Critique of Rational Expectations

Economics Economics Questions
1

This is my critique of Rational Expectations. If you have something to add, or If I’ve made some kind of mistake, please let me know. This thread is extremely lengthy.

The most vocal criticism of the Austrian theory of cycles comes from the rational expectations crowd. They assert that individuals simply do not systematically err in the aggregate, and are therefore immune to arbitrary alterations and manipulations in the price mechanism, especially lowered interest rates. Let’s ignore, for the moment, that this argument entirely ignores the function of the price mechanism, the fact that market interest rates are only “high” and “low” depending on their positions relative to the natural rate[1], and the interrelated microeconomic effects of inflation. Rather, I think it would be more productive if we first actually focused on the theory of rational expectations itself, and its extensions, namely the efficient market hypothesis, both the weak and strong versions.

Rational Expectations:

Prior to the rational expectations “revolution” of the 70s economists regularly employed a one-dimensional theory of expectations solely based on past historical data (adaptive expectations). So, for example, if past inflation rates averaged 5%, expectations of future inflation would be 5% as well. If, on the other hand, inflation rose to a rate of 7%, then inflation expectations would gradually rise to 7%. Clearly, this is problematic. Individuals don’t solely rely on past historical averages when forming their expectations and calculations; they attempt to incorporate as many relevant variables as possible. So, for example, if the Federal Reserve announces that it will triple the supply of high-powered money, and if individuals have some basic understanding of economic theory, then they should expect inflation rates to exceed past historical averages, and they will factor this into their calculations. They will do this because failing to incorporate all relevant variables is very costly (bond holders, for example, will get crushed if they don’t understand inflation).

This is all well and good, and I doubt that many would seriously contest this line of reasoning. But John Muth, the father of rational expectations, went one step further, he asserts: expectations will be identical to optimal forecasts using all available information**[2]**. There are two major implications from this conclusion:

  1. If there is a change in the way a variable moves, the way in which expectations of this variable are formed will change as well: So if the interest rate, for example, is “high,” then individuals will expect it to return to its “normal level.” If it stays “high,” then individuals will expect it to remain “high.”
  2. The forecast errors of expectations will, on average, be zero: This was already mentioned, but the formal statement of the theory is X*e* = X*of* (the expectation of X equals the optimal forecast using all relevant information).

It is important to note that rational expectations does not assert, as many claim, that individuals have perfect information; that is, it admits that some information is simply unavailable, and it actually claims that some individuals may choose to ignore relevant variables because it may require too much effort to identify (too costly). Thus, Rational Expectations admits much of its theoretical deficiencies, and this already casts doubt on its theoretical tenability and usefulness. The truth of this concession is most evident within the political sphere, where individuals (a) are unaware of the true intensions of politicians, and (b) where they simply refuse to educate themselves politically (purposely ignore relevant variables when they make political decisions/vote).

It’s true, though, that the market is unlike the political sphere in many ways. For example, individuals actually have power in the market, and the intensions of market actors are immaterial; only results and performance matter (assuming that the system is free from arbitrary advantages and disturbances). But in the market there is a substantial difference between what individuals attempt to do and what actually happens (the inevitable result of extreme complexity and uncertainty). Individuals may attempt to use all of the relevant information, the same way that the entrepreneur attempts to engage in profitable productions, but distinguishing between relevant information and irrelevant information is an extremely difficult endeavor (much more so than in the political realm), especially when the relevant information is contained within prices (expressed by the price mechanism).

The entrepreneur, for example, needs to understand much more than his own personal preferences and the preferences of one or two actors; he needs to understand the marginal technical rates of substitution amongst various heterogeneous goods with varying degrees of complementarity; he needs to understand the subjective desires of billions of individuals, which are in continuous flux; he needs to understand the ramifications of government intrusion into various markets, ect ect.

The degree of competency required to obtain such information without a functional and accurate price mechanism is beyond the scope of human cognitive abilities. In other words, prices (not intuition) guide production. Hayek explains,

Let’s quickly reexamine the first implication of rational expectations, since there are many assumption already built into it:

  1. Each individual is fully aware of the time preferences of all other individuals and knows what the natural rate of interest is. In other words, they are able to see what the normal return on investment would be in a purely theoretical barter economy.
  2. They fully understand the effects of a suppressed market rate of interest below the natural rate of interest and choose not to capitalize (for lack of a better term) on potential short-term profits because they can see into the future.
  3. Essentially, individuals have some intuitive connection to some illusory general equilibrium; that is, they know where the interest rate “should be” even if the market does not express it.

