The Stock Market

You have not seen much about the stock market on this 'blog recently. I have been taking an asset allocation approach. The general idea is to divide your investments into a fixed percentage of different, uncorrelated, asset classes. One class might go down but another would probably go up, smoothing things out. Every now and then - maybe once a year - you should re-balance the portfolio selling and buying to maintain your target allocation percentages.

With this approach, you do not try to predict whether "the market" will go up or down. You just rely on the fact that historically it has gone up over the long run.

(Of course, as Keynes said, "In the long run we are all dead." I was born seven years after the stock market crash of 1929. I was a Sophomore in college by the time that the Dow Jones average recovered to the 1929 level. But wait! Taking inflation into account, it took 30 years for the Dow Jones to recover. By then I had graduated and was married. That's a long time to wait to get your money back.)

I have explained this in simple terms, but there is a lot of academic study and common sense that says it is a good strategy. It's called Modern Portfolio Theory and you can find it in investment textbooks. My target allocations are: 12.5% energy stocks, 12.5% gold stocks, 20% international stocks, 12.5% raw material stocks, 12.5% REITs, 25% other US stocks, 5% cash, and 0% bonds.

Historically, these asset classes have been uncorrelated. The ups and downs of the different classes should more or less cancel each other out and it should be easy to sleep at night. Well, that has not been so recently - at least in the short term. If the China stock market goes down suddenly, US markets do also. Gold and oil seem to be chained together.

So, I'm confused. It could be that the mysterious actions of hedge funds have made previously uncorrelated asset classes more correlated. If so, an asset allocation approach to investing will not work very well. On the other hand, it could be that asset classes that appear to be correlated on daily or weekly basis are much less correlated on a longer term basis.