Separation Of Ownership And Control (AQA 1983) PDF

SEPARATION OF OWNERSHIP AND CONTROL* EUGENE F. FAMA and MICHAEL C. JENSEN University of Chicago University of Rochester I. INTRODUCTION ABSENTfiat, the form of org...

SEPARATION OF OWNERSHIP AND CONTROL* EUGENE F. FAMA and MICHAEL C. JENSEN University of Chicago University of Rochester I. INTRODUCTION ABSENTfiat, the form of organizationthat survives in an activity is the one that delivers the productdemandedby customers at the lowest price while covering costs.' Ourgoal is to explain the survivalof organizations characterizedby separationof "ownership" and "control"-a problem that has botheredstudents of corporationsfromAdam Smithto Berle and Means and Jensen and Meckling.2 In more precise language,we are con- cerned with the survival of organizationsin which importantdecision agents do not bear a substantialshare of the wealth effects of their de- cisions. We argue that the separation of decision and risk-bearingfunctions observed in large corporationsis common to other organizationssuch as large professional partnerships, financial mutuals, and nonprofits. We contend that separationof decision and risk-bearingfunctions survives in these organizationsin part because of the benefits of specialization of * This paper is a revision of parts of our earlier paper, The Survival of Organizations (September 1980). In the course of this work we have profited from the comments of R. Antle, R. Benne, F. Black, F. Easterbrook, A. Farber, W. Gavett, P. Hirsch, R. Hogarth, C. Holderness, R. Holthausen, C. Horne, J. Jeuck, R. Leftwich, S. McCormick, D. Mayers, P. Pashigian, M. Scholes, C. Smith, G. Stigler, R. Watts, T. Whisler, R. Yeaple, J. Zimmer- man, and especially A. Alchian, W. Meckling, and C. Plosser. Financial support for Fama's participation is from the National Science Foundation. Jensen is supported by the Man- agerial Economics Research Center of the University of Rochester. Armen A. Alchian, Uncertainty, Evolution and Economic Theory, 58 J. Pol. Econ. 211 (1950), is an early proponent of the use of natural selection in economic analysis. For a survey of general issues in the analysis of organizations, see Michael C. Jensen, Organiza- tion Theory and Methodology, 50 Accounting Rev. (1983). 2 Adam Smith, The Wealth of Nations (Cannan ed. 1904) (1st ed. London 1776); Adolf A. Berle & Gardiner C. Means, The Modem Corporation and Private Property (1932); Michael C. Jensen & William H. Meckling, Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure, 3 J. Financial Econ. 305 (1976). [Journal of Law & Economics, vol. XXVI (June 1983)] ? 1983 by The University of Chicago. All rights reserved. 0022-2186/83/2602-0011$01.50 301 This content downloaded from 141.218.001.105 on August 04, 2016 23:19:35 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). 302 THE JOURNAL OF LAW AND ECONOMICS and risk bearingbut also becauseof an effectivecommon management approach to controlling the agency problems caused by separation of decision and risk-bearingfunctions. In particular,our hypothesis is that the contract structures of all of these organizations separate the ratificationand monitoringof decisions from initiationand implementa- tion of the decisions. II. RESIDUALCLAIMSAND DECISIONPROCESSES An organization is the nexus of contracts, written and unwritten, among owners of factors of productionand customers.3These contracts or internal "rules of the game" specify the rights of each agent in the organization,performancecriteriaon which agents are evaluated,and the payoff functions they face. The contract structurecombines with avail- able productiontechnologies and external legal constraintsto determine the cost functionfor deliveringan outputwith a particularform of organi- zation.4 The form of organizationthat delivers the output demandedby customers at the lowest price, while covering costs, survives. The centralcontractsin any organizationspecify (1) the natureof resid- ual claims and (2) the allocationof the steps of the decision process among agents. These contracts distinguishorganizationsfrom one another and explain why specific organizationalforms survive. We first discuss the generalcharacteristicsof residualclaims and decision processes. We then presentthe majorhypotheses aboutthe relationsbetween efficientalloca- tions of residual claims and decision functions. The analysis focuses on two broad types of organizations-those in which risk-bearingand deci- sion functions are separatedand those in which they are combinedin the same agents. We analyze only private organizations that depend on voluntarycontractingand exchange. A. Residual Claims The contract structures of most organizationalforms limit the risks undertakenby most agents by specifyingeitherfixed promisedpayoffs or incentive payoffs tied to specific measures of performance.The residual risk-the risk of the difference between stochastic inflows of resources and promisedpaymentsto agents-is borneby those who contractfor the rights to net cash flows. We call these agents the residual claimants or residualrisk bearers. Moreover, the contractsof most agents containthe 3 See Jensen & Meckling, supra note 2. 4See Michael C. Jensen & William H. Meckling, Rights and Production Functions: An Application to Labor-managed Firms and Codetermination, 52 J. Bus. 469 (1979). This content downloaded from 141.218.001.105 on August 04, 2016 23:19:35 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).  OWNERSHIP AND CONTROL 303 implicit or explicit provision that, in exchange for the specified payoff, the agent agrees that the resources he provides can be used to satisfy the interests of residual claimants. Havingmost uncertaintyborne by one groupof agents, residualclaim- ants, has survival value because it reduces the costs incurredto monitor contracts with other groups of agents and to adjust contracts for the changing risks borne by other agents. Contracts that direct decisions towardthe interests of residualclaimantsalso add to the survivalvalue of organizations.Producingoutputs at lower cost is in the interestsof resid- ual claimants because it increases net cash flows, but lower costs also contribute to survival by allowing products to be delivered at lower prices. The residual claims of differentorganizationalforms contain different restrictions.For example, the least restrictedresidualclaims in common use are the common stocks of large corporations.Stockholdersare not requiredto have any other role in the organization;their residualclaims are alienable without restriction;and, because of these provisions, the residual claims allow unrestrictedrisk sharingamong stockholders. We call these organizations open corporations to distinguish them from closed corporationsthat are generally smaller and have residual claims that are largely restrictedto internaldecision agents.5 B. The Decision Process By focusing on entrepreneurialfirms in which all decision rights are concentratedin the entrepreneur,economists tend to ignore analysis of the steps of the decision process. However, the way organizationsallo- cate the steps of the decision process across agents is importantin ex- plainingthe survival of organizations. In broadterms, the decision process has four steps: 1. initiation-generation of proposalsfor resourceutilizationand struc- turingof contracts; 2. ratification--choice of the decision initiativesto be implemented; 3. implementation-execution of ratified decisions; and 4. monitoring--measurementof the performanceof decision agents and implementationof rewards. Because the initiation and implementationof decisions typically are 5 The terms "public corporation" and "close corporation," which are common in the legal literature, are not used here. "Closed corporation" seems more descriptive than "close corporation." The term "public corporation" best describes government-owned corporations such as Amtrak and the TVA. In contrast, what we call "open corporations" are private organizations. This content downloaded from 141.218.001.105 on August 04, 2016 23:19:35 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). 304 THE JOURNAL OF LAW AND ECONOMICS allocatedto the same agents, it is convenient to combine these two func- tions under the term decision management. Likewise, the term decision control includes the ratificationand monitoringof decisions. Decision managementand decision control are the components of the organiza- tion's decision process or decision system. III. FUNDAMENTAL RELATIONS BETWEEN RISK-BEARING AND DECISION PROCESSES We firststate and then elaboratethe centralcomplementaryhypotheses about the relations between the risk-bearingand decision processes of organizations. 