In Power and Progress, Massachusetts Institute of Technology economists Daron Acemoglu and Simon Johnson argue that for technology to be beneficial to society and create no unemployment, inequality, or poverty, it needs to be controlled by social power (especially trade unions) and government regulation. “Progress is never automatic,” they say; it must be directed.

According to them, the first phase of the industrial revolution impoverished workers, and we are now facing a similar threat. But, they assure readers, the trajectory of technology is a choice, and it does not have to be left to the market. The shift of power to government regulation and organized labor starting in the late 19th and early 20th century translated into fast growth after World War II and a better sharing of the “rent” of technology between capital and labor. This shared prosperity crashed in the 1980s with the new information technology, automation, and free-market doctrines like Milton Friedman’s. Today’s advances in artificial intelligence threaten to make the situation worse. Moreover, social media works against democracy. We need to reorient technology to a “socially beneficial trajectory.”

The book’s economic and historical scope is wide and ambitious. One finds a few good points in its 500 pages—for example, the danger of technology used for government surveillance. However, many of the authors’ claims are doubtful and often puzzling.

Acemoglu and Johnson give a very wide definition of power as “the ability of an individual or group to achieve explicit or implicit objectives.” They tell us that Lord Acton’s aphorism about government’s coercive power—Power tends to corrupt and absolute power corrupts absolutely—“applies just as aptly to persuasion power, including the power to persuade oneself.” They don’t seem to think that there is a significant difference between a government command imposed by force and a private choice reached by an individual or a voluntary group of individuals. This fuzzy notion of power is everywhere in the book.

Industrialization / The authors review the history of the Industrial Revolution in the United Kingdon, where it started around the middle of the 18th century. During the first hundred years or so, they claim, it was detrimental to ordinary workers. From then on, they admit that conditions improved, a reversal ascribed to the workers’ new political power. The authors’ stance on the exploitation of powerless workers for a hundred years or so does fit well in their ideological narrative.

However, their view is hotly debated among historians. Emma Griffin of Queen Mary University argues that ordinary workers benefited very early in the Industrial Revolution from the opportunities offered by steady jobs. As a result of increasing incomes, for example, the urban age of marriage started declining by the end of the 18th century. More generally, she writes:

Yet even with a government that did nothing, there is an uncomfortable truth that we should confront: industrialisation had remarkable power to put food on the table. And for the first generation, that generation which had expected the hunger of their own childhood to be experienced once more by their children and their grandchildren, food on the table was all that really mattered. … Critics will argue that the material gains for most families were small. But they were just enough to drag wage-earners out of the servile submission that poverty had forced upon them since time immemorial.

Progressivism / Acemoglu and Johnson laud the Progressive movement in America’s late 19th and early 20th century. They admit that the movement had “unappealing elements, including the overt and covert racism of some of its leading lights (including Woodrow Wilson), ideas of eugenics …, and Prohibition.” Well, it was not only ideas of eugenics that should be mentioned, but the 65,000 women who were forcibly sterilized in America between 1907 and 1980. The main weapon of government intervention is not roses.

“Narrow vision and selfish interests” were challenged. Our authors praise “Nordic choices” and the “Scandinavian social democratic system” and its “corporatist model,” which already worked better than the American economy. They have remained fans of the Scandinavian and German models. They even quote Rexford Guy Tugwell, a member of Franklin D. Roosevelt’s “Brain Trust,” who advocated “a strong government with an executive amply empowered by legislative delegation.” They don’t mention that Tugwell also thought that new industries should not “just happen” without government permission. (See “Total Regulation for the Greater Whole,” Fall 2014.) That Johnson teaches “entrepreneurship” at MIT is somewhat intriguing.

Power and Progress lauds Roosevelt for recognizing “the right of workers to collectively organize,” helped by special coercive power granted by government. And they seem surprised that Fr. Charles Coughlin, a former populist, New Dealer, and creator of the National Union for Social Justice, later became a supporter of Mussolini and Hitler.

To be clear, I don’t criticize Acemoglu and Johnson for sympathizing with workers’ voluntary associations. I do, too. But coercive bargaining power and compulsory membership are another matter. The coercive formula prevents a market test of the unions’ efficiency.

