2. Fictions and aggregates in the theory of the firm

Since Coase's (Reference Coase1937: 393) statement of the nature of the firm as a contractual ‘system of relationships’, the contract, or the transaction as Williamson (Reference Williamson1975) later emphasized, has become the basic unit of analysis in many branches of economic theory related to human interaction and organization. In this literature, the term ‘contract’ is broadly understood as any voluntary agreement involving some sort of exchange, including legally enforceable relations and both incomplete and implicit contracts. The essence of all forms of economic organization is said to be contractual (Alchian and Demsetz, Reference Alchian and Demsetz1972: 794; Jensen and Meckling, Reference Jensen and Meckling1976: 310). This means that the firm is a voluntary coalition or team of individuals, and a set or collection of contracts between various owners of productive resources (Fama, Reference Fama1980: 289; Alchian, Reference Alchian1984: 34; Alchian and Woodward, Reference Alchian and Woodward1988: 66).

On this view, the difference between the firm and the market is a difference in degree and not a difference in kind. Not surprisingly, Jensen and Meckling (Reference Jensen and Meckling1976: 311) state that ‘it makes little or no sense to try to distinguish those things that are “inside” the firm (or any other organization) from those things that are “outside” of it’. Since everything is contractual, ‘the boundary of the firm is fuzzy: a bright line distinguishing “inside” and “outside” is missing’ (Alchian and Woodward, Reference Alchian and Woodward1988: 76). It follows, as Barzel (Reference Barzel1989: 54–55) argues, that the exercise of ‘classifying transactions into those in the market and those within the firm . . . is not very illuminating’. Insofar as they see the world as a ‘contractual continuum’, these theorists agree that the concept of firm boundaries is next to useless.

Frequently, the next step is to deny the usefulness of the concept of the firm itself. Many argue that definitions of the firm should be abandoned in favor of concepts of ‘firm-likeness’ (Demsetz, Reference Demsetz1988: 155), ‘firmness’ (Lamoreaux, Reference Lamoreaux1998: 70), or ‘firmishness’ (Klein and Coffee, Reference Klein and Coffee2002: 20) characterizing various sets of contractual relations. Alchian (Reference Alchian1984: 46) writes: ‘the “firm” loses some analytical significance as attention is focused on competition among individuals, their particular resources and on types of contractual relationships’. Cheung's (Reference Cheung1983) classic paper on the contractual nature of the firm is equally explicit: ‘what is or is not a firm is immaterial’ (p. 10); in general, ‘it is futile to press the issue of what is or is not a firm’ (p. 18); better, ‘we do not exactly know what the firm is – nor is it vital to know’ (p. 3). The word ‘firm’ has ‘little substantive content’ (Jensen and Meckling, Reference Jensen and Meckling1976: 311) and is ‘simply a shorthand description of a way to organize activities under contractual arrangements’ (Cheung, Reference Cheung1983: 3).

Although such statements suggest that economic analysis should look at individuals interacting as teams, given a distribution of opportunity costs and incentive compatibility constraints, these economists do provide a definition of the firm. The firm is defined as a ‘nexus of contracts’ (Jensen and Meckling, Reference Jensen and Meckling1976: 310–311; Fama, Reference Fama1980: 293; Fama and Jensen, Reference Fama and Jensen1983: 302; Demsetz, Reference Demsetz1988: 154). This expression means that the firm is the only common signatory of a set of contracts. How can a non-defined, non-bounded thing which is simply a word be such a common signatory of a set of contracts? Surely, this could only be done in imagination. Proponents of the nexus of contracts view provide a straightforward answer.

Jensen and Meckling (Reference Jensen and Meckling1976: 310) claim that firms are ‘legal fictions’, meaning that firms are legally treated as ‘persons’ signing the various contracts involved in the nexus. This is so only in imagination since real people do the actual signing in the name of the firm. Firms do not act, have objective functions, or responsibilities. Only individuals can act, whether responsibly or not. Jensen and Meckling (Reference Jensen and Meckling1976: 311, emphasis in original) warn us that ‘the personalization (sic.) of the firm . . . is seriously misleading. The firm is not an individual.’ One should be wary of the ‘reification illusion’ (Klein and Coffee, Reference Klein and Coffee2002: 110) that leads to the common mistake of holding firms responsible for things that firms as such cannot commit. As Easterbrook and Fischel (Reference Easterbrook and Fischel1985: 89) explain, ‘the liability of “the corporation” is limited by the fact that the corporation is not real’. What holds for liabilities also holds for rights which ‘apply to individuals [and] not to imaginary “entities”’ (Hessen, Reference Hessen1979: 1328).

