One of the interesting anomalies of asset trading in secondary markets is that it is not usually possible to observe actual market clearing prices. Rather, we can only observe transactions as bid and ask price? and infer that the market equilibrium price lies between them. Professor Smidt analyzes continuous transactions data for individual NYSE listed securities during 1977. From it he deduces (1) the behavior of transaction prices, (2) the average size of bid-ask spreads, and (3) the movement of market prices. Given that neither the bid-ask spread nor the actual equilibrium price is observable from the available data, this is an ambitious undertaking. The problem is further complicated by the fact that the transactions data contain three distinct types of trading activity: matching trades at the opening or reopening of the auction market, auction trades (the bulk of activity by number of transactions), and block trades (the second largest activity by dollar volume).

Thomas E. Copeland, A Probability Model of Asset Trading, The Journal of Financial and Quantitative Analysis, Vol. 12, No. 4, Proceedings of the 1977 Western Finance Association Meeting (Nov., 1977), pp. 563-578

The performance of Value Line Investment Survey recommendations made between 1965 and 1978 is evaluated by applying a future benchmark technique. The future benchmark technique avoids selection bias problems associated with using historic benchmarks as well as known difficulties of using Capital Asset Pricing Model benchmarks. Potential problems (implicit in the technique) are discussed and resolved within the conduct of the experiment. Results indicate statistically significant abnormal performance when future benchmarks are computed using a market model.

The authors develop a model of market efficiency assuming private information is partially revealed to uninformed traders via the behavior of those who are informed. This partial revelation of information model is tested in fourteen computerized double auction laboratory markets. It explains the market value and allocation of purchased information, and asset allocations, better than either a fully revealing information model (FRE strong-form efficiency) or a nonrevealing expectations model; but it takes second place to FRE in explaining asset prices. The authors conjecture that refined versions of partial revelation of information may provide insight into "technical analysis" and minibubbles in securities markets.