Stephan Schrader.

Barter in newtworks : a revived form of inter-firm exchange online

. (page 1 of 3)
Font size
QR-code for this ebook



Barter in Networks:
A Revived Form of Inter-Firm Exchange

Stephan Schrader
December, 1991 WP# ^3^9-^1





Barter in Networks:
A Revived Form of Inter-Firm Exchange

Stephan Schrader
December, 1991 WP# ^3^9-^1


Massachusetts Institute of Technology

Alfred P. Sloan School of Management

50 Memorial Drive, E52-553

Cambridge, Massachusetts 02139

(617 253-5219)

FEB 2 1 1992

Barter in Networks:
A Revived Form of Inter-Firm Exchange *^


Firms frequently enter network type relationships such as strategic alliances
and partnerships to barter difficult-to-price goods like technical know-how, market
understanding, and management practices. The paper proposes that barter of
difficult-to-price goods is easier to realize than a money-based exchange — as long as
the double coincidence of wants is given. Under such circumstances, barter requires
decisions of less cognitive complexity than are necessary in the context of a money-
based exchange system. Networks support barter through allowing temporal
separation of transactions, facilitating identification of transaction possibilities, and
establishing mechanisms to enforce cooperative behavior. The barter-supporting
role of networks complements other explanations for the existence of networks,
especially the ones derived from a transaction cost framework.

^ I thank Klaus Brockhof, Jack Brittain, Nicolaus Henke, William M. Riggs, Jill D. Teplensky as
well as the participants of the 1991 Academy of Management Review Theory Development
Workshop for their insightful comments and suggestions. An earlier version of this paper was
presented at the 1991 Academy of Management Annual Meeting, Miami Beach.



Inter-firm networks have received considerable attention in recent years.
Popular business publications proclaim that the formation of networks constitutes
an integral part of a successful business strategy (e.g. Economist, 1990). Numerous
academic articles attempt to conceptualize core issues of inter-firm networking (e.g.
Barley & Freeman, 1990; Furukawa, Teramoto, & Kanda, 1990; Gemiinden, 1990;
Hakansson, 1987; Jarillo, 1988; Luke, Begun, & Pointer, 1989; MacMillan & Farmer,
1979; Ouchi & Bolton, 1988; Porter, 1990; Saxenian, 1989; Thorelli, 1986). This paper
investigates a characteristic of networks that has been widely neglected by the
literature: the support of barter as a mode of exchange.

Networks provide the frame for a multitude of transactions, many of those
involving barter. Firms engage in strategic alliances or partnerships to barter, for
example, technical know-how for management experience or market understanding
for technology (Hamel, Doz, & Prahalad, 1989; Leadbeater, 1990; Roberts, 1980).
Several authors have described informal inter-firm networks in which goods such
as technical information are not traded for money but bartered for other like goods
— even if the goods are of considerable economic value (Rogers, 1982; Schrader,
1991, von Flippel, 1987).

In this paper, I propose that situations exist in which a barter system offers
distinct advantages over a trade-for-money system. These advantages exist
especially if it is difficult to put a monetary value on the goods to be exchanged, as is
frequently the case for goods like information or market access. Under such
circumstances, networks help to establish barter systems and to capture the benefits
of such systems.

To characterize situations prone to barter, I introduce a distinction between
the ability to specify the attributes of a good and the ability to price the good in


monetary terms. Most authors discuss only one of these dimensions (e.g.
Williamson, 1975) or view them as closely linked (e.g. Arrow, 1971). Based on this
distinction, four unique transaction modes are characterized: discrete purchase and
discrete barter as well as relational purchase and relational barter. I argue that
networks support both relational barter and discrete barter through allowing
temporal separation of transactions, facilitating identification of transaction
possibilities, and establishing mechanisms to enforce cooperative behavior. This
characteristic of networks augments the traditional transaction cost and
communication oriented explanations for the existence of networks. Furthermore,
the argument that barter is the preferred transaction mode if prices are difficult or
impossible to determine implies that those conventional control and decision
practices relying on prices and related quantitative measures are likely to fail in
barter-supporting networks. This conclusion has considerable management
ramifications and should encourage the study of how to govern such situations.

The paper consists of four parts. First, I review briefly some of the relevant
literature on inter-firm networks. Next, I introduce a framework to distinguish
barter from trade-for-money. Then, I discuss why barter appears to be the preferred
mode of exchange in some settings. Finally, I investigate how networks support the
emergence of barter systems.

Networks as Governance Structures for Inter-Firm Relationships

Firms face several alternatives for organizing their vertical and horizontal
relationships. The two ideal types discussed by Coase (1937) and Williamson (1975)
are markets and hierarchies. In market relationships, transactions take place
between independent entities and are mediated by a price mechanism. In
hierarchical relationships, the transaction partners are part of one corporate body


which somehow mediates the relationship through such mechar\isms as
surveillance, evaluation, and direction.

The market-hierarchy typology has served several authors as a starting point
for more refined typologies. Ouchi (1980), for example, discusses three types:
markets, bureaucracies, and clans. The latter is characterized by a considerable
correspondence of organizational and individual goals.

Jarillo (1988) extends this typology by differentiating markets further into
classic markets and strategic networks. Classic markets are characterized by a
competitive relationship between transaction partners. Not so in strategic networks.
Strategic networks are based on a potential for cooperation. The possibility of
capturing long-term benefits from cooperative behavior induces firms to overcome
short-sighted opportunism and to enter long-term relationships while maintaining
most of their organizational independence.

MacNeil (1978) differentiates market contracts even further. He introduces
three types of contracts: classical contracts (the identity of the transaction partner is
irrelevant and each party's obligation is well-defined), neoclassical contracts (mostly
long term contracts that provide governance structures for solving potential
conflicts instead of explicitly anticipating every contingency), and relational
contracts (agreements on governance structures for entire long-term relationships
that may evolve well beyond the originally planned transactions).

This paper adopts Thorelli's definition of networks as "two or more
organizations involved in long- term relationships" (Thorelli, 1986, p. 37), and
thereby includes both neoclassical and relational contracts. These relationships can
be of contractual as well as non-contractual nature. In most cases, a mixture of both
will be present (Macaulay, 1963). The definition used is sufficiently open to
encompass several forms of inter-firm relationship, ranging from explicitly
organized networks such as supply networks (Jarillo, 1988) to implicitly evolved

networks such as informal information trading networks between competitors (von
Hippel, 1987; Schrader, 1991).

The Hterature discusses networks primarily in the context of vertical
relationships. Note, however, that organizations frequently engage not only in
vertical but also in horizontal networks. Research consortia with competing firms
as members are a case in point (Dimancescu & Botkin, 1986). Firms participating in
such consortia have developed formal horizontal ties to gain advantages from
resource pooling and risk sharing. Informal information trading between
competing firms is another example (Carter, 1989; Schrader, 1991; von Hippel, 1987).
Firms within one industry are linked by informal communication networks.
Employees exchange information in these networks without explicit contracts.

Why do networks exist? The conventional explanation as provided, for
example, by MacMillan and Farmer (1979) argues that networks capture core benefits
of markets while enjoying some transaction costs advantages of hierarchies. First,
networks enable firms to realize economies of scale and specialization advantages
(Jarillo, 1988). One network member may specialize on the production of a specific
good and distribute the good to the other members. Thus, the network partners
benefit from the organizational advantages of small specialized firms. Risk is spread
between different entities. And the market test is still applicable. If better trading
terms can be obtained elsewhere in the long run, no permanent tie (as in the case of
vertical integration) stops either party from forming new relationships outside the
existing network (MacMillan & Farmer, 1979, p. 283). Jarillo (1988, p. 35) concludes:
"Networking introduces a cost discipline that may be absent in an integrated firm,
with its captive internal markets." Also, in times of considerable technical change,
networks allow a less expensive restructuring of a firm's ties to specific technologies
than would be possible in the context of hierarchies (DeBresson & Amesse,
forthcoming). Second, networks capture benefits of hierarchies. Networks help to


