The observation that large changes in asset prices occur in the absence of changes in fundamentals has led to many laboratory market experiments that have established the robustness of 'bubbles' in numerous controlled settings and stimulated the search for underlying causes. We report on 12 experiments with a single dividend at the end of 15 trading periods. Previous experiments and analysis have shown that the tendency to follow the trend (momentum) is an important factor in the formation of bubbles and have suggested that the magnitude of liquidity, or the available cash per share of the asset, may be important in terms of trading prices and price changes. In these experiments we find strong evidence for the role of liquidity and quantify its effects. In particular, there is a 0.8 correlation between price and liquidity during some of the early periods of the experiment. Each dollar of available cash raises the price of the asset by about 29 cents during the initial periods. Another factor in the formation and magnitude of bubbles is the traders' inadequate information about the full bid-ask list (or the ''specialist's book'') during trading. Access to complete information may temper the enthusiasm about rising prices as it becomes evident that bids are associated with progressively thinner volume and fewer traders, and thereby reduce the effects of momentum. The complex interaction between liquidity, momentum and fundamental value is analyzed statistically and through differential equations modeling.

Co-authors David Porter and Vernon Smith