Why conventional economic theory is wrong about technological change
Apr 22, 2024Society as a whole has a critical interest in the direction of technological innovation. This cannot be left uniquely to a limited group of capitalist bosses. Consultation with all the key interest groups and government regulation have a critical role to play in ensuring future economic growth and a fair go for all.
This article argues that the simplified assumptions underpinning conventional economic analysis do not allow an accurate assessment of how the benefits of technological progress will be shared.
The two previous articles in this series of three, summarised the findings by Acemoglu and Johnson in their new book, Power and Progress: Our Thousand-Year Struggle Over Technology and Prosperity, about how the benefits from technological progress have been shared in the past (first article) and how AI might affect the labour market in future (second article).
Finally, in this third article, it is interesting to consider how Acemoglu and Johnson’s conceptual framework differs from the conventional wisdom in economics. In a discussion at the end of their book, they find that their model is different in four critical ways.
First, as summarised in my first article, most of the book examines how the benefits of technological change have been shared through history. This examination leads Acemoglu and Johnson to question the validity of the conventional economic theory which assumes that all are supposed to share in the benefits of increased productivity.
Although the conventional economic model does accept that technological progress can raise inequality, it nevertheless assumes that it will increase the wage levels of all types of labour, even if not equally. On the other hand, while Acemoglu and Johnson accept that if technology raises productivity by enough, it can increase demand for the new products and/or for inputs from other sectors and thus wages generally, they argue that if productivity gains and cost reductions are small, as they mostly are, these beneficial gains will not be shared by everyone.
Second, most models of economic growth either take the path of technological change as exogenous or assume that it proceeds along a given trajectory. By contrast, Acemoglu and Johnson differ by emphasising the malleability of technology and thus how new techniques will economise on different factors of production and the change in their productivities is therefore a matter of choice.
Third, most conventional economic analysis, even when it recognises important deviations from the benchmark of competitive labour markets, do not consider how these deviations will influence how productivity increases will translate into wages growth. In addition, new technologies can also change the balance of power against workers thus further changing the link between higher productivity and increased wages.
Fourth, Acemoglu and Johnson argue that the future choices of technology are largely shaped by a vision of what it can achieve. They then offer a theory of how a vision of future technology is developed, how it will impact on society, and the role of social power exercised on behalf of a broader community in shaping such visions. Specifically, because technology is malleable, who and how the direction of technology is determined, and thus who wins and who loses, become central.
In sum, the analytical approach proposed by Acemoglu and Johnson takes account of institutions and social factors and is much more likely to describe what happens in the real world than the simplifying assumptions’ underpinning conventional economic growth analysis.
However, the conventional economic growth analysis provides the foundation for much of the neo-liberal economic agenda favouring smaller government and allowing the market to rule. Instead, what Acemoglu and Johnson have demonstrated is that when the choices about future technologies are made by powerful vested interests there can be no guarantee that all will benefit and certainly not equally.
Thus, they provide a powerful argument in support of government intervention to ensure a more balanced outcome. In addition, this balanced outcome is also more likely if there are other countervailing forces such as trade unions involved in the decision making.
Furthermore, this more inclusive decision-making will not necessarily slow the rate of technological progress. Indeed, it may even accelerate it, for example by ensuring that workers are retrained for the new tasks and can then benefit too.
But what we have witnessed since the early 1980s is an increase in inequality driven by technological choices made in favour of automation. This emphasis on automation has led to the hollowing out of middle-level jobs, without the retraining and creation of new jobs in too many of the advanced Western economies.
This changing job structure has also led to a loss of consumer and thus aggregate demand. This lack of consumer and aggregate demand over the last decade or more is the reason for the economic stagnation that these countries have experienced, while productivity growth has been the lowest since the 1940s, notwithstanding the spread of digital technologies.
In short, society as a whole has a critical interest in the direction of technological innovation, and this cannot be left uniquely to a limited group of capitalist bosses. Consultation with all the key interest groups and government regulation have a critical role to play in ensuring future economic growth and a fair go for all.
What we should be concentrating on is not smaller government, but good government, while recognising that good government can include how to make markets work better.
For Part 1 and Part 2 on this topic: