The previous lesson catalogued the cognitive biases documented in laboratory experiments and large-sample brokerage studies: loss aversion, overconfidence, anchoring, representativeness, mental accounting. Documenting biases is necessary but not sufficient for asset pricing. The critical question is: under what conditions do investor-level biases survive in equilibrium prices, and which anomalies do they explain?
This note covers three formal models that translate specific biases into asset return predictions — the DHS model (Daniel, Hirshleifer, and Subrahmanyam 1998), the BSV model (Barberis, Shleifer, and Vishny 1998), and the Hong-Stein (1999) gradual-diffusion model. Each makes different structural assumptions about which bias dominates and how information propagates. Together they explain why stock prices underreact in the short run (generating momentum) and overreact in the long run (generating reversals) — the two most prominent patterns in the anomaly literature.
The lesson also covers the broader asset pricing implications: investor sentiment as a priced risk factor, limits of arbitrage as the prerequisite for persistent behavioural pricing, and the demand-curve evidence that identifies moments when rational arbitrage is overwhelmed.
Learning objectives. By the end of this note you should be able to:
- state the core mechanism of the DHS model and derive the prediction that overconfidence generates short-run momentum and long-run reversal;
- explain the BSV model's two-regime Markov switching structure and map it to the conservatism/representativeness bias;
- describe the Hong-Stein model's two-agent structure and explain how gradual information diffusion produces momentum;
- explain why rational limits of arbitrage are a necessary condition for behavioural pricing to survive in equilibrium;
- describe the demand-curve evidence (S&P index additions/deletions) and the Baker-Wurgler sentiment index.
In this lesson
-
- Prerequisites: limits of arbitrage and behavioural pricing
-
- The DHS model: overconfidence and biased self-attribution
-
- The BSV model: conservatism and representativeness
-
- The Hong-Stein model: gradual information diffusion
-
- Investor sentiment and asset pricing
-
- Connecting the models: a unified view
Before you start
The full lesson, with worked examples and exercise solutions, is for members.
Read the full lessonNew to SelfMadeQuant.com? See membership options or browse the Portfolio Theory and Quantitative Investment syllabus.