Wednesday, February 08, 2012

Why rational expectations models can be wrong


Washington University professor David Levine has a pair of articles (article 1, article 2) in the Huffington Post about why financial crises can't be predicted, and why rational expectations theories are the only good theories in economics. Although I disagree strongly on both counts, Levine's arguments are very thought-provoking.

(As an aside, just to get this out of the way, Levine uses the Heisenberg Uncertainty Principle as an analogy to explain why observer effects make non-rational economic theories fail. This is not a great analogy, since the Heisenberg principle can only be related to observer effects in certain specific situations; also, the Heisenberg principle deals with conjugate pairs of observables, which are different from the kind of observer effects Levine is talking about. This isn't super-relevant, just remember not to get your physics intuition from interdisciplinary analogies. But anyway, back to Levine's ideas...)

Levine's first argument is that financial crises, by their very nature, can't be predicted:
Take an example: how we might predict stock market crashes? Suppose that two behavioral psychologists, call them "Kahneman and Tversky," produce a model of "cognitive biases" that predicts when crashes will occur. The model tells us that the stock market will crash on October 28. Since the model is reliable and has a perfect track record, we naturally believe this prediction. So what would you do? You would sell all your stock on October 27. But of course if enough people do this the stock market will crash on October 27 and not October 28. So this apparently reliable model will be proven wrong.
This is similar to an argument made by Levine's colleague Stephen Williamson. I've addressed this type of argument in the past, but here is a brief recap:

Conditional predictions are different than unconditional predictions. If your modeling goal is to say with confidence that "A financial crisis will occur at 9:01 A.M. on February 17, 2012," then Levine is right; you are probably not going to succeed. However, if your modeling goal is to say: "Unless X policy is taken first, a financial crisis will occur at 9:01 A.M. on February 17, 2012," then you have a chance of succeeding. Why? Because unless investors can predict whether policy X will be taken, then knowing that the model is correct doesn't allow them to make riskless profits. And of course, conditional predictions are the kind policymakers usually care about. So Levine is wrong in cases where policy is decisive.

As a side note, even in cases where Levine is right, this does not make modeling crises a useless exercise. It may be that increasing our understanding of the causes of crises leads to a decrease in the amount of crises rather than an increase in our ability to predict the timing of the crises that do occur. But that's fine! It means that the benefit of better crisis modeling accrued to society instead of to the modelers. That just means that research into the causes of crises generates a positive externality, and should therefore be subsidized by the government and/or some other collective public-goods provision mechanism.

And it means that critics of the econ profession who say "You economists didn't pay enough attention to models of financial crises" can still be right, even if Levine is right!

Levine's next argument is that rational expectations models of the economy are the only models that can be right over the long term:
In simple language what rational expectations means is "if people believe this forecast it will be true." By contrast if a theory is not one of rational expectations it means "if people believe this forecast it will not be true." Obviously such a theory has limited usefulness. Or put differently: if there is a correct theory, eventually most people will believe it, so it must necessarily be rational expectations. Any other theory has the property that people must forever disbelieve the theory regardless of overwhelming evidence -- for as soon as the theory is believed it is wrong.
But notice the logic behind this idea. It assumes that economic theories penetrate the public consciousness and are used by economic actors. Levine postulates that over time, as non-rational-expectations theories are shown to be wrong, economic decision-makers (i.e. you, me, and other folks) will gravitate toward rational expectations theories. This will mean that in the long run, only rational expectations theories will describe the economy. But it also must mean that in the long run, only rational expectations theories will be believed by the public. So if Levine is right, we should see more and more regular people gravitating toward the belief that other people are rational.

Do we, in fact, see this? I am not sure we do. On one hand, index funds are becoming more popular, signifying greater public acceptance of the Efficient Markets Hypothesis. On the other hand, the recent recession probably decreased faith in rational-expectations models of the macroeconomy, both among economists and among the public at large.

So if we don't see people accepting rational expectations theories, then the theory/behavior convergence that Levine talks about is not happening. Meaning that rational expectations theories may well be wrong.

Note that one easy way for rational expectations theories to be wrong is for the average person to simply be dumber than the people making the theories. For example, it's easy to reliably predict the mistakes of rats running in mazes, simply because rats aren't smart enough to learn our theories. Now, obviously, humans are not that far apart in intelligence, but if the difference between different humans is big, it could mean that behavioral theories could consistently be right.

(Does this make behavioral economics elitist? Maybe. But not necessarily. It could be that theorists are smart when making their theories but are dumb at other times. I can personally attest to the fact that this really does happen!)

In his second article, Levine asks how long it takes rational expectations to prevail, and cites the actions of the passengers on Flight 93 on 9/11 as an example of a quick rational response to new information (in this case, the info that hijackers would use planes as weapons). But does this mean that crowds are always rational? There is ample evidence for inefficient herd behavior in the psychological literature.

Anyway, Levine's logic in arguing for the supremacy of rational expectations is subtle and powerful. But I think reality is even more subtle, and even more powerful.

52 comments:

  1. Rational expectation, perfect foresight, or any of the other math tricks such as convexity to make the models solve or get reduced forms are a joke. Economics is now a religion, requiring faith. It has lifted off from reality. The public does not place much trust in the profession. Economics has lead the political parties astray with their mythologies. Noah you are on the truth path. Do not stray.

    ReplyDelete
  2. I had two thoughts.

    First, I wonder how long "the long run" is for Levine. I'm reading "Thinking Fast and Slow" now and it seems that even trained statisticians are prone to systematic errors about probability. This suggests simply knowing about errors people make in thinking is not enough to avoid them. If that's the case, it could take a very long time for behavioral economics to fail the test of time.

    Second, there's a lot of discussion that the market will discipline our irrationality. I'm not so sure. Presumably, the idea is that rational firms drive irrational ones out of business. But couldn't this process get gummed up if most firms make the same systematic errors, and success or failure is largely determined by luck?

