.
Levy and Murnane’s project is designed to get at the consequences underlying the new division of labor among people and computers. Just as in the 19th and 20th centuries the advent of ubiquitous engines and motors altered the value traditionally placed on human strength and endurance, so computers in the 20th and 21st centuries are displacing skills possessed by a good proportion of the population that, until recently, were enough to earn a good living.
“In an increasingly computerized world,” they ask, “what well-paid work is left for people to do both now and in the future? How can people learn to do the skills to do this work?”
The authors have a history of this sort of thing. Levy, a professor of economics in the department of Planning and Urban Studies at MIT, is the author of Dollars and Dreams, a classic survey of the changing distribution of income in the United States. Murnane, professor at Harvard University’s Graduate School of Education, wrote Who Will Teach?, a study of the forces that are rapidly changing the composition of the public teaching force.
Together, they wrote Teaching the New Skills, a manifesto on education reform. And with MIT economist David Autor, they co-authored “The Skill Content of Recent Technological Change: An Empirical Exploration,” a 2003 article in the Quarterly Journal of Economics that forms the backbone of this book.
Their conclusion: a new kind of “hollowing-out” of the labor force is taking place. As recently as 1970, half of all US adults worked in two broadly-defined occupational classifications: blue collar jobs and clerical jobs. “Few people got rich off these jobs, but they supported middle- and lower middle-class living and many were open to high school graduates.
Today, they say, the total is under 40 percent, and many of these jobs require at least some college education. The trend is expected to continue. Of course new opportunities are opening up as well, but the new jobs are of two quite different sorts, they say.
There are jobs for janitors, cafeteria workers, security guards and the like, that pay poorly and offer little chance of advancement. There are more of these jobs than there used to be, but the greater growth has been among higher-paying jobs — managers, doctors, lawyers, engineers, teachers and technicians.
Two facts about these better-paying jobs stand out, the authors say. They require extensive skills. And most involve the use of productivity-enhancing computers, even if that means no more than learning to read in rapid fashion from a series of drop-down menus designed to elucidate a series of choices — in an automotive garage, say, or in the office of a health insurance company.
About the lowest-paying jobs, Murnane and Levy have relatively little to say. It has long been noted that technological change creates losers as well as winners, and mechanisms whereby winners can make life easier for losers without seriously diminishing their own gains. Often these take the form of tax-financed retirement and health insurance systems. Combined with a little day-to-day respect, such measures can go a long way in conferring dignity on menial work.
Why should anyone worry about basic fairness? Because “Our market economy exists in a framework of institutions that requires the consent of the governed,” they write. “People doing well today have a strong interest in preserving this consent. If enough people come to see the US job market as stacked against them, the nation’s institutions will be at great risk.”
It is on access to those better-paying jobs that Levy and Murnane spend most of their effort. They identify two sets of skills in particular whose value has been increased by the advent of computers.
They are what the authors call “expert thinking,” meaning identifying and resolving uncharted problems, and “complex communication,” meaning conveying not just information but a particular interpretation of information. Most of their readable book is devoted to illustrating what this means, and offering suggestions as to how the skills can be better taught.
One chapter begins with an anxious Victor Hugo querying his publisher with a single pen-stroke about the success of his latest book — “?” The publisher, not to be outdone, answers with a pen-stroke of his own — “!” Communications among humans are rarely so efficient, the authors say.
Hence complex communication involves teaching understanding, gaining trust and negotiating outcomes. The chapter that begins with Hugo visits a circuit board designer in New Hampshire, a second-grade classroom, a Chicago investment brokerage, a renovation specialist in Boston — and ends with a discussion of the indispensability of the telephone help desk that is reached by an icon on the Lands End Web site.
The shims most needed today, in other words, are human. Chances are they will be in demand for many generations to come.