The multidivisional structure: Competition in the Capital Market

1. Frictionless Capital Markets 

As the remarks of Bork and Bowman cited in Section 2.2 of Chapter 6 make clear, advocates of received microtheory are loath to concede that capital markets may fail to operate frictionlessly. Partly for this reason, the fiction that managers operate firms in fully profit maximizing ways is main- tained. Any attempt by managers to opportunistically promote their own goals at the expense of corporate profitability would occasion intervention through the capital market. Effective control of the corporation would be transferred to those parties who perceived the lapse; profit maximizing behavior would then be quickly restored.

Parties responsible for the detection and correction of deviant behavior in the firm would, of course, participate in the greater profits which the reconstituted management thereafter realized. This profit participation would not, however, be large. For one thing, incumbent managements, by assumption, have little opportunity for inefficiency or malfeasance because any tendency toward waywardness would be quickly detected and costlessly extinguished. Accordingly, the incremental profit gain occasioned by takeover is small. In addition, since competition among prospective takeover agents is presumably intensive, the gains mainly redound to the stockholders. Peterson’s sanguine views on corporate behavior are roughly of this kind. He characterizes the latitude of managers to disregard the profit goal as “small” (1965, p. 11) and goes on to observe: “Far from being an ordinary election, a proxy battle is a catastrophic event whose mere possibility is a threat, and one not remote when affairs are in conspicuous disarray.” Indeed, even “stockholder suits … may be provoked by evidence of serious self- dealing.” On the principle that the efficacy of legal prohibitions is to be judged “not by guilt discovered but by guilt discouraged,” he concludes that such suits, albeit rare, may have accomplished much in helping to police the corporate system (Peterson, 1965, p. 21; emphasis added).

While I do not mean to suggest that such deterrence has been un- important, Peterson’s observations appear to me to be consistent with the proposition that traditional capital markets are beset by serious problems of information impactedness and incur nontrivial displacement costs if the incumbent management is disposed to resist the takeover effort. Why else the reference to catastrophic events, conspicuous disarray, and serious self- dealing? Systems that are described in these terms are not ones for which a delicately conceived control system can be said to be operating As recent military history makes clear 85 controls that involve a discrete shock to the system are appropriate only when an offense reaches egregious proportions. The limits of opportunism are accordingly wider than Peterson seems prepared to concede.

The reasons, I submit, why traditional capital market controls are relatively crude are because an information impactedness condition exists with respect to internal conditions in the firms and. because of sorting out difficulties, the risk of opportunism on the part of would-be takeover agents is great. Given information impactedness, outsiders can usually make confi- dent judgements that the firm is not adhering to profit maximizing standards only at great expense. The large firm is a complex organization and its performance is jointly a function of exogenous economic events, rival behavior, and internal decisions. Causal inferences are correspondingly difficult to make, and hence, opportunism is costly to detect. Moreover, once detected, convincing interested stockholders that a displacement effort ought to be supported encounters problems. Inasmuch as time and analytical capacity on the part of stockholders are not free goods, which is to sa\ that their information processing limits must be respected, the would-be takeover agent cannot simply display all of his evidence and expect stock- holders to evaluate it and reach the ”appropriate” conclusion. Rather, any appeal to the stockholders must be made in terms of highly digested inter- pretations of the facts. Although this helps to overcome the stockholder’s bounded rationality problem, it poses another: How is the interested stock- holder (or his agent) to distinguish between bona fide and opportunistic take- over agents.

The upshot of these remarks is that the transaction costs associated with traditional capital market processes for policing management, of the sort described by Peterson, are considerable. Correspondingly, the range of discretionary behavior open to incumbent managements is rather wider than Peterson and other supporters of the frictionlessness fiction concede.86

2. The M-Form Firm as a Miniature Capital Market

In a general sense, the most severe limitation of the capital market is that it is an external control instrument. It has limited constitutional powers to conduct audits and has limited access to the firm’s incentive and resource allocation machinery. One should not. however, conclude that mere di- visionalization. by itself, is sufficient to correct the inefficiencies and goal distortions that the large U-form firm develops. To emphasize this, the limits of the holding company form of organization are examined below. Attention is thereafter shifted to consider strategic controls of the sort appropriate to an M-form enterprise.