The absurdity of such a position is obvious. Thus, rational expectations, within the realm of economics, are entirely contingent upon a price mechanism that is not continuously manipulated by external authorities. And since prices are in fact continuously altered, we must therefore dismiss rational expectations as a valid critique of the ABCT.

The Efficient Market Hypothesis:

I mention this extension only because I wish to elucidate the point that simple theoretical mistakes have a tendency to turn into unforgivable abominations. The weak version of the EMH merely incorporates RE within the field of finance (current prices in a financial market will be set so that the optimal forecasts of a security’s return using all available information equals the security’s equilibrium return[3]). But it’s the strong version that is truly remarkable: an efficient market prices securities so that they reflect the “true intrinsic” value of the securities. Thus, prices always reflect market fundamentals, so that any investment is just as good as any other investment. And thus we have returned to the classical framework, where value and prices are identical.

Conclusion:

If individuals truly had some mystical connection to some general equilibrium, where their expectations were identical to equilibrium results, then the price mechanism, i.e., the explicit expression of opportunity costs, would be entirely superfluous. We would merely need to find the most intuitive individuals and have them centrally plan our economic system. Recessions and endogenous price rigidities simply could not exist under such circumstances (which is why some proponents of rational expectations deny the existence of bubbles contrary to all empirical and theoretical evidence). But this is merely a utopian fantasy that ignores reality.

Furthermore, Rational Expectations, and its extensions, have been an expedient tool for all those who oppose the free market system and free market economics. Liberal professors continuously refute these straw men while ignoring the indubitable arguments put forth by the likes of Bastiat, Say, Mises, Rothbard, ect.

[1] In other words, a 4% market interest rate may be “too high” and a 15% market rate of interest may be “too low.”

[2] John Muth, “Rational Expectations and the Theory of Price Movements,” Econometrica 29 (1961): 315-335.

[3] Mishkin, Frederick S., Money, Banking & Financial Markets, 9th edition. (2009): 157-158

2

Thought provoking. I’ll say two things that immediately come to mind:

It might be valid to extend that RE is a claim about intentions (their quality), and is therefore immaterial.

Is RE circular? It assumes people will behave rationally BECAUSE they have rational prices and information (which could only have been formed rationally in the first place…)?

3

Esuric, do you believe markets clear? Or, do you believe in non-clearing-markets or specific exceptions? If so, what are they?

4

Esuric, let me present a quote from Todd Knoop: rational expectations “does not mean perfect information because some information may not be publicly available and some information may be prohibitively costly to obtain (i.e., the marginal cost of such information is greater than its marginal benefit). . . . What if some segments of the public are not rational and do not use all available information when setting expectations? Rational expectations are not invalidated. Those who are rational will take advantage of the profit opportunities created by those who are consistently making predictable mistakes. . . . The notion–that individuals learn and do what is in their own best interest–is at the heart of the discipline of economics. To assert that individuals do not act rationally is to assert that most of our economics is wrong.”

Do you disagree with any portion of that and, if so, what?

5

This is an interesting point.

6

I deal with this in my post.

7

You conceded the fact that rational expectationists don’t assert “perfect information,” but then you concluded that such admissions undermine the “tenability and usefulness” of the hypothesis, whereas Knoop stated, “Rational expectations are not invalidated.” So, I’m not sure you fully dealt with that issue.

8

This,

to this,

Is a non sequitur. The RE crowd always retreats to its initial and almost reasonable assumptions. They never actually defend their absurd conclusions, which are not the logical extension of the premises. Furthermore, this comment directly contradicts the formalized expression of RE (Xe=Xof). If RE asserted that individuals attempt to use all relevant information, then I wouldn’t find it problematic at all, and the ABCT would hold. But it doesn’t say this: it says that individuals, in the aggregate, always use all relevant information (if its available). Never mind the fact that it entirely ignores the function of the price mechanism (which contains, or is supposed to contain, the relevant information).

Also, I would appreciate direct and explicit arguments from you. That is, it would be nice if you could quote passages from my thread and deal with them rather than spaming quotes from other economists.

9

From Tyler Cowen

10

Esuric, if expectations aren’t rational, then what are they? Irrational? Inefficient?

11

Expectations of markets depend on market fundamentals. Entrepreneurs are only one element.

12

Uhmm… okay?

13

I’ve been trying to think of a good example that would illustrate how RE applies to the real world, because the mathematical concept can only be strictly applied under completely unrealistic assumptions, e.g., all information must be given to a group of individuals (not to each individual but to them as a whole). This example is a little lengthy but I think it illustrates all of the relevant problems that can arise if you do not carefully consider what RE means.

