شکست دادن ترس منطقی:دامپینگ سهم استراتژیک و طغیانهای سهامداران

In this paper, we develop a dynamic model of institutional share dumping surrounding control events. Institutional investors sometimes dump shares, despite trading losses, in order to manipulate share prices and trigger activism by “relationship” investors. These institutional investors are motivated to trade not only by trading profits but also by a desire to protect the value of their inventory and to disguise the quality of their own information. Relationship investor profit from targeting firms both by improving firm performance and by generating private information.

Trade in the shares of a troubled firm is characterized both by institutional shareholder share dumping and share acquisition by relationship investors.1 This pattern of trade in troubled firms cannot be completely explained without developing a dynamic model that incorporates both relationship investing and informed trading. This paper develops a dynamic model relating relationship investing and informed trading. In our analysis, one type of strategic investor, whom we refer to as the “institutional investor,” is sometimes able to acquire private information about the quality of a firm's management.2 Whether a given institutional investor really has private information, or rather, trades simply on market priors, is itself private information. Another type of strategic investor, whom we refer to as the “relationship investor,” has no ex ante information endowment. However, in some circumstances a relationship investor can create value through activism. When management quality is good, activism does not create value; when management quality is bad, activism creates value. Institutional investors do not intervene in governance themselves, but instead follow the “Wall Street rule,” expressing their preferences through trading. The problem for a given institutional investor is that, from the perspective of maximizing trading profits, she wants prices to be uninformative. However, uninformative prices may cause the relationship investor to abjure activism, which lowers the value of her portfolio. The relationship investor's problem is that he is never sure whether any given firm has poor management and thus needs a dose of activism. Direct communication between the institutional investor and the relationship investor is not incentive compatible: Because the institutional investor bears none of the costs of activism but she reaps buy-side profits from lowering the stock price, has an incentive to trigger activism by reporting adverse conditions even when they are absent. Thus, information transmission can occur through the observation of security prices. In the first-period, the institutional investor and the liquidity investor trade and the marketmaker sets prices based on the pattern of their order flow. This first-period order flow, which reveals both net order quantities and trading volume, is observed by all agents, including the relationship investor. In the second period, the relationship investor, armed with the information from the order flow, then decides whether to trade and/or intervene in the second period. The institutional trader and the liquidity traders also decide the trades they will submit. The trading decisions of the institutional trader, relationship investor, and the liquidity trader determine second-period prices. Each first-period trade by an institutional investor has three effects; namely, it generates an immediate trading profit or loss, it affects the posterior beliefs of the relationship investor, and it affects the posterior beliefs, and thus future price determinations, of the marketmaker. When the institutional trader has superior information, she maximizes her immediate trading profits by trading on this information. Thus, the trading profit incentive, ceteris paribus, leads to the standard informed trading strategy—buy when endowed with good information, sell when endowed with bad information, and hold when uninformed. Given this pattern of informed trade, selling leads the marketmaker to lower share prices. Lower share prices signal to the relationship investor that management is weak and thus activism may be profitable. Because the marketmaker cannot distinguish between informed and uninformed sales, even uninformed sales lower share prices. Thus, selling increases the intrinsic value of the uninformed institutional investor's post-trade portfolio. This value conservation incentive provides uninformed institutional investors an incentive to “dump” their shares. This dumping leads to “herding” equilibria in which the institution trades on the sell-side even when uninformed. In fact, lowering the ex ante precision of institutional information, by lowering the trading losses from dumping, can actually encourage uninformed institutional investors to dump. Thus, in contrast to the predictions of standard microstructure models, reductions in the ex ante quality of information can actually increase trading activity. The third incentive for trade is marketmaker belief manipulation. Institutional traders have information on the quality of their own private information. Although rational expectations imply that, averaged across all liquidity trader actions, the marketmaker's assessment of the probability that the institutional investor is informed is correct, the marketmaker's assessment of actual information embedded in a given order flow may not be correct. For example, if an institutional trader sells when uninformed, the marketmaker will adjust the price of the security downward. However, the institutional investor knows that she has really not received any adverse information about the firm. Thus, the price reduction in the first period provides the institutional investor an incentive to buy back in the second period. These three incentives, trading profit, value conservation, and belief manipulation, determine the structure of the equilibrium outcome. When the cost and benefits of activism are such that the likelihood of activism by the relationship investor is zero or nearly zero, the trading profit incentive dominates. The institutional investor buys with positive information, sells with negative information, and holds when uninformed. Although an uninformed institution realizes positive