® Academy of Management Review1999, Vol. 24, No. 1. 49-63.

CONTRACTUAL COMMITMENTS, BARGAININGPOWER, AND GOVERNANCE INSEPARABILITY:

INCORPORATING HISTORY INTOTRANSACTION COST THEORY

NICHOLAS S. ARGYRESJULIA PORTER LIEBESKIND

University of Southern California

We extend transaction cost economics by arguing that prior contractual commitmentsmade by a firm can limit its ability to differentiate or change its governance arrange-ments in the future—a condition we term governance inseparability. Changes inbargaining power between a firm and its exchange partners also can result ingovernance inseparability. Consequently, governance choice may be more particu-laristic than the current version of transaction cost economics allows. We provideseveral testable propositions.

Transaction cost economics (TCE) is becominga predominant economic theory of the firm. Thistheory holds that transactions with particularcharacteristics are governed efficiently by cer-tain types of organizational arrangements andnot by others. In particular, firms are formed togovern transactions that are characterized byuncertainty combined with economic spilloversresulting from asset indivisibilities, asset spec-ificity, and asset extensibility (Alchian & Dem-setz, 1972; Coase, 1937; Grossman & Hart, 1986;Klein, 1983; Klein, Crawford, & Alchian, 1978;Teece, 1980; Williamson, 1979). Thus, in the TCEtheory of the firm, an individual transaction isthe unit of analysis for predicting organizationalform (Williamson, 1985).

In this article, however, we argue that focus-ing on the characteristics of isolated transac-tions can be insufficient to explain the scope ofthe firm. The reason for this is that the gover-nance of any new transaction in which a firmmay seek to engage may become linked insep-arably with the governance of other transac-tions in which the firm is already engaged. Inthese cases, what TCE might predict to be effi-cient modes of governance for the new transac-tion may, in fact, be excessively costly, or eveninfeasible altogether, for that particular firm.Thus, firms are subject to a condition that we

We thank three anonymous reviewers for many usefulcomments and suggestions. We contributed equally to thewriting of this article.

call governance inseparability—a condition inwhich a firm's past governance choices signifi-cantly influence the range and types of gover-nance mechanisms that it can adopt in futureperiods.

Governance inseparability can constrain afirm's governance options in two ways. First, itmay constrain a firm from switching from onegovernance mode to another for the same type oftransaction. Second, governance inseparabilitymay obligate a firm to use an existing gover-nance arrangement for a new transaction, evenif that particular transaction would be governedmore efficiently by other means. This is a con-straint on governance diffeientiation.

Governance inseparability has important im-plications for the theory of the firm. It impliesthat there will be variance in observed gover-nance mechanisms among firms for a giventransaction, since different firms will tend toincur different levels of costs for any particulargovernance arrangement. Also, governance in-separability can help to explain the limits tofirm scope, because it may be too costly for agiven firm to internalize a given transaction,owing to its arrangements in place, even if in-ternalization were the optimal governance solu-tion for that transaction considered in isolation.Similarly, governance inseparability may makeit more difficult for a given firm to reduce itsscope.

We seek to extend TCE theory by drawingattention to two factors that serve to produce

49

50 Academy of Management Review January

governance inseparability. The first of these fac-tors is contractual commitments. Firms are mostefficiently engaged in long-term exchange rela-tionships (Williamson, 1979). However, thesetypically require long-term contractual commit-ments. Yet, contractual commitments serve toengender governance inseparability because(by our definition) they are costly, if not impos-sible, to reverse. As a result, a firm's outstandingset of contractual commitments may signifi-cantly restrict its governance options in the fu-ture. The second factor that serves to producegovernance inseparability is changes in the bai-gaining power of the other parties to a firm'scontractual commitments—such as employees,suppliers, or customers. Here, these parties canuse unanticipated increases in their bargainingpower to obligate a firm to adopt locally subop-timal governance mechanisms in the future.

Because firms, by definition, engage in long-term transactions, no firm can entirely avoidmaking contractual commitments. Hence, firmsmust always assume some risk of governanceinseparability. As a result, governance insepa-rability implies that most firms will become con-strained over time by their existing arrange-ments in place, which will limit both their scopeand their strategic flexibility.

Because it focuses on the individual transac-tion as' the basis of analysis, TCE does not fullyconsider the possibility that parties to transac-tions—in particular, firms—may be constrainedin their choice of governance mechanisms bypast governance choices.' Incorporating gover-nance inseparability into TCE does not requiremajor changes in the theory's basic behavioralassumptions (i.e., bounded rationality and op-portunism), nor does it imply that firms makegovernance choices that are inefficient in aglobal sense. Rather, governance inseparabilitycalls for tempering the use of the transaction asthe unit of analysis (without abandoning it en-tirely) in order to capture the effects of historical

' Nickerson and Silverman (1997) also consider the rami-fications of transactional interdependencies for the TCE the-ory of governance choice. They emphasize "hazard interde-pendencies," in which the investment made for onetransaction affects the investment made or governancestructure used in another transaction. In contrast, our notionof governance inseparability emphasizes situations wherethe governance of one transaction affects the governance ofanother transaction.

constraints. As we show below, governance in-separability also places somewhat more em-phasis on the bounded rationality assumption,at the expense of the "far-sighted contracting"assumption employed in TCE.

THE TCE THEORY OF THE FIRM

According to TCE, firms are a particular formof organization for administering exchanges, or"transactions," between one party and another(Coase, 1937). In this conception of the firm, thefirm itself is characterized as a "managerial hi-erarchy" and is contrasted with other forms oforganization, most notably markets, in whichtransactions take place without managerialoversight (Williamson, 1975, 1985). The basic in-sight is that firms exist because they can some-times reduce the costs of negotiating and enforc-ing terms and conditions of exchange relative tomarket transacting (Coase, 1937). This will bethe case especially when uncertainty about fu-ture business conditions makes contracts in-complete and when transactions are character-ized by economic spillovers, such as those thatresult from cospecific investments, asset indi-visibility, and asset extensibility (Alchian &Demsetz, 1972; Coase, 1937; Grossman & Hart,1986; Hart & Moore, 1990; Klein, 1983; Klein et al.,1978; Teece, 1980; Williamson, 1975, 1979, 1985).Hierarchies allow for better adaptation thancontracting between autonomous parties. Adap-tation is important because efficiency consider-ations often require that adjustments be madein the distribution of the gains from a tradingrelationship when trading conditions change.Such adjustments are likely to occur in a morecooperative and, hence, less costly way if thetransaction occurs within a hierarchy, ratherthan through autonomous contracting. This isbecause hierarchies are able to resolve tradingdisputes by fiat as a last resort, whereas fiat isunavailable for governing market contracts(Williamson, 1975, 1985). The use of fiat withinhierarchies is supported by contract law, since,courts generally forbear from hearing cases in-volving disputants who are part of the same firm(Williamson, 1991). In contrast, in market con-tracting a party may use the law and the legalsystem opportunistically to "hold up" a contrac-tual partner (Klein, 1993; Klein et al., 1978). In

1999 Aigyres and Liebeskind 51

addition, placing the ownership of the assets ina given transaction into the hands of a singleparty (i.e., organizing the transaction within afirm) improves the incentives for making effi-cient transaction-specific investments whencontracts are incomplete and the cost associ-ated with a hold-up is significant (Grossman &Hart, 1986; Hart & Moore, 1990).

