Both are commonly held beliefs which seemingly contradict each other. Which is it? Do we always choose our favorite food/toy/song or do we opt for less-desirable alternatives so we don’t get bored?
The answer is yes.
When faced with multiple simultaneous options, people diversify their choices, even if they end up with less-desirable alternatives. However, if the same options are presented sequentially over time, people will stick with their preferred choice. Carnegie Mellon University researchers used a Halloween-night experiment to vividly depict this so-called diversification bias.
In the experiment two sets of trick-or-treaters visited two houses. The first set of children had to choose between taking either a single Three Musketeers or a Milky Way bar. The other set of kids were told to ‘choose whichever two candy bars you like’. There were large piles of both candies available to ensure the kids didn’t think it was rude to take two of the same.
More than half of the kids in the first set (the sequential choice condition) picked the identical candy at both houses, presumably the one they preferred. However, nearly every child in the second set selected one of each candy at both houses, demonstrating the diversification bias. Since the candy is typically consumed much later, this result is even more striking. It is end-of-the-evening portfolio that matters, not the portfolio selected at each house.
It’s not just kids and candy that are susceptable to the diversification bias. When employees are simultaneously offered n funds to choose from when signing up for a retirement plan, researchers found they are likely to divide the money evenly among the funds offered. If the fund options are presented over time, employees concentrate their money into a smaller number of funds. The researchers use the term 1/n heuristic to describe this extreme version of the diversification bias.
A plan administrator can greatly influence investments by the number and order of funds that are presented. In a plan with one stock fund and one bond fund, the average allocation would be 50% stocks. However, if another stock fund were added, the stock allocation would be 66%. In fact, the researchers analyzed a large sample of pension plans and found strong evidence to support this prediction.
I have no way of knowing if plan administrators are using this knowledge to influence investments but it certainly would be the ultimate trick-or-treat.