The Organization of Work: A Comparative Institutional Framework

1. Assumptions 

Marglin contends that the nonexperimental naturelof the social sciences con- tributes to the continuing neglect of internal organization. Were that not the case, alternative inodes of organization, including egalitarian work modes, would be designed and tested experimentally (Marglin, 1974, pp. 33-34). While I agree that experimental testing of that kind has great merit, I submit that a great deal can be discovered about the efficacy of alternative work modes by an abstract assessment of their transactional properties. At the very least, a priori analysis of the transactional attributes of alternative modes should permit the empirical issues to be greatly delimited.

So that alternative modes will be on a parity in technological and locational respects, it will be useful first to specify the common manufacturing characteristics associated with each. One of the more serious problems with the work mode literature is that such assumptions are rarely made explicit The following assumptions will be maintained in this and the next two sections and, except where noted to the contrary, will apply across all modes:

  1. Specialized equipment, provided that it can be utilized at design capacity, facilitates low-cost pin manufacture. Nontrivial setup costs are incurred in putting the equipment in place.
  2. Workers acquire dexterity by repeated operations of the same kind, though this is subject to diminishing returns.
  3. It is economical, so as to save on transportation expense, that all pinmaking operations be completed at a common location, so that, the putting-out system excepted, all work is performed under one roof.
  1. The common building is leased and, whatever the station ownership and utilization arrangements, no problems arise with respect to building lease payments.
  2. Successive stages of manufacture are separable in the sense that placing a buffer inventory between them permits work at each stage to proceed independently of the other.
  3. The production line is balanced in the following very special sense: Work stations are designed such that, absent untoward events, a steady flow of intermediate product between stations is assured by placing a single, fully occupied worker at each station.
  1. Market transactions for intermediate product are very costly.
  2. The workers employed under each mode are a random sample of the technically qualified population of which they are a part.
  3. Replacement investment occurs routinely and investment for expansion purposes is ignored.

The first four assumptions are relatively uncontroversial. The fifth as- sumption (separability) means that differences among work modes turn on transactional rather than technological considerations. Coupling this with the one-man-each-station condition (assumption 6) effectively means that the tech- nology associated with the putting-out system is not inferior; rather, the same technology is feasible for and is common to all modes.

As noted, the one-man-each-station assumption is very special. It serves to concentrate attention on transaction cost issues, which have hitherto been neglected, and suppresses technological consideratiqps, the importance of which have previously been exaggerated. Redressing the imbalance by way of the one- man-each-station device scarcely yields a “representative” outcome. It is nevertheless noteworthy that the very same transaction cost attributes of work organization that this device serves to isolate also appear in the multiperson station context. The assumption will accordingly be retained throughout the chapter. Pat Hudson’s remarks regarding organization and technology, made in conjunction with her assessment of proto-industrialization, are instructive— “considerable economies in costs could be achieved . . . without technical change” (1981, p. 46). Indeed she asserts, and thereafter demonstrates, that “much early factory development occurred in order to achieve organizational economies and efficiencies and not according to technological dictates” (Hudson, 1981, p. 46).

The assumption that intermediate product markets work badly focuses attention on the transactional properties of internal organization. If market alternatives to internal exchange could be exercised at slight cost, choice among alternative internal modes becomes less important, since market relief can always be obtained when internal modes threaten to break down. Assumption 7 forecloses that possibility.

The assumption that the workers employed under each mode are a random sample of the population precludes the possibility that workers will match preferences toward work modes in a discriminating way. Thus although certain work modes may be competitively viable if they are staffed with workers with special attributes, that is foreclosed by the random assignment stipulation wherein all modes are assessed with respect to a common workforce.

Assumption 9 permits new investment issues to be set aside; attention is focused on the operating and adaptive attributes of alternative modes instead. That has two advantages. First, the investment properties of alternative ownership arrangements can be and have been investigated within the neoclassical framework. The studies of Vanek (1970), Meade (1972), and Furubotn (1976) all confirm that collective ownership models are beset with investment problems. Second, the operating and adaptive attributes of alternative work modes have been relatively neglected in the prior literature. Omitting investment from the performance attributes under scrutiny serves to compensate for that imbalance.

So much for the assumptions; I turn now to a description of alternative modes. Six different modes are described, first in ownership and then in contracting terms. Both for transaction cost purposes and for purposes of studying hierarchy, the latter is more basic. Ownership, however, is the more familiar way of describing work modes and/will be employed first.

2. Alternative Modes/ Ownership 

Three types of station ownership relations—entrepreneurial, collective ownership, and capitalist—with two variants within each will be considered.


Entrepreneurial modes are ones in which each station is owned and operated by a specialist.

