*305 The pooling of proposals and perspectives breaks down the distinctions between mutually ignorant specialists, each tempted to exploit the ignorance of the other. In simultaneous engineering and error correction by the five whys, for example, actors must teach each other important elements of their respective specialties and reveal the logic of their intentions in order to make themselves comprehensible at all. Where hierarchy assumes and produces the information asymmetries of mutual ignorance, learning by monitoring in effect creates an information‑symmetricizing machine in which actors must keep one another abreast of their intentions and capacities in order to advance the one and develop the other. The assignment of property rights in the sense of rights to residual control, accordingly, loses its centrality as a structuring and coordinating mechanism for the economy. If the supplier is continuously helping to modify the customer's equipment to make better use of the parts supplied (for example, by locating one of its own engineers at the latter's plant) and the customer reciprocates, then the collaborators are in effect jointly exercising residual control over the assets pooled in production; or, rather, they are in some sense partners or co‑owners.
Thus, in the emerging pragmatist economy, the necessity of collaboration means that the benefits to individuals or teams of holding up production will rarely outweigh the costs. From the older perspective, this result is as counterintuitive as the results of the new logic of efficiency. Indeed, so pervasive were fears of holdups in mass production that habitues of that world initially assumed that Japanese firms braved the risks of intimate collaboration only because of certain peculiarities of the Japanese setting, and hence that the Japanese organizational model was unlikely to thrive outside of Japan. According to one view, these peculiarities were cultural. Among the obligations of the Japanese to each other is the duty to forebear from exploiting the vulnerabilities of (Japanese) partners. The mutual expectation of such forbearance is trust, and in those few and fortunate places where historically there happens to exist a culture of trust, that culture, by definition, protects the dealings of those it embraces from the shadows of opportunism that normally darken transactions among the mutually vulnerable. [FN98] From another perspective, the peculiarities of the Japanese were institutional. By a (different) historical accident, the flexibility of the Japanese economy derived from economic structures that encouraged long‑term collaboration between the factors of production. Firms with lifetime employment that cannot cut costs by firing workers have incentives to retrain for new, productive tasks. Banks monitoring corporate performance are more likely to help a distressed debtor restructure if there is no prospect of recovering a loan by forcing liquidation. Economies that could not count on the loyal workforce and patient capital that these institutions produced could not build the other *306 collaborative institutions at the workplace or in customer‑supplier relations that depended on these as foundations. [FN99] Either way, the prospects were slight for the diffusion of the Japanese system outside its territory of origin and those few places with accidental similarities. Nowhere were the prospects slighter than in the United States, with our culture of individual self‑reliance, not mutual trust, labor‑market institutions favoring hire‑and‑fire strategies (with some corrections for seniority), and corporate monitoring by equity markets with a sharp eye for quarterly results.
But developments have confounded these expectations. Japanese‑derived organizational methods are now widely practiced in countries within the developing world [FN100] as diverse as Brazil [FN101] and Malaysia, [FN102] as well as in economically advanced countries as different as Ireland [FN103] and Germany. [FN104] In few places, indeed, have these methods been adopted with more innovative enthusiasm than in the United States, where they have spread from automobile, computer, and semiconductor industries that first came to grips with Japanese competitors adept at their use, to industries as varied as garments and meat processing. [FN105] It is difficult to measure the precise extent of the diffusion of the new methods. [FN106] But *307 diffusion in the United States has progressed far enough so that well‑informed proponents of the institutional‑limitations view now see our institutions as hospitable to the "Japanese" system. [FN107] Yet there is no evidence that our system of corporate governance has shifted power from corporate managers (with incentives to boost quarterly earnings) to banks (supposedly concerned about the firm's long‑term fiscal health) or otherwise encourages patience more than it used to. [FN108] Indeed, there is *308 much evidence that our industrial relations have become more "American" than before. [FN109] What accounts for the capacity of the American economy to adopt the innovations in the absence of the preconditions for doing so?
