Originally Posted by RockOn
I agree with that.
Using a highly diversified and uncorrelated AA tends to reduce risk and reduce return verses a highy correlated portfolio. The idea that a highly diversified and uncorrelated AA should increase return, as I have read more than once, is not a valid premise. (even though it might have happened in the past history)
Sorry, another long post coming . . .
The restatement that you made is not what I intended to communicate. Here's what I believe:
-1) In general, each asset class has a return that is proportional to its risk. (i.e. Highly risky asset classes provide higher returns, if we look at a very long time period.)
-2) Many people are trying to pick the next "hot" asset, and information that might affect the future performance of these assets and the means to analyze this information are easily available to everyone. As a result, it is very unlikely that an individual can consistently pick the next asset class that will significantly outperform the others.
Now, here are the conclusions that follow from this:
To get maximum return over time, an investor needs to invest in risky (i.e. volatile) assets (by rule 1 above). By rule 2, there's no reliable way to decide which particular asset class will do best in the near future. But, I know that, historically, some of these highly volatile (that are, by rule one, also the ones that produce the best payoff over time) increase and decrease in value in ways that are unrelated to, or even opposite of, the ways other asset classes perform. So, I divide up my money and I put a little of it in each of these asset classes.
Result: If I follow this, I've invested all my money in the highly volatile asset classes (the only ones that can produce high returns), but my overall portfolio will have low volatility.
High return, low volatility. That's what we're after. And I didn't have to try to guess winners and losers.
I think it's obvious how this approach reduces volatility compared to plunking all your dough into whatever asset class I "predict" will do best. Does this approach reduce returns? Well, if you accept rule 1 than this approach will certainly produce better results over time than putting all the money on a single low-volatility (low return) pot (e.g. money-market funds). If an investor rejects rule two, then he's likely believe he can pick the winners (consistently, over many years), and will want to concentrate his money on his bets. This investor is likely to believe that spreading assets among many high-risk, high return classes will lower overall returns.
I think he's wrong. The research indicates he's wrong. The firsthand experience of many investors indicates he's wrong. And I'm fairly sure I'll end up supporting many of these people through my taxes, which makes me grumpy.