ABSTRACT. This paper presents an argument forthe democratic (or ‘labor-managed’) firm based onordinary jurisprudence. The standard principle ofresponsibility in jurisprudence (‘Assign legal respon-sibility in accordance with

de facto responsibility’)implies that the people working in a firm shouldlegally appropriate the assets and liabilities producedin the firm (the positive and negative fruits of theirlabor). This appropriation is normally violated dueto the employment or self-rental contract. However,we present an inalienable rights argument thatdescends from the Reformation and Enlightenmentwhich argues that the self-rental contract, like the self-sale or voluntary slavery contract, is inherently invalid.The key intuition of the inalienable rights theory isthat one cannot in fact voluntarily transfer de factoresponsibility for one’s actions to another person. Onecan only voluntarily co-operate with another person,but then one is de facto jointly responsible for theresults. Just as the legal authorities legally reconstructthe criminous employer and employee as a partner-ship with shared responsibility, so justice demands thatevery firm be legally reconstructed as a partnership ofall who work (working employers and employees) inthe enterprise, i.e., as a democratic firm.

KEY WORDS: corporate ownership, democraticfirm, inalienable rights, responsibility principle

Introduction

Who is to be the firm?

Ordinary jurisprudence provides an answer to thefundamental question: ‘Who is to be the firm –Capital, the State, or Labor?’ It is conceptualanalysis that is required, not new principles, inorder to answer that basic systemic question.Once the conceptual framework about the struc-ture of property rights in the firm is developed,then the answer follows from the responsibilityprinciple of jurisprudence that people are to beheld legally responsible for the results of theirintentional deliberate actions.

I will approach the theory of the firm fromthe viewpoints of a descriptive and a normativetheory of property. This approach is ‘non-economic’ in that it is not based on efficiencycriteria. The usual ‘economic’ approach to thefirm is to emphasize cost minimization or, moregenerally, allocative efficiency.1 “The mainhypothesis is that contractual designs, bothimplicit and explicit, are created to minimizetransaction costs between specialized factorsof production” (Holmstrom and Tirole, 1989,p. 63). While this economic approach has itsdomain of applicability, it operates within aframework of norms for the initiation and ter-mination of property rights and of constraints onthe transfer of property rights. That frameworkis deontic or rights-based, not utilitarian orefficiency-based. The rights-based argumentsprovide ‘trumps’ (see Dworkin, 1978) or nor-mative ‘side-constraints’ (see Nozick, 1974) thatlimit the area where optimizing arguments mightbe applied.

Several examples of rights-based side-con-

The Democratic Firm:An Argument Based onOrdinary Jurisprudence† David Ellerman

*

Journal of Business Ethics 21: 111–124, 1999.© 1999 Kluwer Academic Publishers. Printed in the Netherlands.

David P. Ellerman, Economic Advisor to the ChiefEconomist, World Bank, has two Masters degrees inPhilosophy and in Economics, and a doctorate inMathematics. He taught for 20 years before joining theWorld Bank. He has published articles in economics,mathematics, philosophy, and law journals as well as fourbooks: Intellectual Trespassing as a Way of Life:Essays in Philosophy, Economics, and Mathematics(1995), Property and Contract in Economics(1992), The Democratic Worker-Owned Firm(1990), and Economics, Accounting, and PropertyTheory (1982).

straints are readily available in economics. Forinstance in human capital theory, the economicefficiency or inefficiency of selling human capital(e.g., in a self-sale contract) is not activelydebated; it is simply ruled out.

Since slavery was abolished, human earning poweris forbidden by law to be capitalized. A man is noteven free to sell himself: he must rent himself at awage. (Samuelson, 1976, p. 52)

Political vote-selling by the ultimate votersprovides another example.

Any good second year graduate student in eco-nomics could write a short examination paperproving that voluntary transactions in votes wouldincrease the welfare of the sellers as well as thebuyers. (Tobin, 1970, p. 269)

Yet instead of following the dictates of efficiency-based reasoning in law and economics, the ideaof a market in votes is also ruled out.

I shall argue that the non-economic rights-based constraints are stronger than is ordinarilythought, and accordingly that the domain of effi-ciency-based arguments is more confined. Thereare a number of quite plausible economic argu-ments for the democratic firm. For instance,Margaret Blair (1995) has developed arguments,largely within the methodology of the economictheory of the firm, that human capital shouldincreasingly be treated as a residual factor – afactor with net income and control rights asso-ciated with “ownership” of the corporation. Ishall argue, however, that ordinary jurisprudence(not to mention democratic principles) appliedto the firm rules out non-democratic firms so theefficiency-based arguments for one or anothertype of firm (e.g., capitalist versus labor-managedfirms) are ultimately as irrelevant as an analysis ofthe relative efficiency of cotton being producedby free or slave labor.

What is it about human capital that makes it sodifferent from physical or financial capital that itis forbidden by law to buy and sell humancapital? If the reason for that prohibition ofcertain voluntary market transactions is wellunderstood, might there be other implications?After all, the usual employment contract is essen-

tially the contract for renting or hiring humancapital, i.e., for buying the services of humanbeings.2 The forbidden contract to buy a person’shuman capital is essentially a long-term (up toretirement age) version of the still permittedshort-term rental contract. This prohibition inthe long-term case could hardly be based on effi-ciency grounds since economists routinely ignorethe prohibition against lifetime labor contracts inorder to have the complete future markets in allcommodities necessary to show the allocativeefficiency of competitive equilibrium.

Now it is time to state the conditions under whichprivate property and free contract will lead to anoptimal allocation of resources. . . . The institu-tion of private property and free contract as weknow it is modified to permit individuals to sellor mortgage their persons in return for presentand/or future benefits. (Christ, 1975, p. 334)

If the rights-based case against buying humancapital turns out to also apply to renting humancapital, then the case against the non-democraticfirm is made without further appeal to efficiencyarguments. This paper pursues this ‘non-economic’ line of thought based on the distinc-tive features of human capital. The line ofapproach is through descriptive and normativetheories of private property. The prohibitionagainst the sale of human capital is approachedthrough the theory of inalienable rights thatdescends from the Reformation and theEnlightenment and that was pivotal in the devel-opment of anti-slavery thought.3

The structure of property rights in a firm

The life-cycle of a property right

A positive or descriptive theory of property fora private property market economy shoulddescribe how property rights are created orinitiated, how they are transferred from partyto party, and finally how property rights areterminated.