    ReplyDelete
  3. foosion11:37 PM

    index funds are becoming more popular, signifying greater public acceptance of the Efficient Markets Hypothesis

    Even those who don't accept the EMH should buy index funds. Only those who have sufficient (justified) belief in their ability to pick winning securities or winning managers, after costs, should avoid index funds.

    Inefficient markets would be a greater motivation to index, as there's more of a chance of buying a mispriced security.

    ReplyDelete
  4. Rational models assume some high level of information being consumed by actors. But not everyone has the same level of information. And even if they did, some can use it better than others. I only have experience with rational choice models in political science. But I'd guess that the same problems exist in economics, maybe even more so in some instances.

    ReplyDelete
  5. Two points:

    1) Last I looked, all economics-like theories including rational expectations made massive simplifying assumptions such as unbounded computational power, representative agents, complete information, etc. Believing that a model like that tracks the world is quite a leap of faith.

    2) As long as we're working with leaps of faith, there are lots of "rational" equilibria -- for example if nearly everyone believes that businesses that don't pay homage to the flying spaghetti monster will fail, then they will, because they won't be able to get loans, customers, etc.

    I suppose if we could answer the question "How stable are these equilibria with much weaker assumptions and under significant perturbation?" then I might find this line of argument worth thinking about. As far as I know what answers there are along these lines aren't encouraging but I would be happy to find I'm wrong.

    ReplyDelete
  6. 1. If an economy evolves along a deterministic path the stock market can not exist in its current sattus and thus collapses are not possible. This is a real alternative to rational expectations.
    2. Why chaning opinions and expectations should change this deterministic path? For example, the Gini coefficient (BLS data) has not been changing in the US since 1947 with all types of wars against poverty. The mobility and influence of individuals in cooperative behavior is significantly exaggerated.The economic structure is fixed and frozen.

    3. rational expectations are a couterproductive way of scientific research. It says that there are no deterministic laws because we have failed to find them (this is a direct statement from Nobelist Prescott). When applied to physics it would stop it at the level of Ancient Greece.

    ReplyDelete
  7. "Any other theory has the property that people must forever disbelieve the theory regardless of overwhelming evidence -- for as soon as the theory is believed it is wrong."

    Yuck. Theories aren't true or false. Binary thinking is a trap.

    I actually discovered a wonderful paradox, that I think is relevant here. Noting that natural gas prices are diverging from oil prices, it struck me that as natural gas and oil are for many purposes close substitutes, this can only happen if people do not believe that these prices can diverge over the intermediate term (i.e. enough time to make significant technological adjustments). So prices of close substitutes can only diverge if people do not believe they can diverge. WOW! (This is a bit like the old argument about the $20 note that can't be there.)

    ReplyDelete
  8. Rational expectations models must necessarily be wrong because -

    1) perfect information is unavailable.

    2) information access is not symmetric.

    3) people are semi-rational at the best of times, and wildly irrational at the worst. This is the real and unavoidable flaw in RET.

    4) economic crises are arguably the worst; see 3).

    5) herding instinct is powerful.

    6) 49.9% of people are below average intelligence.

    7) the implicit assumption of RET is that people are also always forward thinking. That is absurd on its face.

    Plus, what foosion said about index funds is dead on. Besides, the entire stock market is highly correlated (herding) so your chances of picking a winner are not markedly better than picking a loser. They are both out-liers. Cost-benefit does not favor stock picking for the vast majority of people.

    Cheers!
    JzB

    ReplyDelete
  9. Conditional predictions are different than unconditional predictions.

    you are splitting a non-hair and confusing rational expectations and perfect foresight.

    in you example, where there are outcomes A and B, the market will price the expected value (prob A * impact + prob B * impact). One of those will be revealed to be the true state (in which case the market may further rally or sell off on the certainty), but that does not disprove rational expecations ex-post.

    The "expected" (as in, mathematical expecation) path under rational expectations does not need to be an actual state of the world ex-post . in fact, the probability of observing the "expected" path is zero - the probability of observing the "perfect foresight" path even under rational expecations is ZERO. that does not disprove rational expecations.

    In fact, it is extremely hard to disprove rational expecations because it requires counterfactual analysis (because the real world is probabilistic).

    Would you say Italian bond yields are "irrational" or are pricing a small probability of a very large catastrophe? If that catashtrophe does (does not) happen are the prices today irrational? The only way to know is to measure the probability and impact through 1000 counterfactual simulations, which you cannot do in the real world.


    Second, saying "agents act irrationally" is a very strong statement because it means not just one person is acting irrationally, but millions in aggregate are acting irrationally that adds up to poor average behavior. If 33% are "irrational" and 67% are not, the outcome is the same as if everyone is rational, except there might be some welfare redistributed to the rational. And how exactly do we know its "irrational" and not some perversly poor market design whereby everyone is acting on some incentives (or disincentives)? the answer, is, we do not.

    Again, outside a laboratory setup, its extremely hard to disprove. There are no "arguments" to be made here to disprove it. One has to come up with a single, concrete example.

    The fact that reasonable people can disagree is proof that there is a lot of uncertainty that requires counterfactual analysis, not proof that people are irrational.

    ReplyDelete
  10. Jazzbumpa,
    you missed my main point I'm afraid. Models aren't true or false, they are just models which may be more or less usefull according to circumstances. But they are never complete. And that is the key. You never know, which of the things you omitted might be important and give non-random errors. That is why people will not just accept the status quo - they will keep looking for something better. And that destroys the hypothesis.

    ReplyDelete
  11. Levine's argument seems, as in many discussions of the EMH, to boil down to a bunch of question-begging. "We can't have non-rational expectation models because rational expectation models predict we can't."

    ReplyDelete
  12. @dwb:

    My example regarding conditional vs. unconditional predictions was meant to be deterministic, not stochastic. If we add stochasticity it is another wrinkle, but I prefer to always think about a deterministic model first and then add stochasticity later. Yes, in deterministic models, rational expectations is the same as perfect foresight, and no one thinks that perfect foresight holds in the real, stochastic world. But Levine's argument regarding predictions of financial crises applies equally well in a deterministic world as it does in a stochastic world.

