Embedded Files
Job Market Paper
Deceptive Visualizations in Financial Markets SSRN Revise & Resubmit at Management Science
Abstract:
Digital charts have become ubiquitous in financial markets. While intended to make price paths easier to read on screen, these charts often feature a default auto-scaling function that imposes varying visual scales across stocks. I provide experimental evidence that auto-scaled charts can mislead investors’ risk perceptions. Several interventions show this influence is difficult to eliminate. Even explicit, educational communication of numerical volatility restores the overall accuracy of volatility assessments without fully removing the distortion induced by visual scale. Moreover, I show that stocks’ exposure to visual scaling helps explain the weakening of the positive risk-return relationship, which was prominent before the 1980s but has since become less evident.
In my experiment, participants using fixed-scale charts accurately perceive risk consistent with historical volatility. In contrast, auto-scaled charts distorted this perception, as participants did not judge stocks with higher standard deviation to be riskier.
In stock return data, the premium associated with higher volatility diminished after the 1980s (sigma). However, once accounting for visual scaling exposure (VSE), high-volatility stocks consistently carried a positive premium across all sample periods(sigma_within_VSE). A strategy exploiting this visual deception (visual) began earning a premium after 1980, but this effect has weakened since around 2005, likely due to investor learning and the rise of algorithmic trading.
Working Papers
Luck, Flow and Value Destruction in Active Asset Management
(with Dunhong Jin and Juan Yao)
Abstract
We develop a model in which luck has two distinct components: an exogenous shock and an endogenous exposure chosen by the manager. We show that the majority of variation in fund realized alpha can be explained by the luck shock, which misleads investors’ capital allocation, and that funds endogenously choose excessive luck exposure, leading to overfunding and value destruction in equilibrium. Incorporating
luck provides a unified micro-foundation for understanding fund performance, capital allocation dynamics, the limits of skill measurement, and the design of performance evaluation metrics. We further discuss potential solutions for better fund governance and performance evaluation under the influence of luck.
Mutual funds often grow too large beyond their capacity and destroy investor returns. But this isn't caused by 'evil' managers or 'foolish' investors, it is the natural equilibrium of a system driven by management fee incentives, luck, and performance chasing investors.
Discrete Performance Evaluation and Reference-Dependent Preferences: Evidence from the Star Rating System
(with Stephen Satchell and Juan Yao)
Abstract
Discrete performance evaluation is common in business and organizations. We examine how it affects evaluatees’ risk appetite among active mutual fund managers under star rating system. We show that rating thresholds serve as reference points that induce asymmetric risk-shifting behavior: funds just below an upgrade threshold increase risk, while those just above a downgrade threshold reduce risk. The causal relation is further supported by Morningstar’s 2002 methodology reform, which serves as a quasi-natural experiment. Importantly, risk-shifting funds do not exhibit higher abnormal returns or a greater likelihood of future rating improvements, ruling out strategic rating management.
Discrete rating systems lead to unintended risk-shifting behavior. Funds nearing a rating upgrade (purple) increase portfolio risk, while those close to a rating downgrade (blue) become risk-averse.
Page updated
Google Sites
Report abuse