The previous chapter showed that there is a productivity rationale for integrated production within a market setting based on specialisation through the division of labour. This conclusion is in line with the extant literature on the firm, from Ronald Coase’s ‘The Nature of the Firm’ and on, in which the firm is seen as a means to overcome specific shortcomings in the market. However, our discussion indicated that innovative production processes are distinct and integrated by default as there is initially no market support for the new production structure. As the specific ‘internal’ allocation of resources that makes up the productive innovation constitutes a novel use of specialisation, broadly perceived, those involved in the implementation cannot rely on the market for core functions. This means that integration is not something the actor chooses for overcoming a specific problem in the market (such as transaction costs, control of valuable resources or knowledge, and so on), but is an unchosen effect of implementing a new production structure. The firm is here not a means but an effect, and is not directly subject to a choice of organisational form. Productive innovation through the introduction of intensive specialisation always takes place outside the present extent of the market.
This conclusion turns the conventional logic of economic organising as a means for market actors to deal with and ultimately overcome market imperfections or failures on its head. The literature on organisational economics and the theory of the firm tends to begin with the market and hypothesises that integration is a special phenomenon in it. It is an assumed deviation from standard price mechanism coordination that is intended to solve a specific problem or shortcoming inherent in the market. The argument, then, is that market actors choose to form a firm because of its specific problem-solving property, such as transaction cost avoidance. Stating this type of rationale for the firm does not, however, explain how it, in contrast to the market, is able to produce this exploitable property. As this problem-overcoming property is core to how we recognize and can understand the firm, this missing piece of the puzzle – how it accomplishes what it is perceived to do – ties into what the firm in reality is and therefore also how it can be explained. This may be a reason why individual Coasean questions, and foremost among them the one about delineating the firm’s boundaries or where firm ends and market begins, have received much scholarly attention. Yet despite this vast literature, we do not seem to be much closer to answering the questions than when they were first posed.
Part of the reason is the tautological nature of the argument for the firm as a choice of governance. In the case of Coase’s transaction cost explanation, the firm is simply defined as the type of production organising that supersedes the costly price mechanism by relying on directional authority. But defining the firm as the transaction cost saving device it is sought out to be does not escape the fundamental problem in finding satisfying answers to the Coasean questions of the firm’s rationale, boundaries, and internal organisation. Coase himself was able to answer the question about the firm’s boundaries by simply asserting the firm’s rationale and by assuming its internal organisation is very similar (if not identical) to the market’s allocation of resources. The Coasean firm integrates individual transactions until the marginal net saving on transaction costs (the firm’s asserted rationale) is zero. The boundary is therefore set and then continuously adjusted by the manager dealing at the transactional margin through integrating more or fewer transactions when the cost structure fluctuates. Indeed, it is ‘always possible to revert to the open market’, writes Coase, if the cost of management becomes higher than the cost of transacting in the market.
The logic of the Coasean transaction cost story is internally consistent, but has ambiguous implications because of its reliance on assertions about the institutional structure of production and definitions formulated to support the wanted outcome. Coase’s assumption of an efficient market allocation of resources subject to costs of transacting that can be avoided through integration in firms, which aim to ‘reproduce’ market allocation internally by non-market means, is puzzling and raises quite a few questions that are still to be answered in the literature. Treating the firm as an outcome or result, as we do here, rather than a means avoids the ambiguity inherent in the literature by not begging the question. This also has implications for how we perceive of the firm’s internal structure.
