Musings on Tyler’s Technological Thoughts

Tyler Cowen, in his recent book Average Is Over, argues that computer technology is creating a sharp economic and class distinction between people who know how to effectively use these “genius machines” (a term he uses over and over) and those who don’t, and is also increasing inequality in other ways. Isegoria recently excerpted some of his Tyler’s comments on this thesis from a recent New Yorker article.

I read the book a couple of months ago, and although it’s worth reading and is occasionally thought-provoking, I think much of what Tyler has to say is wrong-headed. In the New Yorker article, for example, he says:

The first (reason why increased inequality is here to stay) is just measurement of worker value. We’re doing a lot to measure what workers are contributing to businesses, and, when you do that, very often you end up paying some people less and other people more.

The second is automation — especially in terms of smart software. Today’s workplaces are often more complicated than, say, a factory for General Motors was in 1962. They require higher skills. People who have those skills are very often doing extremely well, but a lot of people don’t have them, and that increases inequality.

And the third point is globalization. There’s a lot more unskilled labor in the world, and that creates downward pressure on unskilled labor in the United States. On the global level, inequality is down dramatically — we shouldn’t forget that. But within each country, or almost every country, inequality is up.

Taking the first point: Businesses and other organizations have been measuring “what workers are contributing” for a long, long time. Consider piecework. Sales commissions. Criteria-based bonuses for regional and division executives. All of these things are very old hat. Indeed, quite a few manufacturers have decided that it is unwise to take the quantitative measurement of performance down to an individual level, in cases where the work is being done by a closely-coupled team.

It is true that advancing computer technology makes it feasible to measure more dimensions of an individual’s work, but so what? Does the fact that I can measure (say) a call-center operator on 33 different criteria really tell me anything about what he is contributing the the business?

Anyone with real-life business experience will tell you that it is very, very difficult to create measurement and incentive plans that actually work in ways that are truly beneficial to the business. This is true in sales commission plans, it is true in manufacturing (I talked with one factory manager who said he dropped piecework because it was encouraging workers to risk injury in order to maximize their payoffs), and it is true in executive compensation. Our blogfriend Bill Waddell has frequently written about the ways in which accounting systems can distort decision-making in ultimately unprofitable ways. The design of worthwhile measurement and incentive plans has very little to do with the understanding of computer technology; it has a great deal to do with understanding of human nature and of the deep economic structure of the business.


Taking the second point–that automated workplaces require higher skills–it’s not really that simple. Sometimes automation requires higher skills, sometimes it results in de-skilling of jobs. One example is a cashier, who not so long ago had to be able to do simple math. Now, many of them are lost if the POS terminal can’t do their calculations for them. Another example: in a factory, running a CNC machine tool (that has been programmed by someone else) is generally a lower-skill job than operating a manual machine tool. Finally, being a manager in a chain retail store where inventory & ordering are handled centrally (by automated ordering systems with some input from centralized buyers at HQ) is a lower-skill job than being a manager in a retail store where the manager has to manage inventory control directly.( See my post Myths of the Knowledge Society for further thoughts on skill requirements and technological evolution.)

The degree to which automation is used to reduce the skill requirements of a given job is not a function of technological determinism; it is a consequence of management choices. I understand that Costco, for example, gives individual store managers a higher degree of empowerment than does the typical chain store in which these functions are more highly centralized.

Tyler talks a great deal about “genius machines,” and returns repeatedly to the example of a program that acts as an assistant to human chess players. I would have liked to see him focus some attention and analysis on actual supposedly-smart systems that are actually now being used or recently have been used in business and other fields. One example might be the models that have been used to evaluate the risk of mortgage and other loans, both on an individual-loan basis and for tradable pools of mortgages. Another might be the autopilot and flight-control systems used in aviation, and the dispatcher-assistant programs used in the freight-rail industry. And then there are the Enterprise Resource Planning systems used in a wide variety of companies. (I have heard many comments by people about their companies’ ERP systems, but have never once heard anyone refer to “our genius ERP system.”)