It is important to notice the difference between a ‘set of contracts’ and a ‘nexus of contracts’. If individuals in a coalition contract with each other, it is easy to see that there are considerably more contracts than in a nexus in which parties do not contract with each other but do so with a central party, that is the legal fiction of the corporate person.Footnote ² Since nexus of contracts theorists often use the two expressions interchangeably (e.g., Jensen and Meckling, Reference Jensen and Meckling1976; Fama, Reference Fama1980), it seems that when they speak of a nexus of contracts they are actually referring to a set of individuals and contracts. The latter are real, while the former is a fiction.

The nexus of contracts setup is a ‘fictionalist’ view of the firm involving an outright denial of the firm's existence based on three ontological claims. The first is that firms have vague boundaries or that firms are not determinate. The second is that firms are simply shorthand expressions as opposed to real entities. The third, that firms are legal fictions, is supposed to reinforce the claim that firms are ontological fictions. However, the links between the three claims are disputable. The step from vague boundaries to denial of existence is a non sequitur (Hodgson, Reference Hodgson2002: 53; Dejnožka, Reference Dejnožka2006: 74ff). Similarly, the step from legal fiction to ontological fiction is also a non sequitur. Legal fictions need not be ontological fictions (Dejnožka, Reference Dejnožka2006: 4).

One interesting aspect of the ‘collection of assets’ view is that it was developed in reaction to the nexus of contracts view. Hart (Reference Hart1989: 1764) writes: ‘the nexus of contracts approach does less to resolve the questions of what a firm is than to shift the terms of the debate’. For Hart, the nexus of contracts view is question-begging, and its dismissal of firm boundaries is simply unreasonable. Moore (Reference Moore1992: 494, n.2) sums up this position: ‘One can . . . sidestep the issue entirely, by arguing that everything is contractual, and that firms are a mirage . . . This is the view proposed by Jensen and Meckling . . . But if firms are a mirage, it is difficult to explain the enormous resources that firms expend merging and breaking up.’

Addressing this problem, Grossman and Hart (Reference Grossman and Hart1986: 693) ‘define the firm to consist of those assets that it owns or over which it has control’. The advantage of this definition is that one can ascertain firm boundaries in a straightforward manner. When looking for the firm in a contractual continuum world, the rule is simple: ‘identify a firm with the assets it possesses’ (Hart and Moore, Reference Hart and Moore1990: 1120). Property rights thus provide a simple account of firm boundaries: assets that belong to the firm are ‘inside’ firm boundaries, while assets that do not are ‘outside’. Of course, this property rights view excludes human assets or human capital since these are inalienable, that is they cannot be bought or sold.

This view of the firm seems very different from the nexus of contracts setup. Importantly, firms are said to be determinate. Nevertheless, the two views share much more in common than is usually acknowledged. For instance, in Hart and Moore's (Reference Hart and Moore1990) analysis, agents form coalitions depending on the expected value of their participation, which in turn depends on the effects of asset ownership. Contractual incompleteness renders ownership allocations relevant, since property rights condition incentives to invest in specific assets, thereby influencing the bargaining power of the parties involved. Alchian's (Reference Alchian1984) account of specificity, specialization, and coalitions is based on a similar story.

More fundamentally, legal personality clearly plays an important role in both accounts, albeit an implicit one in the collection of assets setup. The nexus of contracts view implies that the fictitious central contracting party is the necessary link between the resource owners composing the coalition. The collection of assets view achieves the same result: since ‘a firm is identified with the collection of physical assets over which the owner . . . has residual control rights’ (Moore, Reference Moore1992: 496), the link is the concentration of property rights in the hands of one agent. The coalition of owners disappears and we are left with ‘the owner’ that personifies the aggregate of owners and becomes the sole central contracting party, much like the legal fiction of the corporate person. However, since both views focus on private contract alone, no further consideration is given to this legal feature of the firm.

