Important interview with Robert Shiller on behavioral economics. Must-read. He cover the high points and shows how the way economics is pursued today is doomed to failure, especially at turning points. Models based on "rationality" and "utility are elegantly simple but too simplistic to model human events. Keynes had already observed the intractability of uncertainty and fallacies of composition. George Soros points to reflexivity. Behavioral economics adds the dimension of findings in the life and social sciences.
Why neo-classical economics, which is based on assuming a rational representative agent maximizing utility over a lifetime, is bonkers. The problem? According to Shiller it seems to be physics-envy.
Read or download transcript, or listen to podcast at Social Science Bites
Robert Shiller on Behavioral Economics
Interview by Nigel Warburton
(h/t Cullen Roche at Pragmatic Capitalism)
My comment at Pragmatic Capitalism:
The interview is a must-read in its entirety. Shiller sums up neatly what wrong with mainstream economics. His summary analysis also implies why the conclusions most economists draw are pretty useless for trading, and why listening to them can be dangerous to your portfolio. But most traders have likely figured this out already.
7 comments:
A couple of points:
There was a nobel speech in eco that tried to draw analogy between eco and physics/science. I can't remember who it was now, but someone here might recall. Pretty amazing set of statements highlighting both hubris and ignorance.
Second, the EMH hypothesis continues to get creamed for its inapplicability to ... well ... anything. Markets are mostly mathmatically efficient (through arbitrage), but not informationally efficient, as common eco theory tries to assume. How can it be, when not only does it NOT contain all possible relevant information, but actively discounts extremely important information such as the way people behave when confronted with changing circumstances (or even the same circumstances presented in a different manner, as psychology so aptly describes - see 'Kahneman' as just one name in this field of work). The market prices in consensus, that's all, which isn't very exciting.
apj
Yes and the consensus about pricing value is often way off the market as volatility and bubbles and busts reveal.
What is the actual value wrt markets? It's what statisticians attempt to determine in terms of price movement over time, on one hand, and what analysts examine wrt to financial and non-financial factors.
Value not an observable like price, so many people take the easy way out and conflate price with value, even though the consensus on this fluctuates widely for reasons that are now becoming better understood through scientific research.
Most market participants know this and the successful ones are the ones best at gauging relationship of price and value and making portfolio adjustments in terms of it changing relationships. This is the Buffet strategy. Technical analysis gives the entrance and exit points.
The fact that some fund managers like Warren Mosler have long-running winning track records suggests this is not randomness and luck.
Nice shout out to Kahneman.
Can't recommend Thinking, Fast and Slow enough.
It's as clear and well-written as The Black Swan is incoherent and poorly written.
Engineers are trained in thinking fast and thinking slow. It's the difference between heuristics and analysis. Engineers are very good at ball-parking using rule of thumb, and they don't confuse heuristics with analysis.
Conversely, many people err in one direction or another, relying on intuitive judgments where analysis is called for, or wasting time on over-analyzing. But the problem is usually using fast thinking where slow is required in order to reduce transaction cost. That can be costly.
The other problem is correct v. incorrect analysis, and this often depends on data selection, assumptions and methodology. Economists don't seem to have this one down yet.
Where are the engineers?
Tom,
The diagram on page 12 in this document illustrates the current official take on the process by which most if not all truly new technologies in the west have come into existence:
http://www.dtic.mil/whs/directives/corres/pdf/500002p.pdf
Not sure if it applies or should apply to the development and implementation of economic policy tho... I guess some of it could...
rsp,
Tom:
Problem-solving can be broken down to two parts:
... first define the problem and then solve it.
Solving the wrong problem doesn't help much.
I think that is what mainstream economics gets so wrong and MMT gets so right.
paul, "first define the problem"
Human problems, unlike the natural science where the unit are the same and unchanging, must be defined in terms of context. Models assuming a rational representative agent maximizing utility are basically imitations of models in the natural science and they fail practically because they do not take into account the human context and the prevailing context.
Post a Comment