Property rights are created in firms and house-holds, and property rights are also terminated byfirms and households. In between is the market

112 David Ellerman

where property rights are transferred. The instru-ment for transferring property rights is the vol-untary contract which is fulfilled by the transferof the de facto possession and control of theproperty from the old owner to the new owner(usually in exchange for other property going inthe opposite direction).

Theory of appropriation

Property rights are transferred by voluntarycontract, but what is the legal means by whichproperty rights are initiated and terminated in aprivate property market economy? It will behelpful to first establish some notation and ter-minology. From economics we can borrow asimplified description of a productive opportu-nity Q = f(K, L) where the outputs Q areproduced by applying the labor L to the capitalservices or non-labor inputs K. Many differenttypes of labor and non-labor inputs could beconsidered but that would only complicate thenotation and would not change the underlyinglogic. The initiation of a property right isthe ‘appropriation’ of the property right.Symmetrically the termination of a propertyright might be called the ‘expropriation’ of theproperty right but that word has now been cor-rupted to mean the forced transfer or ‘taking’ ofproperty by the state.4 The word ‘expropriation’can be avoided by referring to the ‘expropriationof the assets X’ as the ‘appropriation of the lia-bilities –X.’

In the productive opportunity Q = f(K, L),the property rights to Q are initiated and the

property rights to K and L are terminated so thiscould be described as the appropriation of theassets Q as well as the appropriation of the lia-bilities –K and –L. It will be useful to adopt asimple vector notation with three componentsrepresenting outputs, non-labor inputs, and laborinputs respectively. Thus the vector of assets andliabilities appropriated in the productive oppor-tunity is (Q, –K, –L) which I will call the ‘wholeproduct.’ The ‘whole product’ is simply aproperty theoretic name for the productionvector familiar to economists in the productionset treatment of production where the outputsare listed positively and the inputs are listednegatively. When the assets and liabilities in thewhole product are evaluated at their respectiveprices, the result is the residual or pure profit,and the whole product appropriator is called the“residual claimant.”

Who is to be the whole product appropriator?

Our original question of ‘Who is to be theFirm?’ can now be sharpened to ‘Who is to bethe whole product appropriator?’. How does aprivate property market economy in fact deter-mine who appropriates the whole product?There is a ‘laissez-faire’ or market mechanism ofappropriation. A non-market assignment ofliabilities by the legal authorities takes place in atrial for property damages when the defendantis found guilty and held liable for some propertythat was destroyed. But what happens when thereis no trial? To understand the market mechanism,one must consider who would appropriate theliability –X (i.e., terminate the property right toX) if the property X is used-up, consumed, orotherwise destroyed when the legal authoritiesdo not intervene to hold a trial. In that case, theliabilities –X are borne by the last legal ownerof X. Thus one could say that in the absence ofany state intervention to reassign liabilities, theliabilities –X for used-up or destroyed propertyX is ‘laissez-faire appropriated’ by the last legalowner of X. Furthermore, that same legal partywho bore the liabilities would have the defen-sible legal claim (in the absence of any reassign-ment of the liabilities) to any new property that

The Democratic Firm 113

Figure 1. ‘Life-cycle’ of a property right.

might be created in the process of using up theold property.

Market Mechanism of AppropriationLet the liabilities for the used up inputs

lay where they have fallen (i.e., in the hands of the last owner of the inputs), and then

let that party have the claim on any produced outputs.

It should be noted that this mechanism of appro-priation only works for produced outputs asopposed, for example, to gifts of nature.

In terms of our example, the last legal ownerof the non-labor inputs K and labor inputs Lwould laissez-faire appropriate the liabilities or‘negative product’ (0, –K, –L). In the absence ofany state intervention to reassign those liabili-ties, that same party would have the defensiblelegal claim on the produced assets or ‘positiveproduct’ (Q, 0, 0). Putting the two productstogether, one has the market mechanism for theappropriation of the whole product (Q, –K, –L).The legal party who appropriates the wholeproduct of a productive opportunity will becalled the ‘firm’ (with respect to that opportu-nity). The market mechanism of appropriationis a descriptive theory; it describes how thewhole product is actually appropriated in aprivate property market economy.

The ‘fundamental myth’ about property rights

The ‘market mechanism of appropriation’ mightall seem like a fancy way to restate the obvious,but it has quite strong implications. For instance,it shows that in order for a legal party to be the‘firm’ (i.e., residual claimant) with respect to agiven productive opportunity, it is sufficient forthe party to be the last legal owner of all theinputs used up in the production process. Thenthat party has the defensible legal claim on theoutputs that emerge in production so that partywould appropriate the whole product. Since itis the fact-pattern of the input contracts (e.g.,whether capital hires labor or labor hires capital)that determines who is the last legal owner ofthe used-up inputs, the identity of the firm(‘firmhood’) is contractually determined. There is

no need for the legal party to additionally ‘buy’or ‘own’ the production function or productionset. Indeed, there is no such thing as the ‘own-ership’ of a production function or productionset in a private property market economy.