What, then, is the link between the owner (e.g., employer) and the other agents necessary for production (e.g., employees)? The existence of such a link is important for a theory of the firm, and Hart (Reference Hart1995: 57) rightly stresses that ‘without something to hold the firm together, the firm is just a phantom’. Hart (Reference Hart1995: 57–59) says:

Clearly, Hart provides a wrong answer to a good question. Hart makes a logical mistake by stating that a collection holds itself together. Far from being the sort of thing that could bind anything together, a collection is itself in need of being bound together if it is to form a whole. Without some sort of ‘glue’, a collection is no different from a heap of sand easily blown away on a windy day. Arguably, Hart also makes a theoretical mistake by excluding human assets or people from his definition of the firm. Given that property rights hold nonhuman assets together, the glue question makes sense only if it is about what holds human beings together.

The collection of assets conception is an ‘aggregationist’ view of the firm, since the firm is said to be identical to its parts or ‘nothing but’ the collection of its parts. Debates on which parts are the firm's proper parts do not change this diagnosis. Hence, the set of contracts view is equally aggregationist. Alchian (Reference Alchian1984: 47, emphasis in original) says: ‘it is not silly to consider the entry of a new stockholder to be the creation of a new firm’. Clearly, both approaches are variations on the same theme. The denial of the firm's real existence and the reduction of the firm to an aggregate of its parts, whatever these may be, arguably amount to the same thing. The former is simply more explicit than the latter. In both cases, the only real constituents of the world are said to be individuals, contracts, and assets, and firms are said to be nothing but aggregates of these parts. We drive these points home in Section 4.

Interestingly, recent papers have called for new foundations for the theory of the firm. Expanding on Hart, Zingales (Reference Zingales2000: 1649) argues that the firm is a ‘nexus of specific investments’ built over time and held together by strong complementarities between human and non-human capital. Blair (Reference Blair, Blair and Roe1999: 86) observes that the continuity and integrity of the firm may be enhanced by attributing key property rights to a separate legal entity. Accordingly, corporate law plays a greater role than is commonly acknowledged. Indeed, nexus of contracts theorists hold that corporate features such as separate entity status and limited liability can be achieved by private contract and market forces alone, and that the role of the law is limited to enforcing contracts and property rights (Hessen, Reference Hessen1979; Butler, Reference Butler1989; Easterbrook and Fischel, Reference Easterbrook and Fischel1991). On the contrary, the role played by the law in creating separate entity status for the firm needs to be recognized instead of being downplayed or ignored (Orts, Reference Orts1998, forthcoming; Hansmann and Kraakman, Reference Hansmann and Kraakman2000; Hodgson, Reference Hodgson2002; Blair, Reference Blair2003, Reference Blair and Grandori2004; Stout, Reference Stout2005; Hansmann et al., Reference Hansmann, Kraakman and Squire2005, Reference Hansmann, Kraakman and Squire2006).

In this context, Blair (Reference Blair, Blair and Roe1999: 87) believes that we should ‘reconsider the merits of an older school of legal scholarship that emphasizes that a corporation is a separate entity, and more than the sum of its parts’. Blair explicitly suggests that ‘the entity view of the firm should be brought back to the center stage’ (p. 59). Echoing this opinion, Orts (Reference Orts1998: 267) further suggests that ‘a combination of legal and economic analysis opens the door for other disciplines besides economics to describe the social nature of the entities called firms’. The rest of this paper is devoted to these tasks.

3. Real entity theory reconsidered

Discussion of the nature of the corporation has been often conflated with the issue of corporate personality, which itself has more often than not been related to the problem of the locus of legal rights and duties. These questions revolve around the legally fundamental distinction between subjects and objects, or persons and things (Iwai, Reference Iwai1999; Naffine, Reference Naffine2003). A ‘legal entity’ or ‘legal person’ is a legally recognized subject of rights and duties, capable of holding property, entering contracts, and so on. Not all human beings are legal persons (exceptions include infants and slaves) but all corporations are such persons. Debates on the ‘right-and-duty-bearing unit’ (Dewey, Reference Dewey1926: 565) must themselves be related to debates on the nature of legal rights. Savigny's (Reference Savigny1841) ‘will theory of rights’ states that rights result from the right-holder's power of choice. However, Savigny's fiction theory clearly denies the reality of collective or corporate will. Regardless of volition, Jhering's (Reference Jhering1852) ‘interest theory of rights’ sees rights as interests sufficiently important to be legally protected. Here, there is no logical need to decide whether corporations are real entities or merely fictions.