avoid some of the transaction costs that emerge in a pure market system due to
opportunistic behavior and asset specificity (Jarillo, 1988; Joskow, 1987; Klein,
Crawford, & Alchian, 1978; Thorelli, 1986; Williamson, 1985). Klein, Crawford and
Alchian (1978), for example, demonstrate the negotiation problems that arise in
supply relationships if either of two contract partners has to invest in relationship-
specific assets. Such problems can be mediated by networks. For this, the network
partners have to expect that they will benefit from an ongoing cooperative
relationship and that this benefit overcompensates any short-term gains that would
result from opportunistic behavior (Jarillo, 1988). Consequently, network partners
are willing to invest in network-specific assets even if this increases their
dependency and thereby vulnerability to threats. They trust that other network
members will not use this vulnerability to cause them undue harm.

In this paper, I propose that the conventional theory of networks should be
augmented by another characteristic of networks. Networks support the emergence
of barter systems . 1 argue that this is an important characteristic of networks since
some goods can be bartered more effectively than they can be traded for money.
Before this argument can be explored further, however, the concept of barter needs


Barter is a form of "reciprocal dealing" (Thorelli, 1986, p. 44) in which one
good is exchanged for another good without using a separate unit of account or
medium of exchange (Pearce, 1986, p. 36). These reciprocal dealings may occur
simultaneously. For example, IBM gives Siemens access to (parts) of its chip
production technology and at the same time Siemens provides IBM with access to
its technical expertise (Markoff, 1990). Alternatively, a temporal separation of


transactions can occur (von Weizsacker, 1985). Take information trading in the oil-
exploration industry as an example. In this industry, employees have been
observed to exchange well-logs informally (Schrader & von Hippel, forthcoming).
In some instances, several months can pass between giving one well-log and
receiving another well-log in return. The party that first provides a good to the
other party acquires a claim to receive a comparable good in the future.

In the following, three types of barter are distinguished: discrete barter,
relational barter, and pseudo barter. Pseudo barter refers to a transaction that at first
sight appears to be barter but that can be readily decomposed into two good-for-
money transactions. Bartering crude oil of a specific grade for an equally well-
defined amount of wheat with both parties having the alternative of buying and
selling the respective commodities on the world market is a case in point. Using
world market prices, the barter can be broken up into two good-for-money
transactions — including a side-payment if the respective amounts do not cancel
out.^ Since pseudo-barter can be interpreted as a set of separate purchasing
agreements, the term barter as used in the remainder of the paper will not include

Discrete barter refers to the exchange of one well-defined good for another
well-defined good in a situation in which commonly accepted valuations are not
readily available. Assuming self-interested actors, such barter occurs if the post
barter distribution of goods is preferred to the pre barter one. Yet, the parties need
not to be able or willing to agree on precise monetary prices for the goods. Even if
each party would be capable of valuing the goods in exact monetary terms, the
valuations do not necessarily have to be such that they determine a non-empty
negotiation space. Barter necessitates only that ordinal rankings of goods are such

^ In this case, an "objective" valuation of the goods exists insofar as the price is determined in a way
that is wridely independent of the individual market participants' preferences (Weizsacker, 1985).


that an exchange is preferable, not that cardinal evaluations would justify an

In the case of discrete barter, it is not possible for a third party to deduce
without additional information the parties' valuations of the goods exchanged.
What can be determined is that x units of one good have been exchanged for y units
of another good. This relation, however, only conveys the information that the
parties have reached a specific agreement. It does not convey any further
information on the valuation of the goods. It is not possible to conclude for how
many units of a third good the parties would have exchanged the goods in question.