This TCE theory of the firm has been used toexplain not only the existence of firms but thescope of the firm, in terms of both vertical inte-gration and diversification. That is, the degreeof vertical integration and/or diversification of afirm is determined by uncertainty and by assetspecificity, asset extensibility, and asset indi-visibility (Klein et al., 1978; Teece, 1980; William-son, 1975, 1985). Moreover, the firm itself is un-derstood to comprise not only assets that arefully internalized—in the form of wholly ownedproperty, physical plant and equipment, andemployees (Coase, 1937; Hart & Moore, 1990;Masten, 1988)—but also long-term contractswith external parties, such as buyers, suppliers,joint venture partners, and other such "hybrid"arrangements (Pisano, 1989; Williamson, 1991).Thus, the scope of the firm can be understood tobe the scope of both its internal contracts—withits managers, employees, and shareholders—and its external contracts with other parties thatsell to, buy from, or otherwise are parties to thefirm's non-market arrangements.

CONTRACTUAL COMMITMENTS ANDGOVERNANCE INSEPARABILITY

We define a contractual commitment as anagreement between two or more parties that isbinding on those parties, to the degree that torenege on the agreement will be costly. Thisimplies that an exchange agreement (i.e., a con-tract) cannot be described as a contractual com-mitment unless enforcement mechanisms are inplace that ensure both detection and punish-ment of reneging. With enforcement mecha-nisms, contractual commitments become, to asignificant extent, irreversible. Without enforce-ment mechanisms, a contract can be changedwithout cost by either party. In this section weseek to demonstrate, using a number of exam-ples, how contractual commitments can lead toa condition of governance inseparability.

The contractual commitments of a firm fallinto two categories: (1) formal contracts and(2) informal or nonlegally enforceable contracts.

1. Foimal contiactual commitments. A firmcan commit to a contract by preparing and sign-ing a legally enforceable document that sets outthe terms and conditions of future transactionsbetween the firm and one or more other parties,be they employees, suppliers, buyers, or lend-ers. If a firm were to breach such a contract, itwould be legally liable for damages or restitu-tion payments. Moreover, seeking favorable re-negotiating contracts can require side paymentschemes that are difficult to implement (Dem-setz, 1966). Examples of such legal contractswith stakeholders are union wage agreements,long-term supply contracts, exclusive dealer-ship and franchise agreements, debt covenants,and customer warranties.

2. Informal or nonlegally enforceable con-tracts. A firm can also enter into informal orunwritten agreements with various other par-ties. To the degree that these agreements can beenforced, they also constitute contractual com-mitments.

One common type of informal contractualcommitment is "self-bonding," wherein a firminvests in firm-specific (i.e., sunk) assets whosevalue can only be recouped if it behaves incertain ways (Ghemawat, 1991; Sutton, 1991). Forexample, a firm may invest in assets that arecospecific to an exchange partner, such as com-mitting irrecoverable capital to a joint venture,investing in an exclusive dealership, or financ-ing a supplier's firm-specific production facili-ties. If the firm ceases to trade with any of theseparties according to agreed-upon terms andconditions, the other party can cease tradingaltogether, and the firm will lose the value of itsinvestments. Williamson (1983) refers to ar-rangements such as these as "hostage-taking,"drawing an apt analogy between nonlegal en-forcement mechanisms in medieval politics andpresent-day commerce.

Firms are also bound by informal "socialcontracts," through which they are committed (atleast to some degree) to respecting and uphold-ing the norms of society. For instance, it may belegal for a firm to lay off many thousands ofworkers, but to the degree that society at largeconsiders such actions unethical, the firm canexpect to be sanctioned. Examples of such sane-

52 Academy of Management Review January

tions include consumer boycotts,^ campaigns toinvestigate corporate actions,^ and refusal of so-cial consent to other actions that firms may wantto take.*

In general, the contractual commitments thata firm enters into are particularistic; each con-tractual commitment is entered into with a par-ticular and identifiable outside party. Thus,whereas shareholders may come and go at theirown will, the parties to a firm's contracts are tiedto the firm, and the firm is tied to these otherparties, because these commitments are nottransferable. Moreover, contractual commit-ments, by their very nature, cannot be revokedat low cost, (jfovernance inseparability is engen-dered because a firm's contractual commitmentstie it to specific other parties who have rights inrelation to the firm. We consider two types ofgovernance inseparability that may stem fromcontractual commitments: (1) constraints on gov-ernance switching and (2) constraints on gover-nance differentiation.

cant barriers to further agreements, and theyalso limit moves toward forward integration bythe franchisor.

For example, in the past, Coca-Cola enteredinto a number of exclusive franchising agree-ments with independent bottling companies.These agreements prevent Coca-Cola from for-ward integrating, because they do not allow anycompany other than the franchisee to bottleCoca-Cola products within a given territory. Asa result, when Coca-Cola decided to forwardintegrate in the 1980s, it could do so only bybuying out its own franchisees in the open mar-ket for corporate control, where it had to pay ahefty price premium relative to the costs of sim-ply forward integrating through internal expan-sion. Moreover, in some cases Coca-Cola hasbeen unable to buy out its franchisees. In others,it has elected not to do so, since the costs ofacquisition are too high relative to the benefitsof intemalization.

Contractual Commitinents and Constraints onGovernance Switching

Governance inseparability in the form of aconstraint on governance switching exists if afirm cannot efficiently enter into a governancearrangement of Type Y in future periods for aparticular transaction because it already has agovernance arrangement of Type X in placewith another party for that transaction. Fran-chising agreements provide a very straightfor-ward example of this type of constraint on gov-ernance modes. Outstanding contractualcommitments in the form of franchise agree-ments and exclusive dealerships pose signifi-

^ For instance, during the 1980s, General Electric was boy-cotted for producing nuclear bomb components, and in 1996consumers boycotted The Gap for producing garments in"sweatshop" factories in Guatemala.

^ For example, Arco Chemicals recently has been the sub-ject of an investigation following the explosion of its plant inLouisiana; Exxon also is being monitored closely since theExxon Valdez disaster. In both instances society apparentlyhas judged that these firms failed to exercise due care forworkers in their business operations.

* Following its widely criticized layoffs of many thou-sands of workers, AT&T can rationally expect to encountermore social opposition than previously to user rate in-creases and to changes in telecommunications regulationsthat would favor its businesses.

Contractual Commitinents and Constraints onGovernance Differentiation

Governance inseparability in the form of aconstraint on governance differentiation existswhen a firm is obligated to enter into a gover-nance arrangement of Type X with one partybecause it already has a governance arrange-ment of Type X in place with another party. Thistype of inseparability commonly arises whenfirms want to differentiate their internal organ-izational arrangements.