With these debates in mind, we now turn to the entity view that strongly influenced legal, social, and political theory, as well as institutional economists such as Commons (Reference Commons1924), at the turn of the twentieth century. Some call this view the ‘natural entity theory’ (Millon, Reference Millon1990; Horwitz, Reference Horwitz1992), while others refer to it as the ‘real entity theory’ (Mark, Reference Mark1987; Hager, Reference Hager1989; Phillips, Reference Phillips1994). Sometimes both expressions are used interchangeably (Blumberg, Reference Blumberg1990; Harris, Reference Harris2006). For early entity theorists, the terms ‘real’ and ‘natural’ were equivalently used to oppose the then conflated terms ‘fiction’ and ‘artificial’. Although all entity theorists regarded corporations and similar groups as real socio-economic entities, as opposed legal fictions existing only in contemplation of law, we identify three generations of entity theory based on important differences concerning the nature of the corporate entity and the issue of rights. It is unlikely that Blair refers to the first generation in her call for the revival of entity theory.

The entity view is usually associated with Gierke's four-volume Das deutsche Genossenschaftsrecht.Footnote ³ Maitland's famous translation of parts of Gierke (Reference Gierke and Maitland1900) disseminated his work in the Anglophone world. Gierke and many of his followers explicitly espoused Savigny's will theory of rights but rejected Savigny's fiction theory. In his translator's introduction, Maitland (Gierke, Reference Gierke and Maitland1900: xxvi) writes of the corporation that it

Gierke's and Maitland's organicist and vitalist position holds that corporations are real beings with real volition. Against Savigny, statements such as ‘the personality of the corporation . . . is in no sense . . . artificial or fictitious, but is every whit as real and natural as is the personality of man’ (Maitland, Reference Maitland1900: 335–336) explicitly seek to establish that the corporation can be an ‘ultimate moral unit’ (Maitland, Reference Maitland1905: 199), that is a natural subject of rights.

A second generation of entity theorists argued that there is no supra-individual geist: the ‘corporate entity is not a rational being [and it] has no will’ (Machen, Reference Machen1911: 265). For Machen, fiction theory contains a grain of truth, namely that the proposition that ‘a corporation is a person’ is either a metaphor or a legal fiction. However, Machen continues, ‘although corporation personality is a fiction, the entity which is personified is no fiction’ (p. 266). Machen (Reference Machen1911: 258) is careful to distinguish between two questions conflated by the earlier generation:

Machen rightly stresses that the ontological question contained in the first proposition needs to be addressed separately. Several insights can be found in the writings of the second generation.

According to Dicey (Reference Dicey1905: 154), ‘whenever men act in concert for a common purpose, they tend to create a body which, from no fiction of law, but from the very nature of things, differs from the individuals of whom it is constituted’. Given this concerted action and common purpose, Brown (Reference Brown1905: 369) argues, ‘the group becomes, or tends to become, a unit . . . A mere sum of individuals as such can no more become a unit than a heap of sand can become a statue’. In this spirit, one of the clearest statements made by entity theorists is Freund's (Reference Freund1897: 47) list of three ‘salient characteristics of the body corporate: its unity, its distinctiveness, and its identity in succession’.

For Freund, if these features are in fact present in a given association, then one can speak of a real entity. The difficulty is to show how common purpose and collective action produce a level of unity, distinctiveness, and durability sufficient for the group to be a real entity without appealing to any literally volitional or moral features.Footnote ⁴ It is important to notice that Freund, Dicey, and Brown clearly associate existence, identity, and unity of groups in general. Indeed, ‘the inquiry is one which leads us on from the subject of corporations to the wider subject of human association in general’ (Brown, Reference Brown1905: 368).

