This paper studies the impact of the disposition effect on the return predictability before and during the global financial crisis periods. It uses data of 104 French financial institutions’ stocks over the period January 2001-December 2009. To construct behavioral factor, the capital gain-related proxy of Grinblatt and Han (2005) is estimated. The study of the cross-sectional determinants of this factor indicates a significantly positive impact of past return at short and intermediate horizons on the unrealized capital gains. The empirical analysis of the link between disposition effect and excepted return, after controlling for market anomalies, particularly size, momentum and reversals, shows a significantly positive cross-sectional relation between a stock’s unrealized capital gains and its expected returns during the tranquil period. Moreover, momentum and reversal effects, mainly documented in French market, disappear when the disposition effect is controlled for. Further, during global financial crisis period, the presence of disposition investors does not influence the expected return.

This paper deals with modeling brokered segments of a ‘fast’ security market, meaning a market characterized by three interrelated features:

• A large number of participants who place orders within a narrow time span;
• Very short-lived motives for trade (frequent changes of desired positions); and
• Uncertain terms of individual trade for any given participant due to a high concentration in time of other participants’ actions.

The three named elements are typical for electronically brokered trading in many upperend stocks, bonds and major currencies. (An electronic FX broker for a frequently traded currency pair is among the best examples of a fast market.) Electronic brokerage systems with fully or partially1 observable books are now implemented by most exchanges around the world. A market for any top-tier security traded there would meet our criteria.2 In addition, there exists a special category of ‘high-frequency traders’ who exploit short-lived arbitrage opportunities with the help of computerized trading, making the market for the involved instrument even faster. It is known that all but the biggest high-frequency traders operate through brokers, i.e., contribute to increased trade frequency in order-driven markets. Therefore, formalizing the physiology of a continuous double auction under fast market conditions without resorting to the zero-intelligence simplification should be of interest for practitioners. No less pertinent is this task for understanding asset price deviations from fundamentals, the phenomenon actively researched by financial economics, e.g., in the context of exchange rate (or risk-free bond yield) disconnect conundrum. Although, at present, the majority of transactions in this asset class happen in fast order-driven markets, the latter have been largely left unexplored by theoretical literature, at least that part of it which draws on individual rationality. Consequently, existing theories of continuous double auctions say very little about the properties of price discovery resulting from high frequency of orders and trades.

Financial Risk Management Journal, Discounted Cash Flow, DCF, Net Present Value, NPV, DCF Techniques, Monte-Carlo Simulation Method, Cyprus Telecommunications Authority, CYTA, Information Technology, IT, Methodological Issues, Cash Flow, Weighted Average Cost of Capital, WACC, Decision Making.