Von Weizsacker (1985) argues convincingly that barter systems break down as
soon as the parties evaluate specific transactions monetarily. If such valuations
occur, the parties will perceive, at least temporarily, inequalities and will require to
be compensated by side-payments. The resulting negotiations will lead to an
introduction of money-based exchange systems. He proposes that valuation is
either required in an exchange relation (money-based exchange) or that it is
explicitly or implicitly forbidden (barter).

Relational barter ^ differs from discrete barter by the transaction content not
being well defined are well definable. Technical know-how, for example, frequently
falls into this category. Know-how can be tacit, ill-understood, and embedded in a
complex set of subjective knowledge (Polanyi, 1958). If such know-how is to be
exchanged, any attempt to precisely stipulate the content of the exchange is doomed
to fail. Consequently, other (explicit or implicit) contractual mechanisms have to be
employed. Relational contracting (MacNeil, 1978; MacNeil, 1980) provides an
alternative. Here it is not the substance of the exchange that is determined, but
rather procedural aspects and relationship characteristics. Examples of procedural

The notion of relational barter is closely linked to MacNeil's (1980) concept of relational contracts.


regulations are control meetings, documentation requirements, and reporting
obligations. Sometimes these procedures are never agreed on explicitly but evolve
out of an ongoing relationship. An extreme case of relational barter is academic
collaboration. In such collaborations, it is frequently impossible to identify or to
monetarily value individual transactions (what is a transaction in the context of an
academic discourse?). Nevertheless, the partners may have w^ell-formed
perceptions of wrhether the cooperation is beneficial to each of them and of whether
the exchange should be continued in the future.


Discrete Barter

Substantive dimensions
of exchange can be
Valuative dimension
remains vague.

Substantive contracting
Figxire 1: Types of Barter

Relational Barter

Substantive dimensions
of exchange cannot be
Valuative dimension
remains vague

Relational contracting

Pseudo Barter

Decomposition into two
good-for-money transactions
plus sidepayment easily

=> Purchase

Figure 1 summarizes the proposed distinction between discrete barter and
relational barter. Both are characterized by the parties not having to agree on a
precise monetary price for each barter component. A decomposition of the barter
into several good-for-money transactions cannot be done without strong
assumptions, and — as von Weizsacker argues — probably is not done by the parties
even if it would be possible. This characteristic distinguishes barter from money-
based exchange. In money-based exchange, the parties involved agree on precise
prices. Frequently, such agreement is achieved by adopting established market


prices. Under other circumstances, prices result fronn negotiations. This requires
each party to generate at least a rough monetary valuation of the good in question
and that the parties are able to agree on a specific monetary value (the price) that fits
with the valuations by all parties.

Two types of money-based exchanges can be distinguished, again using the
ability to specify the attributes of the goods to be transferred as classification
criterion. The two types are discrete purchase and relational purchase. Discrete
purchase refers to an exchange in which the good can be determined precisely. This
is the classical purchasing agreement as it is considered by traditional
microeconomic theory. Relational purchase provides a more interesting case. In a
relational purchase, the parties agree on a price for the goods to be exchanged, but
fail to precisely specify the substance of the exchange. Consulting contracts are a case
in point. Frequently, well established hourly rates exist and the parties can refer
back to those rates when determining the exchange price. Yet, the substance of the
consulting work cannot be agreed on in detail before the transaction — otherwise
the consulting work would be superfluous. Therefore, a relational contract is
established, using means such as reputation, control procedures, and iterative
interactions to ensure that both parties are satisfied with the outcome of the

In conclusion, the two dimensions discussed — ability to specify attributes of
the goods to be transferred and ability to determine prices — can be used to
distinguish four types of exchange relationships that should be preferred under
different circumstances, as shown in Figure 2. Whether or not the transaction
partners are able to specify ex ante the attributes of the goods to be exchanged
determines whether or not they will utilize substantive contracts or relational
contracts. Their ability to determine prices impacts their inclination to trade the
goods for money or to barter them.


