The blog yesterday provoked a lot of healthy debate about my claim that industrialization is mainly market-driven rather than state-driven, using Korea, China, and India as examples of industrialization out of poverty. I know I am going against the conventional wisdom of the great Asian “developmental state,” authoritarian and heavily involved in planning industrialization. So let me explain why. When I said we can only test what works on average, I am talking about what propositions are testable and falsifiable, using Karl Popper’s definition of what is “real” science. There is no way to test policies if you allow what works to be different in every year and every country, since a hypothesis about ANY policy will always fit the data perfectly under this assumption. I am not implying imposing the same blueprint everywhere, since “what works” is usually too general and can only guide general policy orientation. Of course, I agree that context matters a lot and so policy-makers should use whatever alternative sources of information or political instinct they have available to adjust the policy orientation to local circumstances.
So my general claim is that heavy reliance on markets is associated with long-run success, using as data the Asian successes, the earlier European and North America/Australia/New Zealand successes, the failure of non-market central planning in the Communist Bloc, and the failures of statist policies in Latin America, the Middle East, and Africa. It is true that Asian successes used state intervention more than the earlier European examples, but on average state intervention does poorly across all countries, so we have no Popper-standard evidence that state intervention contributed to their success. So my claim is based on evidence, not ideology.
We could also test industrial policy using within-country data. A well-known old study by Korean economist Jong-Wha Lee ("Government Interventions and Productivity Growth," Journal of Economic Growth, 1(3), September 1996, 391-414) found that government-favored sectors in South Korea actually had worse productivity growth than those that were not government-favored.
There is also the fact that South Korea (which had populist policies in the 1950s), India, and China had rapid growth after they shifted towards much LESS state intervention in the economy. I’m not sure that this one would pass the Popper test, however, since economists’ attempts to explain short- to medium-run shifts in growth have not been very successful world wide.
Now, let's go back to country data and look at the suggestion that we focus only on the success stories in East Asia. This has indeed been the predominant approach and has reinforced what I think is the fallacious conventional wisdom on the "industrial policy success" in East Asia. Looking only at the successes causes "survivor bias" about what really works.
Suppose we have a group of drivers leave New York at the same time to drive to Washington, and we interview the first 5 drivers who arrive in Washington. We find that they drove Lamborghinis at 150 mph, weaving in and out of traffic down the New Jersey Turnpike and I-95, out-running Highway Patrol cars who tried to stop them. Are they models for success getting from New York to Washington?
No, because since we only studied the “successful” first 5 drivers to arrive, we didn’t know about the vast majority of Lamborghini “failures” – the drivers who got into fatal accidents or were caught by the Highway Patrol and jailed for insanely reckless driving. On average, this approach was a disaster. On average, soccer moms driving mini-vans outperformed the Lamborghini drivers, if we study BOTH successes and failures.
So Asian success either happened in spite of statist industrial policy, not because of it, or industrial policy was an incredibly risky strategy that usually fails but occasionally has big successes, possibly in East Asia.
Either view would help explain why a huge amount of effort spent imitating East Asian success stories has NOT successfully replicated that success anywhere else.
So I stand by my claim that the 66-year-old idea of state-promoted industrialization has failed, and that it was irresponsible of Collier and UNIDO to resurrect it as a “major conceptual breakthrough.”