The Firm as an Authority Relation
The literature since Coase asserts and depends on what Herbert A. Simon called the ‘authority relation’, a hierarchical power relationship that allows for low-cost direction of resources within the firm. This assumption has been subject to heavy criticism, notably in an article by Armen A. Alchian and Harold Demsetz,[i] but remains core to the definition of the firm in the literature. Alchian and Demsetz argued that the firm ‘has no power of fiat, no authority, no disciplinary action any different in the slightest degree from ordinary market contracting between any two people’. There is no real difference between a manager firing an employee and a consumer firing his grocer. In both cases, the parties have only two means to ‘punish’ an exchange partner, either ‘by withholding future business or by seeking redress in the courts for any failure to honor our exchange agreement’. Their conclusion, which I have substantiated elsewhere,[ii] is that ‘To speak of managing, directing, or assigning workers to various tasks is a deceptive way of noting that the employer continually is involved in renegotiation of contracts on terms that must be acceptable to both parties’.[iii]
Alchian and Demsetz’s view is that, economically speaking, a contract is a contract. Whether we prefer to call a specific type of contractual relationship ‘market exchange’ or ‘employment’ is of no relevance to the economic implications of that contract: its function is to equally bind parties to what was voluntarily agreed, presumably for their mutual expected benefit, when the contract was set up. A manager has no power to force a future employee that the consumer does not have vis-à-vis a grocer. This may seem unintuitive considering how we commonly perceive of the firm as a hierarchy and that we as employees ‘follow orders’ from higher-ups in the workplace. But this perception is unfounded. An employer (purchaser of labour) has no other means of punishment than the employee (seller of labour), as Alchian and Demsetz point out. This is easily illustrated by considering as example a specific long-term contract establishing the terms for exchange of services for payment. Whether we call the buyer of these services a customer or employee, and whether the seller is called supplier or employee, does not change the contract or its implications. The effect is the same as the contract is the same: it sets the terms for the exchange. What may differ depending on what we call the contract is the legally instituted obligations of the parties by decree of government, which indeed have economic implications. But this does not amount to an economic argument since it does not address how and whether the nature of the firm as an economic phenomenon, as opposed to the legal firm, provides a basis for internal authority. Calling exchange contracts between buyers and sellers of services within the firm ‘employment’ contracts, while similar or identical contracts with external parties are ‘market’ contracts, does not change the economic reality of these contracts or their implications.
The organisational economics literature remains vague on the causes of hierarchical authority within firms but relies on the conception of the firm as hierarchy to theoretically separate within-firm transactions from undirected market exchange. As in Coase’s original analysis of the nature of the firm, the authority relation establishes the market-firm boundary that in turn makes possible the comparative institutional analysis of organisations contra markets. The argument changes, however, if we see the firm as the result of implementing an innovative, specialised production process outside the limits of the market. The boundary question is of little concern here, since the firm is naturally separated from the market it is embedded in by utilising a different degree and kind of division of labour. The firm consists of the coordinated production process that is established outside the reach of the regular market’s price mechanism and thus is unaffected by its high-powered incentives. It is recognised as an ‘island of specialisation’, an intensification of specialisation beyond the productive powers of the division of labour already established and exploited in the market. This means there is indeed a contrast between firm and market, but also that they are interdependent. Karl Marx perspicaciously made this point, building on Adam Smith’s insight about the division of labour, that specialisation in the market and the firm ‘differ not only in degree, but also in kind’.[iv] He further noted that the firm’s comparatively intensive division of labour depends on the market’s:
division of labour in manufacture demands, that a division of labour in society at large should previously have attained a certain degree of development. Inversely, the former division reacts upon and developes and multiplies the latter. Simultaneously, with the differentiation of the instruments of labour, the industries that produce these instruments, become more and more differentiated.[v]
This is very similar to what we established in the previous chapter, where productive innovations are implemented within the institutional structure of production of the market and challenge the status quo. The question here, however, is whether this implies authority and formal hierarchy, as is maintained by both Marx and the contemporary theories of the firm. To Marx, the intensive kind of specialisation established within the firm by increasing the division of labour ultimately serves to both increase the productivity of and alienate labour workers involved in the production process by separating them from the fruits of their labour. This alienation and the resulting lack of knowledge of the true worth of their exerted labour makes it possible for capital owners to extract the labourers’ produced surplus value. Herein lies the Marxian exploitation dimension of capitalist production and the oppression of the working class.
