Michael Munger on Industrial Policy
What this covers
Michael Munger of Duke University joins Russ Roberts to examine why industrial policy—government steering of capital toward favored industries—fails in democracies despite potentially sound economic reasoning. Munger grants the analytical opening that justifies such policy: once market failures are acknowledged, neither profit nor loss proves an activity is socially worthwhile. Yet he argues that even if experts genuinely knew which industries to champion, the political obstacles are insurmountable. The conversation turns on a hard political claim: legislators, looking ahead to what an institution would actually do, will vote it down if they dislike the outcome—so the very experts and insulated committees that Dani Rodrik and others propose as solutions face the same political death as the policies they would implement.
Munger deploys William Riker's framework of democratic instability and the Riker Impossibility Argument, along with institutional-change theory from Cox, North, and Weingast, to show that power distributions in law must match power distributions on the ground or fail through non-passage, coup, or non-enforcement. The consequence is blunt: democracies default to cronyism—subsidized agriculture, Wall Street bailouts, protection for politically concentrated interests—because diffuse majorities cannot coordinate against organized minorities. Roberts pushes back, noting that political equilibrium models, while provocative, are not iron law; people vote against economic interest, movements produce surprises, and change remains possible. The discussion also covers whether experts truly understand industrial policy's political perils, whether the profit-and-loss system is itself a policy choice, and whether carbon taxes might accomplish what industrial targeting cannot.
Munger argues that even granting experts know the optimal industrial policy and that it would beat markets, it is politically impossible to implement in a democracy, because legislators looking down the agenda tree will never adopt the institutions that would put policy beyond their control; the default outcome is therefore cronyism.
- Riker's heritability-of-disequilibrium logic shows institutions cannot solve policy problems, since the institution will be voted down for the outcome it produces
- Industrial policy advocates like Rodrik already understand the political-incentive problem; their proposed fix (insulated expert committees) fails the same political test
- The Cox-North-Weingast proportionality theorem says institutional changes that diverge from the de facto distribution of power fail
Policy outcomes express the rational interests and concentrated power of affected groups, not error or inefficiency.
- Per Gary Becker's view, complaining that agricultural subsidies are inefficient misses the point: they are politically very efficient where farmers have concentrated power (Japan, the US) and absent where farmers are many but powerless, so policy outcomes simply reflect the underlying distribution of political power.
“it's politically very efficient--in certain countries and not in others. So, in Japan, agriculture is highly subsidized, and in the United States highly subsidized, and in other places where there are lots of farmers with no political power, not subsidized at all”
- Argentina's politicians persist with high inflation not out of ignorance but as a rational choice: lacking a functioning tax system and tax-compliance culture, printing money is the easiest way to raise revenue, so even though inflation is bad for the country they prefer it to the worse-for-them alternative of having no money for the government.
“they prefer the world of high inflation with all of its horrors... because even though it's bad for the country, they think there's something that's worse. And that worse thing is not having any money for the government because they have a really bad tax system.”
Social science models of political equilibrium are far less reliable than physical models, so beneficial change remains possible.
- Equilibrium analysis in political science is provocative and useful but is not truth: people don't always vote their economic interest, social movements and leadership produce unanticipated change, and our models of how political outcomes emerge are nowhere near as reliable as physical models like planetary motion, so change—sometimes for the better—is more possible than the impossibility framing implies.
“Political science is not a science. Equilibrium analysis is useful; it's provocative; it's not truth.”
EconTalk can meaningfully improve public discourse on controversial topics by reaching listeners willing to learn.
- Roberts concedes it may be naive, but he likes the idea that EconTalk could improve discourse and conversation about controversial topics by reaching even a small number of listeners who learn something.
“I think it actually could improve discourse and conversation about controversial things”