1. Separationof residualrisk bearingfrom decision managementleads to decision systems that separate decision managementfrom decision control. 2. Combinationof decision managementand decision control in a few agents leads to residual claims that are largely restricted to these agents. A. The Problem Agency problemsarise because contractsare not costlessly writtenand enforced. Agency costs include the costs of structuring,monitoring,and bondinga set of contractsamongagentswith conflictinginterests.Agency costs also include the value of output lost because the costs of full en- forcement of contracts exceed the benefits.6 Controlof agency problemsin the decision process is importantwhen the decision managerswho initiateand implementimportantdecisions are not the majorresidualclaimantsand thereforedo not bear a majorshare of the wealth effects of their decisions. Withouteffective control proce- dures, such decision managersare more likely to take actions that deviate from the interests of residualclaimants.An effective system for decision control implies, almost by definition, that the control (ratificationand monitoring)of decisions is to some extent separatefrom the management (initiationand implementation)of decisions. Individualdecision agents can be involved in the managementof some decisions and the control of others, but separationmeans that an individualagent does not exercise exclusive managementand control rights over the same decisions. The interestingproblem is to determine when separationof decision management,decision control, and residualrisk bearingis more efficient 6 This definition of agency costs comes from Jensen & Meckling, supra note 2. This content downloaded from 141.218.001.105 on August 04, 2016 23:19:35 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).  OWNERSHIP AND CONTROL 305 than combiningthese three functionsin the same agents. We firstanalyze the factors that make combinationof decision management,decision con- trol, and residualrisk bearingefficient. We then analyze the factors that make separationof these three functions efficient. B. Combination of Decision Management, Decision Control, and Residual Risk Bearing Suppose the balance of cost conditions, includingboth technology and the control of agency problems, implies that in a particularactivity the optimal organization is noncomplex. For our purposes, noncomplex means that specific informationrelevant to decisions is concentratedin one or a few agents. (Specific informationis detailed informationthat is costly to transfer among agents.)7 Most small organizationstend to be noncomplex, and most large organizationstend to be complex, but the correspondence is not perfect. For example, research oriented univer- sities, though often small in terms of assets or faculty size, are never- theless complex in the sense that specific knowledge, which is costly to transfer,is diffused among both faculty and administrators.On the other hand, mutualfunds are often large in terms of assets but noncomplexin the sense that informationrelevantto decisions is concentratedin one or a few agents. We take it as given that optimalorganizationsin some activi- ties are noncomplex. Our more limitedgoal is to explainthe implications of noncomplexityfor control of agency problemsin the decision process. If we ignore agency problemsbetween decision managersand residual claimants,the theory of optimalrisk bearingtells us that residualclaims that allow unrestrictedrisk sharinghave advantagesin small as well as in large organizations.8 However, in a small noncomplex organization, specific knowledge importantfor decision managementand control is concentratedin one or a few agents. As a consequence, it is efficient to allocate decision control as well as decision managementto these agents. Withoutseparationof decision managementfrom decision control, resid- 7 Specific information is closely related to the notions of "information impactedness" and "bounded rationality" discussed in Oliver E. Williamson, Markets and Hierarchies: Analy- sis and Antitrust Implications (1975) and The Modern Corporation: Origins, Evolution, Attributes, 19 J. Econ. Literature 1537 (1981). Friedrich A. von Hayek, The Use of Knowl- edge in Society, 35 Am. Econ. Rev. 519 (1945) uses specific information to discuss the role of markets in complex economies. See also Thomas Sowell, Knowledge and Decisions 13- 14 (1980). Our analysis of the relations between specific information and efficient decision processes owes much to ongoing work with William Meckling. 8 See, for example, Kenneth J. Arrow, The Role of Securities in the Optimal Allocation of Risk Bearing, 31 Rev. Econ. Stud. 91 (1964); or Eugene F. Fama, Foundations of Finance chs. 6 & 7 (1976). This content downloaded from 141.218.001.105 on August 04, 2016 23:19:35 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). 306 THE JOURNAL OF LAW AND ECONOMICS ual claimantshave little protection against opportunisticactions of deci- sion agents, and this lowers the value of unrestrictedresidualclaims. A feasible solution to the agency problem that arises when the same agents manage and control important decisions is to restrict residual claims to the importantdecision agents. In effect, restrictionof residual claims to decision agents substitutesfor costly controldevices to limitthe discretionof decision agents. The common stocks of closed corporations are this type of restricted residual claim, as are the residual claims in proprietorshipsand partnerships.The residualclaims of these organiza- tions (especially closed corporations)are also held by other agents whose special relations with decision agents allow agency problemsto be con- trolled without separationof the managementand control of decisions. For example, family members have many dimensions of exchange with one anotherover a long horizon and thereforehave advantagesin moni- toringand discipliningrelateddecision agents. Business associates whose goodwill and advice are importantto the organizationare also potential candidatesfor holding minority residual claims of organizationsthat do not separatethe managementand control of decisions.' Restrictingresidual claims to decision makers controls agency prob- lems between residualclaimantsand decision agents, but it sacrificesthe benefits of unrestrictedrisk sharingand specializationof decision func- tions. The decision process suffers efficiency losses because decision agents must be chosen on the basis of wealth and willingnessto bear risk as well as for decision skills. The residual claimantsforgo optimal risk reduction through portfolio diversification so that residual claims and decision makingcan be combined in a small numberof agents. Forgone diversificationlowers the value of the residualclaims and raises the cost of risk-bearingservices. Moreover, when residualclaims are restrictedto decision agents, it is generally rational for the residual claimant-decision makers to assign lower values to uncertain cash flows than residual claimants would in organizationswhere residualclaims are unrestrictedand risk bearingcan be freely diversifiedacross organizations.As a consequence, restricting residual claims to agents in the decision process leads to decisions (for example, less investmentin risky projectsthat lower the costs of outputs) that tend to penalize the organizationin the competitionfor survival.10 9 In contrast, the analysis predicts that when venture equity capital is put into a small entrepreneurial organization by outsiders, mechanisms for separating the management and control of important decisions are instituted. 1o These propositions are developed in Eugene F. Fama & Michael C. Jensen, Organiza- tional Forms and Investment Decisions (Working Paper No. MERC 83-03, Univ. Rochester, Managerial Economics Research Center 1983). This content downloaded from 141.218.001.105 on August 04, 2016 23:19:35 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).  