The Worst Type of Optimizer

Harry Markowitz introduced the concept of diversification into investing back in the 1950s. Investors can reduce portfolio risk by holding combinations of instruments which are not perfectly positively correlated. Using information on the correlation between the returns of the assets in a portfolio you can choose a weighted portfolio to minimize the total volatility for any expected return. A foundation for many results in quantitative finance - a model for portfolio optimization.


In Mark Buchanan's Blog I read. Diversification doesn't work. And why. Correlations change when markets move up or down.

A new paper in Nature Scientific Reports  (the physicist Tobias Preis and colleagues) shows it and M. Buchanan explains it supported by the metaphor of the balanced man in a small boat.

So, if you think you have an optimal position, but you have not?

In general, the theory of complex systems tells you: if you have a theory that looks perfect in the model world you have described and tested it and you derive measures to balance (optimize) in its frame - a changing condition in a "surrounding system" can destroy everything.

Provided you managed for a while to create populations of ice crystals in a pond (they simulate a primitive life) by perfectly balancing with perfect heating / cooling ... if the sun breaks though the clouds or the northern wind begins to blow you might need hectic actions ....

Life only happens at the border between chaos and order and this border moves.

You cannot beat the game of Roulette. But you can make boring bets to play longer (and get some extra free drinks?). But this could change drastically, if you played Roulette on a small sailing boat.

Up to 1987 the world played simple Black Scholes - but then (far) out of the money options were traded and the volatility smile was discovered ....

Oh yes, dynamics really matters.

Portfolio valuation: If you have a system that is blazingly fast you might get some extra insight when you stress-test across portfolio constructions.