Way to go professor Abby! Difficult issue. The creation of life is unlike anything else. What should we think about the possibility that babies will be produced for profit - as in baby-making businesses - like surrogacy on steroids? Still, good blog, thought provoking.

I can't believe this comparable worth stuff is still around. The courts in the 80's decided this was nonsense. http://www.utdallas.edu/~plewin/Equality.html

http://plewin.blogspot.com/2013/05/why-dont-we-learn-from-history.html

An anecdote. Stan Fischer is from Rhodesia (Zimbabwe), which is to South Africa (my country of origin) as Canada to the US. He rose to greatness as a macroeconomic modeler in the Keynesian mold, skills which he honed at MIT. As far as I can tell his talents focus on his technical prowess. I met him when I was in the Ph.D. program at Chicago and he was a visiting professor. I took a course of his in monetary theory. The whole course was an in-depth examination of a macro-model of Dale Mortenson's. It boggles my mind to see how revered he has become (albeit that he is an honorable man). A technician with very limited understanding of and insight into history and social dynamics.

I think what Jerry said is exactly right. The persistence in high unemployment has taken many by surprise (though many also predicted it - many Austrians included). But, after the fact, the surprised are going to resort to "structural" explanations. The rise in productivity is, at least in part, a statistical artifact. Fewer workers has resulted in higher measured output per worker. Also, it is the result and not the cause. Productivity always rises in recessions, statistically, and because businesses economize on labor. Troy, when I say that the NR could not have changed that quickly I am talking empirically-historically, not as a general proposition.

Thanks Joshua. I look forward to reading it.

@Troy. There is probably a lot more that could be said about this. I think we know, more or less, why stable institutions emerge (looking back) - because of network effects - the signals that each individual gets push his behavior in the direction of more coordinated action (the more people accept a particular money, the more people accept it; the more people use a particular language, communication protocol, legal system, ..., the more people use it, and expect others to). Clearly something like this operates in the market process as well - Adman Smith's invisible hand - but, as far as I know, we don't have a good account of the details - except to assume that the price signals are accurate enough (Hayek 1945). But, in disequilibrium (inconsistent expectations), this may not be the case, in theory. And, more importantly, in the market process we NEED to be able to accommodate mistakes, and lots of them, if we are to have progress (trial and error). How come it works so well in practice? That is the true miracle :-).

@Jefferey: I think this is directed at the hosts of this blog - since it is beyond my control. I think you mean they should provide a print-function link as many web sites do. That would be nice. Otherwise cut and paste works :-). Glad you liked the post.

Yes Alex is correct. Unless you classify all of the unique things that you see into categories (classes), you cannot even begin to make sense of the world. All thought and language does this. The fact that these categories are shared with others, allows us to communicate. (Hayek's Sensory Order and many other works.)

Roger: I need to read this book - I like Paul Rubin's stuff. Thanks for elevating my random musings into something profound :-).

Lord Keynes: I was NOT referring to Rothbard or natural rights, simply to the undeniable fact that all government action involves coercion - it implies coercion (try refusing to pay your taxes). In order to justify it you have to justify the coercion. Not rocket science. Where to expand AD? On middle classes and lower income earners and businesses? - just how it's been done successfully many times in the past 65 years in many nations, for which there is overwhelming empirical evidence on how it works. You clearly don't understand the problem. The question was rhetorical meant to draw attention to the fact that not all AD spending is going to have the same effect - something that Keynesians clearly deny or ignore. There is a structure of production and consumption, output is not an amorphous, homogeneous glob. You think in terms of income classes. I was talking much more fundamentally. In any case, even the Keynesians seem to admit that stimulus spending on cronies has a different effect from spending on food stamps. I suppose what matters is that I believe neither increases employment (quite the opposite) and you do. But you still have to prove it, you have the burden - that was my point. [I don't see the overwhelming evidence you claim].

http://www.li.com/newsdisplay.aspx?id=6778&site=4&type=event&keywords&dm_i=V6Y%2CV5QD%2C6U2KG9

Like.

Lord Keynes: (I love your name). What BZ said. The moral case - all government action involves coercion and distortions (taxes have to be levied). This would seem to place a big burden of proof on the proponents to show that its benefits will outweigh the costs. I don't think the evidence supports this - not at all. But then history doesn't speak to us in one voice. No doubt you hear a different academic scribbler. I don't think I implied that Keynesian macro-policy needs to know "exactly" the affects of AD in order to be successful - but it does need to know where to apply the AD, doesn't it? Unless you want to claim that all AD stimuluses are equal. On what basis? And, yes, there is the formidable Public Choice problem - Buchanan-Wagner, Democracy in Deficit.

