Fool's gold: Social proof in the initiation and abandonment of coverage by Wall Street analysts - Statistical Data Included

Administrative Science Quarterly, Sept, 2001 by Hayagreeva Rao, Henrich R. Greve, Gerald F. Davis

**********

This paper examines the dynamics of social influence in the choices of securities analysts to initiate and abandon coverage of firms listed on the NASDAQ national market. We show that social proof--using the actions of others to infer the value of a course of action--creates information cascades in which decision makers initiate coverage of a firm when peers have recently begun coverage. Analysts that initiate coverage of a firm in the wake of a cascade are particularly prone to overestimating the firm's future profitability, however, and they are subsequently more likely than other analysts to abandon coverage of the firm. We thus find evidence for a cycle of imitation-driven choice followed by disappointment and abandonment. Our account suggests that institutionalization rooted in imitation is likely to be fragile.

A core proposition in organizational theory is that imitation is a characteristic response to uncertainty in decision making (Cyert and March, 1963). Imitation follows from the heuristic of social proof, that is, looking to the actions of others for clues as to what constitutes appropriate action (Cialdini, 1993). Neoinstitutionalist theory holds that decision makers imitate peers, especially peers perceived to be successful and legitimate, to minimize search costs and to avoid the costs of experimentation (DiMaggio and Powell, 1983). If a practice is prevalent among competent actors, then it must be sensible. But neoinstitutional research also presumes a model of human behavior in which people persist in a course of action that they have copied from others even when it is against their own interests. Such a model implies that actors are cognitive dopes rather than cognitive misers. Cognitive dopes blindly follow others and stick to a course of action, whereas cognitive misers use heuristics to reduce search costs but are quite capable of abandoning a choice in light of new evidence about its value (Fiske and Taylor, 1991).

Reliance on heuristics such as social proof can often lead to overvaluation of the choice and regret about the decision, and the cognitive miser model may therefore be a more realistic description of the dynamics of adoption of a practice. Organizations routinely adopt programs intended to improve performance, such as quality circles, only to abandon them when the programs fail to live up to their promise (Abrahamson and Fairchild, 1999), resulting in waves of adoption and abandonment (Barley and Kunda, 1992). Even hybrid corn, the canonical setting for understanding the adoption of innovation since the work of Ryan and Gross (1943), did not diffuse once and for all. Apodaca (1952) found that the percentage of adopters in a population of New Mexico farmers rose from 0 to 60 percent between 1945 and 1947 but fell back to 3 percent by 1949. The wet-kiln process of mixing cement was mimetically adopted and came close to replacing the dry-kiln process in the U.S. cement industry but later fell out of favor and wa s widely abandoned (Anderson, 1999). Recently, matrix structures have been adopted and abandoned (Burns and Wholey, 1993), and conglomerates have been built and disassembled (Davis, Diekmann, and Tinsley, 1994). Thus, imitation-based institutionalization appears to be fragile empirically because social proof triggers adoptions that are later judged to be erroneous: hybrid corn tasted bad, the wet-kiln process wasted energy, matrix organizations were complex, and conglomerates were inefficient. Once adopters have updated their information based on their own experience, they can abandon a course of action if their decision to adopt was not irrevocable in the first place.

Neoinstitutionalists rarely study adoption and abandonment decisions concurrently because they examine courses of action that have either diffused widely or been abandoned wholesale--a form of selection bias (Strang and Soule, 1998). Yet it is clear from the examples above that adoption is sometimes followed by regret and abandonment: the more highly one evaluates an uncertain course of action ex ante, the more likely one is to be disappointed. To the extent that adoption is based on imitation, we should expect to see overvaluation, disappointment, and abandonment. We explore this idea by studying how equity analysts employed by investment banks initiate and cease coverage of the securities of firms listed on the NASDAQ stock market. Analysts evaluate the prospects of securities issuers and render both summary judgments (recommendations to buy, hold, or sell shares) and regular estimates of expected earnings (Zuckerman, 1999). Thus, analysts make choices of whether to add or drop a firm from the portfolio of firms they cover and judgments of the firms' expected profitability. Analysts typically specialize by industry, and limits on time and energy mean that they must be selective in which firms they choose to follow. Incentives in investment banks favor covering firms whose stock market performance is expected to be good in the future and avoiding poor performers, and analysts are punished for being inaccurate in their earnings forecasts (Hong, Kubik, and Solomon, 2000). Yet predicting which firms will be winners is an impossible task, according to the efficient market hypothesis, which suggests that future price movements follow a random walk pattern; thus, "financial forecasting appears to be a science that makes astrology respectable" (Malkiel, 1996: 169). In this context of extreme uncertainty, selecting which firms to follow in one's portfolio is a plausible context for mimesis. It is also an appropriate context for enhancing our understanding of institutionalization by studying adoption, evaluation, and aba ndonment dynamically.