The short answer is that the construction of Japanese production systems does not suppose the existence of long‑term relations, because the system produces them in the course of its operation. Their firms' chief reservation would be fear of engaging an incompetent or unreliable partner. However, the information exchanges intrinsic to learning by monitoring would alert them to this danger before the consequences were ruinous. The same process that allows firms and their internal or external suppliers to agree on the definition of a subsystem or its components allows joint evaluation of target prices, target rates of return for collaborating partners, and a target rate of productivity improvement to be expressed in periodic price decreases. [FN110] Given the targets, simple sharing rules apportion the gains and losses from superior or inferior performance. For instance, the supplier typically keeps at least half the gains from innovations leading to productivity increases in excess of the target *309 rate, with the share declining as the innovation matures. Persistent incapacity to meet price reduction or product‑improvement targets despite continuing, joint efforts to surmount problems is often penalized by stepwise reduction in the supplier's share of the customer's total purchases of the affected product. [FN111] Suppliers that do exceptionally well in one or more rounds can then be delegated more extensive responsibility in the codesign of subsequent models; those that do exceptionally poorly will eventually be dropped from the pool of collaborators. Similarly, workers could be motivated by the prospect of acquiring general‑purpose skills (especially the ability to work in the new kind of teams) under conditions where managerial incapacity or bad faith is easily detected, and superior performance can be rewarded with more responsibility in teamwork.
Another more general way to put the point is to say that learning‑by‑ monitoring systems can be constructed by bootstrapping: the process of incremental change in which a favorable balance of risks and returns encourages first steps from many diverse starting points, and each move points the way down one of several paths that eventually leads to a roughly similar outcome. [FN112] Thus, a large firm can begin adopting the new methods simply by establishing various operating units as work teams or project groups responsible for achieving agreed‑upon goals, and rewarding or penalizing them (with, for instance, larger or smaller budgets) according to results. As these teams and groups choose, in turn, their collaborators from inside and outside the corporation, and adjust their internal organization accordingly, reorganization proceeds in ways that could not have been anticipated by central headquarters, yet are consistent with its (developing) purposes.
But accounts of innovations that permit local yet generalizable efficiency increases at little or no institutional risk and under the most varied *310 background conditions sound too good to be true, and, unqualified, they are too good to be true. We complete this synopsis of the new collaboration, therefore, by correcting omissions in the story so far and noting two limiting concerns in the diffusion of pragmatist institutions.
The first correction concerns the costs of shifting from the old world to the new. The shift to learning by monitoring, like any large change, produces winners and losers. But changing things piece by piece, eventually changing everything, makes it much more difficult to establish the distinction between winning and losing than in the case of punctual, once and for all changes. Abandoning the familiar system one step at a time, many will miss the attractive security that is stripped away long before they catch sight of the novel opportunities they may gain. To continue the automobile example, if the new‑van design team prefers an engine which the corporation's engine division cannot manufacture at an acceptable cost, the designers turn to an outside supplier. If this happens repeatedly, and the internal unit cannot replace the lost business with orders from outsiders, its survival is in doubt. If the central engineering division cannot offer designs for new plants as attractive as those furnished by specialized outside consultancies, it shrinks or is disbanded. The idea of having to compete again and again to create and maintain a fragile version of relations, once taken for granted as constitutive of the work setting itself, is terrifying to managers and blue collar employees alike.
Concerns of economic security aside, moreover, the reconceptualization of productive activity required to move from the old organization to the new is itself daunting. Old‑style managers are accustomed to making investment decisions on the assumption that most of the production system is fixed, and therefore that a good investment is one that returns large savings in production costs per dollar invested, assuming market conditions continue as in the past. For these managers, the notion of justifying projects by the cash flow they will likely generate in emergent markets, essentially without regard to the history of the firm's investments in related areas, is all but incomprehensible. [FN113] For these reasons, the new methods are typically greeted with skepticism and suspicion by those habituated to the routines and security of the hierarchical, integrated corporation, and they are implemented in established firms only under dire competitive threat.