Frank, I look forward to reading the posts at your links. Deweaver, when I teach Milton Friedman on monetary policy I tell my students to imagine trying to steer a boat with a faulty rudder across a lake. The lag in response is long and variable (cannot be predicted) - the result is to amplify the fluctuations (weaving) and most likely miss the target on the other bank. So far so good. But the analogy of a dynamic control problem only takes us so far. If we knew the number of markets, could specify all of the products involved, and could estimate the lags, even probabilistically, we could theoretically devise an optimizing algorithm - one that would probabilisitcally minimize the mean squared error or similar. In fact, we are not really dealing with this kind of problem. We are dealing with a much more complex system in which the number of variables is both large, unknown and continually changing. It is much more akin to a biological system than a mechanical one. It is complex and adaptive for some configurations of the enclosing institutions (Hayek). An evolutionary mindset is more accurate than a mechanical one. Either way, we may agree that discretionary macro-policy of the kind envisaged by Keynesians is impossible.

The full text of my keynote address: What do we know for certain about uncertainty, which relates to what Mario is talking about, will be available soon and I will post the link.

Troy, I am optimistic about ABM precisely because one can create Austrian or Keynesian outcomes. What one should do, however, is then compare the kinds of agents one would have to have to get those outcomes to how real human beings act. I would call ABM a first step in one's investigations. You can give your agents a variety of characteristics and see what happens rather than use, say, perfect rationality and perfect knowledge because that's the only thing that lets you do the calculations.. Certainly reasonable, and I absolutely am not advocating a rejection of ABM on grounds of a multiplicity of possible outcomes. The proof of the pudding will be in the eating. But, here is my expectation. I expect that with not-so-different, plausible assumptions about agent-behavior and institutional-framework one will be able to get very different results - notably, Keynesian v. Austrian. I suspect it will depend much on the magnitude of the behavioral parameters and much less on the type. Even this would be quite informative, suggesting that there is something else going on, and would beg more, varied investigation. How far one can go with this remains to be seen. I counsel guarded optimism.

Barkley, Ah ha! So, you are using this blog to help you make comments on a paper for a conference, eh? Fair enough, I guess, :-) Not really. The paper is much more narrow than this discussion, and I had already written my comments. But what you said will help me add a few things. :-).

@Troy. I envy your optimism. But I don't see it at all. The thing is, it is possible to produce an ABM for whatever macro-result you want, and whatever policy you support. What independent method or criterion do we have for deciding between the Keynesian and Austrian ABMs and the policies they imply?

Barkley, Glad to see I am not the only one awake at this "strange" hour. I think what you say makes perfect sense and will keep it in mind in my approach to ABMs. I have a paper that I am commenting on at a conference next week that I will subject to those criteria. Coming back to what I said at the top, this is really no different from the way we evaluate traditional statistical models in practice - and ABM seems to have unexplored promising potential.

Actually, just looking at what I wrote, in response to what you wrote, I think I may have missed something. You are suggesting that the behavioral assumptions are already plausible or have some prior presumptive value - as, for example, being used in standard neoclassical models. And the ABM then checks the "sufficiency" of these assumptions. This makes me feel a little better, but ... .

@Per: (further to McKinney). Yes, but the ABM method is not about finding the right causes ("right" as in real-world existing), but about analyzing the sufficiency of already known or hypothesized micro foundations or causes of macro phenomena. Epstein & Axtell discuss this in Growing Artificial Societies (1996) "The surprise consists precisely in the emergence of familiar macrostructuresfrom the bottom up--from simple local rules that outwardly appear quite remote from the social or collective phenomena they generate. In short, it is not the emergent macroscopic object per se that is surprising, but the generative sufficiency of the simple local rules." (pp. 51-52; emphasis in original) Interesting. I will look at this article – I saw it referenced in the RAE articles on ABM. My comments were not necessarily to criticize ABM, as to raise a concern. Your comments don’t really assuage that concern. Why the criterion of sufficiency? Why not plausibility? Is this some sort of Occam’s Razor idea. If we can “explain” (generate) the outcome we seek with minimal agent-behavior assumptions, is the is a good thing? Why? Naively, don’t we want to know what is really going on? I can tinker with the values of agents' characteristics as well as their behavioral rules to figure out if the TC need to be high or low, what other conditions are necessary, etc. And then? How do you judge if the model is plausible – does it “explain” or simply show one kind of (many?) processes that could yield similar outcomes? Just asking, genuinely curious. ------------------------------- On the question of facts and values. I intend to address this when I discuss the burden of proof question. In the meantime see: http://www.utdallas.edu/~plewin/factsValuesandtheBurdenofProof.pdf