The idea that there is a property right (vari-ously called ‘ownership of the firm’, ‘ownershipof the production function’, ‘ownership of theproductive opportunity’ and so forth) whichdetermines which party legally appropriates thewhole product of a productive opportunity issuch a pervasive and important idea that it willbe called the ‘fundamental myth’ about propertyrights. An understanding of the fundamentalmyth is important to our argument because theargument tries to ‘trump’ the efficiency-basedeconomic theories of the firm with a right-basedtheory about the norms for appropriating assetsand liabilities in production (see below for thenormative theory). But if it is thought that thewhole product is already owned by the ‘ownerof the firm’ then the entire question of appro-priation (initiation and termination of propertyrights) in production is not even formulated orasked. That is indeed the usual case in the eco-nomics literature.5

The ownership of a corporation and the fundamentalmyth

What are the origins of the fundamental myththat firmhood is established by an ownershipright (‘ownership of the firm’) rather than by thecontractual status of being the last legal ownerof all the inputs to production? Perhaps the mostcommon origin is a misinterpretation of the‘ownership of a corporation.’ Before turning tothat, it might be noted that economists usethe notion of ‘ownership of the firm’ in moregeneral contexts independent of corporations. Inan abstract model, entrepreneurs are “biddingfor ownership of the firms” (Hirshleifer, 1970,p. 124) and become the “owners of the produc-tive opportunity” (p. 125). A proprietor may sell“the rights to the transformation function” or“his rights to the venture” (Fama and Jensen,1996, p. 341) to another proprietor. The entre-preneur is the “owner of a production function”

114 David Ellerman

(Haavelmo, 1960, p. 210), and Robinson Crusoeeven “owns the firm” (Varian, 1984, p. 225).

The most common or ‘standard’ origin of the‘ownership of a firm’ notion is the (mis)inter-pretation of the ownership of a corporation thatis currently undertaking a production opportu-nity Q = f(K, L) (by virtue of its contractualposition) as being ‘ownership’ of the productiveopportunity. But this interpretation can beeasily defeated by changing the contractualposition of the corporation without changing itsownership. For instance, if the capital services Kwere hired out rather than the labor services Lbeing hired in, then the ‘firm’ in the sense ofwhole product appropriator would shift but theownership of the corporation would be in thesame hands. The role of the corporation wouldshift from being the firm (with respect to thatopportunity) to being an input supplier tothe firm (recall that ‘firm’ is defined hereas ‘whole product appropriator,’ i.e., residualclaimant in production, and is not a synonym for‘corporation’).

This argument might be better understoodby considering a productive opportunity bothoutside and inside a corporate form. Consider asimplified process where the labor L is appliedusing the services K of a widget-maker machinein order to produce the widgets Q during eachtime period. If the machine is owned by an indi-vidual, then it is clear that the person could hirein the labor L and produce Q – or could hire outthe services K to another party. The pattern ofcontracts determine whether the individualoperates as the firm (with respect to that oppor-tunity) or as an input supplier to the firm. Nowsuppose that the individual incorporates acompany and issues all the stock to himself inreturn for the machine. This legal repackagingchanges nothing in the market logic of theargument that separated capital ownership fromresidual claimancy. The corporation (rather thanthe individual) would own the widget-makermachine and, depending on the direction of thehiring contracts, may or may not appropriate thewhole product of the productive opportunityusing the widget-maker. The process of incor-poration does not miraculously transubstantiatethe ownership of a capital asset into the owner-

ship of the whole product vectors that might beproduced using the capital asset.

In actual markets, there are likely to be largetransaction costs to rearranging the input con-tracts. The incumbent corporate residual claimanthas sizable first-mover advantages so that anychallenging party would have to incur such hightransaction costs to redirect the input contractsthat it might be just as cheap or cheaper tosimply buy the corporation and thereby take overthe residual claimant’s position in the existingpattern of input contracts. These transaction costbarriers create the image that the existing cor-porate residual claimant ‘owns’ the productionopportunity. Transaction cost barriers are onlythat; they are not property rights. For instance,as transaction costs change it might become morefeasible to acquire residual claimancy by rear-ranging input contracts rather than by purchasingthe corporation.6

The responsibility principle as a normativetheory of property

The principle of imputation in jurisprudence

The market mechanism of appropriation answersthe descriptive question of how does the marketinitiates and terminates property rights. The cor-responding normative question is the question of‘Who should legally appropriate the wholeproduct?’. To whom should the legal liabilitiesand assets created in production be assigned?

We will answer the normative question byusing the standard responsibility principle fromjurisprudence. To see the underlying juridicalprinciple, let us see how the assignments aremade in an explicit civil or criminal trial. Thestandard principle for the assignment of legal orde jure responsibility can be easily seen in a civiltrial to assign legal liability for damages tosomeone’s property. The standard principle is toassign the legal or de jure responsibility for thedamages to the party who was in fact respon-sible (i.e., de facto responsible) for the damages.The purpose of the trial is to ascertain if thedefendant was in fact responsible for causing thedamages (e.g., if the defendant knowingly caused

The Democratic Firm 115

the damages by his or her intentional actions).There are, of course, many shades and degreesof factual responsibility and there may be miti-gating and extenuating circumstances in anygiven case. But the basic principle is always toassign legal liability to the de facto responsibleparty (if any) and that is the natural principle ofresponsibility used here.

The principle is clear in a trial for damages butthere is no reason for the basic principle tochange when property is consumed, used up, orotherwise destroyed in the absence of a trial.Moreover, the same principle would consistentlyapply to the opposite cases where property iscreated or produced. That is, the legal responsi-bility for produced property should be assignedor imputed to the party (if any) who was in factresponsible for producing the property. Thus thebasic juridical principle that will be taken as thebasis for the normative theory of property appro-priation used here is:

Principle of Responsibility: Assign de jure responsibility in accordance

with de facto responsibility.

This is a clear principle within the domain ofintentional human actions, but that domain ofdeliberate actions is far from universal. The prin-ciple gives only limited help in the domain ofaccidents, and it gives no help whatsoever in theassignation of legal responsibility for the propertythat is created or destroyed solely by naturalforces. As the domain of the responsibility prin-ciple recedes, other non-responsibility-basedprinciples of an economic or contractarian naturemight be applied to the assignment of propertyrights. For our purposes here, our attention canbe restricted to the results knowingly producedby intentional human actions – where theresponsibility principle can be clearly applied.This is because our topic is the human activityof production which is perhaps the epitome ofplanned, deliberate, and intentional humanactions.

Only persons can be de facto responsible foranything. Things can bear no responsibility.Responsibility is imputed back through thingsto their human users. Neoclassical economistsare fond of interpreting the causal efficacy of

nonlabor inputs in an animistic way as a type of“economic responsibility” (Wieser, 1930, p. 76)but that is only a metaphor. A knife has a certainefficacy in a crime, but it is only the knave whocan bear de facto responsibility for the crime. Thede facto responsibility for the intended results ofdeliberate human actions using various instru-ments is borne solely by the human users, and itincludes the responsibility for using up theservices of the instruments.

