سازندگان بازار به عنوان فراهم آورندگان اطلاعات: تجربه طبیعی STAR

Market makers are financial intermediaries who are supposed to provide additional liquidity, but do not have any information-related obligation. This paper studies the unique case of the Italian Stock Exchange, where market makers are also obliged to facilitate information disclosure about the firms they cover. We focus on a group of small/medium capitalization stocks (STAR) that are assigned a designated market maker (DMM) starting from 2001. We show that their liquidity requirements are not binding during the sample periods and that the main impact of DMMs' introduction is due to their obligations on information provision. We find that DMMs' activity as information providers reduces spread and price volatility, the probability of informed trading (PIN), and the adverse selection component of the spread. An event study provides evidence that the information released through DMMs is perceived as useful by market participants.

In financial markets information on less traded stocks is generally supplied by firms or by analysts. This paper investigates an alternative channel of information disclosure by considering the role of market makers as information providers. Market makers are financial intermediaries that are supposed to provide additional liquidity but do not usually have any information-related obligation. We here study the unique trading environment of the Italian Stock Exchange, where market makers have obligations aimed at facilitating information disclosure of the listed firms. In April 2001, Borsa Italiana (BIt from now on) started assigning a designated market maker (DMMs from now on) to a group of small-medium capitalization stocks, that were named STAR. The main novelty of this experiment is that DMMs have information disclosure requirements. Information obligations require DMMs to act as analysts on STAR stocks and to produce at least two detailed financial analyses per year; DMMs are also required to organize at least two yearly meetings, named roadshows, with professional investors. The purpose of the paper is to study how these information disclosure requirements affect market quality. Generally in order driven markets DMMs who act as liquidity suppliers are required to comply only with liquidity requirements, the most relevant being the maximum quoted spread (Bessembinder et al., 2008). In the Italian case, instead, for the 59 companies that were assigned a DMM between 2001 and 2005, the maximum spread requirement was not binding,1 and this creates an ideal setting to focus on the effect of the information disclosure provided by DMMs. We use high frequency data covering four sample periods, one before and three after the companies joined the STAR group. We find that after the assignment of the DMM, spread and volatility decrease for STAR stocks compared to a matched sample of control stocks, while volume does not change significantly. In the longer run, spread and volatility decrease substantially and volume increases significantly: we show that this improvement in market quality is associated with a decrease in information asymmetries (and, in turn, in adverse selection costs) induced by the DMMs' disclosure requirements. Accordingly, we find that information asymmetries, measured by the probability of informed trading (PIN) as in Easley et al. (1996), tend to decrease after the companies are assigned a DMM; furthermore, by estimating the model of Glosten and Harris (1988), we document that the component of the spread due to traders' inability to efficiently process information decreases. To verify that the short term reduction in the price impact is due to adverse selection costs and not to inventory adjustments, we estimate a VAR model similar to Hasbrouck (1991) and find evidence of a permanent decrease in adverse selection costs. We show that these findings are consistent with a simple framework of asymmetric information and rational price formation in the spirit of Grossman and Stiglitz (1980) where information is disclosed to uninformed market participants. In order driven markets where both informed and uninformed traders can supply liquidity, this information disclosure reduces adverse selection costs for uninformed traders and hence makes them more willing to supply liquidity. Within this framework we show that when information is publicly disclosed the spread decreases. We also show that information disclosure has two opposite effects on volatility: it reduces the price impact component, and it increases the demand shock component. An extensive literature exists that investigates whether the assignment of a market maker affects market quality. Numerous empirical papers2 have shown advantages and disadvantages of moving from the NASDAQ dealer market (or from the OTC market) to the NYSE auction based specialist market, and at least as many papers have investigated how the specialist's activity can affect trading.3 By contrast, the experiment we study consists of a group of stocks listed on an order driven market organized as a limit order book that at a certain point in time, and all other things being equal, are assigned a DMM. The DMM has to meet the obligations imposed by Borsa Italiana, and does not monopolistically control the order book. This means that we depart from the literature comparing the virtues of NASDAQ vs. the NYSE for at least two main reasons. Firstly, the assignment of the DMM does not come with a change in market structure as in the NASDAQ vs. NYSE experiment. Secondly, the role of the DMM on the Italian stock market, as in any other stock market that works as an electronic limit order book, substantially differs from that of the specialist on the floor of the NYSE, who has an active and key role in managing the order book; by contrast, STAR DMMs act as limit order traders and have no privilege in the access to the order book. The present paper is related to three recent empirical papers that study the effect on market quality4 of the introduction of DMMs with only liquidity requirements. Venkataraman and Waisburd (2007) find that introducing DMMs in the Paris Bourse leads to an increase in liquidity for a sample of stocks traded through a call auction; their analysis also differs from ours as we consider DMMs trading in a limit order book. Anand et al. (2009) document an improvement in market quality after the introduction of DMMs in the limit order book of the Stockholm Stock Exchange; in this case, however, DMMs' maximum spread obligations are binding and, yet again, there are no requirements in terms of information disclosure. Menkveld and Wang (2009) study the introduction of DMMs in Euronext Amsterdam; they find that liquidity increases, stock prices increase and liquidity risk decreases after the event. This analysis is also closely related to the field of research on the relation between analysts' activity and market liquidity. A vast body of literature examines the stock price reaction to analysts' forecasts (for a recent critical survey, see for example, Ramnath et al., 2008)5 but little attention has been devoted to the effects of analysts' information on liquidity and adverse selection costs. Most previous papers, as Brennan and Subrahmanayam, 1995, Roulstone, 2003 and Kanagaretnam et al., 2005 find that liquidity is positively associated with analyst coverage; others (e.g. Chung et al., 1995) document a negative association. There is no consensus on whether analysts' activity fosters liquidity by reducing information asymmetries, or it is instead perceived as a signal of the presence of higher information asymmetries. STAR DMMs differ from the analysts considered in previous research, because they are directly involved in trading on the same stocks about which they provide information.6 Furthermore, previous studies are concerned with the contemporaneous association between analyst coverage and market quality; as a consequence, they cannot clearly identify the causal effect of analysts' activity on liquidity and other indicators of market quality. By contrast, we compare a period before the introduction of the DMMs to later periods and we are able to test the effect over time of the additional information provided to the market. The remainder of this paper is organized as follows. Section 2 describes the dataset and the sample choice, Section 3 outlines a theoretical framework on which the empirical hypotheses are based, and Section 4 discusses the results on market quality. Section 5 focuses on information asymmetries and the market reaction to roadshows and financial analyses, and Section 6 concludes.

