The problem isn’t one of institutions or individuals but of how individuals relate to institutions. Books that concentrate on rules for economic policy are about as useful as books that promise to teach you how to pick up girls.
Rules don’t make societies any more than rules make games. Games exist in the playing, and since there are no umpires in society who are not also players themselves, we have to trust our playing partners to make the honest call more often then not even when it’s in our favor. Ever play tennis?
Crises in society come about not because the rules break down but because rules are all there are left. The gearbox is fine, but there’s no grease. And what’s grease?
That’s the unasked question.

What percentage of the population in any country takes individualism as the model for behavior, up to and including the sort of sociopathological individualism economic science seems to prefer as it model? Both American political and economic liberals think mostly of social and religious conservatives and looney leftists as anti-individualist. And of course there’s the army. But the Scandinavian model is based on it. Social democracy is based on it. Religious conservatives counter the ideal of individual freedom with limits originating in god, social democracy with limits originating not in the state but in the community of which the state is a creature. That’s not a problem if we think of individuals as creatures of community. I speak and write in English, and I try to do so “well.” That means I do so expecting to be judged by others. As an individualist why would I care what others thought? Again the posters at CT celebrate individuation and bemoan atomization by turns. What can I say?

Europeans aren’t nearly as afraid of determinism as historically Americans have been. The question of free will is seen as an amusing conundrum not a problem with an answer. Cartesian philosophy never stopped being literature. But the formal structures of social democracy are beginning to appear now in US. While the academy is discussing libertarianism from above, academically mandated anarchism as the last hope for modernism, everyday post-modern [second modernist?] social-democracy is coming up from below.

So I’ll ask you: What percentage of the optimism now permeating academic thought can not be explained by reference to social determinism, as pathology? My sense of cautious optimism is based on something else entirely, the sense that people are getting used to there being unsolvable problems and are developing the capacity to accept the ad hoc. The academy is drying out, but the world’s getting greasy.

You can describe these approaches with two features 1) the policy rule and 2) the means by which it’s operationalised.

Thus, old Washington consensus would have as its policy rule 1) “eliminate all distortions” and operationalising method 2) “all areas where there is a government regulation are distortions”.

Chapter 2 Rodrik has 1) “eliminate distortions biggest to smallest, ranked by their impact on welfare”, and 2) “the impact of a distortion on welfare is measured by its Lagrange multiplier in the tau-mu equation on p58”. That’s “strictly in the tradition of neoclassical economics”.

Chapter 6 Rodrik has the same 1), but 2) “the importance of a distortion on welfare is something you have to take an educated guess at based on the particular history and institutions of the country. There’s no systematic way of finding this out, so it has to be a process of trial and error carried out by the local governments”.

Two points here. Chapter 6 Rodrik isn’t “strictly in the tradition of neoclassical economics”. You could, with a following wind and a couple of freaky instruments, get econometric estimates of tau and mu and carry out the optimisation, but this is specifically what Rodrik doesn’t advocate. Second, the Chapter 6 process involves estimating the importance of the distortion directly – it doesn’t have an intermediate stage of estimating tau, as far as I can see, so it’s not even implicitly recognising that the Lagrangian analysis has a role to play, except as a way of sugaring the pill for neoclassicals to persuade them to accept that they’re going to have to do institutional economics.

If things had turned out differently and Sraffians ran the world (with neoclassical economics an odd little heterodox outpost laughed at by all), then Chapter 2 would have said that you should prepare an input-output table for the economy, record the departures from the input output table which reflected the global technological optimum and concentrate on removing those departures from global best practice which had the largest effect on market price production of final consumption goods. If Chapter 6 said “it is impossible to put together this kind of counterfactual input-output table, so you have to look at the institutions and try and direct policy toward sectors which are most important”, then I would also say that Chapter 2 wasn’t really doing much work and that the claim in the title to be “strictly in the Sraffian tradition” was untrue.

Because it’s not irrelevant. In so far as this book is “strictly grounded in the neoclassical tradition” (which I don’t think it really is, but its author does), that means that it uses aggregate production functions and their derivatives with respect to various parameters of interest as its main tool of analysis. That’s what “strictly grounded in the neoclassical tradition” means. If it was always analysing things in terms of big changes from equilibrium, or discontinuous output responses to input changes, or in disaggregated terms, then that wouldn’t be in the neoclassical tradition.

Nothing in this whole freaking thing about any kind of freaking production functions

Yes there is, and the fact that it’s not explicit doesn’t mean there isn’t; look on p58. “… where [mu^{s and mu}p] represent net marginal valuations of activity i by society and by private agents, respectively. Of course, they depend not just on the set tau of distortions, but on levels of consumption, labor supply, asset holdings and so forth [...]”. In other words, the “activities” i have a value because of their contribution to overall production of final consumption goods. This is actually made explicit further up the page; “[these distortions] prevent the best use of the economy’s resources and, in particular, keep the economy far below its attainable productivity frontier”.