For example, efforts to start and/or sustainnew venture units have failed at many largefirms (Burgelman & Sayles, 1986; Fast, 1978; Hla-vacek, 1974; Sykes, 1986). According to TCE, thisfailure results from the inability of firms to makecredible commitments to the "marketlike" gov-ernance arrangements often needed to supportnew ventures. Thus, Williamson (1985) arguesthat corporate management cannot sustainhigh-powered incentives for a venture internalto the firm, because venture managers knowthat corporate management can always abro-gate those incentives in the future and may doso if it becomes in its economic interest. As aresult, those ventures often fail or are never ini-tiated. Hence, in this theory the commitmentproblem arises in the relationship between cor-porate management and venture managers.

1999 Aigyres and Liebeskind 53

Our arguments suggest that a different mech-anism may be at work here. In particular, therelationship between corporate managementand managers of established divisions may alsobe important in affecting the kinds of incentivesthat can be offered and sustained for internalventures. Specifically, implicit contractual com-mitments to managers of established divisionsmay make specialized governance arrange-ments difficult to sustain, because these latterarrangements will be seen as violations of es-tablished commitments. For instance, high-powered incentives imply that a new venturedivision may be permitted to retain a higherproportion of its profits, insulating its manage-ment from profit fluctuations elsewhere in thefirm and depriving other divisional managers ofthe opportunity to share in the venture's success.These new incentive structures, therefore, un-dermine the incentive structures in place of es-tablished divisions, who will oppose themthrough any means available, thereby ensuringthat their outstanding incentive contracts arehonored. For example, in several case studiesFast (1978) documented instances of interdivi-sional conflict, where established divisions at-tempted to gain management control over newventures and (implicitly) to suspend their spe-cial treatment.

With these arguments we do not deny that theproblem of credible commitments to venture di-visions is problematic for internal venturing.Rather, our arguments emphasize that a firm'sprior commitments to its other existing divisionsplay an important role in preventing the firmfrom sustaining the specialized governance ar-rangements needed to develop new ventures.Thus, we can conjecture that problems of gover-nance inseparability may account, to a signifi-cant degree, for the failure of internal ventures.

Similar arguments may apply to firms' inabil-ity to differentiate the ways in which their divi-sional performance is measured, divisionalmanagers are rewarded, or transfer prices areset. For example, firms often maintain a singletransfer pricing rule (Eccles, 1985; Poppo, 1995),even when their internal transactions are differ-ent enough to suggest a need for multiple rules.In all of these instances, outstanding contrac-tual commitments make it difficult for firms togovern individual transactions exclusively ac-cording to their individual characteristics, asprescribed by TCE. Rather, the existing commit-

ments of a firm operate in such a way that newactivities can be administered only through alimited set of governance mechanisms that arein place already.

It is important to our argument that we make aclear distinction between the direcf effects ofcontractual commitments on the actions of afirm and their indirect effects in terms of gover-nance inseparability. For instance, if a firm en-ters into an exclusive dealing contract withParty A, it cannot then enter into another exclu-sive dealing contract with Party B. This is adirect effect (and frequently an obvious one) of acontractual commitment on the actions andscope of a firm. Yet, while this commitment maybe important in terms of limiting the scope of thefirm per se, it has no implications for the exist-ing TCE theory of the firm. However, if the firm'sexclusive contract with Party A engenders a sit-uation whereby the firm becomes obligated tomaintain this type of contract with any and allother parties in the face of alternative and pref-erable governance modes, then a situation hasarisen in which the firm's outstanding contrac-tual commitments are constraining its futuregovernance options. That is, contractual com-mitments have resulted in governance insepa-rability. In this case the transaction-level pre-dictions of TCE as to how transactions should begoverned may no longer hold, because a firmmay find it necessary to employ governancemodes for its transactions that are suboptimalaccording to the established arguments.

Institutional theory also implies that certaintypes of contractual arrangements can becomedifficult to change over time and, hence, difficultto differentiate on a transaction-by-transactionbasis. For instance, certain arrangements canbecome "taken-for-granted" (Zucker, 1987), mim-icked in a search for various types of "legitima-tion" (DiMaggio & Powell, 1983), or compelled bystate action (DiMaggio & Powell, 1991; Fligstein,1990). Habits, pressures for legitimation, andmimicry do not play a role in our arguments.From our point of view, firms are profit-seekingentities, and contractual commitments are en-forced by the threat of financial losses that canbe imposed on defectors by their (actual or po-tential) private contractual partners. Thus, inour theory lack of differentiation arises becauseof economic realities and not because of socialstructures or norms per se.

54 Academy of Management Review January

State action does play a role in our argumentbecause contract enforcement can occur throughactual or potential appeal to the courts. How-ever, in our view the state does not compel par-ties to adopt particular, narrowly defined kindsof contractual arrangements. Instead, it seeks toenforce the arrangements that parties haveagreed upon privately, as long as those arrange-ments conform to certain widely held socialnorms, especially those concerning fairness(e.g., Michaelman, 1967). Therefore, for us, thenormative environment in which contracting isundertaken remains in the background, ratherthan coming to the foreground in terms of di-rectly shaping organizational arrangements, asis the case in institutional theory.

Is Governance Inseparability Associated withContractual Commitments Avoidable?

Given that contractual commitments can en-gender governance inseparability, the questionarises as to why a firm may not be able toforesee this eventuality and to eschew any com-mitments associated with this particular hazard.This would allow a firm to keep its governanceoptions open in future periods, avoiding prob-lems of inseparability and economizing ontransaction costs.

There are at least two answers to this ques-tion. First, a firm cannot exist efficiently withoutcommitments. Firms, as institutions, are costlyboth to establish and to liquidate (Williamson,1979, 1985). Consequently, a firm that does notgovern long-term transactions of one type or an-other is an inefficient form, yet long-term trans-actions typically require contractual commit-ments of one type or another.

The second answer is centered more in strat-egy than in governance cost considerations:contractual commitments are necessary for afirm to earn rents. Broadly speaking, firms de-rive rents from ownership of unique capabilities(Rumelt, 1984, 1987); even arbitrage rents derivefrom information that is closely owned and fromchanges in circumstances that make some con-tracts more valuable than their original pur-chase price (Barney, 1986; Demsetz, 1966). Yet,unique capabilities cannot be purchased inmarkets; instead, they usually derive from spe-cialized and/or cospecific investments whosebenefits accrue over time (Dierickx & Cool, 1989;Kirzner, 1979; Lieberman & Montgomery, 1988;

Rumelt, 1987; Winter, 1987). Such investmentscannot be supported rationally without contrac-tual commitments. For instance, in a competi-tive labor market, a firm must commit to long-term employment or make other contractualcommitments to its employees, if those employ-ees are to make significant levels of firm-specific investments of human capital, such aslearning firm-specific routines. Hence, a firmthat eschews making commitments typicallywill be unable to earn rents and, therefore, willnot survive.

If governance inseparability is indeed oftenunavoidable, the natural question raised byTCE is why a firm cannot take steps to createsafeguards against inseparability when negoti-ating its early agreements. This possibility issuggested by the emphasis on far-sighted con-tracting that TCE features. As Williamson notes,TCE

concedes that comprehensive contracting is not afeasible option (by reason of bounded ration-ality), yet it maintains that many economicagents have the capacities both to learn and tolook ahead, perceive hazards, and factor theseback into the contractual relation, thereafter todevise responsive institutions (1996: 9).