    My distinction between conditional and unconditional predictions was not splitting hairs. Plenty of people recognize this...just look at discussions of whether macro models should be used for forecasting or for policy evaluation!

    I complete agree with you that the rational expectations principle is difficult to invalidate without lab experiments. That's why I'm doing some lab experiments! :)

    ReplyDelete
  13. Econo-speak is like Pig Latin.

    The object of which, according to Wikipedia, "is to conceal the meaning of the words from others not familiar with the rules."

    ReplyDelete
  14. @Luke Lea:

    You are absolutely right. Please wait two weeks and I will write a post about exactly what you just said. :)

    ReplyDelete
  15. Fair enough, but i'll make a stronger statement: all deterministic models are bunk. some rational expectations models are deterimistic, therefore some rational expectations models are bunk. think stochastic.

    ReplyDelete
  16. @dwb:

    Well of course you're right, but the point here is to do a thought experiment to show why Levine's logic doesn't hold, not to identify a realistic model.

    The point is that thinking about the deterministic case lets us understand things that hold in the stochastic case as well. I could have made the argument entirely in terms of conditional expectations, but it just wouldn't have been as clear.

    ReplyDelete
  17. we are in complete agreement that Levine's argument in defense of rational expecations is long-winded and in some cases wrongheaded.


    The point i am trying to make (not very articulately) is that one can neither affirm nor rebut rational expections in a deterministic framework. Rational expectations is just not a meaningful statement in a deterministic (representative agent) world. Its not meaningful along one path. rational expectations is only meaningful in a stochastic world. As soon as you start talking about differential agent's beliefs, you are introducing a distribution, and the notion or potential for multiple equilibria. Somehow those "beliefs" have to be grounded in something objective at the aggregate level, like the fact that gold demand cannot outstrip world GDP, or you end up in a silly place - like an infinite gold price.

    you can't defend (or rebut) rational expecations by resorting to "deterministic logic." Invariably it leads to the same impossible mental gymnastics that get Levine into trouble (herd behavior is rational because everyone believes the same model? what if my model says gold will go up indefinitely? whats the critical mass of agents who belive it???) -i.e. I have to postulate that every agent adopts the same model, or some sililar silly thing, in a representative agent world. Even in your policy example, people form expectations about "If policy X then Y," but surely some people have inside information about whether the policy will be adopted (happened many times before!).

    To which I say: none of the above. rational expectations is not a statement about individual behavior, nor a statement about the one "true" model, nor a meaningful statement along one path. It does not say some things can or cannot "be predicted (see below"*

    It's a statement about whether people, in aggregate respond to incentives. If i beat my head against the wall repeatedly, i am irrational. Now if you see 60% of the population beating their heads against the wall...hmmm sounds like economics??

    Levine is wrong that rational expecations precludes predictions of financial crises - moral hazard is perfectly rational and if allowed exists will lead to more frequent financial crises. But market failure is another thing that should not be confused with rational expectations.

    In fact, the existence of financial crises in some sense validates rational expectations because people are rationally acting on the incentives we have allowed to persist.

    ReplyDelete
  18. I find the question of whether rational expectations are wrong quite unproductive. I would much rather ask under which conditions rational expectations are likely to hold. But to answer this, one must first debunk common misconceptions shared by critics of modern economics, and I am a bit disapointed that you do not do so.

    1. The theory of rational expectations does not postulate that people do not make errors. It postulates that people do not make systematic errors.

    2. Systematic errors means that the outcome you expect falls consistetly short (or long) of the realized outcome. To use a sport analogy, it would be like consistenyly shooting the ball past the basket. Hardly ever getting it in because about half the times you shoot it too hard and the other half not hard enough is not inconsistent with rational expectations. It is easy to see that systematic errors can be true in the long-run only for someone who is incapable of learning and adjusting accordingly.

    3. Rational expectations is an average relationship. It is applicable even if most people make systematic errors, so long as a few are able to recognize this fact and exploit it. So suppose that most stock traders systematically overreact to bad news but a few recognize this and decide to go long when others go short and vise versa. Then stock prices will behave AS IF they are traded by a group of individuals that exhibit rational expectations even if most individuals do not.

    4. Rational expectations implies that making an accurate forecast can be a matter of luck (people who have elevated Krugman, Roubini, and others to devine status should remember that) unless their record is consistent, and unless their specific prediction is derived fron a detailed theory that can be generalized and used by others (rather than on a hunch). Otherwise Ron Paul, who has been prediciting a financial meltdown since the 1980s should be considered a master economist. Keep predicting a crisis and over a course of 80 years chances are you will eventually be correct.

    Understanding all this is important because it leads to interesting questions. Are rational expectations less likely to hold for events that are relatively infrequent (and thus the cost of systematic errors is smaller and the opportunities to learn fewer)? How easy is it to exploit the systematic errors of others ("The big short" by Lewis is quite informative).

    The problem is that conducting this discussion in public is extremely difficult, and gives commenters like Luke Lea one more reason to attack economists. People who have no training or understanding of the terminology leave the discussion with a completely distorted view of the issues. I am a bit puzzled however by their expectations. If I showed up in a discussion over string theory vs. loop quantum gravity in a physics seminar it would be unreasonable of me to expect the participants to make everything crystal clear. Yet this is what is expectated of economists. Of course, the blame should fall on those economists, who shall remain unnamed, that chose to bring this discussion (and whatever issues they have with their colleagues) out on the streets, rathen than follow the proper venues.

    ReplyDelete
  19. reason -

    Actually what I said was intended to be orthogonal to what you said.

    I'm saying RET cannot be right because it is based on erroneous assumptions. I guess that doesn't make it a false theory. Just a theory that cannot apply to real people in the real world.

    Another reality is simple distraction. People fall in love, deal with sick kids, suffer hangovers, get caught up in March Madness or the World Series, Blog, argue subtleties about the trappines of binary theories . . .

