By Tony Ditta
The explosive growth of artificial intelligence has put technology at the center of economic and political debates. We are uncertain about how it will affect individual’s lives as it changes how people work and learn and how it will affect the global economy as a whole. In the face of these uncertainties and the rapidness of technological change, understanding the best policy response is a pressing issue.
Daron Acemoglu, Institute Professor of Economics at MIT, joined the Industrial Policy event for a conversation entitled “The Imperative to Redirect Technological Change” with Dani Rodrik, Reimagining the Economy Faculty Co-Director and Ford Foundation Professor of International Political Economy at the Harvard Kennedy School. Here are some of the main takeaways from the discussion.
- Markets won’t always create socially desirable technological change
Markets are incredibly powerful tools for many purposes, not the least of which is creating technological change. Market actors are able to gather people and resources, coordinate their activities, and convert their innovations into products that people want — all at incredible speed and scale. However, their private incentives don’t always line up with socially desirable outcomes.
Markets may not be the best guide for technological advancement for a few reasons. One is path dependence: entrepreneurs and innovators pursue technologies based on what already exists to follow popular fads or take advantage of lower costs from an existing knowledge base. But popular or cheap technologies may not be the ones that help society the most. Acemoglu gives the example of automation in the 1950s and 60s and AI currently; these were/are popular, but they may destroy many jobs.
The most important reasons markets don’t create the right kind of innovation are externalities: spillovers from economic activity that aren’t accounted for by market prices. The most obvious example is climate change. Climate change imposes a cost on everyone on the globe (including future generations), but this cost doesn’t appear in the market price of carbon emissions. So people emit too much, and not enough is invested in green technologies. Another example is social media and the mental health of children. Social media platforms invest a great deal in technologies that keep children engaged because they benefit from advertising to children and don’t bear the cost of their impact on children’s wellbeing.
- Industrial policy can help…
In the past, many economists and policymakers thought about correcting market failures in technology as simply increasing the amount of innovation. Under this framework, the best policy response was broad subsidies and regulations around intellectual property rights. But the market problems with technology aren’t just about the amount of innovation, they’re also about the direction.
Here, industrial policy can be useful. Industrial policy is more targeted than other approaches. By focusing on a particular social goal, policymakers can tackle a particular market failure and direct innovation in a desirable way. So, for example, support for green technologies in the Inflation Reduction Act isn’t just about encouraging innovation in a general sense; it’s about correcting the fact that markets won’t invest enough in green technologies.
- … but directing technological change requires a multi-pronged approach.
Industrial policies are promising, but Acemoglu emphasized that much more is required. He laid out five policy principles for directing technological change.
- Democratic voice - The socially desirable direction of technology isn’t always clear. By its nature, new technology is full of uncertainty. So, when deciding which direction to pursue, policymakers must get input from the public. There is a tendency to go to business leaders or scientists on the cutting edge — people who are well-informed and have a large stake in the outcome. Obviously, it is important to hear from such people, but they represent a narrow set of interests. Good leaders should listen to a wider set of voices, including labor leaders and consumer advocates.
- Regulation - Sometimes the right response is strict regulation: hard and fast rules promoting good technologies or outright bans on bad ones. This is more forceful than the common recommendation from economists to adjust incentives by subsidizing the good and taxing the bad. The “gentler” approach allows market forces to adjust on the margin and make changes in quantity or quality, but Acemoglu says this isn’t always enough. He went back to the example of social media and children’s mental health. If the evidence continues to show that social media is harmful — even leading children to take their own lives — a marginal adjustment isn’t enough. The market needs strict rules and strict enforcement.
- Correcting prices with taxes - The government can use taxes to “internalize” the costs of externalities. These are called “Pigouvian taxes,” and economists recommend their use for externalities for which the social costs can be reasonably well estimated. If the tax makes the private cost match the social cost, then incentives are better aligned. Pigouvian taxation is the motivation behind carbon emission pricing such as a carbon tax or cap and trade.
- Subsidizing desirable technologies - This is the step where industrial policy comes in. In addition to discouraging bad activities with taxes, the government can encourage good activities with subsidies.
- Maintaining competition - Many incentives, such as tax breaks or some subsidies, mainly benefit incumbent firms. However, a lot of new technologies — especially groundbreaking, disruptive technologies — come from new entrants. So, policies should be designed to ensure that incumbents can benefit from good investments, but they shouldn’t be protected from new firms with big ideas. The new firms should be supported, too.