Losing our way / After World War II, the authors note, union power grew and technology became labor friendly. In the 1960s, President Lyndon Johnson launched the War on Poverty and boosted the welfare state. The authors of Power and Progress ignore some inconvenient facts, such as that welfare-state assistance decoupled the bottom income-distribution quintile from the labor market. In this quintile in 1967, about two-thirds of able-bodied working-age persons who were not full-time students worked, but that figure fell to about one-third by 2017. Power and Progress also seems to accept uncritically the misleading ideas that circulate about inequality. (See “Is Inequality Bad, Large, and Increasing?” p. 53.)

The reader will find many economically intriguing statements in the book. Acemoglu and Johnson argue that, after 1980, “we lost our way” and abandoned worker-friendly technology and “the shared-prosperity model of the early postwar decades.” That was a bad choice: “the bias of technology was very much of a choice—and a socially constructed one.” A bit of postmodern jargon cannot hurt. We find many such catchy statements in Power and Progress; for example, “By the 1980s, many American managers came to see labor as a cost, not as a resource.” What about a resource that has a cost—strange, eh?

It is difficult to imagine that robots could destroy 50 percent of the jobs and the incomes that go with them, as the authors seem to suggest. Statistically, the number of jobs simply grows with the working-age population as new entrants in the labor market find ways to satisfy the unlimited desires of their fellow humans. Our authors paraphrase economist Wassily Leontief who, in 1983, worried that “human labor would go the way of the horses and become unnecessary for modern production.” These fears don’t have better foundations than the similar scares that have popped up regularly since the birth of modern technology.

Magical society / Back to the notion of choice, which is another fuzzy matter in Power and Progress. The authors insist that machines must be useful “to human objectives.” But since human objectives are not identical across individuals, that’s not saying much. Although the book does mention Nobel economist Friedrich Hayek (once, but strangely with no bibliographical citation), its authors don’t seem familiar with his view that a free society is one where each individual is allowed “to use his own knowledge for his own purposes”—which means for his or her own objectives.

The authors also want to put machines “at the service of people.” But the way to make basically anything useful to diversified human purposes is to let private choices be made on free markets and allow demand to determine production. Collective choices will at best be uniform; at worst tyrannical. Private choices on markets are more efficient to determine “where exactly society has the greatest need”—assuming that this way of expression can be interpreted in any other way than social anthropomorphism.

The authors often repeat that technological arrangements and developments are choices. So, who gets to make those choices, political rulers with commands or individuals through contracts? For Acemoglu and Johnson, there seems to be no difference between a “choice” as the unintended result of multiple decentralized private choices or as a decision imposed by political authority, provided the latter acts on behalf of 50 percent + 1 of some group. The only distinction they seem to make is that the former are bad and the latter good.

Contrary to the impression they give, we can’t hide behind words like “social.” Power and Progress uses it as what Hayek called a “weasel word” or what should perhaps be a laundering word. Except for social sins and other negative substantives, everything with “social” appended is supposed good. (See the chapter on “Our Poisoned Language” in Hayek’s The Fatal Conceit.) In Power and Progress, similarly, a mythical “civil society” is called for whenever a deus ex machina is needed. In another book by Acemoglu with James Robinson, The Narrow Corridor, the state becomes more and more powerful but is kept in check by a more and more powerful “civil society” in a manner that seems magical. (See “A Shackled Leviathan That Keeps Roaming and Growing,” Fall 2021.) In a few places in both books, the magical “social” mutates into “societal,” which only has the look and feel of something more scientific.

A major rule of standard economic analysis is to carefully distinguish between, on the one hand, positive analysis—that is, explanations of what is—and, on the other hand, what should be, or the normative values that underlie policy proposals. Acemoglu and Johnson ignore that distinction.

For example, they write that “when there are major decisions about the direction of technology, there is no guarantee that the market-based innovation process will select areas that are more beneficial for society as a whole or for workers.” They also write that some technologies profitable for businesses may “not contribute to or may even reduce social welfare.” This raises so many questions: Who says so? Who calculates what is more beneficial for “society as a whole”? Who is “society as a whole”? How do our authors measure “social welfare,” if only in theory? Perhaps they would answer that they are following some sort of utilitarian cost–benefit analysis, but they would then contradict themselves when they oppose any cost–benefit tradeoffs in the regulation of children’s working conditions.

To be clear, I am not arguing for cost–benefit analysis and interpersonal comparisons of utility. I believe that the moderate classical-liberal principles of a free society are more along the lines of Hayek and James Buchanan’s theories. (See “An Enlightenment Thinker,” Spring 2022.) It would be useful for the reader to know what exactly is Acemoglu and Johnson’s political philosophy.