Entity theorists repeatedly underlined the role played by the law, claiming that the law should comply with the fact of the group's socio-economic existence and attribute legal capacity to an already existing or a potential socio-economic capacity. Accordingly, Laski (Reference Laski1916: 422) argues that ‘the entities the law must recognize are those which act as such, for to act in unified fashion is – formality apart – to act as a corporation’. Legal entity status attributed by the law unifies and reinforces the socio-economic capacity created by concerted action and common purpose. It thus greatly increases the possibilities of collective action. Many entity theorists were political pluralists who believed in freedom of association, and the increasing legal recognition of various groups (associations, trade unions, political parties) sat well with their theory.

Of course, nothing precludes the law from considering that a given group is a legal entity, even if it is not really a unified group. Political or tax considerations may enter the picture and in fact often do (Dewey, Reference Dewey1926; Hager, Reference Hager1989; Avi-Yonah Reference Avi-Yonah2005). Second-generation entity theorists readily admitted this possibility. Vinogradoff (Reference Vinogradoff1924: 604) writes: ‘the life of groups has two sides, i.e., the social contents which are real and produce the union, and the legal form which has to be arranged artificially by the State in order to safeguard public and private interests’. Contrary to the earlier generation, second generation entity theorists adopt Jhering's interest theory of rights. Hence, ‘the essence of juristic personality does not lie in the possession of rights but in subjection to liabilities’ (Machen, Reference Machen1911: 263).

Entity views encountered increasing resistance. Typical attacks include Singleton's (Reference Singleton1912: 291) position that ‘nothing is to be gained from the spinning out of ontological theories incapable of verification’. Wormser (Reference Wormser1912: 496) likewise sees a ‘tempting but profitless discussion’. In this context, an important blow was delivered by Dewey's (Reference Dewey1926) influential instrumentalist discussion. Courts, Dewey argues, should avoid the legally profitless debate regarding essential definitions of the corporate entity. When facing actual decisions, courts should consider only the practical consequences involved and avoid reference to some abstract general theory. From the same point of view, many legal realists regarded entity theory as ‘a matter of literature’ (Radin, Reference Radin1932: 667) and dismissed it as ‘transcendental nonsense’ (Cohen, Reference Cohen1935). With the rise of both instrumentalism and legal realism in the late 1920s and 1930s, by the time of Berle and Means (Reference Berle and Means1932), papers upholding entity views became very rare.

One exception is Rowley's (Reference Rowley1931: 558) suggestion that talk of personality should be abandoned in favor of a better term, ‘individuality’. Recall that Jensen and Meckling (Reference Jensen and Meckling1976) conflate the two terms. As we have seen, personality is a legal term involving rights and duties. Wrongly suggesting that some moral issue is involved or that personality is the mark of responsible human beings only, the term is the source of much confusion (Gindis, Reference Gindis, Biondi, Canziani and Kirat2007: 275). Individuality, on the other hand, refers to distinctive existence or individuation and unity (and sometimes to indivisibility), and Rowley (Reference Rowley1931: 560) rightly states that ‘anything has this individuality, otherwise it is not a thing’. Another exception is Dodd's (Reference Dodd1932: 1160) view that ‘any organized group, particularly if its organization is of a permanent character, is a factual unit’, and that in general, ‘if the unity of the corporate body is real, then there is reality’.

Berle (Reference Berle1947: 344) later agreed that ‘the entity commonly known as “corporate entity” takes its being from the reality of the underlying enterprise, formed or in formation’. Arguing that it is ‘the enterprise, and not the incorporation papers, [that] is the true entity’ (p. 358), Berle applies his theory of enterprise entity to de facto corporations and to corporate groups, that is to cases where economic entity and legal entity do not match. Following Berle's lead, Dix (Reference Dix1953) further extends the entity view to corporate mergers and groups. A final entity view worth mentioning is Stauss's (Reference Stauss1944: 112, emphasis in original) subtle position according to which the firm is both independent from and dependent on its members:

Rowley, Dodd, Berle, Dix, and Stauss are the third generation of entity theorists since the accent is neither on the defense of group-will views (first generation) nor on denials of group-will views and attempts to provide other foundations for entity theory (second generation). Contrary to the first two generations who had concentrated on essential definitions, the third generation abandoned talk of personality and concentrated on consequences of the entity view.