  1. Putting-Out system. A merchant-coordinator supplies the raw materials, owns the work-in-process inventories, and makes contracts with the individual entrepreneurs, each of whom performs one of the basic operations at his home using his own equipment. Material is moved from station to station (home to home) in batches under the direction of the merchant- coordinator.

The Putting-Out system has been described by Landes as follows;

[M]erchant-manufacturers “put out” raw materials-raw wool, yarn, metal rods as the case might be-to dispersed cottage labor, to be worked up into finished or semifinished products. Sometimes the household was responsible for more t an one step in the production process: spinning and weaving were a typical combination. But the system was also compatible with the most refined division of labor, and in the cutlery manufacture of Solingen or Thiers or in the needle trade of Iserlohn, the manufacturing process was broken down into as many as a dozen stages, with each cottage shop specializing in one. Putting-Out was a major step on the path to industrial capitalism. For one thing, it brought industrial organization closer to the modem division between employers who own the capital and workers who sell their labor. To be sure, most domestic weavers owned their loom and nailers their forge. They were not, however, independent entrepreneurs selling their products in the open market; rather they were hirelings, generally tied to a particular employer, to whom they agreed to furnish a given amount of work at a price stipulated in advance. [Landes, 1966, p. 12]

  1. Stations are located side by side in a common facility. Intermediate product is transferred across stages according to contract. So as to avoid the need for supervision or continuous coordination, buffer inventories are introduced at each station. Subject to the condition that buffer inventories do not fall below prescribed levels, in which event penalties are assessed, each worker proceeds at his own pace.

Whether this mode was ever widely used is uncertain and perhaps doubtful. Thus although Landes (1966, p. 14) observes that the practice of “leasing space and power in a mill to individual artisans, each conducting his own enterprise” was common in nineteenth-century England, it is unclear whether intermediate product was traded among stations or if each station, was self- contained.

Hudson likewise observes that “the majority of early woolen mills were occupied and run, if not entirely financed, by small manufacturers … rather than by wealthy mercantile concerns” and explains this by “the fact that the size and cost of a competitive mill remained small in the woolen branch until well into the 19th century. More importantly, tenancy and multiple tenancy of mills was ubiquitous throughout the period” (1981, p. 48). Again it is unclear whether each tenant was self-contained or there was trade between stations. In principle, however, there could have been trade.

Moreover, it is useful to consider the Federated mode as an evolutionary development, even if only of a hypothetical kind. For one thing, it illustrates the use of comparative analysis of a microanalytic kind to investigate the properties of new forms of organization. Once an abstract mode has been described, its incentive and contracting properties, in relation to other modes, are relatively easy to establish. Additionally, the Federated mode has the attractive property that it preserves considerable worker autonomy.83 Egalitarian work relations are presumably favored as a consequence.


Work stations are here owned in common by the entire group of workers.

  1. Communal-emh. Although stations are owned in common, every man has a claim to the output associated with his own labors. So as to facilitate the acquisition of dexterity and economize on setup costs, each worker engages in batch process manufacture. The orderly movement of product is accomplished by having workers move between successive stations at prescribed intervals (hourly, daily, weekly, or whatever appears most appropriate), each bringing his own work-in-process inventory with him and selling his final product in the market.

The suffix “emh” is used to emphasize that this is an every-man-for- himself system. Thus although workers pool their resources with respect to the ownership of plant and equipment and orderly station moves are accomplished by calendar, there is no specialization among workers. Such a joining of common ownership with an every-man-for-himself rule is what Harold Demsetz (1967, p. 54) has described elsewhere as the communal mode. Unsurprisingly, the combination of community ownership with emh appropriability leads to mixed performance results. To conclude, however, that collective ownership is inferior to private ownership because of defects in the Communal-emh mode is unwarranted. If collective modes, such as the Peer Group, can be devised that have better properties than does Communal-emh, they presumably should be considered.9

  1. Peer Groups. The same ownership arrangement obtains as in the communal-emh mode, but workers are not compensated on the basis of their own product but are paid the average product of the group instead.10 Workers may rotate among stations or specialize at one or a few stations. Moreover, so as to avoid the need for full group discussion whenever an adaptation has to be made and/or to assure better coordination among the members with respect to woric breaks, variable rates of production, and the like, Peer Groups may elect. temporary “leaders,” who make operating—but not strategic—decisions on behalf of the group. It is important, however, that leadership rotate among group members if rigid hierarchical relations are to be avoided.11 Ernest Mandel’s (1968, p. 677) proposal for self-management “in which everybody will take a turn to carry out administrative work in which the differences between ‘director’ and ‘directed’ will be abolished’’ is in that spirit. The joining of a nonmarginal productivity sharing rule with democratic decisionmaking is what characterizes Peer Group organization.12


Inventories of all kinds (raw materials, intermediate product, finished goods) as well as plant and equipment are owned by a single party under capitalist modes.