To the contemporary theory of the firm, the cause of a within-firm authority relation is identified as one or a combination of several factors: the particular nature of the employment contract as open-ended and long-term,[vi] that courts of law rely on the forbearance doctrine with respect to firms’ internal affairs,[vii] or a rejection of the reciprocal nature of contracting. Each of these factors is curious for organisational economics. The first one seems to provide a basis for ‘authority’ through open-endedness. The contract specifies payment for the buyer but lacks specific terms for what the seller is to supply for this payment. It thereby establishes the buyer’s right to request specific services, within established limits, at will. But this is not a distinct property of the employment contract, but exists in varying degrees also for market contracts. Examples of open-ended non-employment contracts include the sale of services like consulting, talent resourcing, staffing solutions, and spa resorts. The ‘authority’ of the buyer of consulting or a weekend stay at a spa resort is never considered an economic problem with governance implications. What makes the employment contract distinct is the legal implications.
The first two factors, the employment contract and the forbearance doctrine, are equally based on the legal aspect of firm organizing, which suggests that the firm is primarily a ‘legal fiction’ and not an economic phenomenon. The object of study is therefore an empirical phenomenon with legal causes, much like the effects of regulation are indeed economic whereas the regulations themselves are not.
The third factor asserts that there is an underlying power dimension in all types of relationships, including voluntary market exchange, that skews the outcome of negotiation (primarily between employer and employee). Contracts only establish this pre-existing power in a formal and therefore enforceable authority relation, which serves to strengthen the stronger party. Jeffrey Pfeffer makes this point using remarkably Marxian language, stating that ‘power is, first of all, a structural phenomenon, created by the division of labor and departmentation that characterize the specific organization’.[viii]
Legal matters are of no concern here, since we attempt to explain and understand the economic function of integrating production in a firm. But the Marxian argument needs to be taken seriously, since it addresses the division of labour in a sense that is similar to the discussion above. Indeed, Marx identifies firms as islands of specialisation that through artificially increased density allow for a more intensive division of labour and therefore improved productivity. His argument is made more explicit in recent literature adopting the resource dependence view, which holds that resource heterogeneity and specificity produce dependence that gives rise to or increases social power. To use Marxist-inspired terminology, the social power of the capitalists is derived from their capital ownership combined with capital-deficient workers’ need tosell their labour for income; the latter is dependent on the former’s willingness to buy for their survival, and this translates to capitalists’ social power over workers. This argument appears valid when adopting a static view of the economy and a situation where productive innovation has already been implemented and proven successful. If we instead consider a dynamic view, as we have in previous chapters, the logic fails.
Consider again the example from chapter 4, in which production of some consumption good is carried out in parallel processes with similar if not fully standardised stages. The implementation of specialisation-based innovation can only take place outside the limits of the market, as we have already seen, and those involved are therefore separated from the market in a real sense: the intermediate goods they produce within the new sub process are not saleable. Their mutual specialisation creates a chain of bilateral monopoly relationships (interdependence), which is the strongest possible form of resource dependence. The success of the innovated sub process depends on being compatible with the market setting and therefore the previously established market production processes, without which it would suffer incompleteness and fail. As the incompleteness argument is equally applicable to the new tasks carried out within the sub process as it is for the entire sub process’ placement within the larger context of market production, any dependence and therefore power is reciprocal. The threat of hold-up from any party acting within the newly established sub process is not credible, since the result is that the whole process suffers incompleteness. There are no rents to extract for any individual participant in a production process where the parts are serially interdependent and the compensation of each is dependent on completion. Only where there is a limited market, and therefore some substitutability, can opportunism be a real hazard in production. For an implemented productive innovation, losses are incurred at any type of failure at any point in the sub process and borne by all involved. Profits are earned only when the entire process is completed. The resource dependence argument for authority does not apply when production is organised horizontally through the division of labour and splitting of tasks, since all parties are dependent on the completion of the entire process.