OWNERSHIP AND CONTROL 307 However, because contractsare not costlessly writtenand enforced,all decision systems and systems for allocatingresidualclaims involve costs. Organizationalsurvivalinvolves a balanceof the costs of alternativedeci- sion systems and systems for allocatingresidualrisk againstthe benefits. Smallnoncomplexorganizationsdo not have demandsfor a wide rangeof specialized decision agents; on the contrary, concentrationof specific informationrelevant to decisions implies that there are efficiency gains when the rightsto manageand controldecisions are combinedin one or a few agents. Moreover, the risk-sharingbenefits forgone when residual claims are restrictedto one or a few decision agents are less serious in a small noncomplex organizationthan in a large organization,because the total risk of net cash flows to be shared is generally smaller in small organizations. In addition, small organizationsdo not often have large demandsfor wealth from residualclaimantsto bond the payoffs promised to other agents and to purchase risky assets. As a consequence, small noncomplex organizations can efficiently control the agency problems caused by the combinationof decision managementand control in one or a few agentsby restrictingresidualclaimsto these agents. Such a combin- ing of decision and risk-bearingfunctions is efficientin smallnoncomplex organizations because the benefits of unrestricted risk sharing and specializationof decision functions are less than the costs that would be incurredto control the resultingagency problems. The proprietorships,partnerships,and closed corporationsobserved in small scale production and service activities are the best examples of classical entrepreneurialfirms in which the major decision makers are also the majorresidualrisk bearers. These organizationsare evidence in favor of the hypothesis that combinationof decision managementand decision control in one or a few agents leads to residualclaims that are largely restrictedto these agents. We analyze next the forces that make separationof decision manage- ment, decision control, and residualrisk bearingefficient-in effect, the forces that cause the classical entrepreneurialfirm to be dominatedby organizationalforms in which there are no decision makersin the classical entrepreneurialsense. C. Separation of Decision Management, Decision Control, and Residual Risk Bearing Our concern in this section is with the organizationalforms character- ized by separationof decision managementfrom residualrisk bearing- what the literatureon open corporations calls, somewhat imprecisely, separationof ownership and control. Our hypothesis is that all such or- This content downloaded from 141.218.001.105 on August 04, 2016 23:19:35 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).  308 THE JOURNAL OF LAW AND ECONOMICS ganizations,includinglargeopen corporations,largeprofessionalpartner- ships, financialmutuals,and nonprofits,controlthe agency problemsthat result from separationof decision managementfrom residualrisk bearing by separatingthe management(initiationand implementation)and control (ratificationand monitoring) of decisions. Documentation of this hy- pothesis takes up much of the rest of the paper. 1. Specific Knowledge and Diffusion of Decision Functions. Most organizationscharacterizedby separationof decision managementfrom residual risk bearing are complex in the sense that specific knowledge relevant to different decisions-knowledge which is costly to transfer across agents-is diffused among agents at all levels of the organization. Again, we take it as given that the optimalorganizationsin some activities are complex. Ourtheory attemptsto explainthe implicationsof complex- ity for the natureof efficientdecision processes and for control of agency problemsin the decision process. Since specific knowledge in complex organizationsis diffused among agents, diffusionof decision managementcan reduce costs by delegating the initiationand implementationof decisions to the agents with valuable relevant knowledge. The agency problems of diffuse decision manage- ment can then be reduced by separatingthe management(initiationand implementation)and control (ratificationand monitoring)of decisions. In the unusual cases where residual claims are not held by important decision managers but are nevertheless concentrated in one or a few residualclaimants,control of decision managerscan in principlebe direct and simple, with the residualclaimantsratifyingand monitoringimportant decisions and setting rewards." Such organizationsconform to our hy- pothesis, because top-level decision control is separatedfrom top-level decision managersand exercised directly by residualclaimants. However, in complex organizationsvaluable specific knowledge rele- vant to decision control is diffused among many internal agents. This generallymeans that efficientdecision control, like efficientdecision man- agement, involves delegationand diffusionof decision control as Wellas separationof decision managementand control at differentlevels of the organization.We expect to observe such delegation, diffusion, and sep- arationof decision managementand control below the top level of com- plex organizations,even in those unusual complex organizationswhere residualclaims are held primarilyby top-level decision agents. 2. Diffuse Residual Claims and Delegation of Decision Control. In the more common complex organizations,residual claims are diffused " See Armen A. Alchian & Harold Demsetz, Production, Information Costs, and Eco- nomic Organization, 62 Am. Econ. Rev. 777 (1972). This content downloaded from 141.218.001.105 on August 04, 2016 23:19:35 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).  OWNERSHIP AND CONTROL 309 among many agents. Having many residual claimantshas advantagesin large complex organizationsbecause the total risk of net cash flows to be shared is generally large and there are large demands for wealth from residualclaimantsto bond the payoffs promisedto a wide rangeof agents and to purchaserisky assets. When there are many residualclaimants,it is costly for all of them to be involved in decision controland it is efficient for them to delegate decision control. For example, some delegation of decision control is observed even in the largeprofessionalpartnershipsin public accountingand law, where the residualclaimantsare expert inter- nal decision agents. When there are manypartnersit is inefficientfor each to participatein ratificationand monitoringof all decisions. Nearly complete separationand specializationof decision control and residualrisk bearingis common in large open corporationsand financial mutualswhere most of the diffuse residualclaimantsare not qualifiedfor roles in the decision process and thus delegate their decision control rights to other agents. When residual claimantshave no role in decision control, we expect to observe separationof the managementand control of importantdecisions at all levels of the organization. Separation and diffusion of decision management and decision control-in effect, the absence of a classical entrepreneurialdecision maker-limit the power of individualdecision agents to expropriatethe interestsof residualclaimants.The checks and balances of such decision systems have costs, but they also have importantbenefits. Diffusionand separation of decision managementand control have benefits because they allow valuable knowledge to be used at the points in the decision process where it is most relevant and they help control the agency prob- lems of diffuse residualclaims. In complex organizations,the benefits of diffuse residualclaims and the benefitsof separationof decision functions from residualrisk bearingare generallygreaterthanthe agency costs they generate, includingthe costs of mechanismsto separatethe management and control of decisions. 3. Decision Control in Nonprofits and Financial Mutuals. Most or- ganizations characterized by separation of decision managementfrom residual risk bearing are complex. However, separationof the manage- ment and control of decisions contributesto the survivalof any organiza- tion where the importantdecision managers do not bear a substantial share of the wealth effects of their decisions-that is, any organization where there are serious agency problems in the decision process. We argue below that separation of decision managementand residual risk bearingis a characteristicof nonprofitorganizationsand financialmutu- als, large and small, complex and noncomplex. Thus, we expect to ob- serve separationof the managementand control of importantdecisions This content downloaded from 141.218.001.105 on August 04, 2016 23:19:35 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). 310 THE JOURNAL OF LAW AND ECONOMICS even in small noncomplex nonprofitsand financialmutualswhere, ignor- ing agency problemsin the decision process, concentratedand combined decision managementand control would be more efficient. 