Pre-trade transparency is a timely issue in financial market design and regulation. The amount and precision of the information disclosed to market participants before trading has been at the center of a wide empirical and theoretical literature. Little research instead has so far investigated possible channels of information disclosure other than firms and analysts. This paper raises a new question: should information be disclosed by firms and analysts, or by intermediaries? To our knowledge, there is no evidence on the effects of information disclosure by dealers. The concern of conflict of interest prevents regulators from introducing this disclosure vehicle. However, our results show that information disclosure by designated market makers (DMMs) can improve market quality and hence suggest that reputational concerns might prevent front-running and adverse selection costs.27 We here study the effect on market quality of the introduction of DMMs in STAR, a group of small-medium capitalization stocks listed on the limit order book of Borsa Italiana. The peculiarity of STAR is that DMMs have information disclosure requirements: precisely, they are required to provide financial analyses on the stocks and to interact with institutional and retail investors on a regular basis. Because liquidity requirements were not binding during the periods examined, we are able to focus on the effect of disclosure requirements and, thus, on the role of market makers as information providers. We find that, after the introduction of DMMs, spreads and volatility decrease relatively to a matched sample of control stocks. More interestingly, these changes get stronger as time passes. We also find a decreasing trend in information asymmetries, measured by the probability of informed trading (PIN) which is weakly significant but interestingly consistent with the significant decrease in the adverse selection component of the spread, captured by the Glosten and Harris (1988) model and by a VAR model in the spirit of Hasbrouck (1991). Finally, following an event study approach, we show that the information released with the help of DMMs is perceived as useful for investment decisions. We show that the decrease in information asymmetries observed after the introduction of DMMs is due to an improvement in the degree of information disclosure. As DMMs have the objective of improving and fastening the dissemination of the companies' information, the outcome of this disclosure activity is the general improvement of traders' ability to process information about STAR stocks. This reduces the impact of traders' orders on stock prices and, hence, the adverse selection component of the spread. Analogously, the decrease in the price impact reduces price variations around the fundamental, and volatility can decrease. In this paper we concentrate on a unique trading environment and we show how relevant can be the disclosure requirements implemented by an exchange to improve market quality for small-medium size stocks. It follows that regulators may find this information dissemination mechanism preferable when firm-specific incentives for disclosure are not effective. Future research can tackle the issue of how to optimally design and regulate the contracts among DMMs, companies and customers.