The second limitation concerns the ability of firms using the new disciplines to set and implement large‑scale objectives. The issue becomes how to monitor the viability of whole lines of businesses or divisions, to make choices among incompatible, long‑term development goals, or to respond to abrupt threats or opportunities facing the corporation *311 as a whole. As described so far, search routines that detect the limits of habitual responses to design and operational problems are not necessarily well suited to answering these kinds of questions. Put as paradox, the potential limit is this: Only agents monitoring the (new) corporation day‑to‑day‑‑which is to say participating in its routine project selection and evaluation procedures‑‑could know enough of its highly decentralized operations to correct large errors or grasp transformative opportunities effectively. But just such agents are discredited when the errors come to light; traditional outside owners or stakeholders, whatever bundle of interests they are trying to maximize, simply cannot learn enough fast enough to be useful, as recent economic experience shows. Thus, as contingent corporate monitors, the Japanese main banks are supposed to take control of corporations they finance when sitting managers demonstrate incapacity. But during the current recession, the banks have not demonstrated much capacity to act on such contingencies, despite several decades of experience with the routines of highly decentralized decisionmaking. Firms under the banks' supervision have wasted free cash flow in American style. [FN114] German banks have had notorious difficulties monitoring firms with which they have had long‑term relations as those firms adopt new methods. [FN115] For its part, the American shareholder system of monitoring has never failed at the supervision of Japanese‑style corporations for the simple reason that it has not yet had the chance. Allowing decentralization to proceed produces improved performance insofar as there are gains from decentralization, but such improvement is no guarantee that permissive governance conditions are also suited to early detection of errors in the emergent system. From this perspective, the differences in the limitations between banks, with their view of the corporation as a community, and shareholders, with their vengeful selfishness, are less important than the similarities.
Eventual solutions to the governance problem might extend pragmatist principles to the higher level of monitoring by, for instance, constructing boards of directors or other bodies whose members are at once inside and outside the monitored unit. Venture capitalists with expert knowledge of the industries in which firms they finance operate, investment *312 bankers with knowledge of coinvestment possibilities in related industries, and managers of related divisions of the same or different companies are all examples of such figures. Their position gives them broad and constantly refreshed knowledge of the context within which the firm is operating. This knowledge presumably allows the identification of strategic opportunities or threats, as these emerge in the interplay of internal project selection, continuing benchmarking of performance measures, and external change. Strategy would become a joint result of product design and production. [FN116] But until such solutions or others are realized, there is an interregnum in the succession of governance institutions: For now, advancing forms of co‑ownership or partnership in the day‑to‑day use of resources coexist uneasily and disruptively with receding forms of exclusive property which, however vulnerable, are invoked whenever the new forms fail. So long as the interregnum lasts, the new economic institutions remain incomplete.
Yet for our purposes these costs and fragilities count as so many signs of the new institutions' vitality. Their rapid diffusion suggests that actors in the most diverse settings are sure enough of the limitations of organizations premised on bounded rationality and mass production to pay the enormous costs of adopting an alternative, and convinced enough of the robustness of a pragmatist alternative to adopt it in part, pending completion of its ultimate architecture. It is rare in history that prudence counsels such recklessness; the current massive and costly rejection of the known in favor of a promising but manifestly imperfect alternative recalls in form, if not yet in historical significance, those great innovations that exemplify and define our deepest ideas of collaboration, as the substitution of leaseholds and other forms of private property for feudal tenure, of representative democracy for monarchy, and of mass production for craft.
