Wednesday, November 26, 2014

Technology and Inequality

CAMBRIDGE – Until now, the relentless march of technology and globalization has played out hugely in favor of high-skilled labor, helping to fuel record-high levels of income and wealth inequality around the world. Will the endgame be renewed class warfare, with populist governments coming to power, stretching the limits of income redistribution, and asserting greater state control over economic life?

There is no doubt that income inequality is the single biggest threat to social stability around the world, whether it is in the United States, the European periphery, or China. Yet it is easy to forget that market forces, if allowed to play out, might eventually exert a stabilizing role. Simply put, the greater the premium for highly skilled workers, the greater the incentive to find ways to economize on employing their talents.

The world of chess, with which I am closely familiar, starkly illustrates the way in which innovation in the coming decades may have a very different effect on relative wages than it did over the past three decades.

During the late eighteenth and early nineteenth centuries, a brilliantly inventive chess-playing “automaton” toured the world’s capitals. “The Turk” won games against the likes of Napoleon and Benjamin Franklin, while challenging many great minds to penetrate its secrets. Concealing a human player in a shifting compartment amid a maze of impressive-looking gadgetry, it took decades for outsiders to correctly guess how the Turk really worked.

Today, the scam has been turned on its head: chess-playing machines pretend to be chess-playing humans. Desktop-based chess programs have considerably surpassed the best human players over the past decade, and cheating has become a growing scourge. The French chess federation recently suspended three of its top players for conspiring to obtain computer assistance. (Interestingly, one of the main ways to uncover cheating is by using a computer program to detect whether a player’s moves consistently resemble the favored choices of various top computer programs.)

Of course, there are many other examples of activities that were once thought exclusively the domain of intuitive humans, but that computers have come to dominate. Many teachers and schools now use computer programs to scan essays for plagiarism, an ancient transgression made all too easy by the Internet. Indeed, computer-grading of essays is a surging science, with some studies showing that computer evaluations are fairer, more consistent, and more informative than those of an average teacher, if not necessarily of an outstanding one.

Expert computer systems are also gaining traction in medicine, law, finance, and even entertainment. Given these developments, there is every reason to believe that technological innovation will lead ultimately to commoditization of many skills that now seem very precious and unique.

My Harvard colleague Kenneth Froot and I once studied the relative price movements of a number of goods over a 700-year period. To our surprise, we found that the relative prices of grains, metals, and many other basic goods tended to revert to a central mean tendency over sufficiently long periods. We conjectured that even though random discoveries, weather events, and technologies might dramatically shift relative values for certain periods, the resulting price differentials would create incentives for innovators to concentrate more attention on goods whose prices had risen dramatically.

Of course, people are not goods, but the same principles apply. As skilled labor becomes increasingly expensive relative to unskilled labor, firms and businesses have a greater incentive to find ways to “cheat” by using substitutes for high-price inputs. The shift might take many decades, but it also might come much faster as artificial intelligence fuels the next wave of innovation.

Perhaps skilled workers will try to band together to get governments to pass laws and regulations making it more difficult for firms to make their jobs obsolete. But if the global trading system remains open to competition, skilled workers’ ability to forestall labor-saving technology indefinitely should prove little more successful than such attempts by unskilled workers in the past.

The next generation of technological advances could also promote greater income equality by leveling the playing field in education. Currently, educational resources – particularly tertiary educational resources (university) – in many poorer countries are severely limited relative to wealthy countries, and, so far, the Internet and computers have exacerbated the differences.

But it does not have to be that way. Surely, higher education will eventually be hit by the same kind of sweeping wave of technology that has flattened the automobile and media industries, among others. If the commoditization of education eventually extends to at least lower-level college courses, the impact on income inequality could be profound.

Many commentators seem to believe that the growing gap between rich and poor is an inevitable byproduct of increasing globalization and technology. In their view, governments will need to intervene radically in markets to restore social balance.

I disagree. Yes, we need genuinely progressive tax systems, respect for workers’ rights, and generous aid policies on the part of rich countries. But the past is not necessarily prologue: given the remarkable flexibility of market forces, it would be foolish, if not dangerous, to infer rising inequality in relative incomes in the coming decades by extrapolating from recent trends.

  • Contact us to secure rights


  • Hide Comments Hide Comments Read Comments (3)

    Please login or register to post a comment

    1. CommentedStamatis Kavvadias

      Amazing. Does this person ever think straight? Does he all his opinions?

      "[...]market forces, if allowed to play out, might eventually exert a stabilizing role."

      ??? What is this? Is this guy blind to modern economics? Is he aware of Dr. Stiglitz PhD? "If allowed to play out"??? Whenever there is asymmetry of information --that is always-- markets are imperfect (I suppose, we do not know how imperfect: for all we know, they could lead to the worst possible outcome for the largest proportion of the population!).

      "[...] the greater the premium for highly skilled workers, the greater the incentive to find ways to economize on employing their talents."

      This sounds like ...letting employers "innovate" on how to prohibit technological innovation, since this is the basic engine of highly skilled employment creation. Alternatively, it could mean, allowing employers to "innovate" on ways in which payment of highly skilled workers will stagnate. That is exactly what happened with the US middle class.

      Is there any other way? Does prof. Rogoff agree to allow market forces "to economize on employing skilled workers' talents," but prohibit the previous two choices? Unless, we have to trust the markets to bring about yet another crisis?

      Maybe, the point of prof. Rogoff is exactly to let employers try to stagnate highly skilled worker income, and see if such workers will create firms to sell their skills in organized ways. This way, markets will be created (among these new firms) -- the dream of every economist, which will lead to the "best" price for highly skilled labor.

      If only we could find ways to automate the work of highly skilled financiers.... Their work is such a long way of producing any benefit for society (in fact the opposite is exaggerated because of the financial crisis), and they are extracting huge rents from the economy with their bonuses. Ah! And neither before nor after the crisis, did I see any result from the "incentive to find ways to economize on employing their talents"!!!

      Yes, there can be technological innovation that further pushes the boundary of skill required to have a job, rendering large sectors of knowledge subject to automated procedures and further levels of peoples' skills useless. Automated software production, can and will leave large numbers of plain web and application programmers without a use for their skills and, of course, without a job. This will save some of the need to employ skilled workers. But, such "savings" will always come from the bottom of the scale of skills.

      The only case when automation will be able to "save" humanity from most of skilled labor, will be when technological innovation stops and new knowledge becomes mostly the work of computers scanning big data. At that point, knowledge increase will pause, and automation will eventually catch up with all knowledge before that time. There are no signs of a slowdown in knowledge increase. Still, this could come in the form of diminishing returns from new knowledge. There are no broad signs of that either.

    2. CommentedMukesh Adenwala

      I found this argument to be evolutionary in nature. Indeed the life will assert itself and will become increasingly efficient over time. Automatism has higher appeal than systems requiring human interventions, which are bound to be imperfect. Economics is indeed one example where markets are preferred as organizing agents than governments. However, when we talk of human societies we forget that extinction, which is one of the necessary ingredients and outcomes of automated systems is not an option, and ways have to be find to a ensure smoother ride to a future about which we know almost nothing with any degree of certainty.

    3. CommentedShane Beck

      Too many variables in this skilled worker versus machine contest- price of education for skilled worker, cost and lifespan of technology including planned obsolsence, the fact that current globalization is planned this way to favor the elites whose skills cannot be transferred overseas- the politicians, the lawyers the bureaucrats, the security complex etc. Spain and Greece provide a test laboratory in which the variable of high unemployment will be tested against the stability of the state