4. CommonGeneralFeatures of Decision ControlSystems. Ourhy- pothesis about the decision systems of organizationscharacterizedby separationof decision managementand residualrisk bearinggets support from the fact that the majormechanismsfor diffusingand separatingthe managementand control of decisions are much the same across different organizations. Decision hierarchies. A common feature of the diffuse decision man- agement and control systems of complex organizations(for example, large nonprofituniversitiesas well as large open corporations)is a formal decision hierarchywith higher level agents ratifyingand monitoringthe decision initiatives of lower level agents and evaluating their perfor- mance.12 Such hierarchicalpartitioningof the decision process makes it more difficultfor decision agents at all levels of the organizationto take actions that benefitthemselves at the expense of residualclaimants.Deci- sion hierarchiesare buttressed by organizationalrules of the game, for example, accountingand budgetingsystems, that monitorand constrain the decision behavior of agents and specify the performancecriteriathat determinerewards.13 Mutual monitoring systems. The formal hierarchiesof complex or- ganizations are also buttressed by informationfrom less formal mutual monitoringamong agents. When agents interactto produceoutputs, they acquire low-cost informationabout colleagues, informationnot directly availableto higherlevel agents. Mutualmonitoringsystems tap this infor- mationfor use in the control process. Mutualmonitoringsystems derive their energy from the interests of agents to use the internalagent markets of organizationsto enhance the value of humancapital.14 Agents choose 12 See Max Weber, The Theory of Social and Economic Organization(1947);Peter M. Blau, Bureaucracyin ModernSociety (1956);HerbertA. Simon,The Architectureof Com- plexity, 106 Proc. Am. PhilosophicalSoc'y 467 (1962);and the titles by Williamson,supra note 7. The historicaldevelopmentof hierarchiesin open corporationsis analyzedin Alfred D. Chandler,The Visible Hand (1977);and Alfred D. Chandler& HermanDaems, Man- agerialHierarchies(1980). 13 The separationof decision managementfrom decision control that we emphasizeis reflectedin the auditingprofession's concern with allocatingoperatingand accountingre- sponsibilityto differentagents. For instance,it is recommendedthatan agentwith responsi- bilityfor billingshouldnot have a role in receivingor recordingcustomerpayments.See, for example, Charles Horngren,Cost Accounting:A ManagerialEmphasisch. 27 (1982);or HowardP. Stettler, AuditingPrincipleschs. 4 & 8 (1977). 14 See Eugene F. Fama, Agency Problemsand the Theoryof the Firm, 88 J. Pol. Econ. 288 (1980). This content downloaded from 141.218.001.105 on August 04, 2016 23:19:35 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).  OWNERSHIPAND CONTROL 311 among organizationson the basis of rewards offered and potential for developmentof human capital. Agents value the competitiveinteraction that takes place within an organization'sinternalagent marketbecause it enhances current marginalproducts and contributes to human capital development.Moreover,if agents perceive thatevaluationof theirperfor- mance is unbiased (that is, if they cannot systematically fool their evaluators) then they value the fine tuning of the reward system that results from mutualmonitoringinformation,because it lowers the uncer- tainty of payoffs from effort and skill. Since the incentive structuresand diffuse decision control systems that result from the interplayof formal hierarchies and less formal mutual monitoringsystems are also in the interests of residualclaimants, their survival value is evident. Boards of directors. The common apex of the decision control sys- tems of organizations,large and small, in which decision agents do not bear a majorshareof the wealth effects of theirdecisions is some form of board of directors. Such boards always have the power to hire, fire, and compensate the top-level decision managersand to ratify and monitor importantdecisions. Exercise of these top-level decision controlrightsby a group (the board) helps to ensure separationof decision management and control (that is, the absence of an entrepreneurialdecision maker) even at the top of the organization.15 IV. THE SPECTRUM OF ORGANIZATIONS A. Introduction Organizationsin which importantdecision agents do not bear a major share of the wealth effects of their decisions include open corporations, largeprofessionalpartnerships,financialmutuals,and nonprofits.We are concerned now with analyzingthe data each of these organizationspro- vides to test the hypothesis that separationof decision managementfunc- tions fromresidualriskbearingleads to decision systems that separatethe managementand control of decisions. To motivate the discussion of specific organizationalforms, we also outlinea set of more specializedpropositionsto explainthe survivalvalue of the special features of their residual claims. These more specialized hypotheses about the survival of specific organizationalforms in specific 15 Decision functions can be delegated in two general ways: (1) joint delegation to several agents (as in a committee), or (2) partitioning and delegation of the parts to different agents. Boards of directors are examples of the former approach; decision hierarchies are examples of the latter. This content downloaded from 141.218.001.105 on August 04, 2016 23:19:35 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).  312 THE JOURNAL OF LAW AND ECONOMICS activities are developed in our paper "Agency Problems and Residual Claims."'16 B. Open Corporations 1. Unrestricted Common Stock Residual Claims. Most large nonfinancialorganizationsare open corporations.The common stock re- sidual claims of such organizations are unrestrictedin the sense that stockholdersare not requiredto have any other role in the organization, and their residual claims are freely alienable. As a result of the unre- strictednatureof the residualclaims of open corporations,there is almost complete specializationof decision managementand residualrisk bear- ing. Even managers who own substantialblocs of stock, and thus are residualrisk bearers, may elect to sell these shares. Unrestrictedcommon stock is attractivein complicatedrisky activities where substantial wealth provided by residual claimants is needed to bond the large aggregatepayoffs promised to many other agents. Unre- stricted common stock, with its capacity for generatinglarge amountsof wealth from residualclaimantson a permanentbasis, is also attractivein activities more efficiently carriedout with large amountsof risky assets owned within the organizationratherthan rented. Moreover, since deci- sion skills are not a necessary consequence of wealth or willingness to bear risk, the specialization of decision managementand residual risk bearingallowed by unrestrictedcommon stock enhances the adaptability of a complex organizationto changes in the economic environment.The unrestrictedrisk sharing and diversificationallowed by common stock also contributesto survivalby loweringthe cost of risk-bearingservices. 2. Control of the Agency Problems of Common Stock. Separation and specializationof decision managementand residualrisk bearingleads to agency problemsbetween decision agents and residualclaimants.This is the problemof separationof ownershipand control that has long trou- bled students of corporations.For example, potentialexploitationof re- sidual claimantsby opportunisticdecision agents is reflectedin the argu- ments leading to the establishment of the Securities and Exchange Commission and in the concerns of the modern corporate governance movement. Less well appreciated, however, is the fact that the unre- stricted natureof common stock residualclaims also allows special mar- ket and organizationalmechanismsfor controllingthe agency problemsof specialized risk bearing. 16 Eugene F. Fama & Michael C. Jensen, Agency Problems and Residual Claims, in this issue. This content downloaded from 141.218.001.105 on August 04, 2016 23:19:35 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).  