Dear Dani,
Thanks for your reply to my post.I am a bit frustrated with your statement that industrial policy just has different effects in different countries. If we just say “it works” with good outcomes and “it doesn’t work” with bad outcomes, then there is no way of contradicting this with evidence. ANY policy could pass this test. This kind of “theory” fits past data but cannot predict future outcomes – how do we know what side of the bed industrial policy will wake up on tomorrow?
I know you don’t actually fall for this, but I don’t understand how you actually get around this problem with your “growth strategies” approach. If the effects of every policy are different in every country, what is your evidence base for recommending any particular policy in any country ever?
This jibes with the observation that tons of effort to replicate East Asian Tiger success elsewhere has not actually worked to produce Tiger-like success anywhere else.
Your work has done a lot to convince us all about our inability as “growth experts” to make general recommendations on how to raise growth. But then you seem to recommend more intensive use of “growth experts.” I would go in the other direction and find approaches to development that don’t require these clueless “growth experts,” like systems with a lot of local feedback and accountability – lots of political and economic competition with freedom of choice of consumers, investors, and voters.
Best regards,
Bill
7 Comments
Will you please just read his books to see what he considers the characteristics of good industrial policy (It doesn’t sound like you have. I recommend starting with “One Economics, Many Recipes”) — just because there is uniqueness, doesn’t mean there isn’t also important similarity. Likewise, his books give important characteristics and preparation for constructing good industrial policy.
With all due respect, I think Dani has the superior point in this debate: Let us discuss under which conditions IP works, and under which conditions it does not.
That is indeed a little further than simply saying that the effect of IP differs from country to country, as you critizes Dani for.
“If the effects of every policy are different in every country, what is your evidence base for recommending any particular policy in any country ever”
As far as I can unserstand, Dani do not disagree with this statement. His point seems to be that we should not ban IP althogether because it goes bad from time to time. Rather, we should encourage politicians to figure out the optimal IP for their countries through trial-and-error.
Then, of course, the question is: When will politicians actually try to get IP right, and when will they use it as a proxy for some other objective?
To me, it seems like Dani’s answer to this question is “almost always”, and yours is “never”.
I second Richard – in his books Dani does talk about “systems with a lot of local feedback and accountability” but in the context of ‘experimental’ industrial policy. Admittedly these passages can be a bit vague, but when you are talking about idiosyncratic, context specific policies, a bit of vagueness goes with the territory.
Here’s a related question. Say we went through the data, came up with an acceptable way of identifying countries that had attempted interventionist industrial policies, and then estimated some sort of expected growth rate and associated probability distribution of growth rates, conditional on attempting IP. How should policy makers use that information? Some of the countries that we be classified has having tried IP but who performed badly would have been corrupt and incompetent (that seems uncontroversial), so why should a policy maker that does is not corrupt and incompetent be guided by the average performance of a group that contains venal incompetents?
One obvious response is that no policy maker can know their own competence – perhaps there are political economy type forces at work that mean even the best intentioned policy maker will flounder, or perhaps the information problems involved are so profound that there is no such thing as a competent policy maker, only lucky policy makers.
So, how do we distinguish between these alternative models? How do we know whether we are in a world where all policy makers should be guided by the experience of all previous attempts at IP, or whether we are in a world where there is such a thing as unobserved ‘competency’ which our growth regressions cannot control for, and hence ‘competent’ policy makers should no more be guided by the past experience of all countries that attempted IP than Tiger Woods should form his expectations on the basis of past performance of the average golfer?
This data mining stuff has been a bit over my head lately in your posts, and I’m a numbers guy. I can’t begin to think what less technical people are reading into this stuff.
Will we ever get off of data and refuse to accept that the $1/day standard isn’t universally applicable (only by Western Standards) and go to the ground level and actually ask people what they want and if they want help to begin with and if what has been done has provided a higher quality (not standard) of life.
Perhaps at the level you all are working at, there are only numbers to crunch and people are stats instead of real living flesh and bone with feelings and thoughts, passions and souls.
Am I trying to live in a Utopian world here?
In another 20-30 years, because of the public and the people who make up those big organizations (UN and such) wont care about the numbers. They will want stories. Mark my words.
Whereas Easterly only argues for local ’searchers’ to solve development conundrums in their countries, Rodrik actually goes few steps further and provides them with well specified analytical tools. The only difference I can see is that Rodrik thinks that the searchers can actually be located within local government too.
As to the list of industrial policy sucesses: could you please add the post-communist Europe? OK, it is less impressive to make it from the second world to the first one than it is from the third world. Nevertheless, the development success of the now-EU members is impressive and the industrial policy – run by the investment promotion agencies – played a role. These agencies were as much a government tool to attract investments as tool of foreign investors to deal with (unstable and rather corrupt) governments. It helped attracting the right kind of FDI and reduced transaction costs for foreign investors. It is a pity that the developmetn debate ignores these cases, just because it took place in the ‘old’ Europe and not in Asia or Africa.
Ref.: Industrial Policy/Actually Doing Something
I would still like an answer to my question of whether industrial human waste to fertilizer conversion plants is a good investment for the World Bank to Finance and if it si (as seems obvious) than why is it not financed?
The more the profession wallows in its idle abstract debates the more irrelevant it will become and it is grossly irrelevant already in the third world, china, wall street.
SS
Surely the value of Rodrik’s work depends on its purpose?
Academically speaking, working out how and why industrial policy has worked in the past is an interesting historical question. On the other hand, if we are to be more policy-oriented, isn’t the better question to ask, what can be done to alleviate poverty? In this sense Easterly is right: we do need to step back and consider all options, not just industrial policy, even though this is a harder question to answer.