“Naive Diversification Strategies in Defined Contribution Saving Plans” Benartzi & Thaler (1999)

by vonNudge

Diversification is the cornerstone of good investing. While previous papers (1, 2) have shown that investors tend to insufficiently diversify their portfolios, this paper intriguingly shows the opposite: in an experimental setup participants’ modal choices are to use maximum naïve diversification (splitting their portfolio equally between asset classes). Whereas, we might term sophisticated diversification the selection of weights to put an investor’s portfolio on the efficient frontier. Although insufficient diversification is without a doubt the bigger real-world sin, I think this paper helps to show the causal factors behind this error.

This paper found that participants’ allocation to stocks could be predictably influenced by changing the composition of the funds on offer to them. Give them a stock fund and a bond fund and they’ll usually end up at 50% stocks. Give them a stock fund and a balanced fund (which itself contains 50% stocks) and they’ll often end up at 75% stocks. This diversification is naïve, rather than reflecting innate risk preferences.

The most intriguing aspect is the experimental condition where the authors found that naïve diversification is much reduced. In this condition, participants have to invest all of their funds into one of five funds, which vary in constant increments from all bonds to all stocks. Naive diversification would predict that the 50:50 fund should be the modal choice. In fact, the 100% stock portfolio was chosen by 51% of participants, and the average portfolio was 75% stocks/25% bonds!

I believe that this last condition is most like real-world investing, and might explain why insufficient diversification is so commonly observed. Most people invest small amounts at regular intervals. Therefore, they only consider how best to invest this new bit of savings, rather than considering their entire portfolio (which would most likely lead to more diversification, such as in the naïve diversification experiments). My bet is that people who invest a lump-sum diversify to a much greater extent than those who invest periodically. Diversification is a much more intuitive concept when all the relevant parts can be thought about together. And in fact diversification can sometimes lead to a bias in simultaneous consumer choice.