He also notes that "such a concept of contractpresents healthy tensions . . . for organizationtheory" (1996: 9). •

The view we adopt is that, in many cases,these tensions may resolve themselves more infavor of bounded rationality than in favor offoresight. For example, in industries that featurerapid technological innovation, anticipating fu-ture hazards and opportunities is often very dif-ficult (Nelson & Winter, 1982). Many other indus-tries also experience rapid changes in tradeconditions and consumer behavior. Given thepossibility of unforeseeable change, enteringinto any commitment at one point in time willnecessarily involve the assumption of some riskof governance inseparability on the part of afirm and, hence, some risk of increased transac-tion costs. Consequently, recognizing and incor-porating governance inseparability may requiregre'ater emphasis on the effects of bounded ra-tionality on contracting, and on evolutionarychange, relative to TCE's current emphasis onfar-sighted contracting.

1999 Aigyies and Liebeskind 55

BARGAINING POWER AND GOVERNANCEINSEPARABILITY

A second set of reasons for observing gover-nance inseparability among a firm's transac-tions relates to bargaining power, which can bedefined as the ability of one party to a contractto be able to influence the terms and conditionsof that contract or subsequent contracts in itsown favor. In this section we seek to demon-strate that governance inseparability is engen-dered when the other parties to a firm's contrac-tual commitments have bargaining power, andparticularly so when their relative bargainingpower increases unexpectedly. Again, our argu-ments suggest that the TCE theory of isolatedgovernance choice needs to be modified.

Bargaining Power and Constraints onGovernance Switching

Recall that one effect of contractual commit-ments is to prevent a firm with a governancearrangement of Type X in place for a particulartransaction from switching to another gover-nance arrangement of Type Y in future periodsfor that same transaction. Changes in bargain-ing power also can produce this effect. Considertwo bargaining parties—Seller S and BuyerB—who enter into an efficiently structured long-term contract that serves to support investmentin specific assets by both parties so that there isno anticipated ex post asymmetry in their bar-gaining power. However, because there is someuncertainty about prices and/or demand in thefuture, the long-term contract allows some scopefor renegotiatiori. This arrangement is rationalbecause the parties have equal bargainingpower at the time the contract is signed. How-ever, as time passes, conditions change in sucha way that Buyer B increases its bargainingpower over Seller S. At the time of contract re-negotiation, B may use this increased bargain-ing power in such a way that, in the future, thegovernance options for S for some of its transac-tions are constrained. For instance, B may pre-vent S from forward integrating into its ownbusiness, forcing S to continue to sell acrossmarkets rather than to integrate vertically. Thus,the changes in bargaining power of B relative toS over time have imposed a degree of gover-nance inseparability on S that S did not antici-

pate, and therefore could not prevent, in its orig-inal contract with B.

One common example of this type of gover-nance inseparability is when unionized laboruses its bargaining power to restrict outsourc-ing. In recent years a number of U.S. companieshave attempted to outsource production or ser-vice activities in order to sustain or increasetheir competitive advantage, but they have beenopposed by their own workers. Examples in-clude Deere, Ameritech, McDonnell Douglas,Boeing, United Parcel Service, Chrysler, andGeneral Motors (GM).

For instance, a major strike led by the UnitedAuto Workers (UAW) at GM's brake manufactur-ing plant crippled production of autos andtrucks for over a month in 1996, costing the firm$600 million. The UAW launched the strike afterGM announced its intention to increase thenumber of outsourced parts in its automobiles inorder to reduce its labor costs and circumventrestrictive union work rules. Although GM pi'e-viously had agreed with the UAW not to in-crease outsourcing of certain components, thisagreement was stated in fairly general termsand its enforceability through the courts was insome doubt. Thus, this was not a clearcut case ofunion enforcement of an existing contractualcommitment but, rather, a case of GM's workersusing their bargaining power to change theterms and conditions of their basic contract,with the result that GM could not change theway its parts production processes were gov-erned. Indeed, this incident raises the questionof whether outsourcing at GM was impJicifiyconstrained in earlier periods, during whichunion bargaining power was even stronger thantoday (the late 1960s and 1970s, for example). Insuch a case as this, we cannot expect transac-tion cost logic to apply strictly to a firm's choicebetween internal and external contracts.

Franchising is another arena in which recentchanges in bargaining power have served toconstrain the types of contracts that can be of-fered by many franchisors. In the 1990s franchi-sees in a large number of chains have becomemore organized. According to Harris (1997), theAmerican Franchisee Association grew fromroughly 4,000 members to about 7,500 during theperiod 1993-1997, and the American Associationof Franchisees and Dealers (AAFD) grew from 20members in 1992 to about 6,000 in 1997. Chain-specific franchisee groups also have become

56 Academy of Management Review January

more widespread, doubling to 250 during 1994-1997. One of these groups, the Meineke MufflersDealers Association, recently won a major law-suit against the franchisor—a victory that wasapparently made possible only by the newlywon financial clout of the AAFD (Harris, 1997).This legal decision could affect the ways inwhich rights to control communal advertisingfunds can be allocated in all of Meineke's futurecontracts, and perhaps in future contracts of-fered by other franchisors as well. Thus, in-creased franchisee bargaining power is allow-ing them to impose new constraints on futurecontracts a franchisor might wish to offer.

Even when this increased power is partly an-ticipated and when contractual safeguards areincluded in franchise agreements, subsequentlegal decisions favoring franchisees sometimeshave invalidated the safeguards. In one recentcase Naugles, Inc., a fast-food franchisor, specif-ically did not provide territorial exclusivity forits franchisees, allowing the firm to sell newfranchises close to older ones as populationdensities increased (Harris, 1997). Naugles' fran-chisees originally agreed to this contract provi-sion. However, when Naugles attempted to opena new franchise close to an existing one, itsfranchisees collectively sued; they won theircase in a California federal court. In this caseNaugles was unable to switch from a de factosole franchisee in a given area to multiple fran-chisees, despite its contractual provisions, be-cause of increases in franchisees' bargainingpower. Note, however, that it was Naugles' con-tractual commitments to its franchisees that ex-posed it to the deleterious effects of changes intheir bargaining power.

Bargaining Power and Constraints onGovernance Differentiation

In other instances a firm may be constrainedin its ability to differentiate its organizationalarrangements, owing to unanticipated changesin bargaining power—a case in point being thetrucking industry. During the 1980s, the threemajor U.S. long-distance trucking firms (YellowCorp., Consolidated, and Roadway) attemptedto enter the short-haul trucking market. How-ever, none of these companies has, as yet, beenable to differentiate its organizational arrange-ments to suit this new market. Yellow Corp., forexample, tried to establish a new subsidiary to

conduct its short-haul business and to negotiatea new, more flexible union contract for this firm.According to an industry analyst, "The Team-sters basically said, 'No way'" (Kansas City Busi-ness Journal, 1993: 1).