OWNERSHIP AND CONTROL 313 The stock market. The unrestrictedalienabilityof the residualclaims of open corporationsgives rise to an externalmonitoringdevice uniqueto these organizations-a stock marketthat specializes in pricingcommon stocks and transferringthem at low cost. Stock prices are visible signals that summarizethe implicationsof internaldecisions for currentand fu- ture net cash flows. This external monitoringexerts pressure to orient a corporation'sdecision process towardthe interestsof residualclaimants. The marketfor takeovers. Externalmonitoringfrom a takeover mar- ket is also unique to the open corporationand is attributableto the unre- stricted nature of its residual claims.17 Because the residual claims are freely alienableand separablefrom roles in the decision process, attack- ing managerscan circumventexisting managersand the currentboardto gain control of the decision process, either by a direct offer to purchase stock (a tenderoffer)or by an appealfor stockholdervotes for directors(a proxy fight). Expert boards. Internalcontrol in the open corporationis delegated by residualclaimantsto a boardof directors.Residualclaimantsgenerally retain approval rights (by vote) on such matters as board membership, auditor choice, mergers, and new stock issues. Other managementand control functions are delegated by the residual claimants to the board. The boardthen delegates most decision managementfunctionsand many decision control functions to internalagents, but it retains ultimatecon- trol over internalagents-including the rightsto ratifyand monitormajor policy initiatives and to hire, fire, and set the compensationof top level decision managers. Similardelegation of decision managementand con- trol functions, at the first step to a board and then from the board to internal decision agents, is common to other organizations, such as financial mutuals, nonprofits, and large professional partnerships, in which importantdecision agents do not bear a majorshare of the wealth effects of their decisions. However, the existence of the stock marketand the marketfor take- overs, both special to open corporations, explains some of the special features of corporateboards, in particular:(1) why inside managerboard members are generally more influentialthan outside members, and (2) why outside board members are often decision agents in other complex organizations.18 Since the takeover marketprovides an external court of last resort for 17 Monitoring from the takeover market is emphasized in Henry Manne, Mergers and the Market for Corporate Control, 73 J. Pol. Econ. 110 (1965). 18 See Edward S. Herman, Corporate Control, Corporate Power ch. 2 (1981), for data on the characteristics of corporate boards. This content downloaded from 141.218.001.105 on August 04, 2016 23:19:35 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). 314 THE JOURNAL OF LAW AND ECONOMICS protectionof residualclaimants,a corporateboardcan be in the handsof agents who are decision experts. Given that the boardis to be composed of experts, it is natural that its most influentialmembers are internal managerssince they have valuablespecific informationabout the organi- zation's activities. It is also naturalthat when the internaldecision control system works well, the outside membersof the board are nominatedby internalmanagers.Internalmanagerscan use their knowledgeof the or- ganizationto nominateoutside boardmemberswith relevantcomplemen- tary knowledge:for example, outsiderswith expertise in capitalmarkets, corporatelaw, or relevant technology who provide an importantsupport function to the top managersin dealing with specialized decision prob- lems. However, the board is not an effective device for decision control unless it limits the decision discretion of individualtop managers. The boardis the top-level court of appeals of the internalagent market,19 and as such it must be able to use informationfrom the internalmutualmoni- toring system. To accomplish this and to achieve effective separationof top-level decision managementand control, we expect the board of a large open corporationto include several of the organization'stop man- agers. The board uses informationfrom each of the top managersabout his decision initiatives and the decision initiatives and performanceof other managers.The boardalso seeks informationfrom lower level man- agers about the decision initiatives and performanceof top managers.20 This informationis used to set the rewardsof the top managers,to rank them, and to choose among their decision initiatives.To protect informa- tion flows to the board, we expect that top managers, especially those who are membersof the board, can effectively be firedonly with consent of the board and thus are protected from reprisalsfrom other top man- agers. The decision processes of some open corporationsseem to be domi- nated by an individualmanager,generallythe chief executive officer. In some cases, this signals the absence of separationof decision manage- ment and decision control, and, in our theory, the organizationsuffers in the competition for survival. We expect, however, that the apparent 19 See Fama, supra note 14. 20 For example, Horngren, supra note 13, at 911, describes the role of the audit committee of the board (generally composed of outside board members) as a collector and conduit of information from the internal mutual monitoring system: "The objective of the audit com- mittee is to oversee the accounting controls, financial statements, and financial affairs of the corporation. The committee represents the full board and provides personal contact and communication among the board, the external auditors, the internal auditors, the financial executives, and the operating executives." This content downloaded from 141.218.001.105 on August 04, 2016 23:19:35 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).  OWNERSHIP AND CONTROL 315 dominanceof some top managersis more often due to theirabilityto work with the decision control systems of their organizationsthan to their ability to suppressdiffuse and separatedecision control. In any case, the financialpress regularlyreportsinstances where apparentlydominantex- ecutives are removed by their boards. Corporateboardsgenerallyinclude outside members,that is, members who are not internalmanagers,and they often hold a majorityof seats.21 The outside boardmembersact as arbitersin disagreementsamonginter- nal managersand carry out tasks that involve serious agency problems between internal managersand residual claimants, for example, setting executive compensationor searchingfor replacementsfor top managers. Effective separation of top-level decision managementand control means that outside directorshave incentives to carry out their tasks and do not collude with managersto expropriateresidualclaimants. Ourhy- pothesis is that outside directorshave incentives to develop reputations as experts in decision control. Most outside directors of open corpora- tions are either managers of other corporationsor importantdecision agents in other complex organizations.22The value of theirhumancapital depends primarilyon their performanceas internaldecision managersin other organizations.They use their directorshipsto signalto internaland externalmarketsfor decision agents that (1) they are decision experts, (2) they understandthe importanceof diffuse and separatedecision control, and (3) they can work with such decision control systems. The signalsare credible when the direct payments to outside directors are small, but there is substantialdevaluationof human capital when internaldecision control breaks down and the costly last resort process of an outside takeover is activated. C. Professional Partnerships 1. Mutual Monitoring, Specific Knowledge, and Restricted Residual Claims. The residual claims of professional partnershipsin activities such as law, public accounting, medicine, and business consulting are restricted to the major professional agents who produce the organiza- tion's services. This restrictionincreasesthe incentives of agents to moni- tor each other's actions and to consult with each other to improve the quality of services provided to customers. Such mutual monitoringand consulting are attractive to the professional agents in service activities where responsibilityfor variationin the quality of services is easily as- 21 See Herman, supra note 18, at ch. 2. 22 Id. This content downloaded from 141.218.001.105 on August 04, 2016 23:19:35 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).  316 THE JOURNAL OF LAW AND ECONOMICS signed and the value of professionalhumancapital is sensitive to perfor- mance. The monitoringand consulting are likely to be effective when professionalagents with similarspecialized skills agree to share liability for the actions of colleagues. In both large and small partnerships,individualsor small teams work on cases, audits, and so forth. Because of the importanceof specific knowledge about particularclients and circumstances, it is efficient for the teams to make most decisions locally. At this level, however, decision managementand decision control are not separate.To control the result- ing agency problems, the residual claims in professional partnerships, large and small, are restricted to the professionalagents who have the majordecision-makingroles. This is consistent with our hypothesis that combinationof decision managementand control functions leads to re- striction of residual claims to the agents who both manage and control importantdecisions. 