Pricing Compound Ambiguity in Long-Run Asset Markets: Experimental Evidence

Abstract

This paper studies how compound ambiguity is priced in long-run asset markets. We implement a laboratory continuous double auction in which a long-lived asset pays dividends through a two-stage process that determines whether a dividend is paid and, conditional on payment, the dividend amount. By varying the information disclosed about each stage, we compare a fully risky benchmark with treatments that introduce ambiguity about payment timing, payoff size, or both. All four treatments generate bubble-like overpricing, but compound ambiguity does not produce a large uniform shift in average market prices relative to risk. Instead, treatment effects are concentrated on specific margins. Timing ambiguity robustly increases ask-side price dispersion, while payoff ambiguity robustly increases bid-side normalized mispricing, bid volume, and maker orders. The interaction between the two ambiguity dimensions is not significant, providing little evidence of an additional compound-ambiguity effect. Dynamic and individual-level analyses further show that recent prices are more important than dividend history in predicting mispricing, and that risk and ambiguity attitudes matter only in treatment- and side-specific cases. Overall, the results suggest that ambiguity may have limited aggregate effects on average price levels, but that its different dimensions remain behaviorally relevant in long-run trading environments.

Keywords: Compound ambiguity; Experimental asset markets; Long-run markets; Bubble formation

JEL Classification: C92; D81; D53; G40