Monday, April 13, 2015

Theory vs. Practice

Some things, oddly enough, work better in practice than theory, such as quantitative easing.  Most things, however, work better in theory than in practice, especially within the investment world.

The hot debate within the investment world for the last decade has been the argument that index investing beats active management.  In other words, you should just buy the S&P 500 index without trying to buy any particular stocks to beat the market or reduce the risk.  There have been numerous theoretical studies to show this argument is correct, at least in theory.  (Interestingly, it seems to work better in bull markets than bear markets.)  In practice, however, there are mutual funds and hedge funds who do routinely outperform indexes, disproving the theory.

The conventional wisdom within the investment world for the last three decades has been Modern Portfolio Theory, which says you can improve investment performance over the long term, while reducing risk, by allocating portions of your portfolio to all the asset classes, e.g., large caps, bonds, cash, commodities, international, etc.  The problem in practice is knowing how much to put into each asset class.  This is called the "efficient frontier."  I can graph that frontier for you, but there is one problem -- it changes every day.

The latest fashion within the investment world is something called "robo-advisors."  Your portfolio will be invested in the index of each asset class.  As an example, Schwab has introduced something called "Intelligent Portfolios" which has taken in over $500 million in just three weeks.  The theory of combining index investing with Modern Portfolio Theory sounds great, but what you don't know in practice is how much is invested in each asset class and how often is that allocation changed and how is that decision made.  The lack of a flexible, understandable "efficient frontier" dooms robo-advisors.

I will start using a robo-advisor when I start letting a computer drive my car with my kids inside.  Theoretically, a computer can already drive a car safely . . . but not with my kids!  Some things are just too important for any nice, neat little theory.