Daron Acemoglu and Simon Johnson. Power and Progress: Our Thousand-Year Struggle Over Technology and Prosperity. New York: Hachette Book Group, 2024.
Daron Acemoglu and Simon Johnson won the 2024 Nobel Prize in Economics for their research on how political and economic institutions shape national prosperity. In this book, they tackle the relationship between technological innovation and prosperity.
No one doubts that new technologies have the potential to boost productivity and raise living standards. How and when they actually accomplish this is a more difficult question.
In the introduction and first three chapters, the authors lay out their general theory of technology and progress, considering the role of variations in labor demand, wages, and social power. The next four chapters discuss how these variations have played out in various historical situations, ranging from the failure of innovation to benefit farmworkers and early manufacturing workers before the late nineteenth century, to the more widespread prosperity of the mid-twentieth century. Armed with insights from economic theory and history, the authors then address the more recent revolution in digital technology. Readers who follow the argument all the way through should come away with a better understanding of our current technological age and its discontents. I know I did.
The productivity bandwagon
The conventional wisdom in economics, as well as a lot of public discussion, is that technological advances raise productivity, and higher productivity raises living standards. The authors cite Gregory Mankiw’s popular undergraduate textbook, which says that “almost all variation in living standards is attributable to differences in countries’ productivity.”
But the productivity gains from new technologies can only raise living standards if they improve real wages. What about labor-saving technologies that lower the demand for labor, causing unemployment and lower wages? Mankiw acknowledges the problem, but minimizes it by claiming that “most technological progress is instead labor-augmenting.” Most workers find some way to work with new technologies, and their increased productivity enables them to command a higher wage.
Acemoglu and Johnson call this optimistic view the “productivity bandwagon.” They argue to the contrary:
There is nothing in the past thousand years of history to suggest the presence of an automatic mechanism that ensures gains for ordinary working people when technology improves… New techniques can generate shared prosperity or relentless inequality, depending on how they are used and where new innovative effort is directed.
Rather than accept a broad generalization about technology and prosperity, the authors want to study historical variations and identify the key variables involved. The stories that people tell themselves about technology—including the ones economists tell—can both reflect and affect the historical variations. Writing in the Great Depression, John Maynard Keynes coined the term “technological unemployment.” He could imagine “the means of economising the use of labour outrunning the pace at which we can find new uses for labor.” More recently, robotics and artificial intelligence are raising that possibility again, but the productivity bandwagon remains a popular narrative. Economic elites who profit from the application of new technologies are especially fond of it.
Variations in labor demand
Acemoglu and Johnson maintain that technological advances may or may not increase the demand for labor, depending on whether they are labor-augmenting or just labor-saving.
A classic example of technology that augmented labor, increased labor demand, and raised wages is the electrified assembly line introduced by Henry Ford. It not only raised the productivity of the existing autoworkers; it also enabled auto manufacturers to employ additional workers productively. (Economists call that variable the “marginal productivity of labor.”) By producing more cars at lower cost, car companies created a mass market for what had been a luxury item. In addition, they created additional jobs in related industries, such as auto repair, highway construction and tourism.
The effects of today’s robotics on automobile manufacturing may be very different. Carmakers can make just as many cars with less human labor, so labor productivity goes up. But demand for additional labor may go down, if factories are already turning out as many cars as their market can absorb. The marginal productivity of labor then falls, and the connection between technology and prosperity is weakened.
That is not to say that automation is always bad news for workers. That depends on the balance of labor-saving and labor-augmentation:
For most of the twentieth century, new technologies sometimes replaced people with machines in existing tasks but also boosted worker effectiveness in some other tasks while also creating many new tasks. This combination led to higher wages, increased employment, and shared prosperity.
The problem then is not just automation but excessive automation, especially if it is not really very productive in the fullest sense of the word. In economics “total factor productivity” refers to the relationship between economic output and all inputs, including capital as well as labor. Replacing workers with machines has costs as well as benefits, since machines cost money too, and displaced humans might have contributed something that machines cannot. The authors use the term “so-so automation” to refer to replacement of workers without much productivity gain. In that case, the classic gains of the earlier automobile boom—lower costs, expanded markets, rising labor demand, and widespread prosperity—do not occur.
Variations in wages
Even if new technologies are labor-enhancing, higher wages do not necessarily follow. They have not followed in societies where workers have been coerced to work without pay, or forbidden to leave their employer in search of better pay. The cotton gin enhanced the productivity of cotton workers in the Old South and expanded the areas where cotton could be profitably cultivated. But “the greater demand for labor, under conditions of coercion, translated not into higher wages but into harsher treatment, so that the last ounce of effort could be squeezed out of the slaves.”
In modern, free-market labor systems, wages are freer to rise along with labor demand. However, “wages are often negotiated rather than being simply determined by impersonal market forces.” A dominant employer may set wages for a multitude of workers, while the workers are too disorganized to bargain from strength. It was only in 1871 in Britain and 1935 in the United States that workers gained the legal right to organize and bargain collectively. Opponents of organized labor have continued to find ways of discouraging labor unions to this day. The share of national income going to labor rather than capital was highest when unions were strongest, in the 1950s.
Variations in power
Acemoglu and Johnson argue that the effects of technology depend on “economic, social, and political choices,” and that “choice in this context is fundamentally about power.”
What societies do with new technologies depends on whose vision of the future prevails. The most powerful segments of society have more say than others, although they can be contested by countervailing forces, especially in democratic societies where masses of workers vote. Although plenty of evidence points to the self-serving behavior of elites, they must at least appear to be promoting the common good for their views to be persuasive.
The technological choices a society makes can serve either to reinforce the power of elites or empower larger numbers of workers. This is especially true of general technologies with many applications. In the twentieth century, the benefits of electricity helped power a more egalitarian, broadly middle-class society. We cannot yet say the same about the digital technologies of the present century. The authors apply their theory, buttressed by historical evidence, to explain why.
Recall the subtitle of the book: “Our Thousand-Year Struggle Over Technology and Prosperity.” Making technology work for all of us has always been a struggle, and one that is related to the struggle for true democracy. Looking at it that way is more realistic and enlightening than seeing only a “productivity bandwagon” rolling smoothly toward mass prosperity.
Beneath the Surface 