|
Article Excerpt
1. Introduction
Conscientious managers often solicit information from a variety of sources before they commit their firms to make major investments or other significant strategic decisions. In addition to obtaining information from within their organizations, they may solicit outsiders' reactions to their proposed plans, including members of the investment community, potential customers, suppliers, outside consultants, potential joint venture partners, investment bankers, lawyers, and so on.
Acquiring technological and/or economic information about a new plan from external sources comes at a potential cost: it is more likely to result in the divulgence of the firm's investment plans before the plans can be implemented. Such information leakage can occur in various ways. Rivals of the firm inadvertently may become aware of the firm's plan through such contacts, which would allow them to more readily compete with the firm (consider, for example, the case where the firm's plan involves identifying a takeover target). Or the firm's own customers might learn about new products the firm is developing, causing them to decrease their purchases of the firm's existing products in anticipation of the additional or improved features associated with the new products. Or consumer, environmental, political, or union advocacy groups might be alerted to the firm's new plan, all of which might have more success in either restricting the firm's ability to implement the plan, or getting the firm to make extra conces sions to proceed with the new plan without interference, than they would had they learned of the plan at later stages of the plan's implementation.
This paper studies a firm's decision to acquire additional information from external sources about a new plan under consideration in the presence of potential costs of information leakage ("proprietary costs" in the following). When the firm is not subject to managerial moral hazard, we show that under a broad set of conditions there is a positive threshold such that any new plan with an estimated net present value (calculated using only the firm's internal assessment of the plan's value) below the threshold is implemented only after consulting outsiders about the plan's profitability. Any plan with an (internally assessed) estimated net present value (NPV) above the threshold is directly implemented without consulting outsiders. Thus, if one views consulting outsiders prior to implementing a plan as a form of disclosure, this result contrasts with standard results in the disclosure literature that indicate that disclosure occurs only if the plan's estimated NPV is above some threshold. The primary source of the difference between our results and those found in the extant literature is that, in much of the extant literature, the costs of disclosure are taken to be a fixed exogenous constant. In contrast, we posit that firms incur proprietary costs through the leakage of information about a new plan in the course of consulting external sources about the plan only if the firm decides to implement the plan. We regard this as the reasonable case: if a firm does not proceed with the plan, then the proprietary costs associated with information leakage about the plan disappear, because the reactions of rivals and other parties responding to the plan become moot.
After supplying this analysis in a setting without moral hazard, we proceed to consider how agency problems related to the firm's management affects these results. We study two distinct managerial agency problems. In the first version, which we call the ex ante moral hazard problem, a manager runs a research and development center, a new product innovation center, or something similar, and the manager's responsibilities are confined to generating a new plan worthy of further evaluation. We show that this ex ante moral hazard problem generates additional benefits to consulting outsiders relative to nonmoral hazard settings, because consulting outsiders provides a quicker "read" on the plan's merits (relative to directly implementing the plan, which may entail long gestation lags), and this reduces the risk imposed on the manager. If outsiders quickly confirm the plan's viability the manager need not bear the risk of dying, being fired, having the firm face financial stringency, and so on, in the time that laps es between the decision to implement the plan and the plan's completion. Further, we show that managerial moral hazard not only induces the manager's employer to motivate the manager to consult outsiders more often (before deciding whether to proceed with the plan), it also results in the manager implementing the plan less often compared to the nonmoral hazard setting. These results also imply that the presence of moral hazard increases the incidence of information leakage.
In the second version of the agency problem, we extend managerial moral hazard to include project execution. In the extension, the manager's ex ante unobservable effort affects the identification of viable plans, and the manager's ex post (or operational) effort affects the returns to implementing a viable plan. In many settings, we show that this extra layer of moral hazard reinforces the tendency of the manager to consult outsiders before implementing a plan and leads to an even greater incidence of information leakage.
