Michael Mauboussin: Asks What Do We Need To Fix About Markets?
Introduction (via LMCM)
Over the last 30 years or so, there has been a power shift on many Wall Street trading desks. The trading floors used to be filled with people who had limited formal education but who had worked their way up in the organization. These traders lacked academic pedigrees, but the survivors had street smarts. Then a new breed of trader came in, well-educated and quantitatively-oriented. These traders were book smart and tried to outwit the market by using sophisticated models based on finance theory.
But there has recently been a backlash against quantitative techniques, based on the implication that the faulty use of quantitative models was a major source of the boom and bust. Felix Salmon, writing in Wired magazine, suggested an equation dealing with default correlations “will go down in history as instrumental in causing the Legg Mason Capital Management unfathomable losses that brought the world financial system to its knees.” (correction: Legg Mason Capital Management did suffer unfathomable losses but it wasn’t from the equation) A growing chorus is now criticizing the folks who rely on the seat of their intellect instead of the seat of their pants.
To start, here’s a thought on Taleb’s assertion that theory and practice isn’t an equally-paved twoway street. Like most tricky issues, the validity of this assertion depends on the context. In some domains, good theory leads directly to effective practical application. Think of the atomic bomb. But for theory to usefully inform practice, the theoretical models must be a close approximation of reality. This is the case in many physical systems, for example. But since all models are a representation of the world, the greater the difference between the model’s output and the real world, the greater the risk in relying blithely on the model. Models in economics and finance frequently fail to capture the dynamics of markets, which is why blind faith in theory can lead to disastrous practice.
Additional Excerpts (Via LMCM)
The initial question is whether the trader’s success is simply the result of luck. Investment returns combine skill and luck. And of the two, luck is the more significant over the short term.
Chance and survivorship bias make it hard to equate success with a thoughtful process.
There’s an even more basic problem with blaming fancy mathematical models for today’s mess: we’ve had booms and crashes for as long as markets have existed. You can point a finger at computers, CNBC, the Black-Scholes options pricing model, securitization, credit default swaps, or any other innovation or technology as the source of the problems today. But the same pattern has unfolded time after time before any of these alleged culprits existed.
So What Exactly Needs To Be Fixed (Excerpt Via LMCM)
1. Public goods. These benefit everyone in a population, so regulation has to make sure that everyone contributes so as to avoid the problem of free riding—i.e., enjoying the benefits without bearing any of the costs. National defense is one example.
2. Externalities. An economic transaction that produces costs or benefits to parties not involved in the transaction creates an externality. Since negative externalities are not included in market prices, regulators have to step in to limit their extent. Pollution from a factory is a classic illustration of a negative externality.
3. Incomplete markets. When there is not enough supply or demand for a product or service, a market can be incomplete. In such cases, the government may step in to provide that product or service. Government-sponsored unemployment insurance is a case in point.
4. Behavioral biases. Some patterns of human behavior are undesirable, and regulation can help mitigate their costs. Examples of regulation include making school lunches healthier and cracking down on drunk driving.