On 16/17 February 1996 Institut d'Anàlisi Econòmica, Barcelona hosted a CEPR workshop on `Market Microstructure'. The workshop formed part of the network on `Finance in Europe: Markets, Instruments and Institutions', funded by the European Commission's Human Capital and Moblity programme. It was organized by Günter Franke (Universität Konstanz), Marco Pagano (Università Bocconi, Milano and CEPR) and Xavier Vives (Institut d'Anàlisi Econòmica, Barcelona and CEPR).
In "Transaction Taxes and Financial Market Equilibrium" Avanidhar Subrahmanyam (UCLA) explores the effects of transaction costs in financial markets with asymmetric information. The analysis suggests that transaction taxes can have a beneficial effect of reducing wasteful rent seeking in the "race" to obtain information early. It is also shown that transaction taxes can reduce (increase) the equilibrium proportion of agents who are short-term (long-term) oriented in their information acquisition decision, and thus can enhance long-term price informativeness and the efficiency of long term corporate investment decision.
James Dow (European University Institute and CEPR) noted that the results on market liquidity and informed agents' welfare were very sensitive to the specification of the model. A welfare analysis should be undertaken only if all agents have a well defined utility function (risk-averse hedgers vs. noise traders). Also, it is not clear why, in this model, short-term traders are harder hit than long-term traders by transaction taxes. Günter Franke remarked that transaction taxes reduce information acquisition but also increase the cost of capital for companies and so reduces the level of investment.
Bruno Biais (Université de Toulouse) presented "An Empirical Analysis of Walrassian Tâtonnement in the Paris Bourse" written with Pierre Hillion (INSEAD) and Chester Spatt (Carnegie Mellon). The paper studies the dynamics of order placements and indicative prices during the pre-opening period (which precedes the opening call auction). It is found that, 30 minutes before the opening, learning takes place, the indicative price is conditional on the expected opening price and converges towards the equilibrium price. Both convergence and order placement accelerate dramatically in the last ten minutes of the pre-opening. Traders behave strategically by delaying their order placement until the latter part of the pre-opening period to minimize the reveling private information. Additional evidence of strategic behaviour is offered: large orders are placed later in the pre-opening period, are more likely to be hidden, are placed at less aggressive price levels and are more likely to be modified during the pre-opening period.
Frank De Jong (Tilburg University) noted that there are similarities between the hypothesis tested in this model and the "unbiased expectation hypothesis" in forward markets and stressed the need for alternative tests. Rony Hamaui (Banca Commerciale Italiana) mentioned that in Italy, banks have to execute clients' contracts at the opening price and if the same rules applies in France, this can affect the volume of orders submitted to the market in the pre-opening period.
Thierry Foucault (Universitat Pompeu Fabra and CEPR) presented "Implicit Collusion On Wide Spreads" written with Bruno Biais and François Salanié (INRA, Toulouse). He examined whether price competition between risk-averse market makers leaves room for implicit collusive behavior. He compared the spread and risk sharing efficiencies that arise in different market structures. It is found that, in general, competitive prices do not arise in equilibrium, and there is a conflict between risk sharing efficiency and the tightness of the spread. This conflict can be mitigated by an appropriate market structure design. The limit order market is the only market structure in which the competitive equilibrium is the unique equilibrium.
Frédéric Palomino (CentER and Institut d'Anàlisi Econòmica) noted that in this model, collusion arises when, for a given type of market orders, a market structure allows market makers to have a larger market power than in the limit-order market. Therefore, different market structure may also influence the type of orders traders submit to the market. Günter Franke wondered if trading among market makers in a second period would affect the results ?