It is true that automation and centralization tends to create some jobs which are more abstract and conceptual in their nature than were the jobs they replace–the case of a buyer in a centralized chain store versus the store manager doing his own buying in a very hands-on environment is a good example of this.  This is really more a function of centralization more than of automation per se. And there is real danger in losing the intuitive feel for a marketplace via excessive reliance on purely-quantitative information, whether this information comes from 2013 “Big Data” systems, 1995 mainframe computer systems, or 1938 punched card systems. See Peter Drucker’s description of two old-line, hands on merchants, one of whom he calls “Uncle Henry” and the other of whom was Charlie Kellstadt of Sears. After relating some anecdotes about these two men and their management styles–and also introducing Uncle Henry’s son Irving, a Harvard B-school graduate, Drucker continues:

Fifty years or more ago the Uncle Henry’s and the Charlie Kellsadts dominated; then it was necessary for Son Irvin to emphasize systems, principles, and abstractions. There was need to balance the overly perceptual with a little conceptual discipline. I still remember the sense of liberation during those years in London when I stumbled onto the then new Symblolical Logic (which I later taught a few times), with its safeguards against tautologies and false analogies, against generalizing from isolated events, that is, from anecdotes, and its tools of semantic rigor. But now we again need the Uncle Henrys and Charlie Kellstadts. We have gone much too far toward dependence on untested quantification, toward symmetrical and purely formal models, toward argument from postulates rather than from experience, and toward moving from abstraction to abstraction without once touching the solid ground of concreteness. We are in danger of forgetting what Plato taught at the very beginning of systematic analysis and thought in the West, in two of the most beautiful and moving of his Dialogues, the Phaedrus and the Krito…They teach us that experience without the test of logic is not “rhetoric” but chitchat, and that logic without the test of experience is not “logic” but absurdity. Now we need to learn again what Charlie Kellstadt meant when he said, “How else can I see a problem in my mind’s eye?”

(emphasis added)

Companies that understand the point Drucker was making here are going to do better than those that place exclusive emphasis on “big data” and “genius machines.”

Regarding the overall questions of increasing inequality, and especially reduced class mobility, I would have like to see Tyler give more attention to the influence of government and runaway credentialism.

Anyhow, despite my disagreements a worthwhile book, which readers should find thought-provoking.

Has anyone else here read the book yet?


14 thoughts on “Musings on Tyler’s Technological Thoughts”

  1. I will be reading it very soon. Probably after I finish the few I am reading now. Cowen contradicts the picture Jim Bennett and I paint of the future in America 3.0. We need to engage him.

  2. This subject is one I went to some trouble to learn about when I quit private practice. I went back to school at Dartmouth and spent a year learning methodology. Paul Batalden had just arrived to run a quality improvement project. He had begun, like Paul Elwood, as a pediatrician in Minnesota who got into administration as many of them do.

    The Frist family hired him to run a quality improvement program at Hospital Corporation of America. He was focusing on research in how to accomplish this when Rick Scott worked his hostile takeover of HCA. One of the first things Scott did was get rid of Batalden and his program. After Scott created the scandals at Columbia-HCA, the Frists managed to get control back and Scott wound up, I don’t know how, as governor of Florida. I consider him a crook.n Batalden stayed at Dartmouth and I enjoyed his classes and seminars greatly.

    Anyway, health care is dismantling the “knowledge structure” that it has by making skilled OR nurses clean operating rooms instead of managing the OR. Everything in medicine is being dumbed down, with PAs and nurse practitioners delivering care with no supervision. I have no argument with them if they are part of a structure with ready consultation when they feel they are beyond their skill set. That is not what is happening.

    Obamacare is accelerating the process and the “guidelines” that doctors must obey or be arrested or fined, are written with cost as the prime criterion. The auto industry went through a phase where it was run by accountants, not engineers. Medicine is in the same pattern and the same results can be expected.