Implementation as Discovery of the Process
We must also consider the situation where production within a novel ‘island of specialisation’ is established top-down and therefore implemented vertically (as a hierarchy). Resource dependence can here emerge in the Marxian form of the capital owner’s initial investment to realize the production process (ownership issues will be discussed at length in chapter 7) or through the access to and control of knowledge. Both cases suggest that a central authority, perhaps the innovator, enjoys initial control of the project and its implementation, and then invites labourers to take part in the new production process. It is likely that the idea from which the innovation springs is formulated by one or a few individuals, who may not constitute the entire team needed to implement it and therefore assume a leadership position. The founders are in this sense the initial owners of the enterprise, who employ others for carrying out specific tasks. The question is if this type of leadership, whether based on ownership or superior knowledge, suggests organisational hierarchy with authoritative or directional power.
Let us revisit the example from chapter 4 where the innovative sub process o21–o22 is implemented to replace the standard market stage o2. But rather than l3 introducing the idea to l2 and then collaborate on the implementation effort, l3 secretly develops blueprints of the process and intends to keep it hidden from all others to reap the benefits for as long as possible. The situation is different than previously discussed because l3’s control of knowledge of the process suggests an advantage towards l2, who will be invited to contribute labour services to complete the sub process, as well as others. By controlling the knowledge and only sharing necessary bits of it to l2, it is conceivable that l3 could maintain this informational advantage for some extended period of time and therefore, possibly, benefit from this position of ‘power’. But this conclusion overlooks the fact that the sub process does not yet exist.
No matter the technical expertise of an innovator, a novel production process is unlikely to be properly optimized until it is fully implemented. It is theoretically possible to approximate technical efficiency in the production of capital goods to support an imagined innovative production process, but in reality any productive innovation is still subject to unknown parameters and unforeseen events. Knowledge of this type is unknowable as it does not yet exist – it is generated through the implementation process. This is even more so where the innovation, as in this case, depends on the division of labour, which means first implementation and then completion of the process depends on the craftsmanship of those involved. In lack of perfect foresight and perfect information, it is impossible to accurately determine all details required to carry out the yet to be implemented process. The more advanced the imagined process is, the less unlikely it is that everything can be foreseen and considered. As what we are dealing with here is innovation that uses intensive specialisation through the division of labour, the imagined production structure is always more advanced than what already exists in the market in the sense that it includes more distinct stages (is more roundabout). Only in very rudimentary production can technical efficiency in a productive innovation be accurately approximated, and then only where its originality is so limited that the innovator can rely fully on production experience that has already been established through trial and error in the market. That is, on specific knowledge that has already been generated and is fully available.
For any productive innovation with a degree of originality, which is suggested by the very term innovation, its implementation constitutes a discovery process. Kinks and errors need to be worked out, and some of the problems may not be discoverable until the process is tried empirically through scaled production. Many innovative production processes look promising, but turn out to be failures when tried practically. In addition to trying out the innovation technology, the process also needs to be aligned with the specific expertise and craftsmanship of those invited or hired to carry out the specific tasks. In our example, l2 is employed to carry out a single task in the innovative sub process o21–o22 originally imagined and controlled by l3. Whereas implementing the sub process will allow for the discovery of possible errors, the introduction of l2 to the specific task will reveal new knowledge about how it could be carried out. It is likely that this knowledge is revealed specifically to the factor carrying out the task, which provides l2 with specific knowledge necessary for improving the process.
Also, where the task has not been designed specifically with (full knowledge of the expertise of) l2 in mind, the task itself may need to be adjusted in order to take full advantage of l2’s specific skill set. This, in turn, could effectuate changes to the intermediate good between the tasks o21 and o22 or shift the boundary between the tasks altogether. If we hypothesise that l3 intends to eventually hire a second labourer to carry out the other task, the process may again change as new knowledge is revealed in the intersection of the task and the new labourer’s experience and skill set. Due to strict interdependence, any change to a task part of the process can cause changes to intermediate goods and therefore other tasks. Implementation entails discovery both at the process level and for the individual tasks (and the intermediate products between them).