2. Large Professional Partnerships. The partners in large profes- sional partnershipsare diffuse residualclaimantswhose welfare depends on the acts of agents they do not directly control. Thus, these organiza- tions provide a test of our hypothesis that separationof residual risk bearingand decision managementinduces decision systems that separate the managementand control of importantdecisions. The majordecision control devices of large professionalpartnershipsare similarto those of other organizationswith diffuse residual claims. For example, residual claimants in large partnershipsdelegate to boards the ratificationand monitoringof importantdecisions above the level of individualcases and audits. Moreover, the sharingof liability and residualcash flows among importantdecision agents (the partners)ensures that large partnerships have strong versions of the mutualmonitoringsystems that we contend are common to the decision control systems of complex organizations. The boardsof large partnershipshave special featuresthat relate to the restrictionof the residualclaims to importantinternalagents. The residual claimants are experts in the organization'sactivities, and they observe directly the effects of actions taken by the board of managingpartners. Thus, unlike the stockholders of open corporations,the residual claim- ants in largepartnershipshave little demandfor outside experts to protect their interests, and their boards are composed entirely of partners. The board is involved in decisions with respect to the managementof the partnership,for example, where new offices should be opened, who shouldbe admittedto the partnership,and who shouldbe dismissed. The boardis also involved in renegotiatingthe sharesof the partners.Here, as in other decisions, the boards of largepartnershipscombine the valuable specific knowledge availableat the top level with informationfrom part- This content downloaded from 141.218.001.105 on August 04, 2016 23:19:35 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).  OWNERSHIP AND CONTROL 317 ner-residual claimants. The role of the board is to develop acceptable consensus decisions from this information.Thus, the boardsof largepro- fessional partnershipsare generally called committees of managingpart- ners rather than boards of directors. The idea is that such committees exist to manageagency problemsamongpartnersand to study and deter- mine majorpolicy issues in a mannerthat is less costly than when per- formedjointly by all partners. Since the residual claims in a large professional partnershipare not alienable,unfriendlyoutside takeovers are not possible. Inside takeovers by dissidentpartnersare possible, however, because the managingboards of these organizationsare elected by the partner-residualclaimants. D. Financial Mutuals A commonform of organizationin financialactivities is the mutual.An unusual characteristicof mutuals is that the residual claimantsare cus- tomers, for example, the policyholders of mutual insurancecompanies, the depositors of mutual savings banks, and the shareholdersof mutual funds. Like the diffuse stockholders of large nonfinancialcorporations, most of the diffuse depositors, policyholders,and mutualfund sharehold- ers of financialmutualsdo not participatein the internaldecision process. Thus, financialmutuals provide another test of our hypothesis that sub- stantialseparationof decision managementand residualrisk bearingleads to decision systems that separate the managementand control of deci- sions. 1. The Control Function of Redeemable Claims. For the purpose of decision control, the uniquecharacteristicof the residualclaims of mutu- als is that they are redeemableon demand. The policyholder,depositor, or shareholdercan, on demand,turnin his claim at a price determinedby a prespecifiedrule. For example, the shareholderof an open-end mutual fund can redeem his claim for the marketvalue of his share of the fund's assets, while the whole life or endowmentinsurancepolicyholder,like the shareholder of a mutual savings bank, can redeem his claim for its specified value plus accumulateddividends. The decision of the claim holder to withdrawresources is a form of partial takeover or liquidation which deprives managementof control over assets. This control right can be exercised independentlyby each claim holder. It does not requirea proxy fight, a tenderoffer, or any other concerted takeover bid. In contrast, customer decisions in open non- financial corporations and the repricing of the corporation's securities in the capitalmarketprovide signalsabout the performanceof its decision agents. Withoutfurtheraction, however, either internalor from the mar- This content downloaded from 141.218.001.105 on August 04, 2016 23:19:35 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). 318 THE JOURNAL OF LAW AND ECONOMICS ket for takeovers, the judgments of customers and of the capital market leave the assets of the open nonfinancialcorporationunderthe controlof the managers. 2. The Board of Directors. Like other organizationscharacterized by substantialseparationbetween decision managementand residualrisk bearing, the top-level decision control device in financialmutuals is a boardof directors. Because of the strongform of diffuse decision control inherentin the redeemableresidualclaims of financialmutuals,however, their boards are less importantin the control process than the boards of open nonfinancialcorporations.The reduced role of the board is espe- cially evident in mutual savings banks and mutualfunds, which are not complex even thoughoften large in terms of assets. Moreover,the resid- ual claimantsof mutuals show little interest in their boards and often do not have the right to vote for board members.23Outsideboard members are generallychosen by internalmanagers.Unlike open corporations,the boardsof financialmutualsdo not often impose changesin managers.The role of the board, especially in the less complex mutuals,is largelylimited to monitoring agency problems against which redemption of residual claims offers little protection, for example, fraud or outrighttheft of as- sets by internalagents. E. Nonprofit Organizations When an organization'sactivities are financed in part throughdona- tions, part of net cash flows is from resources providedby donors. Con- tracts that define the share of residual claimants in net cash flows are unlikely to assure donors that their resources are protected from expro- priationby residualclaimants.In a nonprofitorganization,however, there are no agents with alienable rights in residual net cash flows and thus there are no residualclaims. We arguein "Agency ProblemsandResidual Claims" that the absence of such residualclaims in nonprofitsavoids the donor-residual claimantagency problemand explains the dominanceof nonprofitsin donor-financedactivities.24 The absence of residual claims in nonprofitsavoids agency problems between donors and residualclaimants,but the incentives of other inter- 23 See Edward S. Herman, Conflict of Interest in the Savings and Loan Industry, in A Study of the Savings and Loan Industry 789 (Irwin Friend ed. 1969), for documentation of such lack of interest. For example, he describes situations where in more than a decade only four depositors in total attended the annual meetings of two savings and loan associations and other situations where management did not even bother to collect proxies. 24 Fama & Jensen, Agency Problems and Residual Claims, in this issue. See Henry B. Hansmann, The Role of Nonprofit Enterprise, 89 Yale L. J. 835 (1980), for a general discus- sion of nonprofits. This content downloaded from 141.218.001.105 on August 04, 2016 23:19:35 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).  OWNERSHIP AND CONTROL 319 nal agents to expropriatedonations remain. These agency problems be- tween donors and decision agents in nonprofitsare similar to those in other organizationswhere importantdecision managers do not bear a majorshare of the wealth effects of their decisions. Our hypothesis pre- dicts that, like other organizationscharacterizedby separationof decision managementfrom residualrisk bearing,nonprofitshave decision systems that separatethe management(initiationand implementation)and control (ratificationand monitoring)of decisions. Such decision systems survive in donornonprofitsbecause of the assurancesthey providethat donations are used effectively and are not easily expropriated. 