What is referred to here as a holding company form of organization is a loosely divisionalized structure in which the controls between the head- quarters unit and the separate operating parts are limited and often unsystematic. The divisions thus enjoy a high degree of autonomy under a weak executive structure.

Perhaps the least ambitious type of divisionalization to consider within the holding company classification is that in which the general office is essentially reduced to a clerical agency for the assembly and aggregation of earnings and other financial reports. The holding company in these circumstances serves as a risk-pooling agency, but in this respect is apt to be inferior to a mutual fund. The transaction costs associated with altering the composition of the portfolio of the holding company, by selling off existing divisions and acquiring new operating companies, will ordinarily exceed the costs that a mutual fund of comparable assets would incur by its trading of common stocks (or other securities) so as to adjust its portfolio. Little wonder that those academics who interpret the conglomerate as being a substitute mutual fund report that it has inferior diversification characteristics to mutual funds themselves (Smith and Schreiner, 1969: Westerfield, 1970).

Moreover, it is not clear that just a little bit of additional control from the general office will lead to results that are superior to those that would obtain were the various divisions of the holding company to be freestanding firms in their own right. Being part of a holding company rather than an independent business entity easily has umbrella effects. If the holding company serves as a collection agency for unabsorbed cash flows and uses these to shore up the ailing parts of the enterprise, the resulting insularity may encourage systematic distortions (of a managerial discretion sort) among the divisional managements. Being shielded from the effects of adversity in their individual product markets, slack behavior sets in.

This is not, of course, a necessary consequence. The general management might consciously refuse to reinvest earnings but mainly pay these out as dividends. Alternatively, it might scrutinize reinvestment decisions every bit as well as the unassisted capital market could. Indeed, because it enjoys an internal relationship to the divisions, with all of the constitutional powers that this affords, the general management might be prepared to assume risks that an external investor ought properly to decline. (Thus, the general management can ordinarily detect distortions and replace the divisional management at lower cost than can an external control agent similarly detect and change the management of a comparable, free standing business entity. The holding company, in this respect, is less vulnerable to the risks of what might be referred to as managerial moral hazard.) Given, however, that the holding company is defined to be a divisionalized firm in which the general office does not involve itself in strategic controls of the sort described below, it is unclear that the holding company form of organization is socially to be preferred to an arrangement in which the various divisions are each set up as fully independent enterprises instead. Holding companies certainly cannot be expected reliably to yield results that compare favorably with those which I impute to the M-form structure.


If indeed the firm is to serve effectively as a miniature capital market, which in many respects is what the M-form structure ought to be regarded,8 a more extensive internal control apparatus than the holding company form of organization possesses is required. This gets to the core issues. Manipulation of the incentive machinery, internal audits, and cash flow allocation each deserve consideration.

Closer adherence to the goals of the general management can be secured if the special incentive machinery to which internal organization uniquely has access to is consciously exercised to favor operating behavior that is consistent with the general management’s objectives. Both pecuniary and nonpecuniary awards may be employed for this purpose.

That salaries and bonuses can be adjusted to reflect differential operating performance, assuming that such differentials exist and can be discerned, is a familiar application of the incentive machinery. That nonpecuniary rewards, especially status, can also be adjusted for this purpose should be evident from the preceding chapter.

Of course, sometimes a change of employment, or at least of position, may be altogether necessary. The division manager may not have the management capacities initially ascribed to him, conditions may change in ways that warrant the appointment of a manager with different qualities, or he may be managerially competent but uncooperative (given, for example, to aggressive subgoal pursuit in ways that impair overall performance). Changes made for either of the first two reasons reflect simple functional assessments of job requirements in relation to managerial skills. By contrast, to replace a division manager for the third reason involves the deliberate manipulation of the incentive machinery to produce more satisfactory results. The occasion to intervene in this way will presumably be rare, however, if the conditional nature of the appointment is recognized from the outset. Rather, the system becomes self-enforcing in this respect once it is clear that the general management is prepared to replace division managers who regularly defect from general management’s goals.