A parallel set of circumstances has arisen inthe U.S. airline industry. Some large airlinesmaintain separate subsidiary firms to handleshorter routes. The pilots working for these sub-sidiaries earn much lower salaries, allowing (forexample) American Airlines' American Eaglesubsidiary to compete with smaller, nonunion-ized airlines, such as Southwest. Recently,American sought to increase the proportion offlights handled by American Eagle by addingshort-haul jets to its previously turboprop fleets.The pilots of American Airlines, arguing thatthey alone should be allowed to fly the short-haul jets, threatened a strike, rather than allowinternal differentiation of jet pilot's pay.^ Otherairlines have avoided this problem by conduct-ing their short-haul businesses through par-tially owned subsidiaries (e.g.. Delta), throughlong-term contracts or strategic alliances withshort-haul carriers (e.g.. Continental), or throughemployee ownership (e.g.. United), which maybetter align the incentives of the different par-ties.

In each of these examples, union bargainingpower prevented a firm from carrying out achange in organization that it was seeking.Moreover, media reports suggest that in all ofthese instances, management was surprised atthe union's strength. It is likely, however, thatchanges in this strength were difficult to fore-see. It is important to note that, in each of thesecases, parties were able to obtain bargainingpower because initial contractual commitmentswere in place. These commitments made itcostly, if not impossible, for a firm to seek alter-natives to dealing with the parties who had thebargaining power. Hence, contractual commit-ments hot only expose a firm directly to a risk ofincreased governance inseparability, but ex-pose it to the indirect risk of changes in bargain-ing power over time. Consequently, the two key

^ Minutes after the strike began, President Clinton inter-vened to avert a "national transportation crisis." The partieswent into binding arbitration, where the issue of subsidiaryairliner substitution is being negotiated as of this writing.See the New York Times (1997) and the Wall Sheet ]ouinal(1997).

1999 Aigyies and Liebeskind 57

variables that determine governance insepara-bility in our arguments—contractual commit-ments and bargaining power—can interact intheir effects on governance choices. First, thepresence of contractual commitments increasesthe effects of relative bargaining power on gov-ernance choices, especially for firms with lowbargaining power. Second, and conversely, thegreater a firm's relative bargaining power, thegreater its ability to negotiate in a waythat limits contractual restrictions on its futureactivity.

Why Firms Are Unable to Anticipate andSafeguard Against Changes in BargainingPower

Given the TCE view that hazards can, in gen-eral, be foreseen and therefore protectedagainst, the question arises as to why contract-ing parties cannot foresee the buildup of bar-gaining power by one party or the other andconstruct their contractual commitments accord-ingly. Indeed, recognizing the potential forchanges in bargaining power and adjustinggovernance mechanisms to offset them is TCE'sprimary concern. Changes in bargaining power,however, are often difficult to foresee, becausethere is a large number of interrelated factorsthat affect the relative power of contracting par-ties. For example, union power may wax andwane with the state of the general economy,with government labor policies, with unionleadership and organization, and with immigra-tion rates. Thus, during the recession of the1970s and early 1980s, union power declined inthe United States, and union membershipshrank. As the U.S. labor market has tightenedin recent years, unions have become,arguably, less accommodating to the interests ofemployers.

Another difficulty in anticipating changes inbargaining power is that these changes mayoccur very gradually over long periods of timeso that their future importance is difficult to per-ceive. For instance, according to North and Tho-mas's (1973) seminal economic history of Europe,small changes in land/labor ratios in Europeaccumulated over time into a large increase inthe bargaining power of peasants over land-lords, giving rise to the wage labor system. Moregenerally. North (1990) emphasizes that such his-toric changes in relative prices have produced

changes in bargaining power that lead to diffi-cult and costly recontracting—or even full-scaleinstitutional change.

An additional problem with foreseeingchanges in bargaining power is that the en-forceability of many kinds of contracts throughthe courts can be highly uncertain. Contract dis-putes, therefore, can result in sudden changes inrelative bargaining power, as was the case forNaugles in its franchise contracts. Moreover, le-gal rulings may prevent a firm from avoidingeven the most overt attempts by other parties togain bargaining power. For instance, ToyotaMotor Corporation recently sued to prevent Re-public Industries from purchasing any more ofits dealerships, after Republic announced its in-tention to purchase a significant share of Toyo-ta's outlets (Wall Street Joumal, 1997b). In 1996Toyota added terms and conditions to its deal-ership contracts, seeking to limit multiple pur-chases. However, the legality of these addi-tional terms and conditions has not yet beentested in court. It is possible that Toyota will beunable to legally prevent the accumulation ofbargaining power by Republic.

Even if some of these contingencies could beforeseen, however, the range of feasible gover-nance mechanisms is unavoidably restricted intransactions involving employees and consum-ers. In particular, integrating parties' intereststhrough hierarchical governance—the mecha-nism of last resort when parties are highly in-terdependent—does not apply to relations withemployees and consumers. Because human as-sets cannot be "owned" in the same way thatphysical assets can be owned, employees must,perforce, be dealt with through contractual ar-rangements, which will always be somewhatincomplete (Grossman & Hart, 1986; Klein, 1983).

Similarly, a firm cannot vertically integrateall of its final consumers. Hence, contracts withemployees and consumers suffer from a condi-tion of incompiefe intemalization, so a firm'sability to impose hierarchical governance ontransactions with these parties will always beattenuated to some degree, increasing the like-lihood that it will be held up in future periods. Insum, it is difficult for firms to completely avoidhazards owing to changes in bargaining power,just as it is difficult for them to avoid contractualcommitments.

58 Academy of Management Review January

Other Views of Power

In the current version of TCE, firms are gener-ally assumed to operate in a highly competitiveenvironment, in which sources of inefficiency,such as. differences in bargaining power, arecompeted away relatively rapidly. As a result,there are relatively few instances in which thefirm's selection of governance arrangements isconstrained by bargaining power. Moreover, inTCE theory the assumption is that changes inbargaining power can be foreseen at the initia-tion of a contract. For instance, when transac-tions involve a specific investment, the partymaking the specific investment might lose bar-gaining power after the investment is made.However, this party is assumed to be able toforesee the buildup of power of its transactingpartner and to be able to act to effectively pre-vent it by negotiating suitable safeguards in theoriginal contract, such as exclusive dealing pro-visions or holding hostage equity positions(Pisano, 1989; Williamson, 1983). According toTCE, if such safeguards cannot be written be-cause critical contingencies cannot be foreseen,hierarchical organization will replace long-termcontracting in order to forestall the emergenceof asymmetrical bargaining power (Klein et al.,1978; Williamson, 1975, 1985).

While proponents of TCE generally eschewthe concept of power, Williamson (1995) doesadmit that power can sometimes play a role inagreements between firms and workers and be-tween firms and final consumers, especiallywhen consumers and workers face collective ac-tion problems. Our arguments suggest, how-ever, that rather than constituting special andunusual cases, labor and consumer transactionsexert an important and pervasive influence onthe boundaries of the firm and on the contrac-tual nature of those boundaries. Furthermore, afirm's implicit and explicit contracts with man-agers also can constrain the firm's choice oforganizational boundaries so that contractswith employees in general can be constraining.If so, this implies, at the very least, that evalu-ations and tests of TCE hypotheses about firmscope should be made contingent on the level oforganization of workers and consumers in-volved in a given transaction and the nature ofprior commitments to managers. We take theview that although competitive processes maypossibly extinguish bargaining power in the

long run, bargaining power differences can ex-ert important impacts on governance choice inthe short to medium term—impacts that re-searchers and managers cannot afford to ignore.