  3. Cowen is what some people call a “Techno-pessimist”. I found a couple of good popular articles on the subject a couple of months ago. One of them is about Robert Gordon, an economist specializing in macro-economics at Northwestern:

    “The Blip: What if everything we’ve come to think of as American is predicated on a freak coincidence of economic history? And what if that coincidence has run its course?” By Benjamin Wallace-Wells, New York Magazine, 21, 2013

    Here is a link to Gordon’s NBER working paper:

    And a WSJ article:

    “Why Innovation Won’t Save Us: For more than a century, the U.S. economy grew robustly thanks to big inventions; those days are gone” By Robert J. Gordon, Wall Street Journal, Dec. 21, 2012

    The techno-optimist side is taken by one Gordon’s colleagues at Northwestern, noted economic historian, Joel Mokyr:

    “Technopessimism Is Bunk” By Joel Mokyr

  4. This passage from Cowen sounds very much like one or more of our radio interviews about the transition from America 2.0 to America 3.0:

    I don’t accept the view that this new era is so different from every time in the past. Consider the industrial revolution, which starts in Great Britain in the seventeen-seventies or seventeen-eighties. For a long time, you had rising inequality, fairly stagnant living standards, a lot of problems adjusting. Of course, we did eventually get over it in the longer run, and it was much better for everyone. But it took, arguably, fifty or sixty years for us to make that transition. I think this future wave of inequality, which is already underway, will be a lot like that. It will take us decades to make the transition. Those decades will bring a lot of problems. But I think that in the much longer run—which is not what the book is about—it will be much more positive than it will seem during the transition era. I think this period fits quite nicely with historical precedent.

    In other words, the inequality he is concerned about and other problems he is concerned about may be temporary. And of course everything in a dynamic economy is temporary.

  5. “Technopessimism Is Bunk”

    I’d say most predictions are bunk.

    The problem with pessimism in this regard is that it’s limiting. If you’re optimistic you may be wrong but at least you try new things.

  6. I personally (IANAE) am inclined not to believe the tech-pessimist. I think that there is a decent argument that technology must reach the top part of its S curve. However, I think there are more pressing reasons for sub-par growth, including:

    1. Fiscal incontinence: Enormous deficits with no clear path to bring them under control.

    2. Tax increases. Not only did the tax law adopted last New Year increase the top marginal tax rates, but Obamacare imposed new taxes on capital income.

    3. Regime uncertainty. Well, the country is being run by an utter incompetent, who is all the more more dangerous because he is in the grip of the Dunning–Kruger effect. His signature accomplishments are Obamacare and Dodd-Frank. Obamacare has gone from regulatory nightmare to complete fiasco. Regulators are years behind on Dodd-Frank, and there is no reason to believe it will work any better than Obamacare. Further, the EPA is still honing its axe with which to shut down industries all over the country.

    4. The Fed, which has continued its uncertain foray into expanding its balance sheet like the universe expanded after the Big Bang. The Federal Reserve system is now carrying $50 of assets on each $ of capital, a ratio that would cause it to put a commercial bank into receivership. Further its ultra-low interest rate policy acts as a confiscatory tax on savers. How they can climb down from this situation without a catastrophe is not apparent.

    5. Institutional sclerosis in major chunks of the economy. Health care spending is more than one sixth of the GDP. In no other OECD country does it exceed one eighth. It is not a function of wealth. By most measures Switzerland has a per-capita GDP equal to or greater than the US, but its health care system is only 11.5% of GDP. Bringing the US back into line with the rest of the OECD means either an enormous cut back in health care expenditures or an enormous expansion of growth in other sectors. Neither seems likely. Education, Law, and Government all have symptoms of institutional problems. (1T$ of student loans, 40% of grads in law unemployed, bankruptcy of Detroit).

  7. I am not an economist, but John Taylor is. Here is his comment about slow growth on his blog today:

    In my view, deviations from good economic policy have been responsible for the very poor performance over the past decade. Such policy deviations created a boom-bust cycle, and were a significant factor in the crisis and slow recovery.

    Examples include the Fed’s low interest rate policy in 2003-2005 and the lax enforcement of financial regulations—both deviations from rules-based policies that had worked in the past. These were largely responsible for the boom and the high level of risk taking, which ended in the bust in 2007 and 2008.

    Other more recent examples are the hundreds of new complex regulations under Dodd-Frank, the vast government interventions related to the new health care law, the temporary stimulus packages such as cash for clunkers which failed to sustain growth, the exploding federal debt that raises questions about how it will be stopped, and a highly discretionary monetary policy that has generated distortions and uncertainty.