In "Institutional Investors as Monitors: On the Impact of Insider Trading Legislation on Large Shareholder Activism", Ernst Maug (London Business School) examined the role of large institutional investors in monitoring companies. He argued that large shareholders face a conflict in the use of their information. They cannot trade on some information which is subject to insider trading legislation. As a precaution the do not acquire it in the first place in order not to jeopardize their trading strategy. As a result they stay less informed, thereby reducing their effectiveness as monitors. The paper investigates this dilemma and the efficiency implications of insider trading in a simple economy where the large shareholder is a mutual fund. The paper shows that appropriate insider trading legislation can increase efficiency through forcing disclosure of private information, and that the definition of inside information has to be sufficiently broad. It also shows that high market liquidity enhances monitoring activities and that the equilibrium size of the mutual fund is too small to provide optimal monitoring.
Ronald Anderson (Université Catholique de Louvain and CEPR) observed that an implicit major assumption of the model is that "monitoring whatever the state of nature" improves social welfare. As a result, there is a conflict between individual welfare and social welfare. If the mutual fund always monitors, the equilibrium price is strong-form efficient (fully revealing) and, consequently, the private profits of the fund are close to zero. Ronald Anderson argued that, in an alternative setup, trading and value gains could go in the same direction, obtaining the opposite welfare results and the above conflict could disappear.
"Informational Efficiency versus Economic Efficiency in a Common Values Double Auction" was presented by Hyun S. Shin (University of Southampton). Economic efficiency relates to the payoffs of the agents while informational efficiency is concerned with how the prices reflect all available information. It is conventional in applied financial research to consider informational efficiency as a normative criterion in making welfare judgments. However, such judgments must rest on the economic efficiency of the market outcome. The paper shows that, in a common values double auction market, the two notions of efficiency diverge rather sharply. When prices are more informationally efficient, the market outcome is less economically efficient since payoffs are lower for all traders. In addition, a strong positive correlation between price volatility and trading volume is found. In the model, higher price volatility, higher trading volume, and more economically efficient outcome, go together. Finally, the author proposes an explanation of the supposed "excess" volatility of prices which is entirely consistent with rationality.
Sandro Brusco (Institut d'Anàlisi Econòmica and CEPR) observed that all of the results crucially depend upon two features of the model: multidimensional uncertainty and the particular equilibrium notion. Given these assumptions, every equilibrium fails to be informationally efficient and there is a trade-off between economic efficiency and informational efficiency. He suggested considering a dynamic version of the model and the introduction of different trading institutions. Both extensions have potentially important implications for the informational efficiency of the equilibrium price and the trade off between the two notions of efficiency.
Narayan Naik (London Business School) presented "Disclosure Regulation in Dealership Markets: The Case of the London Stock Exchange", written with Anthony Neuberger (London Business School) and S. Viswanathan (London Business School). The paper analyzes the effect of trade disclosure in a dealership market, where the market maker receives additional information and uses it in trading with other market makers. The non-disclosure of details of a particular trade allows the market maker who receives that order to extract profit from the information; a part of this profit is then passed on to the original providers of the order. The paper suggests that immediate disclosure of trade details would convert a dealership market into a standard auction market and would reduce incentives to provide information to the market. Immediate disclosure of all trades would not allow the first dealer, and hence the outside informed investor, to earn any rents for the information brought to the market. Therefore, there is no incentive to provide additional information other than the order size.
Ailsa Röell (London School of Economics, ECARE, Université Libre de Bruxelles and CEPR) noted that the results are quite ambiguous: some traders are better off under prompt disclosure while others are worse off, immediate disclosure reduces spreads for uninformed trades but increases spreads for informed trading. Gunter Franke questioned the result that prices are more informative in a dealership market than in an auction market. He suggested that, with several rounds of trading, the long-run price might be more informative in an auction market. Bruno Biais observed that the fact that market makers discriminate between similar sized orders is consistent with motives that have nothing to do with the information content of the orders.
In "Stock Market Transparency", Ailsa Röell studies the impact of measures that regulate the speed of disclosure of stock exchange trading information. In particular, the paper analyzes the effects of prompt last trade publication. The paper shows that, in the absence of prompt trade publication, uninformed traders face higher transaction costs, and informed traders make more profit. The reason is that the informed trader obtains a better price on the initial deal since the market maker who fills the order gains valuable information not available to his rivals. Conversely, the other market makers, aware that they are competing against a better-informed rival, are forced to protect themselves by widening their spreads. The paper also studies the impact of market transparency on trading costs for orders of different size. It is shown that very transparent markets provide low transaction costs for small traders but may not provide the best transaction costs for large traders. This may explain why very large trades tend to gravitate towards less transparent markets.
