Any fledgling investors would know that a good portfolio should consist of a well diversified set of investments. However, not many are very clear about just what level of diversification is sufficient to reduce the risk inherent in their portfolios. In fact, many people believe that more is always better. If you are one of these people, you may be surprised to hear that over-diversification actually be harmful for your returns. Here, we discuss a few academic literatures on the effectiveness of diversification, as well as how this translates to real life investing practice.
What Scholars Say on Diversification
Historically, both the practitioners and the academics have long believed that it's sufficient to hold only a small number of stocks to be diversified. For instance, Evans and Archer (1968) calculated that 90% of diversification benefit came from just 16 stocks, and 95% of benefit could be captured by just 30 stocks. Similarly, another study by Campbell, Lettau, Malkiel and Xu (2001) argued that an investor needs about 50 stocks to sufficiently diversify their portfolios, the increase from 15-30 to 50 being attributed to the increased volatility of individual stocks. Many other studies like this, this and this confirm these views.
However, a new study in 2007 by Domian, Louton and Racine claimed that even 100 stocks are not enough to truly create a well-diversified portfolio. So what gives?
For an average retail investor like ourselves, the first two studies are the more relevant one. They define risk as the volatility of your portfolio's return: to make your portfolio's returns stable on a daily basis, all you really need is somewhere around 15 to 30 stocks. Adding any more than that has very little impact on your investment's volatility. For instance, once you have 20 stocks in your portfolio, every stock that you add up to 50 reduces the volatility of your portfolio by only about 0.04%.
The third study, on the other hand, defines risk as the probability of underperforming. It argues that you would need a lot more than 100 stocks in your portfolio in order to have less than 1% chance of making less than the risk-free treasury rate. However, if you delve into their data, you can soon see that its results are actually not that dissimilar to the first two's. For instance, once you have 20 stocks in your portfolio, every stock that you add up to 50 reduces the probability of underperforming by only about 0.5%. The benefit of diversification diminishes very rapidly after you've reached 15 to 30 stocks.
Diversification in Practice
When you are actually investing your money, it's very easy to get scared. If you only have a few investments and one of them declines meaningfully, you can easily panic and make a wrong move. In this sense, diversification to reduce both the volatility of your portfolio and the probability of not making much money is a great thing.
However, an investor should always buy only things he understands; otherwise it's very easy to overpay. For example, if you had no idea what a bowl of noodle is supposed to cost, how would you know a hawker that charges you S$50 for a bowl is overpriced? We spend all our lives building an inherent sense of what things are worth and use that as a guide when going through our daily lives. Buying stocks is the same thing. If you don't put in the work to really understand what a business does and what its stocks are worth, your "investment" would be equivalent to gambling, if not insanity.
Diversification's Benefits vs Costs
This is the main reason why investors should not over-diversify. While each incremental stock after 20 can reduce your risk by only about 0.04% or 0.5%, it can burden you with 5-10% more time and effort spent on evaluating its value. Worse, increasing your portfolio from 20 to 50 stocks can only reduce your risk by 1% to 16%, while you would have to increase your effort by at least 150%. If you are professional investor that spends all of his time on studying companies, perhaps you can afford to do this. Assuming you can't easily afford such time, however, you would be stretching yourself very thin.
Therefore, when you are building your investment portfolio, you should do your best to study different stocks and companies, with the goal of handpicking 15 to 30 that you really like. Any more than that is likely to dilute the quality of your stocks as well as your conviction on them, without garnering proportional benefit in terms of derisking your portfolio. Afterall, wouldn't you rather buy 15-30 things you know very well at a bargain than stretch yourself thin and own 50-100 things that you barely understand and can potentially overpay?