1. NonprofitBoards. In smallnonprofitsdelegationof decision man- agement to one or a few agents is generally efficient. For example, in nonprofitculturalperforminggroups, an artistic directorusually chooses performers,does the primarymonitoringof their outputs, and initiates and implements major decisions. Nevertheless, the importantdecision agents in these organizationsare chosen, monitored, and evaluated by boards of directors. Boards with similardecision control rights are com- mon to other small nonprofits characterizedby concentrated decision management,such as charities,privatemuseums, smallprivatehospitals, and local Protestantand Jewish congregations.Boards are also observed at the top of the decision control systems of complex nonprofits,such as private universities, in which both decision managementand decision control are diffuse. Although their functions are similar to those of other organizations, nonprofitboards have special features that are due to the absence of alienableresidualclaims. For example, because of the disciplinefrom the outside takeovermarket,boardsof open corporationscan includeinternal decision agents, and outside board memberscan be chosen for expertise rather than because they are importantresidual claimants. In contrast, because a nonprofitlacks alienable residual claims, the decision agents are immune from ouster (via takeover) by outside agents. Without the takeover threat or the discipline imposed by residualclaimantswith the right to remove members of the board, nonprofitboards composed of internalagents and outside experts chosen by internalagents would pro- vide little assurance against collusion and expropriationof donations. Thus, nonprofitboards generally include few if any internal agents as voting members,and nonprofitboardsare often self-perpetuating,that is, new members are approved by existing members. Moreover, nonprofit board membersare generally substantialdonors who serve without pay. Willingnessto provide continuingpersonaldonationsof wealth or time is generallyan implicitcondition for membershipon nonprofitboards. Ac- ceptance of this condition certifies to other donors that board members are motivated to take their decision control task seriously. This content downloaded from 141.218.001.105 on August 04, 2016 23:19:35 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). 320 THE JOURNAL OF LAW AND ECONOMICS 2. The Roman Catholic Church. To our knowledge the only non- profit organizationthat is financed with donations but lacks a board of importantcontinuingdonors with effective decision control rights is the RomanCatholicchurch. Parishcouncils exist in local Catholicchurches, but unlike their Protestantand Jewish counterparts,they are only advi- sory. The clerical hierarchycontrols the allocationof resources, and the papalsystem does not seem to limitthe discretionof the Pope, the organi- zation's most importantdecision agent. Otheraspects of the contracts of the Catholicclergy in part substitute for the control of expropriationof donations that would be providedby more effective donor-customerconstraints on decisions. For example, the vows of chastity and obedience incorporatedinto the contractsof the Catholic clergy help to bond against expropriation of donations by avoidingconflicts between the materialinterestsof a familyand the inter- ests of donor-customers.In addition, the trainingof a Catholic priest is organization-specific.For example, it involves a heavy concentrationon (Catholic)theology, whereas the trainingof Protestantministers places more emphasison social service skills. Once certified,the Catholicpriest is placed by the hierarchy.He cannot offer his services on a competitive basis. In exchange for developingsuch organization-specifichumancapi- tal, the Catholic priest, unlike his Protestant and Jewish counterparts, gets a lifetime contract that promises a real standard of living. The organization-specificnature of the human capital of the Catholic clergy and the terms of the contractunderwhich it is employedact as a bond to donor-customersthat the interests of the Catholic clergy are closely bound to the survival of the organizationand thus to the interests of donor-customers. AlthoughProtestantismarose over doctrinalissues, the control struc- tures of Protestantsects-in particular,the evolutionof lay councils with power to ratifyand monitorresourceallocationdecisions-can be viewed as a response to breakdowns of the contract structure of Cathol- icism, that is, expropriationof Catholic donor-customersby the clergy. The evolution of Protestantismis therefore an example of competition among alternative contract structures to resolve an activity's major agency problem-in this case monitoringimportantagents to limit ex- propriationof donations. Thereis currentlypressureto allow Catholicprieststo marry,thatis, to drop the vow of chastity from their contracts. We predict that if this occurs, organizationalsurvivalwill requireother monitoringand bonding mechanisms, for example, control over allocation of resources by lay councils similar to those observed in Protestant and Jewish congrega- tions. This content downloaded from 141.218.001.105 on August 04, 2016 23:19:35 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).  OWNERSHIP AND CONTROL 321 3. The Private University and Decision Systems in Complex Non- profits. In complex nonprofitswe observe mechanismsfor diffuse deci- sion control similar to those of other complex organizations. For ex- ample, large private universities, like large open corporations, have complicated decision hierarchiesand active internalagent marketswith mutual monitoringsystems that generate informationabout the perfor- mance of agents. Again, however, the decision control structuresof com- plex nonprofitshave special features attributableto the absence of alien- able residual claims. For example, a university's trustees are primarilydonors ratherthan experts in the details of educationor research.In ratifyingand monitoring decision initiatives presented by internal decision agents (presidents, chancellors, provosts, etc.), and in evaluating the agents themselves, boardsrely on informationfrom the internaldiffuse decision system-for example, reports from faculty senates and appointmentscommittees- and on external peer reviews. Moreover, the structureof internaldiffuse decision control systems is a more formal part of a university's contract structure(its charteror by- laws) than in largefor-profitorganizationssuch as open corporations.For example, unlike corporate managers, university deans, department heads, provosts, and presidents are generally appointedfor fixed terms. The end of a contractperiodactivates a process of evaluation,with search committees chosen according to formal rules and with rules for passing their recommendationson to the board. A more formal structureof dif- fuse decision managementand control is helpful to trustees who do not have specializedknowledgeabout a university'sactivities. It also helps to assure donors that the absence of discipline from an outside takeover marketis compensated by a strong system for internaldecision control. V. SUMMARY The theory developed in this paper views an organizationas a nexus of contracts (written and unwritten). The theory focuses on the contracts that (1) allocate the steps in an organization'sdecision process, (2) define residualclaims, and (3) set up devices for controllingagency problemsin the decision process. We focus on the factors that give survivalvalue to organizationalforms that separate what the literatureimprecisely calls ownershipand control. A. The Central Hypotheses An organization's decision process consists of decision management (initiationand implementation)and decision control(ratificationand mon- This content downloaded from 141.218.001.105 on August 04, 2016 23:19:35 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). 322 THE JOURNAL OF LAW AND ECONOMICS itoring).Ouranalysis producestwo complementaryhypotheses aboutthe relationsbetween decision systems and residualclaims: 1. Separationof residualrisk bearingfrom decision managementleads to decision systems that separate decision managementfrom decision control. 