The Economist's Schumpeter column tackles the emergent love affair of business with team production:
Companies are abandoning functional silos and organising employees into cross-disciplinary teams that focus on particular products, problems or customers. These teams are gaining more power to run their own affairs. They are also spending more time working with each other rather than reporting upwards. Deloitte argues that a new organisational form is on the rise: a network of teams is replacing the conventional hierarchy. ...
The fashion for teams is driven by a sense that the old way of organising people is too rigid for both the modern marketplace and the expectations of employees. Technological innovation puts a premium on agility. John Chambers, chairman of Cisco, an electronics firm, says that “we compete against market transitions, not competitors. Product transitions used to take five or seven years; now they take one or two.” Digital technology also makes it easier for people to co-ordinate their activities without resorting to hierarchy. The “millennials” who will soon make up half the workforce in rich countries were reared from nursery school onwards to work in groups.
Pardon me for being somewhat skeptical. We have, after all, been here before. Back in the late 1990s I wrote several articles on the purported tension between hierarchy and team production, perhaps most notably in Bainbridge, Stephen M., Privately Ordered Participatory Management: An Organizational Failures Analysis; available at SSRN: http://ssrn.com/abstract=38600. Back then I concluded that:
According to conventional academic wisdom, perceptions of procedural justice are important to corporate efficiency. Employee voice promotes a sense of justice, increasing trust and commitment within the enterprise and thus productivity. Workers having a voice in decisions view their tasks as being part of a collaborative effort, rather than as just a job. In turn, this leads to enhanced job satisfaction, which, along with the more flexible work rules often associated with work teams, results in a greater intensity of effort from the firms workers and thus leads to a more efficient firm.
Although this view of participatory management has become nearly hegemonic, the academic literature nevertheless remains somewhat vague when it comes to explaining just why employee involvement should have these beneficial results. In contrast, my article presents a clear explanation of why some firms find employee involvement enhances productivity and, perhaps even more important, why it fails to do so in some firms. Despite the democratic rhetoric of employee involvement, participatory management in fact has done little to disturb the basic hierarchial structure of large corporations. Instead, it is simply an adaptive response to three significant problems created by the tendency in large firms towards excessive levels of hierarchy. First, large branching hierarchies themselves create informational inefficiencies. Second, informational asymmetries persist even under efficient hierarchical structures. Finally, excessive hierarchy impedes effective monitoring of employees. Participatory management facilitates the flow of information from the production level to senior management by creating a mechanism for by-passing mid-level managers, while also bringing to bear a variety of new pressures designed to deter shirking.
Accordingly, as I observed in Why a Board? Group Decisionmaking in Corporate Governance. Vanderbilt Law Review, Vol. 55, pp. 1-55, 2002. Available at SSRN: http://ssrn.com/abstract=266683:
Despite downsizing and the widespread adoption of employee involvement programs (such as quality circles), public corporations remain hierarchical institutions. To be sure, with the growth of team production, many firms are more accurately described as hierarchies of teams rather than of individuals. Yet they are hierarchies just the same.
Hierarchy persists because it remains a high survival value adaptive response to the transaction costs associated with organizing production within a firm. In particular, hierarchy is a very efficient *6 mechanism for information development and transmittal. Both new institutional economics and behavioral economics posit that decisionmakers are rational actors but that their cognitive powers are limited. Among other things, bounded rationality implies that decisionmakers can only gather so much information from so many inputs before being overloaded. In the corporate context, bounded rationality thus specifically implies that an individual manager can gather information about the productivity and capacities of only a limited number of inputs and, consequently, that no supervisor should receive such information from more than a few subordinates.
Branching hierarchies are an efficient adaptation to bounded rationality. They limit the span of control over which any individual manager has supervision to a small number of subordinates. Specifically, branching hierarchies put people into small groups, each member of which reports information to the same supervisor. That supervisor is likewise a member of a small group that reports to a superior and so on up to the top. Such an organizational system gets reliable information to the right decisionmaker more efficiently than any other organizational system. Not surprisingly, some form of branching hierarchy therefore tends to be found in most public corporations; they could not make decisions without it.