expected second-period profits from both “pump and dump”—driving the price up by a buy-side trade in the first-period and then selling the overvalued stock in the second period—and “dump and pump”—selling in the first period and then buying back the undervalued stock in the second period–the first period losses from these manipulative strategies always exceed second-period gains. Thus, the uninformed investor stays on the sidelines. As the costs of relationship investing fall, and the relationship investor begins to make his presence felt, the balance of incentives shifts. Because the relationship investor's likelihood of activism is sensitive to the first-period stock price movements, the value conservation incentive provides the uninformed institutional investor with a strong incentive to dump shares in the first period to provoke activism. When this incentive becomes sufficiently strong, a herding equilibrium emerges in which the institution sells in the first period both when it has adverse information and when it is uninformed. Thus, our model predicts herd behavior by institutions on the sell side but not the buy side. This asymmetric herding prediction generally is consistent with the evidence of Wermers (1999), but inconsistent with a pure labor market-based theory of herding based on reputation protection. Whether the uninformed institutional investor can augment her value conservation gains from first-period dumping through trading profits with second-period buying profits depends on rather subtle features of second-period market structure. When the relationship investor intervenes, unless his holdings are very large, he can only profit by following a random activism strategy, buying when intervening and selling otherwise. Playing such a random strategy exogenously generates private information for the relationship investor regarding his own actions. This private information creates adverse selection losses for the uninformed institution trying to exploit her own private information that she previously submitted an uninformed trade. When the activism probability is high, buy orders are highly informative, and thus engender large upward quote revisions. In this case, the uninformed shareholder who dumped in the first-period will hold in the second period. In contrast, when the activism probability is low, buy orders will not engender such large upward price revisions and the uninformed investor can profitably buy shares in the second period. Thus, some parametrization of our model supports dump-and-pump equilibria. Our analysis of the interaction between institutional and relationship investors in a strategic trading framework produces a number of testable implications for researchers in market microstructure and corporate governance. These implications include: • The likelihood of management shake-ups and governance changes is higher after institutional selling. • Institutional investor herding will be concentrated on the sell side of the market. • The likelihood of management shake-ups is not only affected by stock price drops; it is also correlated with trading volume. • The price effect of institutional share dumping depends on the size of the institution's pre-trade portfolio positions. • Institutional order flows will exhibit asymmetric time-series correlations, with the likelihood of buy orders following buy orders being higher than the likelihood of sell orders following sell orders. Delineating the position of this paper within the financial economics literature is somewhat challenging. Because the paper models relationship traders who profit from information about their own activism, institutional traders who profit by exploiting information endowments, and the effect of informed trading on corporate behavior, the paper is a hybrid of the classic microstructure models of informed trade, models of investor activism, and models of the feedback effect of informed trading on corporate policy. First, consider our lineage on the classical microstructure side (see, e.g., Kyle, 1985 and Glosten and Milgrom, 1985). In most respects, our model of informed trading is standard. However, one point of departure is that informed traders have private information about the precision of the signals they receive. In this respect, our work follows Gervais (2000). In fact, for exactly the same technical reasons as Gervais, we use the Glosten and Milgrom model of order flow rather than the Kyle model. Also like Gervais, our analysis shows that when traders have information about the precision of their own information, aggregate demand flow is not a sufficient statistic for order flow information. Moreover, because the Kyle model imposes a normal distribution structure on both the investors’ private signals and asset values, it requires that the relationship investor contribution to firm value be additive. However, our result depends on the substitutability between the value produced by management and the relationship investor, i.e., the relationship investor's marginal contribution is higher when management quality is low. Without substitutability, the institutional investor has no incentive to deviate from the standard informed trading strategy of buying undervalued and selling overvalued shares. On another branch of the family tree–investor activism–our differences with the extant literature are more profound. Unlike the early literature on activist traders (see, e.g., Kyle and Vila, 1991 and Bagnoli and Lipman, 1996), in which activist trading was a precursor to a hostile takeover offer, we focus on minority shareholder activists that attempt to change corporate policy rather than seize formal control.3 Also, in contrast to Kyle and Vila, in our analysis non-value-improving investors (institutions) move first to affect the expectations of the value-increasing relationship investor. In fact, the central focus of our analysis is the interaction between different types of large strategic investors. Our analysis is also related to the literature on the effect of informational trade on corporate investment policy. In this literature, management uses market order flow to gather information relevant to its investment decision. As shown by a number of researchers, in this setting, management has an incentive to encourage the production of information