In this research project, the PIs investigate the possibility that social rationality explains the emergence of one type of bubble in competitive asset markets: a bubble referred to as a "credit market bubble." The bubble is defined as a situation where (i) the debt is priced above its commonly known intrinsic value and (ii) the debt is rolled over even though each creditor should cash in because everyone knows that the debtor will never be able to repay. Building on evidence from behavioral game theory, the PIs conjecture that such credit market bubbles emerge whenever the debtor's payment ability, although never sufficient, grows over time. Pilot experimental data confirm the emergence of bubbles in this setting. The researchers will conduct experiments to further examine the robustness of bubbles in this environment and test the hypothesis that norms are driving the observed behavior.
In terms of broader impacts, this research will provide a better understanding of price bubbles. Credit bubbles and accompanying asset price run-ups re-occur with alarming frequency in the real world. This research suggests that tension between individual and social rationality is the root cause for their existence. It will lead to a better understanding of this ubiquitous phenomenon in modern capitalist society, and inspire novel and effective government policy and regulation to minimize their negative effects.
The major goal of the experimental projects within this proposal was to provide a rigorous test to existing theories of asset pricing (e.g., Lucas Pricing Model) or develop and test new theories (Credit Bubbles). In all of our studies we study not only prices (field data would suffice for an asset-pricing-only study) but also individual choices, and as such we can address the issue of social optimality of market allocations. During the award period we produced three publications and a working paper. The first publication, in Management Science, studies prices and allocations in markets where individual investors cannot trade on their own behalf but have to delegate investment to portfolio managers. The managers are compensated based on the fund flow, and as such compete for flow. In the theoretical model we demonstrate that competition ensures provision of portfolios that maximize investors' welfare. The experimental design translates this into sharp predictions about managers' portfolio composition (each has to provide security that bets on only one outcome). The data confirm the prediction as long as fund flows remained uniformly distributed across managers. Unlike in the theory, investors chased past realized returns. As a result, funds became concentrated with managers who had "lucky draws" and asset pricing predictions no longer held. Thus, fund concentration appears to destroy optimality of market allocations. The second publication, in the Journal of Political Economy, addresses theoretically and experimentally the issue of the impact of asymmetric reasoning on asset prices and portfolio choices. If investors had asymmetric information, prices would be predicted to aggregate all information. In contrast, under asymmetric reasoning prices should not reflect all types of reasoning. Some agents who observe prices that cannot be reconciled with their reasoning switch from perceiving the environment as risky to perceiving it as ambiguous. If those agents are averse to ambiguity, they then become price insensitive and choose ambiguity-neutral portfolio for a range of prices. As a result they no longer affect prices. In the experimental investigation we find that indeed mispricing decreased with the number of price-sensitive agents, and price-insensitive agents indeed traded to more positions that hedged their exposure to ambiguity. The third publication is a test of the Lucas pricing model in the laboratory. Key design features allow us to emulate the infinite horizon setting in the laboratory and to incentivize the participants to smooth earnings across trading periods. As predicted, prices moved with the aggregate dividend (high dividend implied high prices). Trading improved the earnings of the participants in comparison to no-trade earnings. However, prices were excessively volatile. Traditional asset pricing tests were corrupted as they rely on returns, and returns are computed from excessively volatile prices. In addition, individual investors displayed substantial differences in their choices, to the extent that the average trades and prices did not reflect the experience of any one individual. The last project studies the emergence of bubbles in a credit market setting where what is best for the society conflicts with what is best for the individual. A credit bubble would emerge if a sufficient number of agents act socially rational for at least one period of the market game. We robustly observe bubbles, even after replication of the game with the same cohort. For comparability, we also study environments where social and individual rationality are aligned: no bubbles emerged in those environments. Another goal of our proposal was to present to and convince the greater finance audience in the value of the scientific method more generally, and of experimentation in particular. We have attempted to present our work in a format more familiar to the audience, and in addition to statistical tests that are more appropriate for our experimental data, we have applied widely used field testing methods to our data, which allowed us to demonstrate potential misguided or false inference. We have designed and taught three finance classes based entirely on experimentation— at the Universities of Utah and Melbourne, and at Caltech. The classes are based on eight purposely-designed trading sessions. We have adapted two of the above listed experiments for the classroom—the Lucas and the Credit market experiments. In addition, we have organized a prediction market with the participants of the Utah Winter Finance Conference with the sole purpose of familiarizing the (academic) audience with experimental methods as applied to financial markets. The prediction market concerned the best paper award, voted on by attendees; the security with the highest price was indeed the one corresponding to the winning paper.