This is highly problematic for an innovative entrepreneur l3 seeking to control specific knowledge about the process, since the risk of errors, incompatibilities, and inefficiency increases as the specific information available to labourers is limited and fragmented. Implementing a new degree and kind of division of labour where those carrying out the specific tasks are provided only fragmented knowledge suggests the necessary adjustments, optimizations and solutions to cross-task errors – including the detailed monitoring necessary to identify them – falls on l3. This suggests the process is burdened with significant costs that could easily be avoided under full disclosure and that may put the whole project at risk. Attempts to control knowledge should in any case tend to fail eventually, and will thus be ephemeral, since improvements, adjustments, problem-solving and other changes to the tasks themselves reveal knowledge about the overall process. Observant labour workers should soon be able to figure out specific information about adjacent stages in the process that has been kept from them. They even have an incentive to learn about the process as this helps them carry out their specific tasks more effectively. If workers in the process interact, which at a minimum should take place through the delivery of intermediate goods, information about the entire process would soon be revealed in its entirety.
Even if technical efficiency could be approximated, we saw in the previous chapter that the implementation of a new production process takes place outside the extent of the existing market. This means the process, even after having been implemented, can suffer from allocative inefficiency. Regardless of whether this is the case, there is no way of finding out until market prices for the individual stages are available. This will only be the case when the ‘island’ is located within the extent of the market and therefore can no longer be considered an island.
These problems can be mitigated through full disclosure within the innovative enterprise, which would make implementation a coordinative endeavour and thereby take advantage of the experiences, skills, and knowledge of all involved. This would also minimize the chance of failure, which should be a strong incentive for the innovator to share information.
We should also consider the means by which the innovator entices workers to join the enterprise. To attract the type of skills and experience necessary to support the novel process, wages have to be offered that exceed those already offered in the labour market. Even so, as the process has not been implemented, what is offered to l2 is in effect an unsubstantiated promise since l3 keeps information about the process secret. l2 might be hesitant to accept such a promise at face value, even if l3 proves to have sufficient funds to cover wages for a foreseeable future, and therefore ask for even higher payment to commit. Undertaking a new job and skill set is an investment for both the labourer and his employer. The higher wage would force l3 to initially dedicate even more capital to the project, which further increases the cost burden on the enterprise and therefore escalates the risk of failure. Also, since any returns will be available only upon the completion of the process, assuming it turns out to be a profitable undertaking, there may be significant delay before invested funds are repaid. This too is a cost that falls on the original innovator as sole principal.
Yet even ownership of the process does not imply authority in a meaningful sense. Implementation requires training and the development of specific skills by those instructed to carry out the tasks. As there is no labour market for these skills, they are difficult if not impossibleto replace. This makes them highlyvaluable specifically for the innovative process, especially considering how the failure of any task implies incompleteness and failure to earn returns. It is therefore as likely that employed workers can hold up their employer as the other way around. Interdependence and the threat of incompleteness keeps opportunism at bay both horizontally and vertically. There are no grounds for authority, at least not unidirectional and hierarchical, since credible and costly threats can be made both ways.
[i] A. A. Alchian and H. Demsetz, ‘Production, Information Costs and Economic Organization’, American Economic Review, 62:5 (1972), pp. 777-795.
[ii] P. L. Bylund, ‘The Firm and the Authority Relation: Hierarchy vs. Organization’, in Nell (ed) The Firm and the Authority Relation: Hierarchy vs. Organization (New York: Palgrave Macmillan, 2014), pp. 97-120.
[iii] Alchian, et al., ‘Production, Information Costs and Economic Organization’, p. 777.
[iv] K. Marx, Capital: A Critique of Political Economy (1867) (New York: Charles H. Kerr & Company, 1906), p. 389.
[v] Marx, Capital: A Critique of Political Economy, pp. 387-388.
[vi] H. A. Simon, ‘A Formal Theory of the Employment Relationship’, Econometrica: Journal of the Econometric Society, 19:3 (1951), pp. 293-305.
[vii] O. E. Williamson, ‘Comparative economic organization: The analysis of discrete structural alternatives’, Administrative Science Quarterly, 36:2 (1991), pp. 269-296.
[viii] J. Pfeffer, Power in Organizations (Boston, MA: Pitman, 1981), p. 4.