Here's a statement of the obvious: The opinions expressed here are those of the participants, not those of the Mutual Fund Observer. We cannot vouch for the accuracy or appropriateness of any of it, though we do encourage civility and good humor.
(Quicky summary: Avoid having too many investments, foreign, and holding commodities; forget market cap .... Consider adding individual stocks, e.g. RE, also cash.)
The article failed to point out that long term govt bonds did ok in 2008 and permanent portfolio (stocks gold long term bonds and I think swiss francs did ok) Basically you want to make sure the equities you invest in have good balance sheets and if you have long term govt bonds plus an inflation hedge at least some of your diversification will work
We can start with the fact that stock volatility has increased significantly in the three decades since the graph in the article was drawn. That means that even though market volatility has not increased, you need more stocks now than in the past to achieve the same level of volatility in your own portfolio. Below is a graph from a 2000 EfficientFrontier.com column illustrating a 30 year shift: There is the question of what diversification you're really getting with your 20 stocks. Issue diversification is easy. You want to diversify away the risk that you picked the one bad apple in the barrel, say, the one bank that got caught while all the others did well. (I don't know how bank stocks have been performing, but it's a simple example of how a company can have problems that are specific to that one company.) So you might buy multiple bank stocks.
That works to eliminate issue risk, but leaves you with the risk of having picked the wrong sector, even if you bought 20 different stocks. So you buy a single fund that invests in 20-30 stocks. Even if it is well diversified over sectors, that manager is using particular heuristics (say, looking for high dividend stocks of companies with low debt). This is better than focusing on a single sector, but you're still at risk that the strategy will fail for some extended period of time.
These sector and strategy risks reflect the fact that the higher the correlation between the stocks in your portfolio, the more stocks you need to diversify away some of the risk. That should be obvious. So there's not going to be a single optimal portfolio size. It depends on how you select the securities. ISTM that whatever thinking goes into stock selection may unintentionally tend to identify more highly correlated securities than would random selection.
Then there's matter that averages aren't always that useful (see: one foot in boiling water the other in ice). The risk of losing big doesn't decline as quickly as the volatility of an "average' portfolio as the number of stocks increases. This is another point made in the EfficentFrontier article. (I disagree with its conclusion that the only alternative is to buy the whole market, since IMHO that's rather reckless extrapolation.)
Did you not know you were supposed to diversify only into sectors that went up? This way, depending on the weather and time of one's calendar, the press can issue articles telling you how stupid you are for having diversified, or not diversified, while also changing the definition of diversification at will.
Comments
We can start with the fact that stock volatility has increased significantly in the three decades since the graph in the article was drawn. That means that even though market volatility has not increased, you need more stocks now than in the past to achieve the same level of volatility in your own portfolio. Below is a graph from a 2000 EfficientFrontier.com column illustrating a 30 year shift:
There is the question of what diversification you're really getting with your 20 stocks. Issue diversification is easy. You want to diversify away the risk that you picked the one bad apple in the barrel, say, the one bank that got caught while all the others did well. (I don't know how bank stocks have been performing, but it's a simple example of how a company can have problems that are specific to that one company.) So you might buy multiple bank stocks.
That works to eliminate issue risk, but leaves you with the risk of having picked the wrong sector, even if you bought 20 different stocks. So you buy a single fund that invests in 20-30 stocks. Even if it is well diversified over sectors, that manager is using particular heuristics (say, looking for high dividend stocks of companies with low debt). This is better than focusing on a single sector, but you're still at risk that the strategy will fail for some extended period of time.
These sector and strategy risks reflect the fact that the higher the correlation between the stocks in your portfolio, the more stocks you need to diversify away some of the risk. That should be obvious. So there's not going to be a single optimal portfolio size. It depends on how you select the securities. ISTM that whatever thinking goes into stock selection may unintentionally tend to identify more highly correlated securities than would random selection.
Then there's matter that averages aren't always that useful (see: one foot in boiling water the other in ice). The risk of losing big doesn't decline as quickly as the volatility of an "average' portfolio as the number of stocks increases. This is another point made in the EfficentFrontier article. (I disagree with its conclusion that the only alternative is to buy the whole market, since IMHO that's rather reckless extrapolation.)
Regards,
Ted
Too Many Mutual Funds ?:
http://www.investopedia.com/articles/mutualfund/07/too_many_funds.asp