*313 But this similarity aside, the new pragmatist institutions are distinct in blurring the boundaries between public and private organizations which many of the earlier waves of institutional innovation helped establish. We are familiar with the private economy as, ideally, a realm of competing organizations, each under the control of an exclusive owner who decides how to maximize profits given prices for goods and services determined by independent, identically motivated decisions of other owners. In fact, the private sector is often a realm of huge, cooperating organizations under the control of managers responding first and foremost not to markets, but to the enticements and threats of one another, and only distantly responsible to absentee equity owners. The modern ideal of the democratic polity also differs from its practical instantiation. Organizations in the ideal democracy‑‑the state above all‑‑have exclusive jurisdiction in their respective spheres of action. They are controlled by the public, acting through the legislature in response to broad currents of public opinion as clarified in legislative deliberation. In fact, however, there is competition among them because their jurisdictions typically overlap, and they are often controlled by more or less entrenched bureaucratic interests, sometimes colluding, sometimes contending with shifting factions in the legislature or other oversight bodies. Thus, the actual public and private sectors look as much like each other as each resembles its idealization.
Viewed against this backdrop, learning by monitoring transforms economics and politics, assimilating each to the other, in part by introducing to the one a new variant of features normally associated with the ideal of its opposite, in part by restoring to both ideal aspects apparently sacrificed to reality long ago. Thus, we have seen that learning by monitoring "politicizes" the economy by introducing a kind of workplace democracy. Group deliberation in benchmarking, simultaneous engineering, and error detection become central to all decisions, from improvements in manufacturing process to redefinition of the measures and meaning of strategic success. In obliging disputatious yet collaborative evaluation of how diverse potential products will be used in life, of conflicting ways of making them, and of the contrasting measures of corporate and individual performance, learning by monitoring strips from economic decisionmaking the veiling technicity of maximization of profits given prices, and thus distributes authority from the "rulers" to the "people." At the same time, learning by monitoring also restores an aspect of the familiar economic ideal, "(re)privatizing" the corporation by exposing the internal units of mass‑production firms, through benchmarking, to the competition with external or market suppliers from which they have long been sheltered.
Next, we want to show that the pragmatist disciplines produce a correspondingly transformative assimilation and restoration when they are applied to political institutions in the form of democratic experimentalism. They "privatize" political institutions, not by establishing well‑ *314 defined owners‑‑for we have just seen that the very idea of exclusive ownership is losing its clear contours even in the private sector [FN117]‑‑ but rather by exposing them to the novel "market" of compelling, competitive benchmarking comparisons with the performance of like entities, and thus allowing for the substitution of superior service providers for inferior ones. They also "(re)politicize" political institutions by introducing a novel form of deliberation based on the diversity of practical activity, not the dispassionate homogeneity of those insulated from everyday experience. This form of deliberation, we will see, neither depends on consensus nor results in uniformity of view. Rather, it produces workable cooperation by continuously exploring different understandings of means and ends among those who use, provide, and are affected by government services. Just as the new pragmatist disciplines are creating novel partnerships in the governance of the economy, so their application to politics may result in new publics and new forms of public control of government institutions.
III. Democratic Experimentalism
The intuitive appeal of applying the pragmatist disciplines to democracy derives from these disciplines' potential to create a form of collective problem solving suited to the local diversity and volatility of problems that confound modern democracies, while maintaining the accountability of public officials and government essential to the very idea of constitutional order. In this and succeeding sections, we substantiate this intuition. We start by showing briefly how learning‑by‑monitoring solutions are well fitted to the characteristic problems of modern polities as these appear in the travails of post‑New Deal institutions. Then, we construct the organizational rudiments of local, or, rather, subnational, pragmatist government, by transposing to the public sphere the institutions of benchmarking, simultaneous engineering, and error detection. These transposed problem‑solving institutions, we argue, render public officials in each locale and the service providers they supervise accountable to the citizens, while affording the latter the chance to participate directly in practical deliberations concerning the matters that affect them. These same institutions, moreover, allow local jurisdictions to learn from one another. Arguments in any one jurisdiction, and the performance to which they lead, become considerations in the deliberation of similar jurisdictions. *315 To catalogue the novel features of these arrangements, we call the form of democracy thus created directly deliberative polyarchy. [FN118]