József Sákovics (Institut d'Anàlisi Econòmica) presented "Overconfident Speculation with Imperfect Competition", written with Jordi Caballé (Universitat Autònoma de Barcelona). The paper analyses the consequences of the presence of overconfident speculators in a imperfectly competitive financial market. Overconfidence is modeled via erroneous, optimistic beliefs about the precision of the own private information. The paper provides a partial explanation for the excess volatility of asset prices as well as for the trading volume unexplained by the arrival of new information. The analysis confirms that overconfidence increases the sensitivity of speculators' demands to private information and therefore the depth and competitiveness of the market as well as the information content of prices. Moreover, posterior beliefs display higher disparity among speculators, what leads to a higher trading volume. Finally, since overconfidence increases the size of the orders submitted, it does increase the variance of prices although the market is deeper.
Alexander Benos (Hautes Etudes Commerciales, Jouy-en-Josas) welcomed a model capable of explaining the excess volatility of asset prices and excess trading volume. He noted that the presence of overconfident traders may improve social welfare since it increases market depth and the informational efficiency of prices. Finally, he suggested developing the model to analyze static and dynamic competition between rational traders and overconfident speculators.
Peter Zensky (INSEAD) presented "Multi-Dimensional Uncertainty and Herd Behavior in Financial Markets", written with Chris Avery (Kennedy School of Government). The paper shows that herd behavior, which is defined as trading based on the history of trades at the expense of private information about the asset value, may be fully rational in financial markets when trade is sequential. In general, complex information structures lead to herding and can make price bubbles possible, since market makers and informed traders may interpret the trading history differently. Moreover, if the information structure is sufficiently complex, herd behavior may cause extreme short-run price bubbles, since the market may confuse herd trading for informative trading. However, the price mechanism assures that the market price correctly reflects all (public and private) available information in the long run, and prices adjust over time so that trading is ultimately informative about the true asset value.
Luis Angel Medrano (Universitat Pompeu Fabra) stressed that the main contribution of the paper is its ability to explain price distortions in a context with fully rational agents. The paper shows that if there are several dimensions of uncertainty, price bubbles and excess volatility are consistent with rationality. However, the results crucially depend upon two restrictive assumptions: traders must choose from a discrete space of actions and the timing of actions is exogenous. If traders are allowed to buy or sell any quantity at any time, they could fine-tune their actions to their information and, in addition, the speed at which prices reveal information would be higher. As a result, the effects of herding behavior would be less severe and convergence of prices to fundamentals would be much smoother. Mendrano argued that the model does not capture the functioning of financial markets, since traders have a large flexibility in choosing and timing their actions.
A roundtable was took place on Friday evening on 'The Control of Risk in Financial Intermediaries'. Ben Cohen (Bank for International Settlements) questioned whether central banks act as a lender of last resort for large banks facing bancruptcy. Arnoud Boot (University of Amsterdam and CEPR) suggested that given the diminishing predictability in the banking business, there is need for more regulation or at least a new balance between internal and external supervision of banks. Gerard Gennotte (Long Term Capital Management) spoke about liquidity and counterparty risks faced by investment banks and the difficulties of internal control. The develpment of portfolio insurance would make risks and strategies apparent and thereby facilitate risk monitoring. Joan Antoni Ketterer (Intermoney S.A.) discussed the problems related to the specification of stochastic processes in the market and arising in abnormal periods. Josep Lluís Oller (MEFF Renta Fija) concluded the roundtable saying that it was necessary to distinguish between over-the-counter (OTC) and organized markets. Clearing houses act as circuit breakers and thus limit counterparty risks. Therefore, in order to reduce further such risks, it is necessary to organize some OTC markets.