    Delong has a rather topical rant up today.

    http://delong.typepad.com/sdj/2012/02/barry-ritholtz-heres-why-most-investors-are-guaranteed-to-lose.html#tpe-action-resize-152

    Cheers!
    JzB

    ReplyDelete
  20. Wonks Anonymous4:15 PM

    Levine explicitly distinguishes rational expectations from perfect foresight. If an outcome is completely random, you can't predict it, even if you are completely rational. But if you systematically guess way too low (or too high) and don't update your beliefs when confronted with the actual value, that indicates irrationality.

    False assumptions can easily result in correct theories. Starting with false assumptions in mathematics can allow one to prove any result one likes.

    ReplyDelete
  21. Wonks Anonymous4:18 PM

    A few years back David Levine had a bloggingheads debate/discussion (the site likes to call them "diavlogs") with philosopher of science Alex Rosenberg. A major point of discussion was the degree to which economics can predict things. Levine calls that "forecasting" and says the field's ability to do that has greatly improved, but thinks it forte lies elsewhere such as in saying what the conditional effects of policies will be. In that respect he sounds closer to Noah:
    http://bloggingheads.tv/videos/2285

    ReplyDelete
  22. no, sorry dwb, rational expectations is not an assertion that 'people respond to incentives', it's an evidence destroying feature of models that incorporate it, which are therefore comically incapable of reproducing empirical outcomes. The real world system RE caricatures is profoundly complex, with several layers of important feedbacks, many different classes of important actors and institutions with incentives that look nothing like those the cartoon theory would have, and interactions between all and sundry.

    You want to know what real life money managers are most afraid of? It's not losing a lot of money. It's losing money when their peers aren't. You want to know how real world alpha strategies generate signals? They estimate things like growth and sentiment and value, each of which is powerfully affected by the variables they intend to predict. You want to know what managers that have been successful with a strategy most typically do? Double down (which is to say that not all losses endured by the experts in big time busts are the consequence of fraud. Not that many are not).

    Of course we haven't talked about the way in which the financiers fit into the process, or how the provision of liquidity itself can turn the paper they sell into Giffen goods. And we haven't discussed the fact that it's tough watching your neighbor get rich, or the potential that human being's notorious predilection to play 'greater fool' has its roots in evolutionary psychology.

    Long story short, the purpose of rational expectations is to make a difficult problem tractable by destroying the evidence of its abject speciousness and sweeping the detritus under the carpet Levine style. That's all. Never mind the fact that in so doing it elevates a single inert dynamic to a role it has no business in playing.

    The fact that this discussion is happening after 30 years of serial financial calamities with exponentially increasing intensity and with exquisitely well argued/documented texts that have proven uncannily prescient gathering dust in the libraries of economics departments the world over.... what can you say. It's simply mind boggling to me. Truly, economists will never learn anything.

    ReplyDelete
  23. Anonymous12:25 AM

    Nitpicky I suppose, but I must ask: just conjugate pairs of observables, or any number of observables > 1? Admittedly, the original application was to a pair (position and momentum), but I think an inductive argument where a single pair is the basis case that generalizes the principle to any number > 1, would be fairly straightforward. No?

    ReplyDelete
  24. Anonymous12:35 AM

    BTW, first, I don't know that pairing is an issue, as much as interrelatedness (dependence, interaction) of the observables, such that their observation is inseparable. I.e., the principle is likely to apply (my physics is rusty - this is off the top of my head) to n observables that are all mutually interdependent.

    The fallacy of composition comes to mind - but my background here is math, physics, formal logic, and computer science, not economics.

    ReplyDelete
  25. Anonymous12:51 AM

    "By contrast if a theory is not one of rational expectations it means "if people believe this forecast it will not be true.""

    One of my professors called this "the spinach fallacy" - "if you don't eat your spinach, you can't have desert" doesn't guarantee that you'll get desert if you do eat your spinach.

    Proving a hypothesis doesn't prove its corollary. And there are many rhetorical ways to say this, e.g., a necessary condition is not a sufficient one.

    ReplyDelete
  26. Anonymous12:53 AM

    Contrapositive, not corollary... Sorry (my logic is rusty too).

    ReplyDelete
  27. Anonymous12:59 AM

    Nope... Inverse, not contrapositive... Too many years have passed, alas.

    ReplyDelete
  28. My catch phrase when confronting dumbness in action -- apparent all around us, especially in policymaking -- was "I'm just a high-school teacher in Napa, and I figured this stuff out." (I recently retired.)

    Next: I bought $1k of AOL in 1997, and in 2000, it was valued at $34,000. I then read an article somewhere that spoke of the tulip bubble in the Netherlands three hundred years ago or so. Then I looked at my portfolio and, uh, froze. Five years later, I sold my portfolio of stocks, still at a profit, but without about $32k of AOL stock value.

    My point is I'm smart, savvy, rational, and a GOOD READER, but I couldn't sell AOL at its top to save my life. And I'm a rational actor.

    Later, I told my girlfriend that the real estate market was crazy, and that buying in an atmosphere of seller strength and buyer frenzy was a recipe for disaster. This was 2001.

    Around 2003 or 2004 as best I can recollect, I heard Alan Greenspan say that an adjustable mortgage was a good vehicle for some. Since mortgage rates were at historic lows, I tried to think of why he might say this. I concluded he was insane.

    That didn't stop me from buying a house in 2005 -- before the market could eclipse me -- for 120% more than the buyer had paid in 1999. I did it because the market had become irrational -- which I recognized -- but also because I believed that this irrationality had become embedded in the system. At this point, I might have thought that I wasn't as rational an actor as I had thought.

    While this was happening, my father died. That was 2005. I worked with my oldest brother -- trustee of the estate -- trying to convince him we were in a housing bubble and that the family home we were sitting on, bought in '74 for $34k, was solid gold, if we didn't dawdle. He dawdled, dawdled, and so I sued him, forced the sale and grabbed my filthy lucre. He doesn't talk to me anymore. I can't even ask him if he noticed the popping of the bubble.

    Just to conclude the story, I later bought a condo in my town, Sonoma, CA, at a price 60% below a price level I had thought priced me out, a mere three years from the top. Now it's valued about 20% below what I paid for it.