  1. Inside Contracting. The Inside Contracting mode of organization has been succinctly described by Buttrick in the following way:

Under the system of inside contracting, the management of a firm provided floor space and machinery, supplied raw material and working capital, and arranged for the sale of the final product. The gap between raw material and finished product, however, was filled not by paid employees arranged in [a] descending hierarchy …but by [inside] contractors, to whom the production job was delegated. They hired their own employees, supervised the work process, and received a [negotiated] piece rate from the company. [Buttrick, 1952, pp. 201-2]

The Inside Contracting system permits a capitalist who has relatively little technical knowledge to employ his capital productively while limiting his involvement to negotiating contracts with inside contractors, inspecting and coordinating the flow of intermediate product, and taking responsibility sales.13 Howard Gospel observes that Inside Contracting was widely used in batch process systems in the nineteenth century (undated, p. 7), but was never employed on the railways or continuous process industries (p. 9). Robert Eccles (1981) contends that the construction industry is even now organized on Inside Contracting principles.

  1. Authority Relation: The Authority Relation mode involves capitalist ownership of equipment and inventories coupled with an employment relationship between capitalist and worker. The employment relation is, by design, an incomplete form of contracting. Flexibility is featured as the em- ployee stands ready to accept authority regarding work assignments provided only that the behavior called for falls within the “zone of acceptance” of the contract. Joining an organization under the Authority Relation mode thus entails an agreement “that within some limits (defined both explicitly and implicitly by the terms of the employment contract) [the employee] will accept as premises of his behavior orders and instructions supplied to him by the organization” (March and Simon, 1958, p. 90). Rather than enjoy the contractual autonomy of an inside contractor, who is subject to only very loose performance constraints (e.g. that minimum quality standards be met and that buffer inventories not fall below prescribed levels more than a certain percentage of the time), the worker now is subject o much more detailed supervision.

3. Alternative Modes/ Contracting 

Contractual differences of two kinds should be distinguished. The first and more important compares alternative modes in terms of their degree of reliance on contractual detail to coordinate production. That is the distinction emphasized here and in section 4. The second has reference to the bargaining relation between the contracting agents. That aspect is examined under 3.4, below.

The six alternative modes under examination jn this chapter differ signif- icantly in the degree to which they rely on comprehensive contracting. For three of the modes, contracting (and recontracting) is the exclusive basis by which product is exchanged and interfaces are brought into adjustment. For the other three modes contract is used to provide framework, which is subject to reshaping at the contract renewal interval. Within the context of that framework, however, day-to-day operations are governed by an administrative process. The two different styles of organization will be referred to as continuous contracting and periodic contracting, respectively.


Both types of entrepreneurial modes (Putting-Out and Federated) as well as the Inside Contracting mode rely extensively on contracting. The putter-out and the capitalist serve as the common contracting agent in the first and third instances while the workers in the Federated mode engage in bilateral contracts with the owners of predecessor and successor stations. A common .characteristic of contracting modes is that each worker maintains considerable autonomy and, once the terms of the contract are struck, lays claims to a distinct profit stream. Since the gains of one agent are frequently made at the expense of another, relations among the parties are of a highly calculative kind.

The problems with such contracting modes are of two kinds. First, can the requisite complex contract be described, negotiated, and enforced in a low-cost manner? Bounded rationality considerations preclude comprehensive contracting from being realized. Confronted with the infeasibility of such complete contracting, the hazards of incomplete contracting then have to be addressed.

Since bargaining relations between successive stations are necessarily of a small numbers kind, bilateral monopoly problems abound. To be sure, a long- term, recurring relationship between the parties is contemplated. Unrestrained, myopic subgoal pursuit is accordingly discouraged. But it is unrealistic to expect autonomous parties to adapt to unforeseen, hence unplanned, circumstances in a joint profit maximizing way without first settling their respective claims on profit streams through intensive, self-interested bargaining. Merely to transfer a transaction out of the market and organize it internally does not, without more, harmonize exchange. The prospect and actuality of such recurrent bargaining is a serious impediment to autonomous internal contracting work modes.


There is no exchange of intermediate product among members of Com- munal-emh firms, so there is little occasion for contracting under that mode. Ad hoc contracts might, however, be negotiated if workers were to become disabled, since work-in-process inventories would otherwise stand idle. Also, original investment, reinvestment, and maintenance agreements will have to be worked out. Although those are not trivial matters, the problems of recurring contracting that arise in connection with day-to-day operations in each of the contracting modes described previously do not appear.