So imagine 10 people are asked to guess something about some individual, say the number of sit-ups that he will do next Sunday. This is very private information. The 10 people can make reasonable guesses on the basis of that man’s weight, his personality, and so forth. But if that man has in the past done his sit-ups (if any) inside his home where no one else can see, there is no reason to expect that the average of the 10 individual guesses will be equal or even very near to the actual number of sit-ups that that man ends up doing.

So in short answer to your question, without full information given in some pre-requisite form to a group, there is no reason to imagine that those expectations will on average be equal or even near to the true value, i.e., that they are rational. If anything we can only describe those expectations as subjective.

Now imagine we modify the experiment so that there are rounds. Each Sunday the guessers are accurately told whether they were ‘way high’, ‘high’, ‘just high’, ‘just low’, ‘low’, or ‘way low’. Assuming they want to guess the right value, and assuming that the sit-up guy does the same number every Sunday (this could be modified), each will update his expectation for next Sunday accordingly. Over time their guesses will come to approximate the true value. In other words, over time, the average of their expectations will seem to be ‘rational’. But note that such ‘rationality’ has nothing to do with their incorporating all possible information and utilizing it optimally. As stated before, all of the relevant information will never be given. Thus, there is nothing intrinsically rational about their expectations. What makes it seem that way is the selection process, i.e., the revealing of who was good and who was bad at the end of each round.

Now imagine we modify the experiment one more time. Many would say this last modification is trivial but it’s the most important part. Imagine that for some reason we cannot directly convey by means of telling the guessers how good their estimates were. Rather, we may only convey their accuracy by writing on a small slip of paper and giving that slip to the respective individual. In this case such a modification sounds funny, but in the case of a complex system like an economy it makes sense. Producers wouldn’t be able to be ‘told’ about the accuracy of their “guesses” unless money (a common numeraire) is received. So at the end of each round, each guesser is given a slip and on this basis updates his guess. Whether the average of the guesses approach the true value depends pivotally on whether the slips accurately convey the guesser’s accuracy. In other words, their expectations may only be said to approach rationality if the slips are true to the underlying data. If they are not, then the average of those expectations will not approach the true value (recall that the guessers never have access to the underlying necessary information).

And last, suppose that the guessers were told that the some of the slips were inaccurate in the sense that they would suggest the true value was higher than it actually is. Then if you receive a slip that says “way low” it may actually mean “low” depending on whether your guess would fall into the low range. If it says “low”, it may actually mean “just low”; and “just low” may actually mean “just high”, etc.. Only if the slip says “way high” can the guesser be certain that he is in the accurate region. So the question is, would their general knowledge of this circumstance negate the effects of the inaccurate slips. The obvious answer is no. The average of the expectations will not approach the true value under this situation. First, the guessers don’t know who has the faulty slips. Second, even if the guesser knew his slip was faulty (which he does not) this does not mean it falls in a different category since he does not know whether his guess would cross-over or not.

In order for them to avoid the effects of the faulty slips they would need access to the true value. But of course they can never have such access. In other words, we cannot assume that their expectations are intrinsically rational. This is the grave mistake many REer’s make. They believe, as Esuric put it, that entrepreneurs have some sort of mystical connection to the underlying Eq-m. If they do not assume that mystical connection then they must pre-suppose that monetary exchanges at given prices accurately convey the underlying data. But this too is a mistake when such a process becomes disconnected from the underlying data. And even general knowledge of such a disconnection, in a certain direction, cannot negate it’s effects.

14

edward_1313, do you believe market participants, through their interaction, determine the most reliable prices that could be developed? That is, do you agree with Hayek?

15

StrangeLoop, do you think that people can be “rationally irrational” with their expectations? :stuck_out_tongue:

16

First, let me state that I do believe that rational expectations can be weakened by systematic biases; however, I believe the findings of behavioral economics are typically overemphasized and such biases don’t seriously undermine free markets.

However, such biases can be significant when costs are low (e.g., political environments), as Caplan has demonstrated in regard to the “rational irrationality” of voters.

17

Sounds like a “yes” to me! :smiley:

Furthermore, the economy is partially controlled by political means (e.g. central banking)… need I say more?

18

The key part of rational irrationality is that it’s rational to hold irrational beliefs; why so?

Workers will, I believe, negotiate contracts to match inflation; otherwise, they’re being irrationally irrational! Not all of politics is vulnerable to Caplan’s critique.

19

PLEASE REREAD THE PREVIOUS ARGUMENTS CAREFULLY. This reveals that you’re not grasping the essence of the matter. The possibility of business cycles does not hinge on any assumptions about the relative efficiences of various market structures…

20

If there is anyone who had a truly deep understanding of ‘expectations’ and their place in the market structure, it was Hayek. And he maintained that understanding simultaneously with his expansion of the theory of the cycle. I would not invoke Hayek as being on your side.