Better Average Than Unequal
In his latest book, Tyler Cowen takes up where The Great Stagnation, his penultimate work, left off. If America’s economy has been stagnant, as the previous book contended, is there a way out of the morass? And if so, what will the post-stagnation world look like? Cowen’s answer, which has to do with the future course of human-machine interactions, does not paint a pretty picture.
Before heading into the book’s substance, I want to note that Cowen has truly mastered the art of writing a book for the “intelligent layperson.” When I sat down to start reading it, since it was “work reading,” I told myself I’d read for an hour before turning to “pleasure reading.” Four hours later I had not put the book down: his prose is that engaging.
Cowen’s main theme is the increasing importance of “intelligent” machines to our economy. The book is littered with interesting facts related to the increasing power of computers. For instance, we learn that “Today’s iPhone would have been the most powerful computer in the world as recently as 1985.” On the other hand, we still need plenty of people to get them to do what we need them to: “According to the Air Force, keeping an unmanned predator drone in the air for twenty-four hours requires about 168 workers laboring in the background.” That latter type of fact is crucial to understanding Cowen’s prognosis for the economy. The people who will get ahead in Cowen’s new economy share a common skill: “The ability to mix technical knowledge with solving real-world problems is the key, not sheer number-crunching for its own sake.” They can blend their skills with machines in a symbiosis that is more effective than humans or machines working alone.
Those who lack such skills face a bleak future, the handwriting of which is already on the wall: “But for men, from 1969 to 2009, as measured, it appears that wages for the typical or median male earner have fallen by about 28 percent.” Per Cowen, the fact that this trend will continue is just something we have to come to terms with: “In essence, the American economy is learning that—for structural reasons—it can’t afford as many mid-wage jobs as it used to.” Cowen is fair-minded enough to admit that this “coming world of hyper-meritocracy I’m sketching is not necessarily a good and just way for an economy run.” But we really shouldn’t worry too much about this:
Sure [smart and intelligent and conscientious and talented] are good qualities overall, but we don’t always need to price them above all other human qualities. How we value them does have a moral implication, but such moral judgments can be left to each and every one of us to make.
Is it perhaps relevant that if we demand that social rewards for cleverness must not enter the realm of public morality, people like Cowen will come out ahead? I’m pretty sure that he does not take a similar laissez faire attitude towards rewards going to brute strength and a propensity to violence. And, furthermore, the idea of private morality makes no sense: to declare something is wrong is to make a claim about reality, not about the state of one’s own preferences.
Cowen does not seem to recognize that if his analysis of our economic situation is accurate, and machines are increasingly replacing lower-skilled office workers, then the current push to get everyone into college is sorely mistaken. After all, for what sort of job does a degree from an average college and a transcript with average grades and no specialized STEM training qualify one? Generally speaking, not work as a high-priced consultant or a finance wizard, but a low-skill office job: precisely the sort of jobs that are disappearing due to automation.
But there is a category of job that requires intelligence, can be very fulfilling, and allows workers great autonomy and the opportunity to exercise creativity: high-skill manual-labor or personal-service jobs. If the returns on capital are increasing due to “smart” machines, the people getting those returns are going to be expanding their houses, decorating them, hiring musicians for parties, installing custom tile work, getting massages, having gourmet meals prepared for them at home, and so on. Such jobs cannot be outsourced to India, and none of them require a college degree. Cowen acknowledges this fact: “As workers are displaced by smart machines in manufacturing and other areas, more individuals will be employed [in offering] direct personal services.” Thus if a high school guidance counselor cares about the future of his charges, he shouldn’t be sending a disinterested C student on to become a disinterested C student at a second-tier college. Tell that student to become a carpenter, or a skilled tile worker, or an auto mechanic, or a chef, so they can be good at something they like rather than merely mediocre at something they don’t like.
Far from questioning the conventional wisdom that pushes as many people into college as possible, Cowen appears to support it. He notes that some wealthy metropolitan areas, such as Raleigh, San Francisco, and Stamford, have a high percentage of college graduates and are getting an influx of more of the same, and he recommends more education as an antidote to struggling areas’ economic ailments. But if Detroit increases college subsidies for Detroit residents, what I would expect to see is: 1) the more ambitious beneficiaries of this program moving to Raleigh, San Francisco, and Stamford; and 2) the less ambitious beneficiaries stuck in Detroit with few job prospects. How this would help Detroit Cowen does not make clear.
A common mistake runs through Cowen’s book. He is right to reject such overblown claims for “machine intelligence” as Ray Kurzweil’s nonsense that we are going to “upload” our brains into computers one day soon. But Cowen does not seem to realize that the phrase “machine intelligence” is itself only a metaphor. The intelligence comes from the designers of the machine, just as the intelligence of a rabbit trap comes from the hunter who built it. This misapprehension is apparent in such thoughts as: “Computers are still struggling to master the more complex board games of Shogun and Go.” Imagine these poor computers, staying up late hours in their machine rooms, “struggling” to play Go better. Isn’t it, in fact, humans who are struggling to write programs that do well at Go?
Of particular interest to conservatives who favor protectionism as a solution to middle-class woes, Cowen offers convincing evidence that while off-shoring might contribute to declining real wages, they are minor parts of America’s employment troubles. For instance, imports from China, a bugbear for protectionists, account for only about 1.3 percent of American spending. So even if we banned all imports from China, the impact on American wages would be minuscule. Protectionism will not save the blue-collar middle class.
Cowen offers a puzzling perspective on the future of science. First of all, he thinks the era of major theoretical breakthroughs is over; we have had our Newtons, Lavoisiers, and Darwins, and their day is past. Furthermore, he believes our best “theories” soon will be generated by computers, and humans will not understand them: “Eventually … science will also look more like religion and magic because of its growing inscrutability.” But is it useful to call an incomprehensible computer model that happens to make good predictions a “theory”? And Cowen ignores something like the Langlands program, which is uniting diverse fields of advanced mathematics in one comprehensible, overarching framework. Perhaps he is the modern equivalent of the nineteenth-century physicists who thought their entire field was complete in its essentials, just before it was hit by two radical revolutions.
Cowen is overly confident that today’s trends will simply continue unchanged, and overly sanguine about social cohesion in a society divided between a wealthy, “productive” class and a class of proles kept docile by video games and legalized pot. But while the future he depicts is not attractive, we probably don’t need to worry about it too much: given that someone is forecasting it, it almost certainly isn’t the future we will actually get.
Posted: November 17, 2013