It just goes on and on, this talk of confidence, animal spirits, and what not. It now seems pretty plainly true that the thrust of prescriptive macroeconomics is propaganda. That’s why, in this post, I wasn’t even joking when I asked whether anyone had empirical evidence for the poor economic effects of state-sponsored propaganda, or whether we’ve just given up on it because we think it’s wrong.
Anyway, Robin Hanson, picking up on the macro-as-mind-control meme writes:
The consensus among macro-economists seems to be that people can in fact be fooled by such stimuli, but as Tyler indicates, it is not clear which policies most fool us. In particular, the more public attention we give to the stimuli, the less they might work. We might make people realize that they need to compensate via saving, and the more we scare folks into thinking we need huge stimuli, the more we might scare them away from normal economic activity levels.
So should we stop explaining macro-economics during this crisis, and stop saying how desperately we need stimuli?
The answer, of course, is that no one knows–especially not the macroeconomists!
Will Ambrosini is annoyed by all the macro bashing, and says that one can see that modern macro is successful, if one knows about the best new stuff. But when asked in comments what advice he’d give the president, Will says:
I’d have him give speeches about positive prospects for growth (”your company will turn the corner! your wages will go up!”); I’d have him talk about how this recession is a short term bump in the road. No more talk about how shitty things are (without, of course, denying the statistics… I assume Obama’s a better rhetorician than I am). Pessimism feeds on itself so to the extent the president controls expectations, he should try to improve them. Even though I don’t think the President has that much control over expectations, I think Obama’s rhetoric of transformation is more useful than his rhetoric of crisis.
I’d advise him that any actions he takes should reinforce his positive message.
Well, why not try this?
Or this:
Or this:
Somebody please tell me why anyone thinks Larry Summers is better at mass mind-control than Scarlett Johansson?
I plead guilty to the charges of meeting the market demand for my product.
In a perfect world, macroeconomists could give advice to policy makers, and it would be taken, that perfectly reflected the findings of the science. Because the science tells us inflation expectations are key, policies should reflect that. Monetary policy is the best way for policy to affect expectations, but the President wants to “do something” and the congress want to “do something”. Giving rose colored speeches seems least harmful while being most likely to have an a positive impact on expectations.
Because Scarlett Johansson has a “No Nudity” clause in her contract. Until she renegotiates, I remain unswayed by her doe eyes, pouty lips, ample (alas, concealed) bosom, and mediocre acting.
Now Larry Summers … well let's just say he needs to show some serious man-boob if he wants me to get behind Obamanomics.
Arnold Kling's “Macro out of the mouths of children of the Austrian School” makes the most sense, in my uneducated opinion:
http://econlog.econlib.org/archives/2009/02/my_…
summarized as:
#1) The way out of the recession will be powered by business profits (as usual). A targeted macro solution: tax cut on the payroll tax (profits go up, and only after the profits have been locked in, employment will go up)
#2) Nobody wants leverage. Let the modern finance system built on leverage die, quickly and painfully, because drawing it out will only be more painful (like removing an adhesive bandage). Nobody wants leverage, people will only pay attention to profits (see #1)
Really? I seem to remember Scarlett herself bickering over the sex scene in “The Island” because she thought it was unrealistic to have her top on. Has she changed her mind?
While this is completely inappropriate discussion for an economics blog, I will address it.
She has yet to show the goods, and has had quite a few opportunities to do so. It is the hot thing for actresses to have this in their contracts now (yet another reason I miss the 80's). Rumor has it that she gets nekked in the new chick-flick cum Guantanamo detainee screener, “He's Just Not That Into You” but it may be just that, a rumor, based on the brief glimpse we get of a skinny-dipping scene, which, if I know contemporary Hollywood, will be coy and disappointing.
On that note … back to the economics!
This makes me have second thoughts about Phil Gramm talking about how America is having a “mental recession”!
For example, if I'm the head of Bear Stearns, and it's Friday and I'm telling my shareholders (and anyone else who'll listen) that B.S. is as solid as a rock, that nobody has anything to worry about and we'll still be around Monday morning, people will assume it's true because I'm in charge and I'm supposed to know these things. But when Monday rolls around, it turns out B.S. is in dire straits and looking for a gov't handout, no amount of reassurances I can give is going to change the fact that B.S. is in deep shit, especially since I've already stated that the situation B.S. is in wasn't going to happen.