Ability to specify attributes of good to
be transferred

High Low

Ability to High


price of good to

Discrete purchase


Money based

be exchanged


Discrete barter

Relational barter




Figure 2: Classification of Exchange Relationships

In the following sections, barter is investigated further. First, examples of
barter in networks are provided; then advantages of barter are discussed; and finally
the support of barter through networks is examined.

Characterization of Barter in Networks

Network transactions frequently show characteristics of barter, i.e. occur
without direct monetary compensation. Engineers, for example, exchange
information about how to solve specific technical problems through professional
networks with colleagues in other firms (von Hippel, 1987; Schrader, 1991). A large
German bank provided a competing bank with access to its central computer facility
when the other bank's computer facility was not operational for several days due to
a power failure (Kieler Nachrichten, 1978). U.S. minimill steel companies
frequently exchange metal dies in situations of unexpected need. Motorola provides
Toshiba with access to its microprocessor technology in return for Toshiba's support
in penetrating the Japanese semiconductor market (Hamel et al., 1989).


Valuable goods — information, access to computer facilities, metal dies,
technology and market access — are exchanged within networks without a
compensating flow of money. These exchanges are part of a barter system. The
goods are provided in the expectation of receiving other valuable goods in return.
The bank that opened its computer facilities to a competitor thereby gained the right
to use the other firms facilities in a similar situation. In other words, the bank
bartered access to its facilities for some type of insurance against a breakdown of its
own computer system.^

One type of good that is frequently bartered in networks is information. A
leading user of printed circuit boards (PCBs) and its suppliers shall serve as an
example. The PCB user decided to change its relationships to its suppliers. The
motto was "We are a world class manufacturer, we want to be a world class
customer." The number of suppliers was greatly reduced to be able to develop a
close network with the remaining ones. A system of long-term supply contracts was
established. In addition, the companies bartered valuable, proprietary technical
information. Each firm was contributing some of its knowledge regarding design,
production, and testing and was receiving similar knowledge and other difficult to
price goods such as greater supplier reliability in return.

Barter of information in networks can also be observed between rival firms
(von Hippel, 1987; Schrader, 1991). In some instances, the horizontal exchange of
information is supported by other organizations that are placed in the vertical chain
before or after the horizontal network. For example, some equipment suppliers
encourage their customers to exchange technical information between each other in
so-called user groups. Similarly, manufacturers frequently induce irxformation
trading between their suppliers. Concast's management of its customer's relations

In addition, the bank might have been motivated to help in order to prevent a negative image


demonstrates the vertical support of horizontal information exchange networks.
Concast is a leading supplier of continuous casters, a crucial piece of equipment for
steel minimills and nowadays for other steel companies as well. In the early years of
the industry, Concast required its customers to feed irxformation about equipment
modifications back to Concast. Concast in turn communicated this information to
other customers. This system can be interpreted as an indirect barter system. Each
equipment user contributes technical expertise to the network and gains access to
other users' expertise. The system, however, started to break down when the user
community evolved and the level of technical sophistication started to vary
strongly between firms, leading to a perceived inequality of the exchange.

Advantages of Barter

The observation that many goods are bartered in networks and are not
exchanged for monetary compensation is puzzling at first sight. Traditionally,
economists have assumed barter to be an inferior mode of exchange in comparison
to trade-for-money. A frequently cited argument in support of the supremacy of
money-based exchange is that barter requires finding a partner that wants what one
has and that has what one wants, the so called double coincidence of wants (Pearce,
1986; Samuelson, 1985). A trade-for-money system, however, allows decomposition
of one barter transaction into two transactions that may occur with different
partners. No double coincidence of wants is necessary. This , so the argument goes,

1 3