2. Combinationof decision managementand decision control in a few agents leads to residual claims that are largely restricted to these agents. B. Combination of Decision Management and Control When it is efficient to combine decision managementand controlfunc- tions in one or a few agents, it is efficient to control agency problems between residual claimants and decision makers by restrictingresidual claims to the decision makers. This propositiongets clear supportfrom the proprietorships,smallpartnerships,and closed corporationsobserved in small-scaleproductionand service activities. These organizationsare all characterizedby concentrated decision systems and residual claims that are restrictedto decision agents. C. Separation of Residual Risk Bearing from Decision Management 1. The Role of Specific Knowledge. In contrast, most of the organi- zations characterizedby separationof residualrisk bearingfrom decision managementare complex in the sense that specific informationvaluable for decisions is diffusedamongmany agents throughoutthe organization. Thus in a complex organizationseparationof residualrisk bearingfrom decision managementarises in partbecause efficientdecision systems are diffuse. Benefits from better decisions can be achieved by delegating decision functions to agents at all levels of the organizationwho have relevant specific knowledge, ratherthan allocatingall decision manage- ment and control to the residual claimants. Controlof the agency prob- lems of such diffuse decision systems is then achieved by separatingthe ratificationand monitoringof decisions (decision control)from initiation and implementation(decision management).The efficiency of such deci- sion systems is buttressedby incentive structuresthat rewardagentsboth for initiatingand implementingdecisions and for ratifyingand monitoring the decision managementof other agents. 2. The Role of Diffuse Residual Claims. In most complex organiza- tions, residual claims are diffused among many agents. When there are many residual claimants, it is costly for all of them to be involved in decision control. As a consequence there is separationof residual risk bearing from decision control, and this creates agency problems between This content downloaded from 141.218.001.105 on August 04, 2016 23:19:35 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).  OWNERSHIP AND CONTROL 323 residual claimants and decision agents. Separationof decision manage- ment and decision control at all levels of the organizationhelps to control these agency problems by limitingthe power of individualagents to ex- propriatethe interests of residual claimants. Thus diffusionand separa- tion of decision managementand control have survivalvalue in complex organizationsboth because they allow valuable specific knowledgeto be used at the points in the decision process where it is most relevant and because they help control the agency problemsof diffuse residualclaims. 3. Common Features of Decision Control Systems. What we call separationof residualrisk bearingfrom decision managementis the sep- arationof ownershipand control that has long botheredstudentsof open corporations.We arguethat separationof decision and risk bearingfunc- tions is also common to other organizationslike large professionalpart- nerships, financial mutuals, and nonprofits. Moreover, our central hy- pothesis about control of the agency problems caused by separationof residualrisk bearingfromdecision managementgets supportfromthe fact that the majormechanismsfor separatingdecision managementand deci- sion control are much the same across organizations. The common central building blocks of the diffuse decision control systems of complex organizationsof all types are formaldecision hierar- chies in which the decision initiativesof lower level agents are passed on to higherlevel agents, first for ratificationand then for monitoring.Such decision hierarchiesare found in large open corporations,large profes- sional partnerships,large financialmutuals,and large nonprofits.Formal decision hierarchiesare buttressedby less formalmutualmonitoringsys- tems that are a by-product of interaction that takes place to produce outputs and develop humancapital. The common apex of the decision control systems of organizations, largeand small, in which decision agents do not bear a majorshare of the wealtheffects of theirdecisions is a boardof directors(trustees,managing partners,etc.) that ratifiesand monitorsimportantdecisions and chooses, dismisses, and rewardsimportantdecision agents. Such multiple-member boards make collusion between top-level decision managementand con- trol agents more difficult,and they are the mechanismthat allows separa- tion of the managementand control of the organization'smost important decisions. APPENDIX BIBLIOGRAPHY Alchian,ArmenA. "Uncertainty, EvolutionandEconomicTheory."Journalof Political Economy 58, no. 3 (June 1950):211-21. This content downloaded from 141.218.001.105 on August 04, 2016 23:19:35 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). 324 THE JOURNAL OF LAW AND ECONOMICS Alchian, Armen A., and Demsetz, Harold. "Production, Information Costs, and Economic Organization." American Economic Review 62, no. 5 (December 1972): 777-95. Arrow, Kenneth J. "The Role of Securities in the Optimal Allocation of Risk Bearing." Review of Economic Studies 31, no. 86 (April 1964): 91-96. Berle, Adolf A., and Means, Gardiner C. The Modern Corporation and Private Property. New York: Macmillan Publishing Co., 1932. Blau, Peter M. Bureaucracy in Modern Society. New York: Random House, 1956. Chandler, Alfred D., Jr. The Visible Hand. Cambridge, Mass.: Harvard Univer- sity Press, 1977. Chandler, Alfred D., Jr., and Daems, Herman. Managerial Hierarchies. Cam- bridge, Mass.: Harvard University Press, 1980. Fama, Eugene F. Foundations of Finance. New York: Basic Books, 1976. Fama, Eugene F. "Agency Problems and the Theory of the Firm." Journal of Political Economy 88, no. 2 (April 1980): 288-307. Fama, Eugene F., and Jensen, Michael C. "Agency Problems and Residual Claims." Journal of Law and Economics 26 (June 1983): 327-49. Fama, Eugene F., and Jensen, Michael C. "Organizational Forms and Investment Decisions." Managerial Economics Research Center Working Paper no. MERC 83-03. Rochester, N.Y.: University of Rochester, Graduate School of Management, 1983. Hansmann, Henry B. "The Role of Nonprofit Enterprise." Yale Law Journal 89, no. 5 (April 1980): 835-901. Hayek, Freidrich A. "The Use of Knowledge in Society." American Economic Review 35, no. 4 (September 1945): 519-30. Herman, Edward S. "Conflict of Interest in the Savings and Loan Industry." In A Study of the Savings and Loan Industry, edited by Irwin Friend. Washington, D.C.: Federal Home Loan Board, 1969. Herman, Edward S., Corporate Control, Corporate Power. Twentieth Century Fund Study. New York: Cambridge University Press, 1981. Horngren, Charles. Cost Accounting: A Managerial Emphasis. Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1982. Jensen, Michael C. "Organization Theory and Methodology." Accounting Re- view 50, no. 2 (April 1983): 319-39. Jensen, Michael C., and Meckling, William H. "Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure." Journal of Financial Eco- nomics 3, no. 4 (October 1976): 305-60. Jensen, Michael C., and Meckling, William H. "Rights and Production Functions: An Application to Labor-Managed Firms and Codetermination." Journal of Business 52, no. 4 (October 1979): 469-506. Manne, Henry. "Mergers and the Market for Corporate Control." Journal of Political Economy 73, no. 2 (April 1965): 110-20. Simon, Herbert A. "The Architecture of Complexity." Proceedings of the Ameri- can Philosophical Society 106 (December 1962): 467-82. Smith, Adam. The Wealth of Nations. 1776. Edited by Edwin Cannan, 1904. Reprint. New York: Modern Library, 1937. Sowell, Thomas. Knowledge and Decisions. New York: Basic Books, 1980. Stettler, Howard P. Auditing Principles. Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1977. This content downloaded from 141.218.001.105 on August 04, 2016 23:19:35 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).  OWNERSHIP AND CONTROL 325 Weber, Max. The Theory of Social and Economic Organization, edited by T. Parsons. Glencoe, Ill.: Free Press, 1947. Williamson, Oliver E. Markets and Hierarchies: Analysis and Antitrust Implica- tions. New York: Free Press, 1975. Williamson, Oliver E. "The Modern Corporation: Origins, Evolution, Attri- butes." Journal of Economic Literature 19, no. 4 (December 1981): 1537-68. This content downloaded from 141.218.001.105 on August 04, 2016 23:19:35 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).