Critics have all too often reduced entity theory to the first generation (e.g., Bratton, Reference Bratton1989b: 1512) and have argued that since entity theorists admit all sorts of geists and vital forces, their theory is to be discarded. But it is important to understand that these more dubious aspects were abandoned by later entity theorists themselves. Another important observation that commentators rarely mention is that Dewey's attack was confined to essentialist thinking in the realm of judicial decision only. Contrary to the standard account (e.g., Bratton, Reference Bratton1989b: 1491), Dewey's (Reference Dewey1926: 673) comments on Dicey do not dispute the reality of the unified body created by people acting together for some common purpose. However, like Machen, he simply does not think that issues of reality and legal personality should be conflated. Dewey is closer to later entity theorists than is commonly acknowledged. Dewey (Reference Dewey1925: 163) writes:

Dewey believes that the corporation is neither a conceptual entity nor a physical entity. In modern language, one would say that it is an ‘institutional kind’ (Thomasson, Reference Thomasson2003), that is an object of social ontology and not of natural science. Dewey also believes that corporations exhibit new properties that do not exist in the absence of specific human interactions. Corporations, just like other complex systems, possess emergent properties that have consequences. Freund (Reference Freund1897) makes a similar point, and Brown (Reference Brown1905: 379) argues that individuals, as members of groups, are affected by their membership: ‘qualities have been developed in [them] which cannot be explained save by reference to the union’.

Overall, second and third generation entity theorists held that the firm is more than the aggregate of its parts and different from a heap. Existence, identity, unity, and persistence are associated. Unity is produced by common purpose, concerted action, and organization. There is also an important sense in which the firm is both independent from and dependent on its members. This can be best understood as a problem of emergent properties. Put together, these ontological insights clearly state that the firm is real or at least as real as its members. These, then, are the relevant aspects of entity theory that need to be brought back to the center stage if one seeks to build a theory of the firm in which the firm is neither a fiction nor an aggregate.

4. Advancing a real entity view of the firm

Fictionalism and aggregationism in both economic theory and legal doctrine are varieties of ontological individualism, the view that reality is ultimately composed of individuals only (although contracts and assets may be included), and that firms and similar social entities are ‘nothing but’ the individuals composing them, or ‘nothing over and above’ these individuals. On the contrary, the argument here is that a case can be made for the fact that the firm is a real entity and that ontological individualism is untenable, as is the systematic reduction of the firm to any of its parts. However, further clarification of the notion of ‘real entity’ is needed. We argue our case by advancing several interrelated requirements that a thing needs to satisfy in order for it to be considered a real entity. This approach is similar in spirit to Davis's (Reference Davis2003) examination of conditions of individuality. Davis (Reference Davis2003: 185) says:

This is how we shall proceed in the following. Relevant ‘existence tests’ for a real entity view of the firm include identity, unity, and persistence. The possession of some form of causal power is also important. These features exclude aggregates from the category of real entities.

Real entities, in the language of traditional philosophy, are different from mere concepts, shorthand expressions, or fictitious entities. Traditionally, an entity is said to be real if it exists or can exist independently of our knowledge or language. There are different ways to define independence and these depend on the entity in question. Natural kinds such as rocks and trees are completely independent. Arguably, the entities composing the social realm – institutional kinds such as firms – depend at least in part on us for both their creation and persistence. However, although institutional kinds do not meet the strong realist requirement of entities entirely independent of us for their existence, ‘they certainly are . . . not mere mental constructs’ (Thomasson, Reference Thomasson2003: 605). These differences notwithstanding, a general way to define a real entity is to say that real entities can exist apart from each other.

Importantly, for any thing to be an entity, it must be identifiable, that is it must satisfy some identity criteria. We take Quine's (Reference Quine1969: 23) dictum, ‘no entity without identity’, to be a necessary condition meaning that a candidate entity can be individuated or distinctly identified relative to other things according to some key characteristics. Rowley (Reference Rowley1931) makes the same point. Like Davis (Reference Davis2003: 12–13), we also take identity through time to be important since persistent entities can be re-identified over time and through change according to the same key defining respects. Lowe (Reference Lowe2003: 335, emphasis in original) points out that this implies unity: ‘an individual object is an entity which, quite literally, counts as one entity of some kind, in order to do which it must possess unity’. Indeed, Quine's dictum, as Dejnožka (Reference Dejnožka2003: 184–185) explains, has roots in the past and in particular in the philosophy of Leibniz (Reference Leibniz and McCormack1902 [1687]: 191, emphasis in original) according to which ‘what is not truly a being is not truly a being’.