    And I'm a rational actor. Good luck building a reliable model based on rationality. But don't take my word for it. I'm just a high-school teacher in...

    ReplyDelete
  29. Herman4:42 AM

    Levine's talk of prediction seems to be a red herring. He is making a correct but irrelevant point.

    One can instead more fruitfully talk of robust and fragile systems. Of increasing or decreasing fragility ( susceptibility to shocks) and hence increasing or decreasing chance of a crisis without ever being able to predict when a crisis will occur. This is hardly an original observation.

    ReplyDelete
  30. Herman6:16 AM

    Levine:
    "...if there is a correct theory, eventually most people will believe it, so it must necessarily be rational expectations. Any other theory has the property that people must forever disbelieve the theory regardless of overwhelming evidence..."

    Overwhelming evidence in macroeconomics? That is a contradiction in terms, isn't it?

    Does a single correctly and completely specified model exist? No.

    If a single correct model exists, it will be a ratex model. However it does not follow that any random ratex model will be a better approximation of the actual economy than a non ratex model until Levine's nirvana is actually realised.

    In any case, different people believe different theories because differing knowledge, differing information, limited intellectual capacity and limited evidence. This will always be the case unless we manage to create a society of omniscient beings.

    Honestly, Levine's writings since the crisis broke have been very disappointing. It is becoming increasingly difficult to take him seriously anymore.

    ReplyDelete
  31. David Levine: "Or put differently: if there is a correct theory, eventually most people will believe it, so it must necessarily be rational expectations. Any other theory has the property that people must forever disbelieve the theory regardless of overwhelming evidence -- for as soon as the theory is believed it is wrong."

    I have what I've named the Iron Law of Right-Wing Projection: Anything that they accuse others of, they're doing themselves.

    In this case "people must forever disbelieve the theory regardless of overwhelming evidence " is precisely what the Chicago School is doing.

    ReplyDelete
  32. First let’s debunk common misconceptions.
    1. The theory of rational expectations does not postulate that people do not make errors. It postulates that these errors are not systematic. Errors are systematic when they are predictable. To use a sports analogy, it would be like shooting from the same spot the ball too hard, past the basket, over and over, game after game. It is easy to see that systematic errors can occur in the long-run only if people are incapable of learning.
    2. As an average relationship, rational expectations are applicable even if most people do make systematic errors, so long as a few are able to recognize this fact and exploit it. So suppose that most stock traders systematically overreact to bad news but a few recognize this and decide to go long when others go short and vise-versa. Then stock prices will behave AS IF they are traded by a group of individuals that exhibit rational expectations even if most individuals do not.
    3. Like Levine mentions, making an accurate forecast can be a matter of luck (people who have elevated Roubini to divine status should remember that) unless their record is consistent, and unless their specific prediction is derived from a detailed theory that can be generalized and used by others (rather than on a hunch). Otherwise Ron Paul, who has been predicting a financial meltdown since the 1980s, should be considered a master economist. Keep predicting a crisis and over a course of 80 years chances are you will eventually be correct.

    All this is important because it shows that rational expectations are appropriate under very broad assumptions. Unless people are incapable of learning, which means that educators are scam artists, I can think of only one case where they are less likely to hold: for events that are relatively infrequent (and thus the cost of systematic errors is smaller and the opportunities to learn fewer) and when it is not easy for the few that “know better” to exploit the systematic errors of others ("The big short" by Lewis is quite informative). The problem is that conducting this discussion in public is extremely difficult. People who have no understanding of the terminology leave the discussion with a completely distorted view of the issues.

    ReplyDelete
  33. @CA
    I largely agree with your statement about "systematic errors." I prefer to avoid the term "predictable" though when refering to rational expecations because there is a lot of rational behavior that is predictable even when agents know the correct model. I prefer to phrase it as a statement that people do not (collectively) fail to respond to incentives (in aggregate, not at the individual level).

    Moral hazard and externalities are a good example: it is perfectly predictable that if a lot of banks think someone else will bear the risk, they will underprice lending, which will lead to overpriced assets (e.g. housing where leverage/lending plays a key role in purchasing).

    So I would just be careful: people predictably respond to incentives, and poor market design can lead the observed price to be different from the EMH (full information/riskbearing) price.

    @ Noah: i see you are a former physicist. I think levine got his physics analogy wrong, its statistical thermodynamics not the uncertainty principle. {All the atoms in my car are going in random directions, but on average they are all travelling at the same speed as me as i travel down the highway with my windows closed.} the uncertainty prinpical would be more apt if he had said that in an illiquid market, I can only do price discovery by putting out a bid-ask. but by putting out a bid-ask, I am affecting the demand/supply and therefore the potential price as i move along the supply/demand curve. In an illiquid market, its hard to know the P and Q simultaneously because they covary.

    ReplyDelete
  34. @dwb

    I suppose the issue is whether arbitrage is possible or not. If it is then "responding to incentives" would require that some people short the housing market therefore bringing the price down to the EMH. This issue has important policy implications. How do we deal with such deviations from the EMH? Do we ask policy-makers to identify and correct them or can we improve the functioning of capital markets? We do agree though: if we simply declare that people are irrational, since this includes policy-makers then might as well shut everything down and go home.

    ReplyDelete
  35. CA/dwb, I'm sure you've heard the phrase 'the market is climbing a wall of worry' before, did you guys consider the idea that 'arbitrage' is an essential component of speculative asset bubbles? Ever caught site of a market maker after he got caught on the wrong side of a trading algorithim? Ever tried putting on a short or zero-sum position yourself?

    Long before Lewis, Keynes famously observed that 'the market can stay irrational longer than you can stay solvent', and there are infinite anecdotes of just that phenomenon in the last 30 years. GMO's almost going out of business getting the TMT bubble right was but one quintessential example, though there were professionals shorting credit and getting killed all throughout burgeoning mortgage finance bubble.

    In each case, turning points have been defined by near universal capitulation of the shorts, and the attendant collapse in the costs of holding those positions. You might want to consider whether there are any implications of that for your theory. In spite of what you seem to believe, RE defies all caricature fair and otherwise, and that's not an endorsement.

