Quote from vladiator:
Good post, T4A.
I think it sums up the debate quite well and makes some very good points.
I like your swiss army map example. Except that I'm one of those who believe that the purpose of economics (and any other science) shouldn't be simply discriptive, to purely explain as best as possible what is observed in the bounded-rational behavior/reality 
Albeit that is a very useful objective, economics/finance should also analyze the way reality/behavior should be observed, once the boundedness of that rationality decreases.
Economics does not do a very good job of explaining what is, but in many cases that's not what its aim is, so it shouln't be judged solely by how well it does in this respect.
For instance, it is widely observed that people consistently make errors in choosing between different lotteries with cleverly structured payoffs. But I don't think we should adjust the theories to incorporate the occurence of such errors, b/c we think they would persist in the future
Keep up the good posts.
You raise some very good points, vladiator, perhaps I (and others) are expecting too much from economics.
Is Economics a Hard Science or Just a Science that is Hard?
Hard Science = Accurate Causal Models I suppose I evaluate the scientific merits/basis of economics, or any field, in terms of its ability to create causal models of great predictive accuracy. Thus, I wholeheartedly agree with you that a science should be much more than just descriptive. The entire notion of a scientific experiment is defined by the ability to accurately hypothesize the outcome of some experimental treatment of a system. If the theory cannot predict the outcome of the experiment, the theory is rejected. In this way, true science is extrapolative -- letting one predict outcomes in new or fabricated scenarios.
I'm Just a Cranky Engineer with High Expectations And I suppose that my engineering background helps raise the high bar that I set for economics. To me, theory is only useful and true to the extent that it lets me design something that does
exactly what I want and expect it to do. To the extent that the discipline of economics can lead to accurate decisions (whether in interest rate policy or the trades made by LTCM) is the extent that it qualifies as a true hard science. For this reason, I find economics lacking, although I am now learning that I have made extremely unreasonable demands on the field.
My Excessive Expectations of Economics: I admit that I, and others, probably judge economics too harshly. One problem is that people have such high expectations of economics because it has such a substantive impact on quality of life. Nobody gets mad at meteorologists when they fail to predict next year's weather (OK, maybe farmers get a little pissed). Nobody expects weather predictions to be accurate more than a few days into the future. In contrast, everyone gets mad when inaccuracies in next year's economic predictions drive interest rate decisions that turn out poorly. The problem is that, like meteorology, economics must cope with predicting the behavior of a massive, globally distributed nonlinear system. This severely limits the predictive horizon of economics.
Lyapunov Exponent of the Economy is High and Increasing: At some level economists are trying to do their jobs under far worse conditions than those suffered by meteorologists. All paranoid anecdotes aside, the behavior of solar radiation, water vapor, and air masses don't change when they hear the weather forecast. In contrast, economists face the ugly reality that the very subjects of their predictions make deliberative changes in response to the predictions that economists seek to create. The sometimes atrocious investment decisions made with respect to entities that are "too big to fail" are a prime example of this (people take risks that they really should not on the basis that government will step in and forestall any problems).
If economists could just step outside the system, and the system were incapable of seeing the predictions and theories of economists, the accuracy of economics would improve with time. As it is, economic knowledge filters into the system and induces ever more complex dynamics. While the equations of motion for the atmosphere remain constant regardless of the increases in meteorologists knowledge, the equations of motion for the markets and the economy grow ever more complex in response to participant's growing knowledge of economics.
Regarding Boundedness:
Absolute Boundedness is Decreasing: I do believe that the absolute boundedness of each agent in the modern economy is decreasing (maybe we agree on this?). We (individuals, businesses, and governments) all enjoy much greater visibility on to timely, relevant information than in the past. Moreover, software (whether it is Black-Scholes or ERP) augments our intellectual abilities to improve the consistency (= rationality) of our decisions. But does this mean that we are approaching the EMH-required nirvana state of having rational profit-maximizers?
Relative Boundedness is Increasing: I think not because even as boundedness of the individuals is decreasing in an absolute sense, the overall complexity of the economy is increasing far faster. To extend the chess metaphor: the size of the board is increasing, the number of pieces on the board is going up, the pieces are learning new moves, the topology of the board is become more convoluted, etc. (BTW, this litany of on going changes to the rules of the metaphorical markets=chess is the reason that I don't much care for the chess metaphor for trading/economics -- chess is a closed world, the markets are not). Although the economic equivalent of Big Blue is letting us examine ever greater numbers of future moves and countermoves (and we have smarter algorithms for evaluating board configurations), the growth of the combinatoric complexity of economic systems is rising faster than our ability to predict it. In my opinion the look-ahead depth is actually decreasing as companies and individuals adopt sophisticated methodologies that accelerate cycle times.
Boundedness is Unavoidable: The point is that a market of a million participants is ALWAYS smarter than any one participant. Compared to the markets, the individuals will always be pathetically bounded-rational. Indeed, relative bounded rationality worsens in the face of a denser, more interconnected global economy. Ultimately, I believe that a theory of economics that assumes purely rational participants (not bounded rational ones) is probably nearly useless. That EMH is NOT totally useless is because it does provide a sound theoretical basis for the tendency toward efficiency and an explanation for why the markets are as efficient as they usually are.
In Search of a Better Theory
The Real Issue: What is the Distribution/Dynamics of Inefficiencies It seems that we can all admit that the markets have at least some inefficiencies in them -- we now only seem to be arguing over the number, size, and transience of the microinefficiencies (and whether some threshold number/magnitude of microefficiencies are sufficient to invalidate EMH).
IMO, Perhaps the biggest deficiency of EMH is that it is a first-order theory -- predicting the central tendency (the first-order moment) of the markets to converge towards price efficiency. What I, and other traders, want is a second-order theory that predicts the dispersion of dynamics in the market. (BTW, options traders, who would like to understand the second-order dynamics of the second moment of prices would seek a third-order model of pricing :eek: ).
An improved, second-order version of EMH would consider that the boundedness of participants has a cross-sectional distribution and undergoes long-term changes. Perhaps some merger of EMH and complex adaptive systems would help predict the presence of heavy tails, pricing bubbles, and the quantum foam of a inefficiencies that real markets populated by real people are subject to. Until such a theory appear, funds like LTCM should not base their decisions on a theory that could not predict the potential for the market to be wrong for as long as it was (assuming, of course, that is was the market that was wrong, and not LTCM that was wrong).
Markets Become LESS Efficient if Everyone Believes in EMH? Here's a real mind blower. If everyone believed in EMH, nobody would look for inefficiencies in the market (all us traders would pack it up and get "real" jobs). Inefficiencies would grow in the markets as long-term structural changes in the economy inevitably change the true valuation of companies WRT the prior valuation models (only nobody would notice because under EMH there is no point in looking). The proverbial sidewalk would become littered with $100 bills that nobody is looking for because everyone believes that they do not exist. I find it deliciously amusing that EMH only works to the extent that some participants disbelieve the theory and continue searching for inefficiencies anyway.
This paradox that an equilibrium state of efficiency requires a disequilibrium belief in inefficiency is just the sort of second-order effect that a better economic theory would contain. Of course, part of me hopes that nobody discovers that theory.
Please Believe in EMH
: I agree with what daniel_m said back on page 3 of this thread. I secretly hope everyone becomes an EMH adherent because it leaves more opportunity for us nonbelievers.
Until economics creates a better theory that shows me the error of my ways, I'll be
Traden4Alpha