This leads Leibniz to exclude aggregates from the category of real entities: ‘where there are only beings by aggregation, there are not even real beings’ (p. 189), since ‘beings by aggregation . . . have their unity only in our minds’ (p. 191). Similar arguments can be found in Plato (Sophist 245) and Aristotle (Metaphysics Z, 17, 1041b) who uses the term ‘heap’. Clearly, any candidate entity should have the real unity that aggregates and heaps lack. An important case in point is that of an entity having not only certain characteristic kinds of parts, but also a characteristic constitutive structure. Baker (Reference Baker2004: 101) appropriately says that ‘constitution is a relationship of unity’ explaining the non-identity between the whole and any aggregate of its parts. Baker (ibid.) further claims that ‘constitution is a relation that accounts for the appearance of genuinely new kinds of things with new kinds of causal powers’.

Recall that Plato's Eleatic Stranger (Sophist 247e) famously argued that a sufficient mark of real things is the power to affect other things and to be affected by them. Causal powers, dispositions, or capabilities are significant features of any real entity.Footnote ⁵ Therefore, this is a crucial existence test for entities such as firms. Discussion of this requirement often appeals to two types of arguments: an argument from emergence and an argument from constitutive rules. Both can be found in the writings of second-generation entity theorists. Both seek to distinguish identifiable and persistent ‘social integrates’ (Pettit, Reference Pettit2002) from mere aggregates of individuals. Both support the view that wholes arising from a constitution relation are qualitatively different from their parts taken separately or collectively, and that it is no longer possible to hold that these wholes are ‘nothing but’ the parts composing them, or ‘nothing over and above’ these parts.

The argument from emergence is that emergence is the ‘failure of aggregativity’ (Wimsatt, Reference Wimsatt1997: S372). In other words, resultant properties, or macro-properties of sums and aggregates, are to be distinguished from truly emergent properties that entail ‘qualitative novelty’ (Bunge, Reference Bunge2003: 14). New properties at the emergent level, not to be found among its constituents however considered, are generated by the ‘causal reciprocity’ of proper parts interacting in a constitutive structure (Weissman, Reference Weissman2000: 47). This can help explain collective knowledge and capabilities that are neither reducible nor ontologically equivalent to individual knowledge and capabilities. Importantly, emergents are ‘integrated wholes that tend to maintain some sense of identity over time’ (Goldstein, Reference Goldstein1999: 50) and ‘survive the demise of particular individuals that once related to them’ (Hodgson, Reference Hodgson2004: 12). However, the most important but most controversial aspect of emergents is their possession of some form of causal power.

One recent formulation is the notion of ‘reconstitutive downward causation’ (Hodgson, Reference Hodgson2003: 164–165; Hodgson, Reference Hodgson2004: 184ff). On this view, also held by Brown (Reference Brown1905), emergent entities causally change, even if the change is very slight, those things from which they have emerged. Things that are parts of wholes are different from what they are in the absence of the constitutive relations that account for emergence. This means that parts cannot be taken as given. Individual knowledge and capabilities, for instance, are affected by collective knowledge and capabilities. Furthermore, if the emergent entity can causally affect those things from which it emerges, then it is both independent from and dependent on those things.Footnote ⁶ These ideas are particularly relevant for the analysis of social entities such as firms from the ‘structure and agency’ perspective (Lawson, Reference Lawson1997; Hodgson, Reference Hodgson2004) that usefully links the argument from emergence and the argument from constitutive rules.

The argument from constitutive rules further underlines the constitutive structure that distinguishes aggregates such as crowds from institutional kinds such as firms (French, Reference French1982; Copp, Reference Copp1979, Reference Copp1984). Broadly understood, the argument is that ‘constitution is the essential basis of an organization, the link that ties its members together’ (Vanberg, Reference Vanberg1992: 243). Vanberg (p. 240) argues that ‘the procedural rules that underlie organized or corporate action . . . constitute organizations as corporate actors’. Freund (Reference Freund1897) similarly explains corporate agency without appealing to group-wills. The argument is summed up by Searle's (Reference Searle2005: 10) constitutive rule: ‘X counts as Y in context C’ creates new ‘deontic powers’, that is rights, duties, obligations, and empowerments. Interestingly, Searle claims that private recognition of these powers is insufficient: ‘there has to be official recognition by some agency . . . and there have to be status indicators issued by the official agency’ (p. 15). The law's creation of legal entity status is thus a fundamental institutional fact about the firm.