Unless I have a Big Brother-style stranglehold on access to information (about Bear Stearns' financial matters), most–if not all–people are going to notice that rather inconvenient gap between what I said would happen and what actually happened. And then they start questioning everyone else's judgement, and start assuming everyone else might be as full of shit as I am. Goodbye rational markets, hello financial meltdown.
I don't find Ambrosini's argument particularly convincing, and I find his earlier phlogiston comparison thing somewhere between sad and funny. He claims that replicating real (large N) data is good reason for assuming the causal mechanisms posited by the model are, in fact, doing the work, and that the model is thus explaining the data. This is simply untrue as a general matter; SEP's Scientific Explanation article is nice here. Based on the abstract of the Eggertson model (no, I'm not going to read it, life is too short), it doesn't appear that it attempts to do the work established to show causation; the paper is simply establishing a *potential* mechanism–in other words, a just-so story.
Look: nobody doubts that economists are great at creating just-so stories, mathematically elegant bundles of assumptions and valid conclusions. But economists, particularly macro economists, are *not* very good at demonstrating genuine causation, whether historically or in terms of proposed interventions. This doesn't mean that physicists or sociologists would do any better, of course; the problem is incredibly difficult. But it's far from clear that either, A, the training and tools of modern economics is the right approach to finding the knowledge we're looking for, or B, the structure of academic economics is organized in a way to (collectively) discriminate reliably between good and bad explanations or research programs.
It's worth noting that the profession's barriers to entry have both selection effects (only appealing to those both skilled at math and tempted by the prospect of reducing behavior to tractable systems of equations) and treatment effects (it would be psychologically devastating to give up on the formalization / general equilibrium framework after suffering so much to acquire proficiency in it) that have worrisome implications for precisely the desideratum Ambrosini highlights: an awareness of what is actually known versus not known.
Generating the models involves making assumptions about how people make economic decisions. These assumptions generate model economies that generate model data (i.e. predicted behavior of actual economies). The assumed cause (people attempt to maximize such and such utility under such and such budget constraints) is being tested by comparing the model data to the actual data. This has nothing to do with how much data (N big or not) is generated by the model.
I don't see how the method I'm describing is any different than making assumptions about the behavior of particles of matter, deducing expected behavior and then comparing those measurements against actual behavior.
These aren't just-so stories. Its true macroeconomic data is observational, but unlike in history or evolutionary psychology or whatever, the economy keeps running and new data are generated all the time. Models die when new data contradict their implications.
I think a disconnect here is that macroeconomists study business cycles (or growth) in general but not particular recessions. Studying particular recessions/depressions is what historians do and their stories are just-so. To macro people, historical events are just data. Our models explain or they don't. To the extent they do, they're good models. To the extent they don't, they're bad models.
It occurs to me that my argument that macro folks don't study particular events in history is kinda lame given the paper I linked to was about analyzing the new deal. Fine. We study historical events.
The point is that preexisting models, tested on modern data, are being tested using data from that era. This back and forth between Eggertsson and Cole/Ohanian has taught us that we don't know much about “animal spirits” when inflation expectations are unanchored by monetary policy. These papers have pointed in the right direction though for figuring this stuff out.
Testing whether the model generates a similar distribution of summary statistics is not necessarily the best way to figure out whether the assumptions are in fact doing the work posited. This is especially true given that no one actually believes people think like this. Look, you're not going to convince me; I've put in my time at Chicago; my reservations are the rock-solid ones of the apostate. I only ask that you give some serious thought to the philosophy of soc-sci issues, and recognize the massive disanalogies b/w econ and physics.
Pingback: Wanted: More Scarlett Johanssons — But As For Me
Testing whether the model generates a similar distribution of summary statistics is not necessarily the best way to figure out whether the assumptions are in fact doing the work posited. This is especially true given that no one actually believes people think like this. Look, you're not going to convince me; I've put in my time at Chicago; my reservations are the rock-solid ones of the apostate. I only ask that you give some serious thought to the philosophy of soc-sci issues, and recognize the massive disanalogies b/w econ and physics.