You keep trying to argue that words like “frontier” don’t really mean anything specific and should just be unpacked as meaning “the maximum amount of stuff we can make”. But that’s throwing out the entire point of neoclassical economics, which is that it’s meant to be rigorous; this is the big advantage and disadvantage.

You, very very very obviously.

“One extra what?”

This ain’t the Cambridge Capital Controversy. We’re not talking about factor intensities here. Or heterogeneous capital or the index problem. Why you bringing stuff up that is irrelevant (like continuity, derivatives, and now this)? If your problem with this whole thing is that one can arbitrarily define “Social Welfare” (which % changes in is what this is being measure in) then just come out and say it: “I don’t think one can measure Social Welfare”. If you think that subjective value theory is bunk and it should all be measured in the bushels of corn then … well … I think the framework still applies if one measures Social Welfare in terms of bushels of corn (and in that case your objections are neither here nor there and that’s basically what you should’ve said at the start)

What’s your point?

“Yes they are, because there are a specific set of assumptions about both preferences and production functions which are characteristic of neoclassical economics, particularly to do with aggregation, but not only”

No, there are actually not. Nothing in this whole freaking thing about any kind of freaking production functions. Maybe, just maybe something about preferences. In the sense that at some point or another someone somewhere has to decide what exactly someone somewhere wants (i.e. Social Welfare). Bushels of corn if you like.

Not in the relevant Lagrangian equation. But at some level of analysis, something similar to the tau parameter is describing the set of institutional constraints which are stopping this tribe from getting on a canoe and going to a place where the life expectancy is more than 30, so they clearly do matter, just not in a way that’s very well described by the equation.

I mean, you can answer the question “what would the impact be on the Japanese economy in 2006 from the decision to enter the Second World War on the German side in 1942?” in terms of

you have a society where all kinds of things are screwed up [...] and you change one of the things that’s screwed up then the effect on the overall well being/screwed up-ness etc. of this society is going to be the direct effect of making that one screwed up thing less screwed up weighted by how much that one screwed up thing is important in the overall well being of the society (which could be small) plus the effect of un-screwing that one thing on all the other screwed up things in this society, each one judged by how important their screw ups are in overall well being of this society (some of which’s screwed upness could go up and some which’s could go down). “

but once you’ve written that book, it doesn’t make a lot of sense to translate it back into Lagrangian problem, even though it would be quite easy to do so if you’re helping yourself to tau(feudalism), tau(vintage of capital stock) and tau(return on domestic savings).

no, you talked in general terms about institutions. You didn’t even assert the existence of a production function. And it was only confusing because you were trying to use language that could easily be mapped on to the constrained optimisation problem in your mind. So in other words, you were doing worse institutional economics than you’d normally be capable of because you were trying to ensure that it could be translated into a particular kind of equation that’s characteristic of neoclassical economics.

I might have made mistake though and it’s much easier (simpler, more understandable, more precise and easier to get across) if you write it as a solution to a Lagrangian problem

Easier to talk about it in some ways but not others, easier to avoid some kinds of mistakes but not others.

I mean, consider the Leontief table for some primitive economy. You can either approach the analysis of that table by constructing an aggregate production function and looking at the constraints on overall growth, or by leaving it as a table and start interviewing businessmen to try and guesstimate the inter-industry coefficients. One of these approaches is neoclassical and one of them isn’t.

Are you arguing with me or Dani Rodrik here? Dani agrees (see his response) that the essence of neoclassical economics is that “the importance of a distortion” can be given a specific mathematical representation, and that this representation is going to bear some kind of resemblance to the equations they teach you on economics courses. I agree with him, but think that in lots of interesting cases, “the importance of a distortion” can’t, or that this isn’t a productive way to analyse real world problems.

“It means “one extra of something vs. one extra of something else”.

One extra what? One kilogram of something? One joule of something? One automated lathe versus one bushel of rice? You and I (and Rob Vienneau) know how that one ends up. And this isn’t even the main problem with the NC approach.

If you’re gonna go all “there are phenomenon in this world which can’t be captured by maths” than go all “there are phenomenon in this world which can’t be captured by maths, and writing stuff down in terms of mathematical equations (which have common sense, straightforward translations into late forms of Latin bastardized Germanic dialects) cannot help us! Ye Odin! Poor Harolds!”. But then don’t bring derivatives and continuity and real numbers into it, because none of these are in any way important to the point.

Yes they are, because there are a specific set of assumptions about both preferences and production functions which are characteristic of neoclassical economics, particularly to do with aggregation, but not only. If Dani had written the entire book in terms of input/output tables, then it would have been very mathematical but it wouldn’t have been “strictly grounded in neoclassical economic analysis”, for example, would it?

I find cross-country regression analysis useful as a means of disproving theories. Very Popperian.