Working Paper
Information, Asset Price Volatility, and Liquidity (JMP)
(with Daniel Friedman and Grace Gu)
Management Science
Abstract: How does public information on asset payoffs affect private information acquisition and aggregated market information? In turn, how do private information and market information affect asset price volatility and liquidity? What are the roles of information costs and market competitivity? To investigate such questions, we derive theoretical predictions from a stylized bond market model where investors receive public information on default probability and then can purchase costly private information on the repayment recovery rate should a default occur. We test those implications in the laboratory with human subjects as investors. They trade bonds in markets with three different levels of competitivity. We also control and vary exogenously the information cost schedules and default probabilities. The laboratory data support most of the theoretical predictions, e.g., traders purchase more private information and price is more efficient when information is less expensive, when the market is less competitive, and when the public information indicates more ominous conditions. In turn, more private information purchases and less competitive markets increase the bond’s overall price volatility and liquidity. However, ominous public information can increase volatility and decrease liquidity. This paper highlights the roles of public information and market competitivity, and their interactions with private information. The laboratory data analysis emphasizes the “good” and “bad” components of price volatility, which are difficult to disentangle in field data. Our results also suggest policy implications, e.g., tradeoffs between asset price volatility and liquidity.
Behavioral Insights into Bidding and Asking: Experiments with BDM, Call Market and Continuous Double Auctions
A draft is available upon request.
Built on Friedman, Gu and Zheng (2024), this paper investigates the interplay of neoclassical and behavioral factors in shaping bid-ask spreads under three different market formats in the laboratory settings: Becker-DeGroot-Marschak (BDM) individual choice, Call Market, and Continuous Double Auction (CDA). Through the analysis of both individual spreads and market spreads, we reveal how factors such as information precision, default risk, market size and adverse selection impact spread settings behavior. Our theoretical model predicts that rational, risk-neutral traders would minimize spreads in the BDM model and face lower spreads in CDA than in the Call Market, as the former’s continuous trading format supports improved information aggregation, reducing adverse selection. However, the laboratory results demonstrate a great departures from the rational benchmarks, suggesting that trader’s behavior often diverges from traditional neoclassical expectations. We found that other behavioral factors such as risk aversion also have crucial impact on asset pricing behaviors. This study underscores the importance of integrating behavioral insights into financial models.
Market Visualizations
(with Daniel Friedman and Brett Williams)
Status: Minor Revison at Experimental Economics
Abstract: We introduce a pure point and click user interface for n human traders in a two-good general equilibrium economy. Compared to standard continuous double auction with text entry, we conjecture that our video-game-like Visual Market will converge much faster to a neighborhood of the Pareto set, and will generate final allocations that are at least as efficient. By varying the visualization of the orderbook, the text entry capabilities and visualization of the trade history, we experimentally test such a hypothesis. We find compelling results for both experienced and inexperienced traders, with a positive monotonic relationship between price convergence, allocation convergence and efficiency and market visualization.
Motivated Inference of Ambiguous Information
(with Zhaoqi Wang)
Status: Under Review at
Journal of Economic Behavior and Organization
Abstract: We study the impact of motivated beliefs, the phenomenon that people believe what they want to believe, on the assessment of the precision of ambiguous information. Motivated beliefs are induced so that subjects have polarized preferences regarding the dividend of a financial asset. Our main preliminary results suggest that (i) people tend to believe information that is consistent with their motivation to be more precise, regardless of the ambiguity-attitudes, (ii) when receiving multiple information, motivated beliefs have a significant impact on the assessment of the information precision, and (iii) people tend to report a larger precision if the information contradict each other.
Vivian J. Zheng 