One theory of governance choice that explic-itly features power is resource dependence the-ory (Pfeffer & Salancik, 1978). In this theory or-ganizations choose governance forms, such aslong-term contracts or mergers, in order to avoiddependence upon other organizations for criti-cal resources, or to acquire the critical resourcesnecessary to create dependence on the part ofother organizations. Our theory of governanceinseparability shares this emphasis on the im-pact of power on governance choice, but it dif-fers from the resource dependence approach inseveral respects.

For instance, in resource dependence theoryfirms acquire power by using ploys to dupeother firms. Donaldson (1995) argues that theploys described in resource dependence theoryare often transparent, and he criticizes the the-ory for assuming excessive naivete. In contrast,in our theory firms are alert to power-seekingattempts by others, so power can only be gener-ated in environments where it is difficult to pre-dict changes in bargaining power. A second dif-ference between our arguments and those ofPfeffer and Salancik (1978) concerns the sourcesof power. We follow the standard economic ap-proach in associating bargaining power withmonopolistic or oligopolistic market structurescreated by entry barriers and/or collective ac-tion (Bain, 1956; Scherer & Ross, 1990). In re-source dependence theory, however, opportuni-ties for power are not generated by particularmarket structures but' simply by de facto needsfor inputs or other resources. We take a moreeconomic view of power, in which firms-are as-sumed to be perceptive (boundedly rational) andgain power from difficult-to-anticipate changesin underlying economic structures.

SOME IMPLICATIONS OF GOVERNANCEINSEPARABILITY

Governance inseparability has important im-plications for the transaction cost theory of thefirm. These implications concern three areas.First, governance inseparability has implica-tions for predicting the relationship between thecharacteristics of individual transactions andthe mechanisms that are used to govern them.

1999 Argyres and Liebeskind 59

Second, governance inseparability has implica-tions for the theory of the limits to the scope ofthe firm. Finally, considerations oi governanceinseparability allow a closer theoretical rela-tionship to be forged between transaction costtheory on the one hand and theories of compe-tition and industry evolution on the other.

Use of Alternative Governance Mechanisms

As we have pointed out in our main argument,the condition of governance inseparability im-plies that, in many instances, the governance ofa given transaction will not be based solely onthe characteristics of that specific transaction.Instead, if the set of governance options avail-able to one or both parties in a transaction isconstrained, the transaction may be governed inways that are suboptimal if the characteristicsof that transaction alone are considered. Thus,we would expect to observe a considerable de-gree of empirical variation in the ways in whicha given type of transaction is governed. For in-stance, we might observe that a given type oftransaction sometimes is internalized within afirm and sometimes is governed by a long-termcontract. Hence:

Proposition 1: Different firms may gov-ern identical transactions in differentways, as long as each firm is also aparty to other types of transactions.

We would expect these differences to be mostpronounced in settings where firms of differentages are active in the same transaction set, be-cause older firms tend to be more encumberedwith past commitments, constraining their abil-ity to switch or differentiate their internal gov-ernance mechanisms. This has two implica-tions. On the one hand, it implies that an olderfirm may be obligated to externalize transac-tions involving firm-specific assets, whereas ayounger firm with few outstanding commit-ments may be able to internalize such transac-tions efficiently according to the usual TCE ar-guments. This is because in the older firm it ismore likely that internalizing new types oftransactions will abrogate commitments inplace that are shaped around existing transac-tions. On the other hand, older firms are morelikely to be engaged in various types of contrac-tual commitments with internal parties, limitingtheir ability to outsource activities that do not

involve firm-specific goods or assets, if similartransactions have been internalized in previousperiods. Hence:

Proposition 2a: Compared withyounger firms, older firms more oftenwill be obligated to use market con-tracting to govern transactions featur-ing asset specificity for the same levelof firm bargaining power.

Proposition 2b: Compared withyounger firms, older firms more oftenwill be obligated to use hierarchicalmechanisms to govern generic trans-actions for the same level of firm bar-gaining power.

Note that since younger firms tend to have lessbargaining power than older ones (e.g., becauseof capital constraints generated by uncertainty),our propositions control for firm bargainingpower.

The degree to which a firm can switch or dif-ferentiate its governance mechanisms also willdepend on the legal jurisdiction in which it op-erates. For instance, some countries, such asGermany and France, accord greater bargain-ing power to labor unions than do other coun-tries. This greater power of unions will tend toengender a greater degree of governance insep-arability than where union bargaining power isweaker.

Proposition 3; Firms operating in juris-dictions in which labor unions are ac-corded more bargaining power will beobligated more often to use hierarchi-cal mechanisms to govern generictransactions than will firms operatingin jurisdictions in which labor unionpower is more restricted.

Limits to Firm Scope

One of the most important questions posed byCoase (1937) in his pioneering article on the the-ory of the firm—a question that remains—concerns the limits to the size and the scope of afirm. If firms form because markets are some-times costly to use, why are not all transactionsthat are more economically governed by a hier-archy governed by one large firm? One resporiseto this question, offered by Williamson (1985), isthat incentives within firms become increas-

60 Academy of Management Review January

ingly attenuated as the size of a firm increases.Consequently, beyond a certain size, the costs ofincentive attenuation within a firm will out-weigh the benefits of hierarchical governance.Milgrom and Roberts (1988, 1991) offer an alter-native explanation: firms, precisely becausethey are organized as hierarchies, promulgateinfluence activities by managers and therebyincur costs that markets avoid.

The argument we present in this article offersa third answer to Coase's (1937) question. Be-cause governance inseparability limits the gov-ernance options available to any particular firm,a single firm can only engage in a limited set oftransactions that can be more or less efficientlygoverned by its particular set of feasible gover-nance options. Otherwise, transaction costs willoutweigh the economic surplus earned. Hence,while each firm may strive over time to econo-mize on transaction costs, this process is subjectto a set of historically determined constraintsthat play an important role in determining fu-ture firm growth. As a result, firms will becomespecialized to both particular types of gover-nance arrangements and to particular types oftransactions. This implies that, ceteris paribus,a firm will face transaction cost limits on bothvertical integration and diversification, if theseexpansion paths call for internalizing transac-tions that cannot be governed efficiently by itsexisting organizational arrangements. Hence:

Proposition 4: The greater the differ-ence is between a transaction's opti-mal governance mechanism and afirm's governance arrangements inplace, the greater the cost v/ill be tothe firm of internalizing that transac-tion.