The judge . . . who, in his narrowly-defined task,is only concerned with the legal imputation, confineshimself to the discovery of the legally responsiblefactor, – that person, in fact, who is threatenedwith the legal punishment. On him will rightlybe laid the whole burden of the consequences,although he could never by himself alone –without instruments and all the other conditions– have committed the crime. The imputation takesfor granted physical causality. . . .

If it is the moral imputation that is in question,then certainly no one but the labourer could benamed. Land and capital have no merit that theybring forth fruit; they are dead tools in the handof man; and the man is responsible for the use hemakes of them. (Wieser, 1930, pp. 76–79)

Thus even as Wieser introduces metaphoricalnotions of ‘economic responsibility’ and ‘impu-tation’ in his treatment of marginal productivitytheory, he recognizes that for the ordinary non-metaphorical notion of imputation, “no one butthe labourer could be named” and that the“imputation takes for granted physical causality.”7

Application to appropriation of the whole product

It is this principle of responsibility from jurispru-dence that is being applied to the normativequestion of who should legally appropriate thewhole product. The (stylized) facts are taken tobe that in a given economic enterprise, thepeople working in the enterprise (managers andworkers) cooperate together to use up the inputsin the process of producing the outputs. Thepeople working in the given enterprise will bereferred to as ‘Labor.’ In our simple canonicalexample, Labor, by performing the intentionalhuman actions L, uses up the nonlabor or capital

116 David Ellerman

inputs K in order to produce the outputs Q. Itis no accident. The using-up of the inputs K andthe production of the outputs Q are the delib-erate and intentional results of the actions L ofthe managerial and non-managerial workers inthe enterprise. Thus the people working in theproductive opportunity are jointly de facto respon-sible for producing the vector (Q, –K, 0). I haveemployed the convention used in economicsof reifying the human activity of producing(Q, –K, 0) as the labor L which must then beseen as being both produced and used up in theproduction process. Thus Labor’s production ofQ by using up K is represented as the produc-tion of L and then as the using up of L and Kin the production of Q. In vectorial terms,

Labor’s Product = (Q, –K, 0) = (0, 0, L) + (Q, –K, –L) = Labor Services + Whole Product.

This set of facts coupled with the principle ofresponsibility yields the assignment of the legalresponsibility for Labor’s Product (labor servicesplus whole product) to Labor. Thus Laborshould, by the principle of responsibility, legallyappropriate the whole product (in addition to thelabor services). In short, the people working inan enterprise should ‘be the firm’ (in the senseof whole product appropriator or residualclaimant). The principle of responsibility impliesthat production should be legally organized aswhat are called ‘democratic firms’ or ‘labor-managed firms’ where the people working in thefirm are the residual claimants.

This is a very striking result. Yet it is surpris-ingly robust. The principle of responsibility isclear, for example, in criminal and civil trials, andthere seems to be no reason why the same prin-ciple should not apply when no civil or criminalwrongs have been committed. There is muchcontroversy about the borderline cases of respon-sibility (e.g., impaired capacity and insanity), butthe principle has only been applied in the caseof productive work which epitomizes deliberateand intentional human actions.

Brief intellectual history of responsibility theory ofappropriation

The normative theory of property appropriationbased on the responsibility principle is not new;it represents a reworking and reformulation in amodern and consistent form of what was vari-ously called the ‘natural rights theory ofproperty’ (e.g., in Schlatter, 1951) or the ‘labortheory of property.’ The intellectual history ofthese ideas will be reviewed by considering themain ways in which the classical ideas needed tobe changed or reformulated to arrive at themodern theory:8

– expanding the site of appropriation fromsome Lockean ‘original state of nature’ toany human activity where property isproduced or consumed,9

– including the negative product (liabilities forthe used-up inputs) in the concept of the‘whole product’ so that the question ofappropriation could be posed in an alge-braically symmetric (plus and minus)manner,10

– identifying ‘responsibility’ as the uniquecharacteristic of actions of persons incontrast with the causally efficacious but‘non-responsible’ services of things,11

– identifying the basic normative principle ofthe ‘labor theory of property’ with thenatural principle of responsibility (‘Assign dejure responsibility in accordance with de factoresponsibility’),12

– decisively separating the responsibilitytheory of appropriation from the hopelesslyerroneous ‘labor theory of value’13

– and integrating the responsibility principlewith the theory of inalienable rights basedon the inalienability of de facto responsibilityso that the problem can be located not inthe market mechanism of appropriation butin the contract for the sale of human labor(the employment contract for the renting ofpersons).

I turn now to this last point, the theory ofinalienability which accounts for the distinctivetreatment of contracts for the alienation ofhuman capital.

The Democratic Firm 117

De facto inalienability of human action

It might be argued that one can ‘transfer’ theresponsibility for the results of one’s actions sothe employees in an enterprise, having ‘trans-ferred their labor,’ are not responsible for thepositive and negative results of their actions. Onlythe employer is responsible, according to thisargument, so the ownership should be assignedsolely to the employer. But this argumentconfuses de jure and de facto responsibility. Underthe legal institution of the employer-employeerelationship, legal or de jure responsibility isindeed transferred from the employees to theemployer. The normative principle of responsi-bility states that de jure responsibility should beassigned in accordance with de facto responsibilityso the question is whether or not de facto respon-sibility is transferred from the employees to theemployer. If not, then there is a conflict betweenthe principle of responsibility and the employer-employee relationship (the legal institution forthe renting of people).

