Experience and Confidence in an Internet-Based Asset Market Experiment

Abstract

The experience effect in asset markets is one that was thought to be settled. As subjects gained experience with the interface and each other, they typically exhibit fewer instances of mispricing and at lower magnitudes. But questions regarding trading experience are not easy to address in the lab with the typical subject pool since the kind of experience one can typically generate in the lab is experience with the experimental environment itself—not with external environments. However, in virtual worlds asset markets are highly evolved, providing a subject pool with skilled and experienced traders that can be accessed via the Internet. This study compares experimental asset markets with participants recruited from virtual world trading groups to experimental markets with participants recruited from the virtual world at large. I further examine trader performance and trading behavior within markets. The findings indicate that asset markets with virtual world participants recruited from trading groups are more prone to exhibit bubbles than are markets with virtual world participants recruited at large. Within condition, experienced traders are less likely to follow fundamentals and more likely to engage in strategies that result in loss of earnings. Excess confidence is rejected as an explanation for this pattern, as confidence is found to be related to higher earnings and fundamental value trading strategies.