through informed trading (see, e.g., Subrahmanyam and Titman, 2001, Khanna and Sonti, 2004 and Boot and Thakor, 1993) show that, because of complementarities, feedback effects from information can produce large moves in asset values. In contrast to both Khanna and Sonti and Subrahmanyam and Titman, our analysis features interaction not between informed traders and a nontrading agent, the manager, but rather between two different types of large, strategic trading agents, namely, relationship investors and informed institutional investors. Both types of agents may be willing to suppress information and sacrifice intrinsic value for the sake of trading profits. In addition, in Khanna and Sonti and Subrahmanyam and Titman, investment in new projects is positively correlated with firm quality, while in our model, investment in corporate activism is negatively correlated. Thus, our analysis highlights trade reversal strategies (such as dump and pump), rather than continuation strategies (such as momentum) highlighted in their work. Glostein and Guembel (2002) model an informed investor taking a short position in an asset and then trading in order to send an incorrect signal to management regarding optimal investment policy. Similar to the other papers discussed above, their analysis considers trading by a single type of trader while we model the interaction between two types of strategic traders. Also, in contrast to both our analysis and the other literature, Goldstein and Guembel consider investors with short positions initiating trades that lower rather than increase value. This paper is organized as follows. In Section 2, we lay out the basic features of our model. In Section 3, we present some basic results. In Section 4, we develop and analyze strategic trade, corporate activism, and price determination in the second period of the game. In Section 5, we analyze first-period trade and its effect on price behavior, trading strategies, and corporate governance in the second period of the game. In this section, we outline the conditions necessary for a share dumping equilibrium and the conditions for dumping institutions to recapture some trading profits by reversing earlier sell orders. Finally, in Section 6, we conclude the paper.

In this paper, we develop a dynamic model of institutional shareholder trading and activism. The model combines key elements of a rich model of market microstructure including volume-dependent quotes with a model of activism by outside relationship investors. The analysis shows that voice and exit, which may be substitutes at the level of an individual investor, are highly complementary at the aggregate level, with exiting investors attempting to provoke other activist investors into exercising voice, and activist investors looking to the trading patterns of exiting investors as a guidepost for their activism activity. This model explains the well-documented practice of institutional share dumping, that is, institutional investors selling shares in troubled firms. The model also has a number of implications for trading volume and corporate control, the portfolio positions of institutions, the market reaction to their trades, and the effect of strategic trading on governance. Extending this research would produce even more insights into the interaction of governance and market microstructure. The analysis could be extended either by framing the relationship between shareholder activism and firm value in a more general fashion, analyzing the alternative information structures for the relationship investor, or by considering a more flexible model of liquidity trader behavior. First consider the relationship between activism and management quality. We assume that shareholder activism and good management are substitutes; that is, activists are not needed when management is strong. This model fits the pattern of disciplinary relationship investing usually observed in large firms. However, with smaller firms, relationship investors frequently adopt a developmental role. Such developmental efforts may well bear the greatest fruit when the targeted firm and its management have good prospects. In this case, activism and quality would be complements. Such complementarity could be developed in our framework, and modelling them it might produce an explanation for institutional pumping—institutions herding and buying into a small firm's shares, even when uninformed. This sort of herding phenomenon on the buy side, restricted to high-tech and small firms, has been documented by Wermers (1999). In addition to extending our treatment of the value-to-activism relationship, we might also gain by extending our analysis of the information structures faced by the players. In our analysis, the relationship investor has the same information as the marketmaker; that is, he observes the entire pattern of order flow (absent, of course, the identity of the traders submitting each order). The relationship investor can use this information to structure his activism strategy. An interesting extension of our analysis would be to restrict the relationship investor's information set to a subset of the information in the order flow, say to observing only price or only price and aggregate demand. Such restrictions would affect the relationship investor's activism strategy both by increasing his exposure to trading losses with informed institutions and by changing the initial trading strategies of institutions, who would change their date-1 strategy of affecting the relationship investor's beliefs. A third direction for future research would be to systematically vary the liquidity of markets across periods. Our analysis fixes market liquidity and assumes that liquidity demand is exogenous. In a model with variable and endogenous liquidity demand, date-1 institutional trades would affect not only the informativeness of prices but also, by releasing private information, second-period liquidity. Strategic institutions would factor these liquidity effects into their trading strategies. Increased liquidity can help or retard relationship investing activity depending on the cost of activism and the prior endowment of the relationship investor. Thus, such strategic manipulation of liquidity could provide another layer of incentives for the institutional investor.