A. Good Government Under Conditions of Volatility and Diversity
A central lesson of the limitations of New Deal institutions is that effective government services and regulations must be continuously adapted and recombined to respond to diverse and changing local conditions, where local may mean municipal, county, state, or regional as the problem requires. This adaptability is just what the separate, centralized agencies of the New Deal, and the doctrines authorizing delegation of rulemaking power to them, lacked. The constant effort to adjust programs, regulations, and doctrines to changing circumstances has been the agencies' undoing. [FN119]
More precisely, and with all the advantages of hindsight, the lesson of New Deal distress is that the success of any one government program or regulation depends not only on its local adjustment, but also on the availability of other, equally well‑adjusted services and rules. Successful training for employability, for instance, must be carefully coordinated not only with developments in the local labor market, but also with the provision of day‑care and family‑support services (themselves a composite that must address problems as diverse as substance abuse, domestic violence, and foster care), as well as various kinds of social insurance that likewise reflect the conditions of local life. By the same token, comprehensive health, occupational, accident, and other forms of social insurance will be affordable in the long run only if accompanied and informed by local services that provide the insured with the information and means necessary to reduce the risks to which they are exposed (e.g., preventive medicine, occupational health and safety measures, and training for employability).
Looked at this way, effective government is first and foremost local government; local government itself is a complex service product composed of discrete programs so mutually dependent that difficulties or successes in one may suggest or require changes in the others, or in the connections among them. From this vantage point, the dilemma of government adaptability is just an instance of the general and apparently intractable design problem of continuous, mutual adjustment of parts and wholes to which simultaneous engineering now provides a solution.
But if local knowledge and simultaneous engineering are indispensable to government under diverse and volatile conditions, there is no reason to assume, and many reasons to doubt, that they are sufficient. Locales *316 may be diverse and changing, but they are not unique. To the extent that there are similarities in their current situations or the kinds of changes they face, the efficient search for large improvements to current practice, or for early warning that apparently promising alternatives are in fact dead ends, starts with the experience of units facing analogous problems. Just as discussion of the relation among programs and rules within a single locale reveals strengths and weaknesses concealed when each is considered in isolation, so comparison among individual programs' variant rules and methods of coordinating them allows each jurisdiction to see its viewpoints and its proposals in the light of alternatives articulated by the others. Looked at this way, of course, learning among locales is an instance of the systematic questioning of routines through the circumscribed examination of alternatives now formalized in benchmarking. Beginning with these analogies, we can easily imagine an open, federated structure for local government that could both encourage and respond to the changes‑‑not least in its own routines‑‑prompted by the new pragmatism.
B. Local Government on Pragmatic Lines: Directly Deliberative Polyarchy
The basal cell of this structure is a stylized institution, corresponding to the design team, that we will call the governance council. Nowhere fully established, this construct nonetheless draws together and connects as a whole crucial features of current, partial innovations in state and local government. Its core members are the public officials charged with specifying and organizing provision of the services required in the jurisdiction, and answerable to the jurisdiction's senior elected executive official; ex officio membership is accorded local officials of administrative agencies or other public entities that make rules or provide services relevant to these core activities.
The council's initial task is to characterize the goals of government in its various departments as informed on the basis of past practice, benchmarking examination of relevant experience elsewhere, and the simultaneous‑engineering proposals of council members from complementary service areas. Then, it chooses service providers to achieve those goals within a fixed period, and evaluates their performance and its own initial goals in the light of the experience as captured in the pragmatist disciplines. At the end of the period, if necessary, it redefines its goals and its own organization, suggests changes in rules under which it operates, and selects collaborators for the next round. Just as the design group is free to pick suppliers from outside the firm as well as inside, so, both to experiment at first and solve problems as they are identified, the governance council must be free periodically to choose service providers from among specialist government bureaucracies operating within its own or other jurisdictions, nonprofit corporations, or for‑profit firms. For the purpose of solving particular problems it may choose to federate with other jurisdictions like itself or delegate responsibility to more or