    ReplyDelete
  36. Unfortunately, late is the best I can often do, with little time to spend on blogging, but this is still worth commenting:

    What about just ignorance and asymmetric information in a ridiculously complex world with people worked to death and little time to learn the massive amounts these right wing economists assume. You don't need fancy behavioral explanations. You've got two parents working 50 hours plus, per week, coming home and having to then do housework, and all of the much more intense parenting today. You #%@*!% think they're going to spend the few spare hours they can find studying the governments' budgets and politics, to adjust their personal spending to new expectations of future government spending and taxes??!! And then they'll have to get masters degrees worth of education in political economy to even understand well, and predict well, from the data they did study. And you think people live forever, so eventually they learn everything? As P.T. Barnum said, there's a sucker born every minute.

    Some of these guys I'm sure are rooted enough in reality to understand this, and just purposely lie to gain support for their extreme libertarian philosophy. But others just seem to live in another world. Take Stephen Williamson. He's amazed that anyone can think that the utility a typical person gets from something can depend substantially on the items position/context/prestige, that a person could get substantially less utility from the same house/car/clothing if it's in the 10th percentile as if it's in the 90th, and yet he has no problem by and large accepting the assumptions of Ricardian equivalence. I find it just amazing. What kind of reality do some of these economists live in? Here's the Williamson discussion:

    http://newmonetarism.blogspot.com/2011/12/world-according-to-frank-part-ii.html#comment-form

    And for some of the more formal evidence for positional externalities (aside from the constant barrage of it since age four), see Robert Frank's new book, The Darwin Economy.

    ReplyDelete
  37. @CA,
    no - arbitrage has nothing to do with it. One cannot arbitrage or diversify away systematic market failures including poor market design, moral hazard, adverse selction (externalities) etc.

    @Majorajam

    I've personally seen many traders blow themselves up, and some try to hide their losses. speculation does not cause "asset bubbles." Speculation with someone else's money poor risk control and poor incentives cause traders to take risks they would not take with their own money, and if enough traders are doing that then the market underprices risk and overprices assets. Its not irrational -quite the opposite - and there is no arbitraging that away, because that is the correct equilibrium price given the market design and incentives. heads-i-win tails-you-lose incentive mechanisms are the root cause. If you give me a deal like that, i will take it. most people would, because its rational.

    If the market price seems to be an "irrational speculative bubble" i invite you to think of how it might be rational and what the real root cause is.

    ReplyDelete
  38. @Majorajam

    I largely agree with dwb's response above. Let's go back to LTCM's story. Their initial strategy was to identify deviations of asset prices from equilibrium relationships and exploit them. But as other hedge funds started immitating them such opportunities became scarcer and scarcer, as prices stayed closer to their EMH level. This caused LTCM to take more risks, which led to its collapse. So arbitrage plays a huge role in bringing prices down to EMH levels. How fast it depends of course on the distribution of expectations. I ackowledged first that for events that are infrequent (so people have limited experience in dealing with them) and if arbitrage is difficult precisely because the few that try to short the market cannot do so we may end up with a bubble.

    But this raises several questions: are policy-makers, economists, or anyone else able to identify such a bubble? Why would they not be subject to the same "irrational exuberance" just like anyone else? And if they why don't they share their model with the rest of us so that the bubble would not occur in the first place? Or why do more people also not develop similar models (after all the returns from doing so are huge)? Are RE a bad approximation for describing everyday decisions nd if so is there a better approximation?

    After all this, I believe calling RE "my theory" (I only wish I could take credit for it) is quite unfortunate. And throwing anathemas at it will not make it go away (you may want to read on this Kuhn's "The nature of scientific revolutions").

    ReplyDelete
  39. dwb,

    you take the market as a closed system. But if some market participants make decisions that are clearly "wrong" why will new participants with different operating proceedures not enter and try to exploit these mistakes? (this is an evolutionary argument) What prevents such "mutations"?

    ReplyDelete
  40. dwb,
    So speculation only causes asset bubbles when you’re playing with other people’s money, is that it? CA brought up LTCM- it may interest you to know that their fund was all partner’s money by the eve of their collapse. Round about 5 billion of it if I recall correctly, which one might even consider real money. Were they speculators, or…?

    GMO were betting the right way against the TMT bubble with client’s money, and nearly went out of business for their efforts. Many investment professionals that did try to profit off of the underwriting fiasco well underway by 2004 went bust before 2007 came around. Were these also cases of ‘speculation with someone else's money poor risk control and poor incentives causing traders to take risks they would not take with their own money’…?

    Whatever, heads-I-win-tails-you lose is an institutional arrangement. It doesn’t require that you’re long, which is what you appear to be implying by claiming that it defeats arbitrage. Certainly not if you stop there. Moreover, aren’t all these organizations you’re referring to privately owned? One would think that all the shareholders of all the banks had a stake in not paying people to bankrupt them, and that in a rational expectations world you might not see such systematic ineptitude as led to the multi-billion in losses…? Remember, ‘while the music’s playing you have to get up and dance. We’re still dancing’? Yea, that was printed in the FT.

    Some blanks here to fill in. While we’re at it, I’ll need to know what keeps the mooks that own airline stocks going.

    Course you missed the point in the first place, which was probably my fault for not making it clearer. The point is that running over the shorts is what gives bull markets their mojo. It provides confirmation of one narrative and embarrasses the counter-narrative. But that is logistics. The underlying fuel for asset bubbles of all stripes is credit and lots of it, but that’s a story for another day. Suffice to say, I will agree with you this much- the idea that all the massive bubbles we’ve seen in the last 30 years are a function of collective irrationality is not persuasive to me either.

    ReplyDelete
  41. CA

    LTCM’s initial strategy was to buy rich assets and sell cheap ones, when they bothered. In some cases they made money on these trades by throwing their weight around with the financiers. In other cases, they made money because they were short volatility in a declining volatility environment, and generally taking the other side of asymmetric hedges. You should know that a strategy like that is doomed regardless (people generally do now). It was only ever just a matter of time.