A second implication of our arguments for thetheory of the scope of the firm relates to theoptimal scope of a firm under differing levels ofuncertainty. TCE theory holds that when uncer-tainty is high, intemalization will be more effi-cient than market contracting because marketcontracts will be excessively costly to write, gov-ern, and enforce (Grossman & Hart, 1986; Wil-liamson, 1975, 1979). Considerations of gover-nance inseparability, instead, imply that firmsmay be more costly than markets when certainaspects of the firm's transacting environmentare uncertain. In particular, intemalization maybe highly inefficient if there is a high degree of

uncertainty about (1) whether the relative effi-ciency of one governance form over another willchange significantly in future periods (whichwould call for a switch between governanceforms for a given transaction) or (2) how thescope of a given firm may need to change infuture periods (which would call for governancedifferentiation).

Balakrishnan and Wenerfelt (1986) have ar-gued that in environments characterized by fre-quent technological obsolescence, the incentiveto vertically integrate is reduced because theexpected returns to investments that are irre-versible and transaction specific are lower.Teece (1992) argues that intemalization of activ-ities may be excessively costly in the presenceof technological uncertainty because it is costlyper se for a firm to absorb new activities. Ourarguments suggest that intemalization is alsocostly, because it may be impossible to undo inlater periods and may engender commitmentsthat spill over to other activities.

Hence, following point 1 above, we would ex-pect that firms operating in environments where(for example) production technologies arechanging rapidly in terms of firm specificity,inseparabilities, or other types of economic spill-overs would be cautious about internalizing pro-duction activities that might engender gover-nance inseparability in later periods. Similarly,following point 2 above, we would expect toobserve that firms that possess highly extensi-ble resources, such as firms that carry out dis-covery research, would be more cautious aboutentering into early commitments than firms withfewer diversification prospects. For instance, anew R&D-based firm might prefer to outsourcecomplementary functions until it can identifywhich activities would be most profitably inter-nalized in terms of both rent generation andgovernance costs, taking inseparability into ac-count. This is because a firm that internalizes arelatively low-valued activity early on, andshapes its contractual commitments around thatactivity, may later find it difficult to differentiateits internal arrangements to accommodate othermore highly valued activities. Thus, for reasonsof governance inseparability alone, we wouldexpect the following:

Proposition 5; Greater uncertainty willreduce the vertical and horizontalscope of the firm.

1999 Argyres and Liebeskind 61

Competition and Industry Evolution

It is important to note here that our predic-tions do not depend on an assumption thatcompetitive forces are attenuated. Althoughwe expect governance inseparability to en-gender suboptimal governance, the total valueof a given transaction is the net value of (1) theeconomic surplus yielded by that transactionand (2) the costs of conducting the transaction.Thus, even in a competitive environment, aprofitable transaction will endure, even if it isburdened to some degree by excess transac-tion costs. Of course, over time we would ex-pect to observe a trend toward increased econ-omizing on transaction costs, as argued byWilliamson (1991). However, it may well bethat the sources of surplus (i.e., rents) may beeroded long before a firm is able to optimizeits governance arrangements.

Our view of the firm as a system of commit-ments provides some institutional underpinningto Lippman and Rumelt's (1982) model of compe-tition. In this model the authors assume firms tobe "locked in" to a particular cost function thatis drawn randomly from a given distribution.Our arguments here suggest that this lock-inresults, in part, from contractual commitments.For instance, in the early stages of an industry,some firms may make choices to produce agiven input internally, whereas other firms maychoose to outsource this input. Over time one orthe other choice may prove to yield a higher netsurplus. However, it may be extremely difficultfor any firm to change its governance arrange-ments over time, if its initial governance choicesinvolved entering into commitments that engen-der governance inseparability. Thus, becausecommitments may foreclose future opportunitiesto adjust to more efficient governance modes,the governance choices that firms make early intheir development may well determine theirlong-run competitive success.

Our arguments are also relevant to address-ing the phenomenon of organizational inertia.Population ecologists argue that organizationsare inherently restricted in their ability to adaptto changes in circumstances so that processes oforganizational birth and death will contributemore to industry change than will processes oforganizational adaptation (Hannan & Freeman,1977, 1984). Evolutionary economists have takena similar view (Nelson & Winter, 1982). We take

a somewhat more moderated view of inertia:organizations will be inert according to the de-gree that the contractual commitments they en-tered into in earlier periods constrain their sub-sequent governance options, either by directlyengendering governance inseparability or byexposing the firm to changes in bargainingpower that result in governance constraints. Weassume here that these early commitments aremade in at least a boundedly rational way andare not merely the result of random processes.However, when circumstances change, the con-straints on governance choices that follow thesecommitments may restrict a firm's ability toadapt to changes in circumstances.

Hence, contractual commitments place lim-its on Lamarckian adaptation but do not pre-clude it. Accordingly, we would expect thatchange in organizational arrangements willtake place only very slowly within survivingfirms. Alternatively, when environments arehighly selective, change will take placethrough the death of firms that are burdenedwith inescapable governance inseparabili-ties, to be replaced by other less-constrainedfirms. Nonetheless, it is important to note thatinefficiency is not necessarily part of this ac-count. Rather, our theory suggests that organ-izations in general, and firms in particular,will demonstrate "weak-form path dependen-cy," a condition in which history matters be-cause it is too costly to reverse—not because itgenerates inefficiencies per se (Leibowitz &Margolis, 1995). But, whatever the efficiencyimplications, we submit that governance in-separabilities often have an important impacton governance choices and must therefore beaccounted for in a positive theory of gover-nance.

REFERENCES

Alchian, A., & Demsetz, H. 1972. Production, information costsand economic organization. American Economic Re-view, 62: 777-795.

Bain, J. 1956. Barriers (o new competition. Cambridge, MA:Harvard University Press.

Balakrishnan, C, & Wernerlelt, B. 1986. Technical change,competition and vertical integration. Strategic Manage-ment Journal 7: 347-359.

Barney, J. 1986. Strategic factor markets: Expectations, luckand business strategy. Management Science, 32: 1231-1241.

62 Academy of Management Review January

Burgelman, R. A., & Sayles, L. R. 1986. Inside corporate inno-vation: Strategy, structure, and managerial skills. NewYork: Free Press.

Coase, R. 1937. The nature of the firm. Economica, 4: 386-405.

Demsetz, H. 1966. Some aspects of property rights. Journal ofLaw and Economics, 9: 61-70.

Dierickx, I., & Cool, K. 1989. Asset stock accumulation and thesustainability of comparative advantage. ManagementScience, 35: 1504-1514.

DiMaggio, P., & Powell, W. 1983. The iron cage revisited:Institutional isomorphism and collective rationality inorganization fields. American SocioiogicaJ Review, 48:147-160.

DiMaggio, P., & Powell W. 1991. Introduction. In W. Powell &P. DiMaggio (Eds.), The new institutionalism in organi-zational analysis: 1-25. Chicago: University of ChicagoPress.

Donaldson, L. 1995. American anti-management theories oforganization. Cambridge, England: Cambridge Univer-sity Press.

Eccles, R. 1985. The transfer pricing problem: A theory forpractice. Lexington, MA: Lexington Books.

Fast, N. 1978. The rise and fall of corporate new venturedivisions. Ann Arbor, MI: University of Michigan Press.