Fictionalists argue that firms cannot exist without any individuals. We do not dispute this. Yet the firm's both economic and legal persistence through time, based on its constitutive structure that allows the replication of behavioral patterns and collective routines, implies that any particular firm continues to exist even if all its present human members are progressively replaced. Such independence qualifies the firm as a real entity. This fits an important insight from the evolutionary theory of the firm, namely that collective knowledge and capabilities are typically ‘sticky’, that is retained through progressive change in firm membership (Nelson and Winter, Reference Nelson and Winter1982: 99ff; Winter, Reference Winter1988: 172).Footnote ⁷ If this were not the case, the firm would certainly be a flimsy entity, incapable of survival in an evolutionary setting. Significantly, the firm can only be a ‘unit of selection’ if it is cohesive and durable, that is if its members ‘are bound together in a sufficiently cohesive manner to share a mostly common fate’ (Hodgson and Knudsen, Reference Hodgson and Knudsen2004: 300). Unitary legal status is important in this respect.

Legal entity status also matters for the boundaries of the firm. Firms are ‘indivisible’ legal persons that appear in court as single parties (Iacobucci and Triantis, Reference Iacobucci and Triantis2007: 524). Divisions within the firm, on the other hand, cannot sue or be sued, cannot own property, and cannot contract (unless exercising delegated powers) since they do not have legal personality. Since internal transfers do not generally involve the exchange of legal property rights, as Hodgson (Reference Hodgson2002: 46) similarly explains, there are no markets within the firm. Given the accent on legal indivisibility, the boundaries of the firm are to be drawn such that there are two firms whenever there are legal contracts between two legal persons. Employees and managers act as agents of the firm during their contractual time of employment or according to any other contractual stipulation. As such, their legal identity is different from the legal persons they are when acting as consumers outside working hours, that is ‘outside’ the firm. The fact that boundaries may sometimes be difficult to draw does not mean that there are no firms.

Matters seem less straightforward in cases where an economic entity is formally organized as a multiple legal entity structure. Traditional corporate law sees corporate groups, for instance, as aggregates of legal entities instead of singular economic entities. Multinational enterprises likewise exist as separate legal entities in different countries. In this context, like third generation entity theorists, Blumberg (Reference Blumberg1993) argues that courts should not systematically place ‘form over substance’ or ‘legal entity over real enterprise’, especially in terms of group liabilities. Blumberg (Reference Blumberg1993: 89ff) documents numerous court decisions based on ‘enterprise principles’ that acknowledge the reality of a singular economic entity based on various existence tests emphasizing, among other things, unity of control. Of course, whether corporate groups are ‘considered one entity or a group of entities will depend . . . on the question at hand’ (Orts, Reference Orts1998: 313). Arguably, a real entity view provides a useful framework for addressing such issues.Footnote ⁸

Unlike aggregates, sets, collections, and heaps, the firm is held together by ‘ontological glue’ (Gindis, Reference Gindis, Biondi, Canziani and Kirat2007: 279). This includes: ‘institutional glue’ created by legal entity status, constitutive rules, contracts, and norms; ‘organizational glue’ manifested by structures, processes, functions, and roles; ‘motivational glue’ that ensures loyalty and adherence to common goals through a variety of means; ‘cognitive glue’ accounting for identification, shared beliefs, and representations; and ‘capabilities glue’ that relates to the complementarity between human assets such as knowledge and non-human assets, to productive routines, and so on. Firm heterogeneity is partly explained by the existence of indefinitely many blends and combinations of these types of glue. Overall, glue contributes to the firm's cohesion and unifies the collective action of its members through time, thus contributing to the firm's causal powers vis-à-vis the rest of the world. Although these considerations abstract away from internal conflict, we do not wish to imply that internal conflict is absent or should be ignored, but only that the firm's unity is generally sufficient to prevent dissolution.