    With all due respect, I find the contention that LTCM ‘drove prices to their EMH level’ comical. The most humorous part of this is that is exactly how LTCM would and did describe what they did, and it is exactly that absurd level of hubris and naivete that led directly to their collapse. And here we are not just after LTCM, but after LTCM and quite literally 100s of other examples of just this kind of naivete leading to massive liquidity/financial crises, and you want to argue that these characters bring prices ‘to their EMH levels’. Words fail.

    As to the more general question about policy, you’re making the well worn Greenspan argument, which I once responded to in a letter to the FT: http://www.ft.com/intl/cms/s/0/810f75aa-0697-11dd-802c-0000779fd2ac.html#axzz1mEIz9RGP. It discounts by smiling handwave terrabites worth of evidence to the contrary from work on the degree to which asset bubbles ARE identifiable ex-ante (see e.g. GMO- a model, fyi, well within the public domain), to the implications of differing regulatory regimes in international experience have led to differing outcomes to our own Post-Depression experience. It ignores the groundbreaking and prescient work of Hyman Minsky. And it frankly hasn’t a leg to stand on (didn’t Kuhn have something to say about the how dying fields of science resist to the bitter end…?).

    The question is, remains and always will be ‘how do you manage the endogenously effected, inevitable and profoundly deleterious financial crises that arise in a capitalist system, and how especially in a capitalist system with a modern financial apparatus. We could probably improve on what was in place following the Great Depression. We could not conceivably do any worse than we have done in designing the one that in the last 30 years has arrived the entire global economy to the precipice of despair. You see those riots in Greece? Yea. Watch this space.

    ReplyDelete
  42. Majorajam,

    "With all due respect, I find the contention that LTCM ‘drove prices to their EMH level’ comical"

    You may want to read "When genius failed". It describes how, as more firms got into arbitrage deviations of asset prices from equilibrium levels became harder to identify and LTCM had to rely increasingly on leverage.

    "(didn’t Kuhn have something to say about the how dying fields of science resist to the bitter end…?)."

    Yes. He also explains why such resistance is both appropriate and constructive.

    As for my own views, I believe that for things that people deal with relatively high frequency RE is a good approximation. For infrequent events I do think that people make systematic errors. My favorite explanation is herd behavior (it is less costly to be wrong when everyone else was also wrong). And I do think that shorting the "irrational exuberance" is very difficult because of financing issues. At the same time, I remain skeptical about the ability of policy-makers to predict such bubbles and correct them. Particularly because if their opinion is credible, self-fulfilling prophecies will lead them to always be right (you have an endogeneity problem). My proposal is similar to that od dwb's. Simply have regulation that addresses the perverse incentives, so that when do bubbles occur they remain limited. Of course, in an ever-changing world, regulation will always play catch-up.

    ReplyDelete
  43. PS to Gall's list, we could almost certainly add things like, I don't know, 'off balance sheet vehicles and other regulatory arbitrage- not a good idea' and 'do not sell the most complex indecipherable mortgage products to the least sophisticated/most-vulnerable demographics' and let's not forget 'something is amiss when already massive mortgage assets double in six years'.

    Of course that requires us relaxing the constraint that common sense is verbotten when it comes to fiancial regulation, something that republicans and certain 'economists' are very keen to keep from happening.

    ReplyDelete
  44. CA,

    You may want to reread "When genius failed". Because LTCM did not increase leverage ratios to target investment returns. They were under no pressure to do that. They were not managing against liabilities, nor even client expectations by the end. They did however manipulate leverage to target S&P500 like volatility, which is yet another tell tale sign that the naivete and eye-watering hubris which bankrupted them was born in the faculty lounges of the Chicago school.

    And let's just close the book on LTCM. Yes, some of the trades they made money on became less lucrative over time. Yes, some of that had to do with the crowding of those trades, (a feature of dysfunctional 'money manager capitalism'). But some had nothing to do with that- e.g. euro convergence trades- and some, like selling vol and punting long on emerging market debt were eagerly indulged in the whole way through. So no, the problem wasn’t a changing risk profile.

    On the plus side, the story you are telling about LTCM is not even wrong. Let’s unpack this. So a bunch of prop desks join LTCM in flooding trillions into credit and liquidity trades. This leads directly to (and is reinforced by) a speculative boom that ends in a massive Asian/EM/credit contagion fiasco, which via subsequent policy response leads the blowoff phase of the developed markets TMT/corporate investment bubble. And we know what happens next.

    The idea that the pre-crisis asset prices that resulted from this rent seeking behavior were somehow discovered through some efficient 'rational expectations' process counts as whistling past a world war worth of graveyards.

    You keep giving me reading assignments. Here is one for you. It is a contemporaneous account of the global context in which the LTCM bust went down, and profound in the context of what has transpired since. Of course, voices like Gall's get no hearing notwithstanding their record of prescience while the consistently laughably wrong remain at the head of the table. But that doesn't relegate that voice to non-entity status, much as the VSPs would like it that way.

    As to policy, let's see how much 'catch up' the regulation prescribed in that 15 year old piece would have to meet the challenges of the circumstances that came after it:

    1. Limit short-term lending to a certain percentage, fixed by the BIS, of foreign assets of lending countries and foreign liabilities of borrowing countries, making allowance for normal provision of trade credits.
    2. Limit leveraged trading of financial assets by banking institutions. [a.k.a. Volker rule]
    3. Eliminate offshore financial centers by international agreement, with the United States and Britain taking the lead in refusing to enforce contracts registered in these jurisdictions. [note that all the SIVs and most hedge funds enjoyed offshore status]
    4. Restrict sales and trading of derivatives to public futures exchanges where contracts are registered, records of large positions are kept and prices published under regulated capital-adequacy provisions. Banks and other financial institutions should be required to allocate capital to support derivatives bets, with courts assigning them the same legal status as gambling debts.
    5. Banks should not be allowed by regulators to supervise their own risk profile with "risk control" software, which can generate dangerous macroeconomic effects by failing to anticipate political and credit risk as well as random events

    Long story short, people that want to claim that 'it's just not possible to foresee these things' and 'why don't they get out there and arbitrage it themselves', really ought to stop pulverizing a straw man and deal with the evidence itself. They might wake up to how unambiguously unfavorable it actually is to their policy prescriptions.