Angelic democracy / To these critiques, the authors of Power and Progress may invoke a standard objection: democracy will determine what’s good and what’s bad. But what is democracy? The closest they come to defining it is when they write that “democracy, above all else, is about a multitude of voices, critically including those of ordinary people, being heard and becoming significant in public policy decisions.” But how are these voices aggregated? This is a standard economic problem that Acemoglu and Johnson ignore. What about the opinion of Hayek and many classical liberals that democracy is merely a way to assure peaceful transfers of power? (See “Populist Choices Are Meaningless,” Spring 2021.)

In short, it seems to me that Acemoglu and Johnson espouse a simple and angelic conception of democracy, which may be synonymous with “good” and “social.” I suggest they would greatly benefit from studying the public-choice explanations of how collective choices are made in different forms of democracies—majoritarian versus constitutionally limited, for example.

The government that Acemoglu and Johnson hope will follow their advice is not a government found in the real world, but their ideal government, not of this world. Consider, for example, their discussion of the 2007–2009 Great Recession, for which they blame greedy corporations. The only blame they assign to the federal government is to not have regulated enough. Remember that the crisis started in the market for mortgage-backed securities (MBS). It was Ginnie Mae, a government sponsored enterprise (GSE) created by Congress in 1968, that pioneered the issuance of MBSs in 1970. Up to the MBS crash, numerous politicians were on record claiming the desirability of more mortgages for poorer households. For example, then-congressman Barney Frank wanted to “roll the dice” and “get Fannie and Freddie [two other federal mortgage GSEs] more deeply into helping low-income housing.” As for a lack of government supervision, Stanford economist John B. Taylor noted that hundreds of regulating bureaucrats were working on the premises of large banks before the crash.

Still, Acemoglu and Johnson maintain that a benevolent, powerful, and effective government would lead society to nirvana: “We must find ways of countering power with alternative sources of power and resisting selfishness with a more inclusive vision.” But nirvana is a dangerous mirage.

To be fair to the authors, they do mention the constraint of the rule of law once, and constitutional constraints a couple of times. But they don’t explain how the vast new powers they want to grant to the state are consistent with such constraints.

They notably propose to “redirect technological change,” remake digital technologies, and create broader-based trade unions. They want “worker-friendly technologies,” government-funded worker training programs, data ownership regulations, the break-up of tech giants (Google, Facebook, Amazon), and digital advertisement taxes. Mentioning Sens. Bernie Sanders and Elizabeth Warren, they long for wealth taxes. It nearly goes without saying that “society should strengthen its existing social safety net.”

Perhaps more worrisome is what seems to be the theoretical foundation of the redistribution they are after. They consider any technological innovation and probably any advancement of any sort as creating a sort of “rent” to be shared between “labor” and “capital” under the diktats of some political authority. They don’t seem to realize that there is no such floating rent belonging to nobody and waiting to be politically apportioned. Everything belongs to the actors without whom it wouldn’t otherwise exist, and free markets are the only known mechanism to distribute these rewards so that opportunities and production are maximized. A state with the power to redistribute the alleged floating rent from social cooperation—including the work of academics, no doubt—would have to be a totalitarian state. Leviathan’s monitoring and surveillance would certainly increase, not decrease.

Despite its authors’ good intentions, Power and Progress looks like the work of philosopher-kings à la Plato. German poet and philosopher Friedrich Hölderlin had an answer when he wrote (as quoted by Hayek) that “what has always made the state a hell on earth has been precisely that man has tried to make it his heaven.”

Readings

  • Law, Legislation, and Liberty (1973–1978), by F.A. Hayek, edited by Jeremy Shearmur. University of Chicago Press, 2021.
  • Liberty’s Dawn: A People’s History of the Industrial Revolution, by Emma Griffin. Yale University Press, 2013.
  • The Fatal Conceit: The Errors of Socialism, by F.A. Hayek. University of Chicago Press, 1989.
  • Three Generations, No Imbeciles: Eugenics, the Supreme Court, and Buck v. Bell, by Paul Lombardo. John Hopkins University Press, 2008.
  • Why I, Too, Am Not a Conservative: The Normative Vision of Classical Liberalism, by James M. Buchanan. Edward Elgar Publishers, 2006.