Members of Peer Groups have even less need for contracting. Work left undone by a disabled worker would be completed by his associates. To be sure, membership affiliation and disaffiliation terms would have to be reached. But no bilateral contracting between successive stations on operating matters would occur. Democratic decision-making, effected by the rotating leader or by full group discussion, is used to bring station interfaces into adjustment.

Contracting under the Authority Relation is apt to be somewhat more complete, in that explicit and implicit understandings regarding the zone of acceptance of the employment relation (Barnard, 1938; Simon, 1957) need to be reached. Once agreement has been reached, however, this is an essentially noncontractual mode. Adaptations of an operating kind are made within the framework of that rather general contract, whereby boss and worker essen-tially agree to “tell and be told.” Strategic decisions affecting the overall configuration of the enterprise are mainly left to the boss’s discretion.

4. The Degree of Hierarchy 

The degree of hierarchy is usually assessed in decision-making respects. Where the responsibility for effecting adaptations is concentrated on one or a few agents, hierarchy is relatively great. Where instead adaptations are taken by individual agents or are subject to collective approval, hierarchy is slight. A less common but nonetheless useful way to characterize hierarchy is in contractual terms. If one or a few agents are responsive for negotiating all contracts, the contractual hierarchy is great. If instead each agent negotiates each interface separately, the contractual hierarchy is weak.84 Although there is a strong, positive rank correlation between the two ways of characterizing hierarchy for the work modes investigated here, the correlation is not perfect. What is perhaps more interesting is that ownership is imperfectly correlated with hierarchies of both kinds. Using E, Co, and Cap to denote entrepreneurial, collective, and capitalist modes respectively, and using braces to denote ties (or near ties), the rank ordering of modes from least to most hierarchical in contractual and decision-making respects is as follows:

There is no central contracting agent in the Federated, Communal-emh, or Peer Group modes of organization, so a contractual hierarchical relationship is altogether absent for them. By contrast, there is a central agent for the orther three modes. Although characterizing the hierarchical relation be-tween central agent and workers is not simple, a plausible case for the relations shown between Putting-Out, Inside Contracting, and the Authority Relation can be made in terms of bargaining strength of workers vis-à-vis the central agent at the contract renewal interval. That varies with ( 1 ) the extent to which workers have acquired firm specific skills and knowledge, (2) collective organization among workers, and (3) physical asset ownership.

Skill acquisition is the same under all three central agent modes, since each involves specialization in identical degree. Collective organization may be slightly stronger under the Authority Relation, since workers here are less autonomous than under Putting-Out (where they are dispersed) and Inside Contracting (where they appropriate separate profit streams). Physical assets are owned by each worker under Putting-Out, but the central agent owns the stations in both instances under the Authority Relation and Inside Contracting. The upshot is that the contractual hierarchy is weak for Putting-Out, while the Authority Relation and Inside Contracting are somewhat stronger in contractual hierarchy respects.

Consider now the decision-making hierarchy. There is no command relation whatsoever between the members of the Federated and Communal- emh modes. The former is governed by rules and bilateral contractual relations; the latter is governed by rules and democratic decision-making. A relatively weak command relation exists for Inside Contracting and the Putting-Out modes. The central agent to the contracts can appeal to the workers to adapt in coordinated ways to changed circumstances, but the contracts govern as responsibility for operating matters has been extensively delegated. Thus bargaining and bribes may be needed if interim changes favored by the central agent are to be effected. The,Peer Group acknowledges the benefits of a command structure by designating a leader to coordinate day-to-day affairs. The leadership position turns over regularly, however, and strategic decisions are reached only after a full group discussion. Democratic decision-making effectively prevails. The Authority Relation posits at the outset that a superior-subordinate relation will govern in both operating and strategic respects. To be sure, the zone of acceptance of the employment relation, within which workers will accept orders without resistance, is limited by formal and informal agreement. But a command hierarchy is a prominent feature of the Authority Relation.

Although capitalist modes are more hierarchical than collective ownership modes from a contractual point of view, the more critical hierarchy for performance purposes is the decision-making hierarchy. The observed relation between ownership and hierarchy is very weak in decision-making respects. The least hierarchical modes, Federated and Communal-emh, are of different ownership kinds (entrepreneurial and collective ownership, respec- tively) The Peer Group, Putting-Out, and Inside Contracting modes have intermediate degrees of hierarchy, and each is from a different ownership Class. Although the most hierarchical decision-making mode is the capitalist mode, the next strongest command hierarchy features collective ownership.

Source: Williamson Oliver E. (1998), The Economic Institutions of Capitalism, Free Press; Illustrated edition.

Leave a Reply

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