This question can be illuminated by consid-ering the parable of the criminous employee.Suppose that an entrepreneur hires a van for aweek from its owner in an impersonal markettransaction. The van-owner is not otherwiseinvolved with the entrepreneur. The entrepre-neur also hires a worker as his assistant. Inaddition to employing the van and the workerin normal business, the entrepreneur employsthem to rob a bank. The employer and employeeare caught and hauled before a judge. Theemployee claims that his position is quite analo-gous to that of the van-owner. In both cases,certain services (man-days or van-days) were soldto the entrepreneur, and the later use of thoseservices by the entrepreneur are not the respon-sibility of the original seller of the services. Theentrepreneur telling the employee to do this orthat is like the entrepreneur “telling a grocer tosell [him] this brand of tuna rather than thatbrand of bread” (Alchian and Demsetz, 1972,1994). In either case, the grocer bears no respon-sibility for the subsequent use made of the tunaor bread.

The judge would no doubt be unimpressed bythis argument. He would ‘pierce the veil’ of the

labor contract to point out the factual differencein the alienability of labor services on the onehand and the services of a van or any thing (suchas tuna or bread) on the other hand. The use ofthe van could in fact be transferred from theowner to the entrepreneur so that the entrepre-neur could use the van without the owner beinginvolved. But it is factually impossible for aperson to do the same with his or her ownactions. At best the worker can only agree to co-operate with the entrepreneur, but then theworker shares the de facto responsibility for theresults. Having established those facts, the judgewould hold them both legally responsible for therobbery.14 The servant in work becomes thepartner in crime.15

It is sometimes thought that the worker isresponsible because an employment contractinvolving a crime is null and void, but the logicalorder is the reverse. The employee is legallyguilty because he knowingly committed acrime,16 not because of the bogus contract, andthus the employment contract must be set asidein order to view the worker as a co-venturer andto explicitly apply the responsibility principle.

The parable of the criminous employee illus-trates the non-transferability of de facto responsi-bility for human actions. Yet the facts of thematter do not change when the actions arelegal. Workers do not become some type of non-responsible instrument when they do not commitcrimes. Thus the argument that workers have“transferred” their de facto responsibility fails. Theprevious conclusion that Labor is de facto respon-sible for the whole product is sustained, andthus the responsibility principle implies that thepeople who work in an enterprise should legallyappropriate the positive and negative fruits oftheir labor.

This inalienable rights theory based on thede facto inalienability of human responsibilitydescends from the Reformation and Enlighten-ment (See Ellerman, 1992 and 1995). The de factonon-transferability of human labor implies thatthe entire contract for renting human beings (orrenting human capital) is inherently null and voidlike the longer term self-sale contract. It is nota question of ‘buy or lease’; neither the conceptof renting (hiring) nor buying should be applied

118 David Ellerman

to persons. With the contract for renting personsrecognized as being null and void (as the contractfor buying persons is currently recognized),Labor would always have to hire capital and themarket mechanism would correctly impute thewhole product to Labor.17

The debate about the nature of the corporation

I have outlined a descriptive and a normativeproperty theoretic treatment of the firm. Thedescriptive theory leaves unanswered theeconomic questions addressed in the economictheories of the firm. The normative theory ofthe firm was approached from the responsibilitytheory of appropriation. In the same way that thejoint work of the criminous employer andemployee is legally reconstructed as a ‘partner-ship’, so should the joint work of all who workin each productive opportunity be legally recon-structed as a ‘partnership’ or democratic firm.Moreover it found a fatal flaw in the contractfor the renting of human beings which ficti-tiously pretends that responsibility can be de factotransferred from one person to another. Thatcontract is the rental version of the self-salecontract – which is already recognized in con-temporary norms as being null and void. Indeed,the inalienable rights theory based on the non-transferability of de facto responsibility applies aswell to the self-rental contract as to the self-salecontract. The results about the firm and the laborcontract fit together in the sense that without theemployment contract, Labor would always behiring capital so the market imputation of thewhole product would go to Labor, i.e., the firmwould be a democratic or labor-managed firm.

Conceptual entry point

For a strictly logical viewpoint, the argumentspresented here find no flaw in the conventionalcorporation so long as it has no employees (an‘unstaffed’ company that just relends or leavesidle its capital) – just as the case against slaveryprovides no critique of a ‘slave plantation’ so long

as it has no slaves! But as was pointed out byCoase in his classic article on the nature of thefirm (1937 or 1988), the “legal relationshipnormally called that of ‘master and servant’ or‘employer and employee’ ” (1988, p. 53) providesthe best approximation to the firm in the realworld. Thus one may assume that the employ-ment relation is an essential part of the conven-tional joint stock corporation and that acorporation is not established just to relend orleave idle the shareholders’ capital. The abolitionof the employment relation could then be con-ceptually approached as changing the nature ofthe corporation to redefine the members of thecorporation to be the people who work in it. Inthis manner, the arguments presented hereprovide a conceptual entry point into the debateabout the nature of the corporation.

The shareholders’ property rights

The responsibility argument presented hereapplies to all firms, incorporated or not, and toall corporations, not just to large companies withdispersed publicly traded shares. I focus on thecorporation as opposed to unincorporated legalforms simply because it is the prevalent form ofbusiness organization. There does not seem to beany serious argument that the absentee share-holders are de facto responsible for producing thewhole product. That question is not posed. Butaren’t the shareholders’ control rights just amatter of property rights? Wouldn’t the democ-ratic firm take away “property rights” from theshareholders? This question requires a distinctionbetween two senses of control: the negative (orexclusionary) control rights to forbid anotherperson’s use of one’s property and the positive (ordiscretionary) control rights to direct the otherperson’s activity while using one’s property. Onlythe negative control rights come from propertyownership by itself – the right to exclude theother person from use of one’s property. Thepositive or discretionary control rights over theother person’s activity requires the ownership ofthat person’s ‘time’ – namely the employmentcontract.18 The ownership of the property thatthe other person might be using is only a bar-

The Democratic Firm 119

gaining chip to get the person to agree to theemployment contract.

Thus the rights of ownership in, say, land giveone the right to make another person a trespasserwho is using one’s property (if that use did nothave the owner’s consent) but it does not auto-matically make the other person a servant oremployee. The latter requires the employmentcontract. Since land can be rented out instead ofpeople being rented in, the landlord is not auto-matically the ‘lord’ of the land. This is a recentdevelopment. In medieval times, the authority ofa king over his subjects was seen as part of hisultimate ownership of the land in the country.“Rulership and Ownership were blent” (Gierke,1958, p. 88). Marx (incorrectly) thought thesame idea carried over to capital in more moderntimes and thus he misleadingly gave the name“capitalism” to the system of production basedon the employment contract.

