The myth of the M-Form? Governance, consent, and organizational change
At the time this paper was published, common efficiency-theory wisdom held that the M-Form (multi-division organizational form) was a rational economic approach to minimizing cost, namely that it was less costly to create divisions focused on tactical matters and a central authority responsible for matters of strategy. Freeland attempts to falsify this and propose new theory through an examination of GE’s use of M-Form between 1924 and 1958; Chandler (1962) had previously used GE as an exemplar of efficiency-theory wisdom related to M-Form, so this paper is quite the rebuke. Freeland holds that GE’s adoption of “textbook” M-Form led to poor financial performance, and that GE performed well when it used a modified M-Form that took both cost and consent into account.
Freeland begins by outlining some fundamental problems with an efficiency approach to understanding M-Form. He presumes Simon’s bounded rationality, then quotes Williamson: “The central imperative facing economic organizations is to "devise . . . governance structures that have the purpose and effect of economizing on bounded rationality while simultaneously safeguarding transactions against the hazards of opportunism" (Williamson 1985, p. xiii).” (p 486)
He explains that efficiency theorists explain M-Form as just such a structure, in that they make it very difficult for executives to muck about in tactical matters and make it similarly challenging for managers to get a voice in strategy. The presumption is that the organization will be better off if these coalitions are motivated to focus on their respective domains.
Then Freeland claims theoretical ground: “In a world of pervasive bounded rationality, even the most complete set of sanctions will suffer informational deficiencies that leave it susceptible to distortion and manipulation, making such sanctions "inadequate tools for control" of the organization” (p 488). Given this, Freeland claims that commitment (or consent) to the organization – independent of role, task, structure or function – is the additional, necessary “glue” that keeps an organization together. He does a nice job pulling together the literature to support this point and flesh it out; most of the empirics he cites show the downsides to operation by fiat, and/or show the economic benefits of operation by involvement.
The second section is relatively long and straightforward. Freeland himself does a nice job of summarizing it: “…this section shows that this textbook image of GM's M-form is a myth. For most of its history, GM was characterized by a form of participative decentralization that was deliberately introduced in order to create divisional consent to corporate policies. During other periods, it reverted to administrative centralization in which the general office made both strategic and operating decisions. This centralization arose out of efficiency considerations, but its success or failure also rested on the social relations underlying consent.” (p 492)
Freeland uses new data to analyze GM: historical documents made available by the Hagley Museum. Previous work promoting this “M-Form myth” relied almost exclusively on data drawn from lawsuits.
He identifies four phases: 1. The Struggle for Participative Decentralization, 1924-33 (Sloan pushes hard for participative decentralization; good financial performance) 2. Administrative Centralization and Increasing Coercion, 1934-41 (depression eases, society in general values conformity, low cost and variety more, so managers accept administrative involvement on the tactical level; good financial performance) 3. Participative Decentralization Redefined, 1942-46 (Wartime production requires rapid, high quality strategic decisions and innovation, participative decentralization becomes a matter of law, Sloan et al move to retain some elements of administrative centralization; good financial performance) 4. From Consent to Co-optation, 1946-58 (GE’s president begins to treat Operating Planning Committee as a rubber stamp, begins to “rule by fiat”, consent decays; poor financial performance)
Freeland concludes that the true correlate of GE’s financial performance is the consent of the governed, not the M-Form structure of the organization. He runs back through his history of GE showing at each turn how Transaction Cost Economics either explains GE’s behavior as it relates to M-Form or fails to do so. It mostly fails (surprise, surprise). Freeland goes to some length to show that Sloan’s espoused and actual approach to structuring GE drew from the same theoretical assumption: consent is crucial (see p516, 2nd paragraph for an example).
Freeland makes one final theoretical flourish: when a firm diversifies, tradeoffs between cost and consent become more pronounced (p517-8); leaders will find it “…extremely difficult to identify, evaluate, and implement long-range competitive strategies without the direct participation and initiative of divisional management. It is in this context that the myth of the M-form becomes most dangerous.”