Several streams of research are related to this paper. The decision to proceed with an investment in our model is not a one-time decision, so the investment problem we study can be viewed as a problem involving real options (see, for example, Dixit and Pindyck 1994). To the best of our knowledge, none of the real option literature has been concerned with the information leakage issues we raise. Our evaluation of the investment distortions that result from the introduction of an agency problem is part of the huge literature on that topic initiated by Jensen and Meckling (1976). Consistent with much of that literature, we show that moral hazard in our model leads to under-investment. One problem we study that is not part of this general literature is how moral hazard can affect the timing of investment decisions, which is also the subject of Antle et al. (1995).
There exists a substantial literature in economics (see, for example, Townsend 1979, Diamond 1984, Gale and Hellwig 1985) on the structure of bilateral contracts when one of the contracting parties has superior information about the payoffs from an investment decision to the other, and this information asymmetry can be eliminated by undertaking a costly investigation.1 A common characteristic of such contracts is that there is a threshold payoff level such that investigation proceeds only if the payoff is below the threshold. This result has parallels to results in this paper establishing that a firm's manager will seek to obtain the advice of costly external counsel only if the payoff to the project based on the information available to the manager is below some threshold. There are, however, several differences between our model and those in the literature. As two examples, we note that (1) in our model, the decision to incur the potential cost of seeking external counsel precedes the investment decision, w hereas in the cited papers, investigation occurs only after the investment's output has been realized, and (2) in our model, the cost of seeking external counsel is a contingent cost, incurred only if investment is subsequently initiated (whereas in most of the literature, these investigation costs are uncontingent). Finally, this paper contains features of the classic problem involving the costs and benefits of acquiring extra information before making a decision. There is a voluminous literature in accounting on the costs and benefits of information acquisition for which good summaries can be found in Demski (1980) and Magee (1986).
The paper proceeds as follows. Section 2 presents the basic model setup. Section 3 studies the calculus of the costs and benefits associated with the option of acquiring additional information about a new plan in the absence of managerial moral hazard. Section 4 extends the theory to the setting in which the manager is subject to moral hazard in the design stage. Section 5 augments that study by including moral hazard in plan execution. Section 6 contains conclusions. The Appendix contains proofs of several major results. (2)
2. Model Setup Without Moral Hazard
The basic model we investigate is the following. As of some date 0, a firm operates in one line of business. At some later date 1, a viable new business opportunity becomes available to the firm with probability q [member of] (0, 1). This new opportunity might be a joint venture, a new product, a possible takeover candidate, and so on. We refer to this opportunity in the following generically as a "viable" new plan or project. With probability 1 - q, no viable new plan arrives at the firm. If such a new plan becomes available, the firm is not obliged to undertake it: its availability creates an option, not an obligation. As with other options, the firm can be expected to exercise it only if that constitutes a NPV-increasing action.
Initially, knowledge of the new plan's availability is private information to the firm's manager, along with the manager's estimate [y.sub.i] of the plan's NPV, where the subscript i denotes "inside" information the manager has. Upon learning of the new plan and his estimate of its value, the manager can either immediately implement the new plan or obtain additional information about the new plan's value by consulting outsiders before deciding whether to implement the plan. These outsiders, if contacted, produce another estimate, [y.sub.m], of the plan's value. (3)
As discussed in the introduction, the process of obtaining the outsiders' estimate of the plan's value runs the risk of divulging the availability of the plan to the firm's rivals, or other external parties whose actions affect the plan's NPV, with positive probability. Whatever the identity and motivation of these external parties, we assume that this information leakage diminishes the expected NPV of the new plan, if implemented, by some (expected) amount c > 0. In the text, we take c to be a constant. In the notes, we indicate under what conditions our results hold when c is variable. (4)
A central feature of our model is that, if the plan is not implemented, no proprietary costs are posited to be associated with having consulted outsiders about the plan. We believe this assumption is natural. For instance, if the proprietary costs of consulting outsiders would have arisen from a first-mover advantage, the firm may have lost to its rivals as a consequence of obtaining outsiders' counsel; this lost first-mover advantage is...
|