It is not because he is a leader of industry that aman is a capitalist; on the contrary, he is a leaderof industry because he is a capitalist. The leader-ship of industry is an attribute of capital, just as infeudal times the functions of general and judgewere attributes of landed property. (Marx, 1977,pp. 450–451)

This idea that ‘rulership’ is part of capital own-ership has become a part of the standard under-standing of the “ownership of the firm” in the‘capitalist’ system.

The owner of capital resources, or the agent whoacts on behalf of the owner or a number of asso-ciated owners, controls and determines, in virtueof such ownership, the process of production and theaction of the workers who are engaged in the process.(Barker, 1967, pp. 105–106 emphasis added)

This assumption that the decision-makingauthority over people using an asset is part andparcel of the ownership of the asset (as if theemployment contract were superfluous) alsocreeps into the modern economics literature.19

It is precisely the decision-making version of thefundamental myth that the legal responsibility for(i.e., ownership of ) the product was part andparcel of the ownership of a capital asset (as if it

were superfluous for the capital-owner to buyall the non-capital inputs used in production).

By virtue of their property rights in a con-ventional corporation, the shareholders have(indirectly through management) the positive dis-cretionary control rights over ‘corporate affairs’but it is only by virtue of the employmentcontract that the activities of the workers becomepart of ‘corporate affairs.’ In short, the employ-ment or self-rental contract is necessary to estab-lish the positive control rights over the employeesin the conventional firm. It has been argued thatthese ‘property rights’ based on the self-rentalcontract are invalid like the ‘property rights’slaveowners would have based on voluntary self-sale contracts. Hence the shareholders’ claim tothe positive discretionary control rights over cor-porate employees based on ‘property rights’ failsbecause (1) no such rights are part of the own-ership of capital assets, and (2) the employmentcontract to acquire such ‘property rights’ isnaturally invalid. Without the contract to rentother human beings, the owner of capital canrent the capital out or reserve it for own usage.In the conversion of a conventional corporationto a democratic firm, the shareholders’ rolewould be redefined as a rentier role while theresidual claimant or membership rights would bereassigned to the people working in the company.

Final points

It is clear that the arguments and points of viewpresented here are ‘heresies’ at odds with today’s‘orthodoxy.’ The differences are not in the detailsor only at the edges; the differences are majorand central. Perhaps the most surprising pointis that the differences are mainly on factualquestions rather than normative principles. Forinstance, the normative principle of assigning dejure responsibility in accordance with de factoresponsibility is fundamental to contemporaryjurisprudence and is thus not at odds with today’sconventional wisdom. Few of the orthodoxwill respond to these arguments by launchingan attack on the responsibility principle. Theprimary differences are concerned with the factsabout conventional institutions (e.g., the struc-

120 David Ellerman

ture of property rights in production) and abouthuman beings. In conclusion, I will mentionthree institutional facts and two facts aboutpersons that are basic to the arguments givenhere:

• that the identity of the residual claimant(‘firm’) in a production opportunity isdetermined in a market economy by thecontractual fact-pattern of being the last legalowner of the inputs (and not by somemythical ‘ownership of the firm’);

• that the ownership of the whole product ofa production opportunity is legally appro-priated (as opposed to being an alreadyowned part of the mythical ‘ownership ofthe firm’);

• that the whole product is appropriatedby one legal party, the residual claimant(as opposed to the ‘distributive shares’metaphors which picture each factor asowning part of the product);

• that human actions (as opposed to theservices of things) are the only de factoresponsible factor in production (in contrastto the engineering picture that all inputs arepassive or to the poetic picture that allinputs are active) which entails that thepeople working in a firm are de facto respon-sible for producing the whole product; and

• that the de facto responsibility of persons isin fact non-transferable from one person toanother (in contrast to, say, the responsibilityfor the use of a tool like a truck or van).

If the above factual theses be admitted, theheretical conclusions will follow closely behind,so to resist these heresies, it is incumbent on theobjective social scientist to show wherein thesefactual theses are false.

Notes

† Research on this paper was funded in part througha contract with the Brookings Institution, throughits project on Corporations and Human Capital.Opinions expressed in this article are those of theauthor, and not those of the Brookings Institution, itsofficers, directors, or financial supporters.

* The findings, interpretations, and conclusions

expressed in this paper are entirely those of the authorand should not be attributed in any manner to theWorld Bank, to its affiliated organizations, or to themembers of its Board of Directors or the countriesthey represent.1 For an excellent anthology exclusively on the“economic” analysis of the firm, see Putterman andKroszner (1996). 2 “One can even say that wages are the rentals paidfor the use of a man’s personal services for a day ora week or a year. This may seem a strange use ofterms, but on second thought, one recognizes thatevery agreement to hire labor is really for somelimited period of time. By outright purchase, youmight avoid ever renting any kind of land. But in oursociety, labor is one of the few productive factors thatcannot legally be bought outright. Labor can only berented, and the wage rate is really a rental.”(Samuelson, 1976, p. 569) 3 There are parallel arguments based on democratictheory that look at the governance aspects of the firmrather than the structure of property rights. See Dahl(1985) or Ellerman (1992). 4 “This word [expropriation] primarily denotes avoluntary surrender of rights or claims; the act ofdivesting oneself of that which was previously claimedas one’s own, or renouncing it. In this sense, it is theopposite of ‘appropriation’. A meaning has beenattached to the term, imported from foreign jurispru-dence, which makes it synonymous with the exerciseof the power of eminent domain, . . . .” (Black, 1968,p. 692, entry under “Expropriation”) 5 For example in Putterman and Kroszner anthology(1996) on the “economic” nature of the firm, noneof the papers pose the question of appropriation intheir treatment of the firm. The question of appro-priation in the firm is similarly ignored in the“economics of property rights” (e.g., Furubotn andPejovich, 1974) and in the so-called “property rightsapproach” to the firm (e.g., Hart and Moore, 1990;Hart, 1995). 6 While the lay misinterpretation of ownership of acorporation might be understandable in a world ofsignificant transactions costs, it is less clear why econ-omists should still be wedded to the lay concept inthe “standard model” of general equilibrium in anidealized world free of transaction costs. In partic-ular, the fabled attempt by Arrow and Debreu (1954)to show the existence of a competitive equilibriumwith positive “pure profits which are distributed tothe owners of the firm” (Arrow, 1971, p. 70) is flawedin theory because they incorrectly assume there is theownership of production sets in a private property

