By Robert Metcalfe
It’s well known to new and old investors alike that stocks yield much higher returns than bonds and other riskless securities. In fact, in the last 100 years US equities have seen an 8% average annual real return, compared to only a 1% return for more riskless securities. This gap is called the equity premium puzzle – why are equities valued so much higher than securities?
The puzzle can’t seem to be explained by systemic economic factors alone, so that’s where a behavioral approach comes in handy. One behavioral theory by Shlomo Benartzi and Richard Thaler attributes the equity premium puzzle to what’s known as myopic loss aversion (MLA) – the idea that loss-averse investors (as all investors are) take too short-term a view of their investments, leading them to react overly negatively to short-term losses. Since investors are especially worried about losses, they need to know that equities have high return potential to justify investing in them – hence the equity premium.
This is an intensely behavioral phenomenon, dealing with how individual investors react to the returns they’re seeing right before their eyes. But it’s also difficult to measure this idea, and all past efforts to do so have come from lab experiments. This is a problem: if you want to measure investment behavior, you have to see how investors act in real life, not in a lab.
Here’s where our experiment comes in. Partnering with the technology trading firm Normann, we used a beta test of a trading platform to see how access to price information affected professional investors’ willingness to take on risk, and their ultimate profits. We broke the investors up into two groups: a frequent information group with access to second-by-second changes in price and portfolio value, and an infrequent information group that could only see these changes every four hours. Both groups traded the same mutual fund-like instrument, so both groups received the same price realization, despite different frequencies of being able to view that realization. Importantly, the investors did not know that they were being experimented on and the stakes of the trading were high.
The results are striking, and align well with the MLA theory. Over the 14-day experiment, the MLA effect took time to arise, but by the last two days of the experiment the results were clear. The key finding is that investors with infrequent price information invest around 33% more in risky assets than investors with frequent information do. What’s more, those with infrequent information earned 53% higher profits than those with frequent information did. And these results are consistent across all levels of profit: an investor in the 90th percentile of earnings in the infrequent information group saw 80% higher profits than an investor in the 90th percentile in the frequent information group; a 25th percentile investor in the infrequent information group returned a profit, whereas a 25th percentile investor in the frequent information group returned a loss.
What does this mean about MLA and the equity price premium? For starters, it’s the first natural field experimental evidence to show that MLA exists for professional traders, so that makes the MLA explanation for the equity price premium a lot more empirically justifiable. Investors taking a short-term view will see frequent losses, making them demand that equities have a higher potential return to compensate for these losses.
Further, our study has some important implications for investors and financial markets. In general, people like having more access to information. In the case of investing, however, this is actually detrimental, since our research shows that traders are worse off when they receive more frequent information. Our investment outcomes are in fact better with less information, in a situation where a person would generally expect them to be worse. Maybe it’s true that what you don’t know can’t hurt you. This is the core irony of the MLA findings: in financial markets, investors don’t like to lose, but by trying to avoid every small loss they end up preventing themselves from winning more in the long run.
Lastly, since social security, investments in public services, capital cost estimates, and more are all affected by the equity premium puzzle, the finding that MLA partially explains that puzzle can help us make better public policy. And with this knowledge in hand, perhaps those in financial markets should adjust their communications strategies to deal with the dilemma that people’s desire for information conflicts with their desire to earn higher returns. Though financial markets and public policy are sweeping concepts involving millions of people, our MLA research is a useful reminder that individual behavior – behavior as simple as disliking seeing losses – undergirds them both.