Although the general office does not ordinarily become directly involved in the exercise of the incentive machinery within the operating divisions, its indirect influence can be great. The decision to change (replace, rotate) a division manager is often made for the incentive effects this has on lower- level participants. Employment policies — including criteria for selection, internal training procedures, promotions, and so forth — can likewise be specified by the general office in ways that serve to ensure closer congruence between higher-level goals and the behavior of the operating parts. A more pervasive incentive impact on lower-level participants who are not directly subject to review by the general office can in these ways be effected.

Adjusting the incentive machinery in any fine tuning sense to achieve reliable results requires that the changes be made in an informed way. A backup internal audit that reviews divisional performance and attempts to attribute effects to the several possible causes — distinguishing especially between those outcomes that are due to changes in the condition of the environment from those that result from managerial decision-making — is useful for this purpose.88 As Churchill, Cooper, and Sainsbury observe: “… to be effective, an audit of historical actions should have, or at least be perceived as having, the power to go beneath the apparent evidence to determine what in fact did happen” (1964, p. 258). Of particular importance in this connection is the recurrent nature of this auditing process. Thus, although current variations of actual from projected may sometimes be “accounted for” in plausible but inaccurate ways, a persistent pattern of performance failure can be explained only with difficulty.

The advantages of the general office over the capital market in auditing respects are of two kinds. First, division managers are subordinates; as such, both their accounting records and backup files are appropriate subjects for review. Stockholders, by contrast, are much more limited in what they can demand in the way of disclosure. Even relatively innocent demands for a list of the stockholders in the corporation, much less the details of internal operating performance, may be resisted by the management and disclosed only after a delay and by court order.

Second, the general office can expect knowledgeable parties to be much more cooperative than can an outsider. Thus, whereas disclosure of sensitive internal information to an outsider is apt to be interpreted as an act of treachery,11 internal disclosure is unlikely to be regarded opprobri- ously. Rather, internal disclosure is affirmatively regarded as necessary tc the integrity of the organization and is rewarded accordingly. Disclosure to outsiders, by contrast, commonly exposes the informant to penalties— albeit that these may be subtle in nature.

Not only are internal audits useful for ascertaining causality, they also serve as a basis for determining when operating divisions could benefit from assistance. The general management may include on its staff what amounts to an internal management consulting unit — to be loaned or assigned to the operating divisions as the need arises. Partly the occasion for such an assignment may be revealed by the internal audit. Thus, although the general management ought not routinely to become involved in operating affairs,91 having the capability to intervene prescriptively in an informed way under exceptional circumstances serves to augment its credibility as an internal control agent.92 Self-regulatory behavior by the operating divisions is thereby encouraged.

In addition to the policing of internal efficiency matters, and thereby securing a higher level of adherence to profit maximization than the un- assisted capital market could realize (at comparable cost), the general management and its support staff can perform a further capital market function — assigning cash flows to high yield uses. Thus, cash flows in the M- form firm are not automatically returned to their sources but instead are exposed to an internal competition. Investment proposals from the several divisions are solicited and evaluated by the general management. The usual criterion is the rate of return on invested capital.

Moreover, because the costs of communicating and adapting internally are normally lower than would be incurred in making an investment pro- posal to the external capital market, it may be practicable to decompose the internal investment process into stages. A sequential decision process (in which additional financing is conditional on prior stage results and developing contingencies) may thus be both feasible and efficient as an internal investment strategy. The transaction costs of effectuating such a process through the capital market, by contrast, are apt to be prohibitive.

In many respects, this assignment of cash flows to high yield uses is the most fundamental attribute of the M-form enterprise in the comparison of internal with external control processes, albeit that the divisionalized firm is able to assign cash flows to only a fairly narrow range of alternatives at any one point in time. Even if the firm is actively acquiring new activities and divesting itself of old, its range of choice is circumscribed in relation to that which general investors, who are interested in owning and trading securities rather than managing real assets, have access to. What the M- form firm does is trade off breadth for depth in this respect.94   In a similar context, Alchian and Demsetz explain: “Efficient production with heterogeneous resources is a result not of having better resources but in knowing more accurately the relative productive performances of those resources“ (1972, p. 29).

Source: Williamson Oliver E. (1975), Markets and hierarchies: Analysis and antitrust implications, A Study in the Economics of Internal Organization, The Free Press.

Leave a Reply

Your email address will not be published. Required fields are marked *