Sunday, January 15, 2012

Opposing Cognitive Biases and Rational Behavior

A friend and I were discussing the logical fallacies that often occur when individuals are making decisions or constructing an argument (hat tip to Ryan Miller).  We identified a particular fallacy--the gambler's fallacy--that exists when an a sequence of identical outcomes is mistakenly construed as implying the probability of the opposite outcome increases for the next iteration.  For example, if the roulette wheel results in black outcomes for 5 spins in a row, there exists a common misperception that the next spin is more likely to land on a red outcome, because overall, red and black outcomes should be approximately equal.  According to the gambler's fallacy, red outcomes would need to "catch up" to make equal outcome possible, and therefore red outcomes become "due" when preceded by a series of black outcomes.

We then considered an alternative and opposite bias that can occurs in the exact same situation.  Many other gamblers believe that a series of black outcomes implies that black is more likely to occur going forward.  From this perspective, black is on a streak, or is the "hot" play.  Its common to hear in casinos that certain craps tables are hot, an implicit belief that a series of outcomes is likely to persist into the future.

I find it very interesting that, in certain circumstances like the one described above, we are prone to any erring on either side of the rational conclusion.  Its unclear whether a sample of 100 gamblers (each without a favorite roulette color) would be more likely to bet on black or red if they observed five black outcomes in a row.

Behavior finance teaches that humans have natural heuristics in evaluating issues and making decisions that make us prone to certain biases.  These biases are generally described as unidirectional.  For example, behavioral finance teaches that individuals generally sell winning investments too early, and hold on to losing investments too long.  This may certainly be true, in general.  However, as described above with the gambler's fallacy, the extent of biases can be more complex, with other possible biases working in the opposite direction.  When competing biases exist, the impact of such biases is not clear then. It depends on which bias, among those that exist, dominates in the circumstance.  Moreover, it seems odd to me that many finance and economics analyses assume investors act rationally, but for certain biases that can be accounted for.  Since we are subject to biases--biases that allow us to err on either side of the rational conclusion--it appears more likely to me to simply accept that humans don't always act rationally.  

No comments:

Post a Comment