  8. Technology has been drastically changing the nature of labor — especially for low- and medium-skill workers — for 200 years or so. That’s when the Luddites first appeared. So in one sense the advent of digital technology is just the latest in a long line of changes.

    And like those previous advances, the dividing lines are as much generational as they are educational. Consider that a ten year old of even moderate intelligence is likely to have mastery of tablets and smartphones, while even a highly educated retiree may be completely flummoxed by the same devices.

    I think the big question, which I haven’t seen widely addressed yet, is when and if intelligent machines will get powerful enough to outright replace entire classes of moderately skilled workers, rather than moving them into new jobs.

    When semi-autonomous machines can drive cabs and trucks, deliver packages, stock vending machines, clean floors, assemble computers and appliances, sew garments, and maintain landscaping, will that result in a fundamentally different change to human society compared to what we’ve seen from other technologies, or will it be more of the same? I suspect it will be deeply different, but then that’s what most people said about all the other “revolutions”.

  9. “I’d say most predictions are bunk.”

    “Predictions are difficult, especially about the future.”

    Yogi Berra.

  10. These “scientific” metrics for employee performance aren’t being driven by a perception that they are giving us a better feel for whether an employee is pulling his weight than subjective evaluation by boss and cow-orkers.

    They’re being driven by fear of lawsuits.

    Bosses and cow-orkers have always had a pretty good feel for who’s contributing and who is not. But that feel is subjective and hard to defend in court.

    “Scientific” metrics are objective and defensible in court. Even when they aren’t actually very effective metrics.

  11. Related: forthcoming FAA study about the hazards of excessive reliance on cockpit automation

    Note especially this:

    In addition, the draft recommended training for rare but potentially catastrophic malfunctions “for which there is no specific procedure” or readily available checklist.

    The panel also called on manufacturers to develop cockpit designs that are “more understandable from the flightcrew’s perspective” and specifically guard against technology failures resulting from integration of various onboard systems.

    Kathy Abbott, a senior FAA scientist and one of the committee’s three co-chairs, declined to comment. In the past, she has said excessive reliance on computer aids means pilots “sometimes are not prepared to deal with non-routine situations,” especially when the message from airline management and trainers “is that automated systems can do the job better” than humans.

  12. I don’t think the Uncle Henry and Charlie Kellstadt types are going to disappear from this world any time soon. In 1975- after thirty years of Communist occupation – my Hungarian friends claimed that most factories and businesses had a nominal, Party-sanctified leader – but behind the scenes they kept old Henrik and Karoly,who actually understood how the enterprise worked.

  13. Margaret…yes, Koestler referred to something similar in writing about his travels in the Soviet Union…there were a small number of dedicated people who actually made things work (to the extent that things worked at all).

    But bad systems and bad measurements can greatly reduce the ability of the people with hands-on knowledge to do their jobs properly. There was a famous case in which Hershey implemented a new ERP system, resulting in orders and inventory being so screwed up that a significant part of the Halloween season revenue was lost, and lots of similar cases…in most of these, there were probably people who knew very well that producing 700 Gerbilator Type C11 products and only 50 Gerbilator C15s was a bad mistake, but that’s what the system said to do, so that’s what was done.

  14. I think it’s worth noting, also, the relentless warping effect of collectivist ideology, and the actions of collectivist regimes around the world, that caused the US to move on paths it would never have chosen based on its own theories of government and our national interests.

    How much money, energy, intellectual capital, and general loss of focus has been caused by the statist theories we borrowed from European thinkers and politicians, whether labeled progressive or socialist or some other euphemism.

    The decades long militarization of our culture due to WW2, and then the Cold War, caused enormous damage to our self-image as a society without a large and powerful omni-present state, and the idea of the state as the final voice in all our affairs can largely be traced to those periods of war and threat of war.

    I have long believed that the greatest challenge to our future prosperity is the weight of government, at all levels, and the enormous amount of time and resources taken up, and mostly wasted, in satisfying the demands of all the various rules and regulators that have layered themselves like barnacles on our every private activity and initiative.

    The state does not produce, it only absorbs and re-directs the production of the private sector. It is imperative we act to reduce it in both respects.

Comments are closed.