Econolegends - My Favourite Truly Wrong Legends in Economics
Collected by Marc Oliver Rieger.
All of these legends share the common features:
- They are widely believed and used in publications.
- They refer to a famous paper.
- The famous paper says exactly the opposite.
What do we learn from this? Many people are nowadays too busy to read a paper well, even if it is famous and even if they quote it!
The legend on Prospect Theory
Legend: Prospect Theory, as published by Kahneman and Tversky (Econometrica, 1979), is a decision model for lotteries with at most three outcomes.
Reality: Kahneman and Tversky in their original article:
"The extension to ... prospects with any number of outcomes is straightforward."
(This quotation is at the end of their article, before, they study the special case of two or three outcomes. Some people obviously overlooked their remark at the end, others not: there are several articles using the natural extension of Prospect Theory to arbitrarily many outcomes.)
The legend of the "Rabin paper"
Legend: People are risk-neutral when it comes to small-stake gambles, as has been proven by Rabin (Econometrica, 2000).
Reality: After being misunderstood in this way, Rabin clarified in a letter published in Econometrica:
"We refer the reader who believes in risk-neutrality to pick
up virtually any experimental test of risk attitudes. Dozens of
laboratory experiments show that people are averse to far more
favorable bets for smaller stakes. The idea that people are not
risk neutral in playing for modest stakes is uncontroversial."
(This econolegend is based on the fact that Rabin argues that the assumption that persons behave according to expected utility theory would lead to a risk-neutral behavior for small stake gambles, which is not observed. Hence expected utility theory is not an appropriate model to describe decisions under risk. - People seemed to have stopped reading his paper somewhere in the middle of this argument.)
The legend of the logarithmic utility
Legend: A utility function should be logarithmic, as has been proven by Kenneth Arrow (1970).
Reality: Arrow writes in this paper, that a utility function needs to have a RRA (relative risk aversion) larger than one in order to be compatible with the Super St.Petersburg Paradox, and smaller than one in order to have a finite value in zero.
The logarithm has a constant RRA of one - neither larger nor smaller than one.
In other words: among all functions with constant RRA, the logarithmic function is the worst, namely the only one that is neither compatible with the Super St.Petersburg Paradox, nor has a finite value at zero!
(Arrow uses in the often-cited remark in his paper no mathematical notation, but words. Some people seemed to have mixed up "less" with "less or equal" and "larger" with "larger or equal", that is the only possible explanation I have.)
If you find more such striking examples, please let me know!
Marc