    ReplyDelete
  45. @CA

    you take the market as a closed system. But if some market participants make decisions that are clearly "wrong" why will new participants with different operating proceedures not enter and try to exploit these mistakes? (this is an evolutionary argument) What prevents such "mutations"?

    you have to define for me what "wrong" means in a precise economic sense. not the market per se is closed, but the law, rules, and market design are closed. think of prisoner's dilemma: its optimal for them to collaborate, but if this is forbidden, and they are in separate rooms, the equilibrium is for them to rat each other out.

    @ Majorajam

    sounds to me as though you see bubbles everywhere, and they are whatever you define them to be at the moment. asset valuation is far more complicated than you make it out to be. If it were as easy as you seem to think, it would have already been done by the gazzilion quants i already know.

    ReplyDelete
  46. My first comment today got eaten by the spam filter. Yet another try:

    CA,

    You may want to reread "When genius failed". Because LTCM did not increase leverage ratios to target investment returns. They were under no pressure to do that. They were not managing against liabilities, nor even client expectations by the end. They did however manipulate leverage to target S&P500 like volatility, which is yet another tell tale sign that the naivete and eye-watering hubris which bankrupted them was born in the faculty lounges of the Chicago school.

    And let's just close the book on LTCM. Yes, some of the trades they made money on became less lucrative over time. Yes, some of that had to do with the crowding of those trades, (itself a feature of the bigger 'money manager capitalism' dysfunction). But some had nothing to do with that- e.g. euro convergence trades- and some, like selling vol and punting long on emerging market debt were eagerly indulged in the whole way through. So no, the problem wasn’t a changing risk profile.

    On the plus side, the story you are telling about LTCM is not even wrong. Let’s unpack this. So a bunch of prop desks join LTCM in flooding trillions into credit and liquidity trades. This leads directly to (and is reinforced by) a speculative boom that ends in a massive Asian/EM/credit contagion fiasco, which via subsequent policy response leads the blowoff phase of the developed markets TMT/corporate investment bubble. And we know what happens next.

    The idea that the pre-crisis asset prices that resulted from this rent seeking behavior were somehow discovered through some efficient 'rational expectations' process counts as whistling past a world war worth of graveyards.

    ReplyDelete
  47. continued:

    You keep giving me reading assignments. Here is one for you. It is a contemporaneous account of the global context in which the LTCM bust went down, and profound in the context of what has transpired since. Of course, voices like Gall's get no hearing notwithstanding their record of prescience while the consistently laughably wrong remain at the head of the table. But that doesn't relegate that voice to non-entity status, much as the VSPs would like it that way.

    As to policy, let's see how much 'catch up' the regulation prescribed in that 15 year old piece would have to meet the challenges of the circumstances that came after it:

    1. Limit short-term lending to a certain percentage, fixed by the BIS, of foreign assets of lending countries and foreign liabilities of borrowing countries, making allowance for normal provision of trade credits.
    2. Limit leveraged trading of financial assets by banking institutions. [a.k.a. Volker rule]
    3. Eliminate offshore financial centers by international agreement, with the United States and Britain taking the lead in refusing to enforce contracts registered in these jurisdictions. [note that all the SIVs and most hedge funds enjoyed offshore status]
    4. Restrict sales and trading of derivatives to public futures exchanges where contracts are registered, records of large positions are kept and prices published under regulated capital-adequacy provisions. Banks and other financial institutions should be required to allocate capital to support derivatives bets, with courts assigning them the same legal status as gambling debts.
    5. Banks should not be allowed by regulators to supervise their own risk profile with "risk control" software, which can generate dangerous macroeconomic effects by failing to anticipate political and credit risk as well as random events

    Long story short, people that want to claim that 'it's just not possible to foresee these things' and 'why don't they get out there and arbitrage it themselves', really ought to stop pulverizing a straw man and deal with the evidence itself. They might wake up to how unambiguously unfavorable it actually is to their policy prescriptions.

    ReplyDelete
  48. Substantive dwb. Sounds to me like you are reading comprehension deficient or willfully obtuse. A definition of asset bubbles is a part of about nothing I've said so far, except in that GMO does have a working definition (which is arbitrary, but has some highly robust, as in 'no exceptions' empirical evidence).

    The truth is that monomaniacal focus on 'asset bubbles' misses the point. The question ultimately at issue is how to design a robust functional financial system that fosters a maximally stable real economy, something that resembles the opposite of what we actually have. But by all means, you keep tilting at strawmen if it makes you happy.

    ReplyDelete
  49. dwb,

    I suppose my objection is with the notion that market structure leads to uniform behavior. For any trade to takes place you need two exact opposite positions. For any long position there must be a short position or the market breaks down. So say then that those who go long do so because of perverse incentives. Then there must be something different in the organizational structure of those who go short that makes them immune to these incentives. So if in repeated games those who go short (bet that the price of an asset will decline) end up being correct then this should lead new entities that mimic the organizational structure of the "winners" to enter the market and those who make systematic errors to start exiting. If this is not happening then we must wonder why? Is the type of game repeated too infrequently? Are there barriers to entry and exit? And so on...

    ReplyDelete
  50. I think you're absolutely right in critizing Levine's almost insulting (to my intelligence, at least) attempt to misrepresent what is the real discussion about the scientific status of REH. I wrote a (rather lengthy, I'm afraid) critique of Levine's failure of defending REH, that may be of interest:
    http://larspsyll.wordpress.com/2012/02/14/david-k-levine-is-totally-wrong-on-the-rational-expectations-hypothesis/

    ReplyDelete
  51. Anonymous7:32 AM

    You make sense out of the most complex topics.
    bad credit home loans

    ReplyDelete
  52. The quality of your articles and contents is great.
    dwi

    ReplyDelete