The Democratic Firm 121

market economy. The ownership of a corporation ismisinterpreted as the ownership of a production seteven in the idealized frictionless model where achallenger corporation could bid slightly higherfor the inputs (and get lower but still positive profits)to take over a production opportunity from an incum-bent corporation and thus defeat the purported‘competitive equilibrium’ with positive pure profits.7 Wieser’s development of economic notions of“responsibility” and ‘imputation’ illustrates an impor-tant intellectual strategy of the science of economics.How is economics to protect itself against the ‘night-mare’ of being invaded by jurisprudence and fromhaving some ‘economic’ question addressed byjuridical principles (e.g., as in this paper) outside theprofessional expertise of economists? Offense mightbe the best defense, and thus there have been‘economic’ theories of law, politics, and much else.These intellectual forays might involve developing‘economic’ versions of non-economic concepts (e.g.,Wieser’s metaphors) and then redefining the impor-tant questions as the questions that could be dealt withusing economics concepts (e.g., ignoring questions ofproperty appropriation in favor of wage and pricedetermination by marginal productivity theory).8 See Ellerman (1992) for a more complete treat-ment. 9 Any discussion of the appropriation or initiationof property rights is ordinarily relegated to a rathermythical original state of nature (e.g., in the philo-sophical literature) or to a situation where propertyprevious held in common is being privatized. Forinstance, Harold Demsetz (1967) considers howprivate property in land with fur-bearing animals wasestablished as a result of growth in the fur trade. JohnUmbeck (1981) considers how gold rights were estab-lished in the 1848 California gold rush on landrecently ceded from Mexico. Yoram Barzel (1989)considers how the common property rights tominerals under the North Sea were privatized. Butin Barzel’s book (see particularly Chapter 5 ‘Theformation of rights’, 1989) as elsewhere in the eco-nomics of property rights literature, there is no recog-nition that the appropriation of the outputs (and thesymmetrical termination of rights to the used upinputs) takes place in normal production. Thus thequestion of appropriation at the heart of the theoryof the firm is not even posed in the ‘economics’literature on property rights and the firm. Once thequestion of appropriation is well-posed, the descrip-tive answer of the market mechanism of appropria-tion and the normative answer of the responsibilityprinciple are relatively straightforward.

10 Classical treatments of the labor theory of property(e.g., see Menger, 1899) tended to assert ‘Labour’sClaim to the Whole Product’ without being clearabout the inclusion of the negative product. Thisview of ‘immaculate appropriation’ led to mucheasy criticism of the theory as having neglectedthe other scarce inputs. The algebraically symmetricdescription of production (using a vector withthe outputs as positive and the inputs as negative)came into common usage in recent decades as pro-duction opportunities came to be described withproduction sets rather than production functions.Applying the old label of ‘whole product’ to theseproduction vectors allows the modern and consistentreformulation of ‘Labour’s Claim to the WholeProduct.’ It is perhaps ironic that the usual treatmentof marginal productivity theory also indulges inthe notion of ‘immaculate’ production. A unit ofan input cannot produce its “marginal product” exnihilo. Other inputs need to be used up. Marginalproductivity theory needs to be reformulated usingthe vectorial ‘marginal whole product of a factor’so that the market value of the marginal wholeproduct of a factor equals the price of the factorwhen profits are maximized. See Chapter 5 ‘AreMarginal Products Created ex Nihilo?’ in Ellerman(1995). 11 Even as the legally-trained Austrian economistFriedrich von Wieser introduced the metaphoricalnotions of ‘economic responsibility’ and ‘imputation,’he showed an understanding of the non-metaphoricalnotions in jurisprudence. But the metaphoricalnotions of responsibility and imputation have capturedthe minds of economists and the minds of thosewho see the world through the economics mind-set.For example, the reader as intellectual anthropologistis urged to try to find a single economics text writtenin, say, the last half century (e.g., Samuelson’sprimer and its vast progeny) that shows even Wieser’sappreciation of the non-metaphorical notions ofresponsibility and imputation in the text’s treatmentof marginal productivity theory or the ‘labor theoryof value.’ Thus one has the common picture whereall the factors of production – land, labor, andcapital – are seen as active agents cooperating togetherto produce the product. Alternatively, all the factorsare seen as being passive inputs that are ‘used up’as the output is ‘produced’ in an engineeringdescription of production. The economics mind-setavoids the ‘non-economic’ asymmetry of treatingthe responsible actions of persons as being funda-mentally different from the non-responsible servicesof things.

122 David Ellerman

12 Since the responsibility principle is fundamental toordinary ‘bourgeois’ jurisprudence, this interpretationof labor theory of property removes any bizarre or‘radical’ connotations of the theory and makes itunderstandable to the layperson. Moreover, this inter-pretation completes the decoupling of the principlesof private property from the conventional firm (basedon the employment relation) – a decoupling begunby the understanding of the fundamental ‘ownershipof the firm’ myth. It is the democratic firm that isbased on the responsibility principle in jurisprudence.13 The labor theory of value, particularly in itsMarxian form, is surely one of the most spectacularfailures in the history of economic thought. Once thelabor theory of property is cleanly formulated as atheory of property appropriation, it is seen to havenothing whatever to do with value or prices. Thereis always a strong temptation to try to attack the laboror responsibility theory of property by associating itwith the labor theory of value, Marxism, and all that– an attack that is particularly ironic since the respon-sibility principle is the basis for the just appropria-tion of private property. 14 “[B]oth the principal and the agent, the personwho hires the hit man and the hit man who carriesout the murder, are held liable.” (Coleman, 1982,p. 99)15 “The general thesis in the hit-man case is straight-forward: agents are not held responsible for actionsthat, if taken under one’s own authority, are notcriminal, but they are held personally responsible foractions that are criminal acts as defined by the law ofthe land.” (Coleman, 1982, p. 99) 16 “All who participate in a crime with a guiltyintent are liable to punishment. A master and servantwho so participate in a crime are liable criminally, notbecause they are master and servant, but because theyjointly carried out a criminal venture and are bothcriminous.” (Batt, 1967, p. 612) A hired killer is con-victed of murder because he committed murder, notbecause he engaged in an illegal murder-for-hirecontract. 17 The fundamental theorem of property theoryshowing how the market mechanism of appropriationworks correctly in a private property market economysans the employment contract is outlined in Ellerman,1992.18 An employment contract is not needed if theperson using the property is the owner but I am con-sidering the case where the roles of worker and capitalsupplier are kept separate. 19 For instance, the “rights of authority at the firmlevel are defined by the ownership of assets, tangible