In addition to its information production and transmission functions, hierarchy also provides an important constraint on agency costs within the firm. Although agents ex post have strong incentives to shirk, ex ante they have equally strong incentives to agree to a corporate contract containing terms designed to prevent shirking. In any organization, however, the familiar triad of contracting problems—uncertainty, complexity, and opportunism—precludes the organization and its agents from entering into the complete contract necessary to prevent shirking by the latter. In large organizations, these transaction cost barriers to contracting are compounded by the equally familiar litany of collective action problems. Accordingly, organizations rely not on ex ante contracting but on ex post governance—creating mechanisms for detecting and punishing shirking. Specifically, managers of such organizations are tasked with monitoring the organization's members: management meters the marginal productivity of each member and responds as necessary to prevent shirking.
Have things really changed all that much?
Schumpeter cautions that:
A good rule of thumb is that as soon as generals and hospital administrators jump on a management bandwagon, it is time to ask questions. Leigh Thompson of Kellogg School of Management in Illinois warns that, “Teams are not always the answer—teams may provide insight, creativity and knowledge in a way that a person working independently cannot; but teamwork may also lead to confusion, delay and poor decision-making.” The late Richard Hackman of Harvard University once argued, “I have no question that when you have a team, the possibility exists that it will generate magic, producing something extraordinary…But don’t count on it.”
Hackman (who died in 2013) noted that teams are hampered by problems of co-ordination and motivation that chip away at the benefits of collaboration. High-flyers forced to work in teams may be undervalued and free-riders empowered. Groupthink may be unavoidable. In a study of 120 teams of senior executives, he discovered that less than 10% of their supposed members agreed on who exactly was on the team. If it is hard enough to define a team’s membership, agreeing on its purpose is harder still.
Indeed. In Why a Board, I reviewed at length the literature on group decision making. It is true that groups have advantages with respect to certain types of tasks: "groups are superior at evaluative tasks. ... Group decisionmaking presumably checks individual overconfidence by providing critical assessment and alternative viewpoints .... The proposition that group decisionmaking counteracts individual biases obviously overlaps with the claim that group decisionmaking is an adaptive response to bounded rationality. Numerous studies suggest that groups benefit from both pooling information and from providing opportunities for one member to correct another's errors."
But groups are less well suited to other types of tasks:
The old joke about the camel being a horse designed by a committee captures the valid empirical observation that individuals are often superior to groups when it comes to matters requiring creativity. Research on brainstorming as a decisionmaking process, for example, confirms that individuals working alone generate a greater number of ideas than do groups. Strikingly, this is especially true when the assigned task is “fanciful” rather than “realistic.”
Indeed, as Schumpeter suggests, groups have their own biases and idiosyncrasies that impede effective decision making. As I explained in Why a Board:
A widely cited example is the so-called risky shift phenomenon. Although we might assume that group decisionmaking has a moderating influence, social dilemma experiments demonstrate that groups actually make more extreme decisions than individuals. In early versions of these experiments, individual subjects were pretested by being presented with a story in which they were featured as the central characters. The story placed them in a familiar social setting and asked them to choose between two options, one of which was described as being the riskier of the two, but also as having a potentially higher return. Small groups were then presented with the same problem and asked to make a collective decision. Groups were significantly more likely to select the riskier option than individuals. Given that individuals tend to be risk averse but that shareholder interests often require risk-preferring decisions, the risky shift phenomenon seems useful on first blush. Unfortunately, later experiments demonstrated that group shifts to greater caution could also be induced. There seems to be a polarizing effect in group decisionmaking, so that post discussion consensus is more extreme than the individual pretest results.
The most significant group bias for our purposes, however, is the “groupthink” phenomenon. Highly cohesive groups with strong civility and cooperation norms value consensus more than they do a realistic appraisal of alternatives. In such groups, groupthink is an adaptive response to the stresses generated by challenges to group solidarity. To avoid those stresses, groups may strive for unanimity even at the expense of quality decisionmaking.