Fligstein, N. 1990. The transformation of corporate control.Cambridge, MA: Harvard University Press.

Ghemawat, P. 1991. Commifmenf: The dynamic of strategy.New York: Free Press.

Grossman, S., & Hart, O. 1986. The costs and benefits ofownership: A theory of vertical and lateral integration.Journal of Political Economy, 94: 691-719.

Hannan, M., & Freeman, J. 1977. The population ecology oforganizations. American Journal of Sociology, 82: 929-964.

Hannan, M., & Freeman, J. 1984. Structural inertia and organ-izational change. American Sociological Review, 49:149-164.

Harris, N. 1997. Franchisees get feisty. Business Weei, Feb-ruary 14: 65-66.

Hart, O., & Moore, J. 1990. Property rights and the nature ofthe firm. Journal of Political Economy, 98: 1119-1158.

Hlavacek, J. D. 1974. Toward more successful venture man-agement. Journal of Marketing, 38: 56-60.

Kansas City Business JoumaL 1993. Red hot regionals: Non-union subsidiaries mean big bucks to big trucking firms.November 26: 1.

Kirzner, I. 1979. Perception, opporfunify and profit: Studies inthe theory of entrepreneurship. Chicago: University ofChicago Press.

Klein, B. 1983. Contracting costs and residual claims: Theseparation of ownership and control. Journal of Law andEconomics, 26: 367-374.

Klein, B. 1993. Contracts and incentives: The role of contractterms in assuring performance. In L. Werin & H. Wij-kander (Eds.), Contract economics: 149-172. Cambridge,MA: Blackwell.

Klein, B., Crawford, R., & Alchian, A. 1978. Vertical integra-tion, appropriable rents and the competitive contractingprocess. Journai of Law and Economics, 21: 297-326.

Leibowitz, S., & Margolis, S. 1995. Path dependence, lock-inand history. Journal of Law, Economics and Organiza-tion, 11: 205-226.

Lieberman, M., & Montgomery, C. 1988. First mover advan-tages. Strategic Management Journal, 9(Special Issue):41-59.

Lippman, S., & Rumelt, R. 1982. Uncertain imitability: Ananalysis of interfirm differences in efficiency undercompetition. Bell Journal of Economics, 13: 418-438.

Masten, S. 1988. A legal basis for the firm. Journal of Law,Economics and Organization, 4: 181-198.

Michaelman, F. 1967. Property, utility and fairness. HarvardLaw Review, 80: 1165-1257.

Milgrom, P., & Roberts, J. 1988. An economic approach toinfluence activities in organizations. American Journalof Sociology, 94: sl54-sl79.

Milgrom, P., & Roberts, J. 1991. Bargaining, influence costs,and the organization of economic activity. In J. Alt &K. Shepsle (Eds.), Perspectives on positive political econ-omy; 57-89. New York: Cambridge University Press.

Nelson, R., & Winter, S. 1982. An evolutionary theory of eco-nomic change. Cambridge, MA: Belknap Press of Har-vard University Press.

New York Times. 1997. American Airlines and its pilots reacha tentative agreement. March 20: C21.

Nickerson, J., & Silverman, B. 1997. TransactionaJ interdepen-dencies, resources, and path dependence: Firm hetero-geneity within transaction cost economics. Working pa-per, Washington University, St. Louis, MO.

North, D. 1990. institutions, institutional change and eco-nomic performance. New York: Cambridge UniversityPress.

North, D., & Thomas, R. 1973. The rise of the western world: Anew economic history. Cambridge, England: CambridgeUniversity Press.

Pisano, G. 1989. Using equity participation to support ex-change: Evidence from the biotechnology industry. Jour-nal of Law, Economics and Organization, 5: 109-126.

Pfeffer, J., & Salancik, G. 1978. The external control of organ-izations: A resource dependence perspective. New York:Harper & Row.

Poppo, L. 1995. Influence activities and strategic coordina-tion: Two distinctions of internal and external markets.Management Science, 41: 1845-1860.

Rumelt, R. 1984. Towards a strategic theory of the firm. InR. Lamb (Ed.), Competitive strategic management: 556-570. Englewood Cliffs, NJ: Prentice-Hall.

Rumelt, R. 1987. Theory, strategy and entrepreneurship. InD. Teece (Ed.), The competitive challenge: Strategies forindustrial innovation and renewal: 137-158. Cambridge,MA: Harper & Row.

Scherer, F. M., & Ross, D. 1990. Industrial market structure

1999 Argyres and Liebeskind 63

and economic performance (3rd ed.). Boston: HoughtonMifflin.

Sutton, J. 1991. Sunk costs and market structure: Price com-petition, advertising, and the evolution of concentration.Cambridge, MA: MIT Press.

Sykes, H. 1986. The anatomy of a corporate venturing pro-gram: Factors influencing success. Journal of BusinessVenturing, 1: 275-293.

Teece, D. J. 1980. Economies of scope and the scope of theenterprise, /ournal of Economic Behavior and Organiza-tion, 1: 223-247.

Teece, D. J. 1992. Competition, cooperation and innovation:Organizational arrangements for regimes of rapid tech-

nological progress. Journal of Economic Behavior andOrganization. 18: 1-26.

Wall Street Journal. 1997a. Strike threat at American givesjitters. March 3: Bl.

Wall Street JoumaL 1997b. Toyota asks regulators to blockRepublic's purchase of a franchise. April 3: B7.

Williamson, O. E. 1975. Markets and hierarchies: Analysisand antitrust implications. New York: Free Press.

Williamson, O. E. 1979. Transaction cost economics: The gov-

ernance of contractual relations. Journal of Law andEconomics, 22: 3-61.

Williamson, O. E. 1983. Credible commitments: Using hos-tages to support exchange. American Economic i?evievi',73: 519-540.

Williarnson, O. E. 1985. The economic institutions of capital-ism. New York: Free Press.

Williamson, O. E. 1991. Comparative economic organization:The analysis of discrete structural alternatives. Admin-istrative Science Quarterly, 36: 269-296.

Williamson, O. E. 1995. Hierarchies, markets and power inthe economy: An economic perspective. Industrial and

Corporate Change, 4: 21-49.Williamson, O. E. 1996. The mechanisms of governance. New

York: Oxford University Press.

Winter, S. 1987. Knowledge and competence as strategicassets. In D. J. Teece (Ed.), The competitive challenge:Strategies for industrial innovation and renewal: 159-184. Cambridge, MA: Ballinger.

Zucker, L. 1987. Institutional theories of organization. AnnualReview of Sociology, 13: 443-464.

Nicholas S. Argyres received his Ph.D. in economics from the University of Californiaat Berkeley. He is an assistant professor in the Department of Management andOrganization, Marshall School of Business, University of Southern California. Hisresearch interests include strategy and organization structure, management of tech-nology, contracting, institutions, and organizational politics.

Julia Porter Liebeskind received her Ph.D. in policy and organization from the Ander-son Graduate School of Management, UCLA. She is an associate professor in theDepartment of Management and Organization, Marshall School of Business, Univer-sity of Southern California. Her research interests include firm scope, corporategovernance, restructuring, and intellectual property.