(machines or money) or intangible (goodwill orreputation).” (Holmstrom and Tirole, 1989, p. 123)

References

Alchian, A. and H. Demsetz: 1972, ‘Production,Information Costs, and Economic Organization’,American Economic Review 62, 777–795.

Arrow, K. J.: 1971, ‘The Firm in General EquilibriumTheory’, in R. Marris and A. Woods (eds.), TheCorporate Economy (Harvard University Press,Cambridge MA).

Arrow, K. J. and G. Debreu: 1954, ‘Existence ofan Equilibrium for a Competitive Economy’,Econometrica 22, 265–290.

Barker, Ernest: 1967, Reflections on Government(Oxford University Press, London).

Barzel, Y.: 1989, Economic Analysis of Property Rights(Cambridge University Press, New York).

Batt, Francis: 1967, The Law of Master and Servant,Fifth ed. G. Webber (Pitman, London).

Black, H.: 1968, Black’s Law Dictionary (WestPublishing, St. Paul MN).

Blair, M. M.: 1995, Ownership and Control: RethinkingCorporate Governance for the Twenty-First Century(The Brookings Institution, Washington DC).

Christ, Carl F.: 1975, ‘The Competitive Market andOptimal Allocative Efficiency’, in John Elliott andJohn Cownie (eds.), Competing Philosophies inAmerican Political Economics (Goodyear, PacificPalisades CA), pp. 332–338.

Coase, R. H.: 1937, ‘The Nature of the Firm’,Economica IV (Nov.), 386–405.

Coase, R. H.: 1988, The Firm, the Market, and the Law(University of Chicago Press, Chicago).

Coleman, J. S.: 1982, The Asymmetric Society (SyracuseUniversity Press, Syracuse).

Dahl, Robert: 1985, Preface to Economic Democracy(University of California Press, Berkeley).

Demsetz, H.: 1967, ‘Toward a Theory of PropertyRights’, American Economic Review 57 (May),347–359.

Dworkin, R.: 1978, Taking Rights Seriously (HarvardUniversity Press, Cambridge).

Ellerman, David P.: 1992, Property & Contract inEconomics: The Case for Economic Democracy(Blackwell, Cambridge MA).

Ellerman, David P.: 1995, Intellectual Trespassing as aWay of Life: Essays in Philosophy, Economics, andMathematics (Rowman & Littlefield, LanhamMD).

Fama, E. and M. Jensen: 1996, ‘Organizational Forms

The Democratic Firm 123

and Investment Decisions’, in L. Putterman and R.Kroszner (eds.), The Economic Nature of the Firm,Second edition (Cambridge University Press,Cambridge), pp. 336–344.

Furubotn, E. and S. Pejovich (eds.): 1974, TheEconomics of Property Rights (Ballinger PublishingCo., Cambridge MA).

Gierke, Otto von: 1958, Political Theories of the MiddleAge. Trans. F. W. Maitland (Beacon Press, Boston).

Haavelmo, T.: 1960, A Study in the Theory of Investment(University of Chicago Press, Chicago).

Hart, O.: 1995, Firms, Contracts, and Financial Structure(Clarendon Press, Oxford).

Hart, O. and J. Moore: 1990, ‘Property Rights andthe Nature of the Firm’, Journal of Political Economy98, 1119–1159.

Hirshleifer, J.: 1970, Investment, Interest, and Capital(Prentice-Hall, Englewood Cliffs NJ).

Holmstrom, B. and J. Tirole: 1989, ‘The Theory ofthe Firm’, in R. Schmalensee and R. Willig (eds.),Handbook of Industrial Organization Vol. I (North-Holland, Amsterdam), pp. 61–133.

Marx, Karl: 1977, Capital. Volume I. Trans. B.Fowkes (Vintage Books, New York).

Menger, Anton: 1899, The Right to the Whole Produceof Labour: The Origin and Development of the Theoryof Labour’s Claim to the Whole Product of Industry.Trans. M. E. Tanner. Intro. by Herbert S. Foxwell

(Macmillan, London) (Reprinted by AugustusKelley).

Nozick, Robert: 1974, Anarchy, State, and Utopia(Basic Books, New York).

Putterman, L. and R. S. Kroszner (eds.): 1996, TheEconomic Nature of the Firm. 2nd edition(Cambridge University Press, Cambridge).

Samuelson, P.: 1976, Economics. Tenth edition(McGraw-Hill, New York).

Schlatter, Richard: 1951, Private Property: The Historyof an Idea (Rutgers University Press, NewBrunswick).

Tobin, J.: 1970, ‘On Limiting the Domain ofInequality’, Journal of Law and Economics (13 Oct.).

Umbeck, J.: 1981, ‘Might Makes Right: A Theoryof the Formation and Initial Distribution ofProperty Rights’, Economic Inquiry 19(1), 38–59.

Varian, H.: 1984, Microeconomic Analysis. 2nd ed.(W.W. Norton, New York).

Wieser, F. von: 1930, Natural Value. Trans. C.A.Malloch (G.E. Stechert and Co., New York).

World Bank,1818 H St., NW

Washington, DC 20433,U.S.A.

124 David Ellerman