economics

Is It Time for an American Industrial Policy?

American Compass
Political Science, Duke University
Genesis
Response
Penultimate
Finale

Michael Munger

Political Science, Duke University

May 27th, 2020
There are two kinds of arguments about industrial policy:
1. Does the system get the right COMPOSITION of investment in infrastructure, research, and product development?
2. Does the system get the right LEVEL of such investment?
In sequential debates one must respond to the initial argument. Oren Cass chose an example—BooBoo vs SteelSeam—about composition, saying that markets are dumb, so the state should choose investment. I had to respond to that argument; anything else would have been a non sequitur.
State investment would need two features to work: Information about correct investment composition, and incentives to enact that composition, ignoring political motivations. Actual states satisfy neither of those conditions, and certainly not both.
In his response, Mr. Cass sensibly abandons his hopeless original position, and switched to level. And that is a more interesting question. There is no reason to expect market systems to focus on the public interest. There are problems of externalities and positive spillovers that create new opportunities, including profits for subsequent private investment.
Examples range from quotidian (roads, court systems, military protection) to esoteric (the 1714 "Longitude Prize" or the 1960s space program). We might identify promising categories as "infrastructure" and "basic materials science," because those (a) dramatically increase the productivity of other investments, but (b) do not themselves promise directly capturable returns to their investors.
The problem is that industrial policy is not about levels of "infrastructure and materials science," correcting the shortcoming of markets. Industrial policy is about composition, the selection of winners and losers, targets of investment based on political considerations, animated by strictly political concerns for rewarding supporters and smiting opponents.
Consider so-called "green energy." Suppose you believe global warming is a problem, and the state must act. There is a bad, and a good, approach.
BAD: US industrial policy focused on "investing" in new energy, subsidizing the most promising companies. We "invested" in Solyndra, a company that never made any profits, because it never produced anything.
GOOD: Impose a carbon tax, since the goal is to make new energy alternatives viable. Invest in improving the efficiency of solar panels in laboratories, as well as other basic scientific force multipliers. Then let competition identify viable commercial applications.
Note the difference: the BAD approach assumes that officials elected to two-year terms have the information and incentives to foresee what forms of alternative energy will be best a decade from now. The GOOD approach assumes no one can see the future, and the process of discovery and trial and error must be tried on a massive scale, but with the best materials and techniques.
Going last imposes a responsibility to offer a synthesis. As I have written, the key to a successful industrial policy is "permissionless innovation." If the state can create an infrastructure where myriad private actors have secure incentives to search the space of innovations, we can make faster progress on the problems that vex us. But that solution requires both broad action, and focused humility, on the part of state actors.
Your support is vital to keeping Pairagraph free for all.
If you enjoyed this dialogue, we hope you will consider becoming a patron!
Your support is vital to keeping Pairagraph free for all. If you enjoyed this dialogue, we hope you will consider becoming a patron!
0 Comments