Information Cascade or Direct Benefit? Peer effects in Financial Decisions
Paper Cited: Bursztyn, L., Ederer, F., Ferman, B., & Yuchtman, N. (2014). Understanding mechanisms underlying peer effects: Evidence from a field experiment on financial decisions. Econometrica, 82(4), 1273-1301.
Link: http://onlinelibrary.wiley.com/doi/10.3982/ECTA11991/abstract
People often make choices similar to choices made by their friends. I buy this t-shirt because my friend has one, we have “what they have”, we use facebook because our friends are using it. The same phenomenon can be observed in the financial industry, where people tend to purchase the same financial assets as what their peers own. Such peer effects is easy to test econometrically, but disentangling the channel through which the “herding” occurs (information cascade or direct benefit) is challenging. This paper uses a cleverly designed randomized experiment to identify the mechanism and provides clear explanation on why the two channels might be at work.
The authors cooperated with a financial brokerage to offer an experimental financial asset to 150 pairs of investors who have identified each other as social contacts. Then the brokerage randomly informed one member of the peer pair, investor 2, of the investment made by the other member of the pair, investor 1 (assignment to the roles of investor 1 and investor 2 was random). To disentangle the effect of investor 1’s possession (direct benefit channel) from the effect of the information conveyed by investor 1’s revealed preference (information cascade), the authors implemented a lottery to determine whether individuals who chose to purchase the asset would actually be allowed to possess it. Among investor 1’s who chose to purchase the asset, they implemented a second, independent randomization to determine whether or not investor 2’s will receive information on investor 1’s decision. Thus, among investor 1’s who chose to purchase the asset, the associated investor 2’s were randomly assigned to one of three conditions:
(A) No information about investor 1’s decision was provided;
(B) Investor 2 received information that investor 1 made a decision to purchase the asset, but was not able to consummate the purchase (information cascade only);
(C) investor 2 received information that investor 1 made a decision to purchase the asset, and was able to consummate the purchase (information cascade + direct benefit);
The authors then compared take up rates by investor 2’s in the above 3 conditions to determine the marginal effect of information cascade and direct benefit separately. They find that in condition A, 42% chose to purchase the asset; in condition B, the take-up rate increased to 71%; finally, in condition C, the rate increased to 93%. Therefore information cascade leads to an increase in take up rate by 29% and direct benefits further increases take up rate by 22%.
To better understand investors’ incentives and the underlying mechanism of the two channels, the authors conducted a follow-up survey of the investors in the study. They find that information cascade effects are greatest when the first (second) investor is financially sophisticated (financially unsophisticated). Regarding the direct benefit channel, many investors reported that the expectation of discussing the asset’s performance with their peers has motivated them to purchase as well.
I find this paper interesting not only because of its smart randomization design (which led to it’s appearance in Econometrica), but also because it is closely related to what we have recently learned (in chapter 16 and 17) in class. We have learned in class that “crowding” may be information-based (people update their beliefs by observing previous people’s choices) or direct-benefit based (people go to a party if many of their friends go because it will be more fun). The two dynamics is also applicable to the “crowding” in financial markets, and this paper, with its lottery design, clearly identifies which channel was at work. An information-based crowding in financial asset purchase is due to the fact that an observation of a friend’s purchase conveys a relatively good signal of the asset’s return, which increases the probability that the observer purchases the asset, relative to making a purchase decision in isolation. A direct-benefit based crowding, on the other hand, may arise for a variety of reasons widely discussed in the finance literature. First, investors may be concerned with their incomes or consumption levels, relative to their peers. Second, a friend’s possession of an asset may enter the observer’s utility function through “joint consumption” of the asset: peers can follow and discuss financial news together, track returns together, etc. (very similar to the partying or using facebook cases we discussed in class). Identifying these reasons, however, remains empirically challenging.