“This is an exciting topic for me, because one of the problems in asset pricing research is that you never really know what drives the results. You have two camps: the neoclassical one, that says everything is related to risk, and the behavioral one – of which I am a proponent – which says it might be partly related to risk, but it could also be linked to human behavior. These two camps are very difficult to disentangle.”
“But I think we found a way to do that through a quasi-natural experiment, whereby we used the online transfer market in the FIFA 19 video game.1 Here, the quality or fundamentals of these players are known and fixed. That means there is no fundamental newsflow. So, if you were to compare it to the stock market, prices in the latter move either because of fundamental news or because of whatever craziness of the market. Like Elon Musk tweeting, for instance.”
“In this virtual market, the fundamentals are constant, and you know that every price move stems from the behavior of individuals because, by definition, there is no news. So that's one great characteristic of this market. Also, it's very similar to an equity market in terms of structure and players.”
“First of all, it's huge. Literally billions of transactions take place in this market. It's also not just teenagers who play the online game. We have some insight on the demographics. Sure, the average gamer is a bit younger than the average equity market participant, but it's not that far off. The majority of players are male and college-educated, which is similar to what we see in equity markets. The trading process is also similar as it has a limit-order book, for example. It's very advanced.”
“What we find in this setting is that the return dynamics in the FIFA market are similar to what we see in equity markets. In equity markets, there are several prominent factors, and we find these exact same factors in the FIFA market. For example, there is a size premium, whereby players with a smaller market value have higher expected returns. Then there is the book-to-market factor. We can calculate a book value or a fundamental value for players and we do see that there is mean reversion towards this figure over time. We also find momentum and mean reversion patterns. We see reversals in the returns of players that did well in the previous week. But if we look back further, we can also see patterns of momentum.”
Significant portions of equity price movements are mainly driven by behavior, not risk
“So, these are all very similar characteristics to equity markets. And now we know that the FIFA market is purely driven by behavior, not fundamentals. So, if we can compare it to an equity market, then it could imply that significant portions of equity price movements are mainly driven by behavior, not risk.”
“These are typically very stark examples, or say deviations from efficiency. In our paper, we point towards an 800% mispricing.2 Of course, this is an extreme anecdote. But if these things can happen, then, I think it also says something about everyday market behavior. If there are individual instances with such huge mispricing, then it cannot be the case that there is no mispricing elsewhere.”
“Again, it remains very difficult to disentangle mispricing from rational pricing. These anecdotes are useful because they represent examples where we are sure that it is mispricing. That's why I think they are important, because they are so clearly identifiable.”
“What I see happening a lot is the introduction of machine learning into asset pricing, especially with respect to factors. There are hundreds of factors that have been identified, some better than others, I guess. But of course, that does trigger the question: is this really true? Maybe we're picking up the same things. Is there time variation or country variation in factors? So, I think an exciting development is that machine learning techniques allow you to better pinpoint which factors are important, at which point in time, and whether they are picking up the same thing or not. I think it's a great development.”
Machine learning techniques allow you to better pinpoint which factors are important
“Also, behavioral finance proponents were viewed as being negative in the past. Maybe there was some truth in that, because the school of thought only pointed out things that went wrong without really explaining them. On the other hand, researchers that looked into individual behavior, typically experiments showing individual choice behavior, also came up with alternative explanations, like prospect theory, for example. Another point of criticism on behavioral finance was that there is a huge number of biases. And whenever you find something that goes wrong in equity markets, you can just pick a bias and connect it to the issue. So, there's always a bias that fits your anomaly.”
“What I think is great is that this is improving of late. The profession is making progress and really finding the driving mechanisms of anomalies. To give you an example, more and more papers are coming out that show that prospect theory has an effect on asset pricing on the preferences side. Even more recently, we see that expectation formation is being included into asset pricing, like trend extrapolation. These experimental studies show that when people form expectations about future returns, they just look at the recent past. That’s a very strong human reaction. Researchers are now really incorporating trend extrapolation into asset pricing in mainstream finance literature.”
“So, you are seeing the development of a proper alternative to the efficient market hypothesis. This progress is filtering out this big bag of biases to pinpoint the few that are important to asset pricing in general.”
Read the full interview which also includes commentary on retail investor participation and sustainable investing.
1 Montone, M., and Zwinkels, C.J., April 2021, “Risk, return and sentiment in a virtual asset market“, working paper, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3787339.
2 Van den Assem, M. J., Van Dolder, D., Zwinkels, C.J., and Schauten, M. B. J., October 2020, “Can the market divide and multiply? A case if 807 percent mispricing “, Review of Behavioral Finance.