858117

Upload: rajendrasamy

Post on 13-Apr-2018

215 views

Category:

Documents


0 download

TRANSCRIPT

  • 7/27/2019 858117

    1/13

    Introduction

    Supply chain management is a rapidly evolv-

    ing area of interest to academics and business

    management practitioners alike. Aspects of

    marketing, economics, logistics and organiza-

    tional behaviour are all important for develop-ing insights into how and why different supply

    chain management arrangements emerge and

    for understanding the consequences of these

    arrangements for industry efficiency and

    competitiveness. When undertaking any

    analysis, it is helpful to have a framework

    within which to work and from which testable

    hypotheses can be drawn. A theoretical frame-

    work enables predictions to be made about the

    likely outcomes of different business strategies

    and public policy initiatives. I t enables

    observed business behaviour to be evaluated

    and therefore provides better explanations of

    the motivations for firms behaviour and the

    consequences for efficiency within a supply

    chain.

    This paper discusses a theoretical frame-

    work for the study of supply chain manage-

    ment which is drawn from the economics

    literature. Transaction cost analysis (TCA)

    represents one possible approach to under-

    standing and evaluating supply chain manage-

    ment and has the potential to be combined inan interdisciplinary setting with the insights

    provided by the marketing, logistics and

    organizational behaviour literatures. The

    purpose of this paper is to provide those inter-

    ested in supply chain management, but who

    are not familiar with TCA, with an overview

    of this approach and to discuss methods of

    applying the theory empirically.

    Econom ics and supp ly chainmanagement

    Business people, academics who teach and

    study management and others interested in

    the operation of supply chains often express

    frustration with how few insights economics

    appears to provide them. Although it is not

    always obvious, the source of this frustration

    lies in the assumptions underlying the ruling

    neoclassical paradigm used in economic

    analysis. Most economic studies of markets,

    industries and firms use this theoretical

    approach. Central to neoclassical theory is the

    concept of a single product firm, operating in

    a perfectly competitive industry with a large

    number of competitor firms all producing the

    15

    Supply Chain M anagement

    Volume 1 Number 2 1996 pp. 1527

    M CB Un iv ersi ty Press I SSN 13 59 -8 54 6

    Research pa perA t ransact ion costapp roach t o supply

    chain m anagem entJill E. H obbs

    The author

    Jill E. Hob bs is based at Excellence in th e Pacifi c Research

    Institute, University of Lethbridge, Lethbridge, Canada,

    and Faculty o f M anagement, The University of Calgary,

    Calgary, Canada.

    Abstract

    Observes that supply chain managem ent is a rapidly-

    evolving subject w hich offers many insights into how

    industries are organized and int o the effi ciency gains

    w hich can be made under different organizational struc-

    tures, pointing out t hat it is an interdisciplinary concept,

    draw ing on aspects of m arketing, economics, logistics,

    organizational behaviou r, etc. Presents a fram ewo rk from

    the economics literature whi ch may be useful for those

    interested in understanding and exploring th e concept of

    supply chain managem ent. Describes the origins and

    development o f transaction cost analysis and explains the

    key concepts of the fram ewo rk. Discusses the potenti al

    effects of transaction costs on vertical co-ordination w ithin

    an industry and, hence, on supply chain management.

    Finally, suggests methods for empiricizing transaction cost

    analysis, result ing in recommendation s for closer co-

    operation betw een researchers and business managers.

  • 7/27/2019 858117

    2/13

    same product under the same cost conditions

    and all facing the same market demand curve.

    (Of course, neoclassical theory has been

    successfully extended to cover monopolies

    and, with less success, to other intermediate

    forms of industrial organization such as

    monopolistic competition and oligopoly.)

    The standard neoclassical transaction

    involves the exchange of a homogeneous

    product there are no quality variations

    between products and consequently no costs

    involved in measuring the value of a product.

    Where products do exhibit quality differences,

    they are regarded as distinct products serving

    separate markets. Economic agents are

    assumed to possess perfect information,

    hence, there is no uncertainty regarding

    prices, product characteristics, or the behav-iour of competitors and trading partners. The

    neoclassical transaction occurs in the current

    time period between multiple buyers and

    sellers, thereby ruling out the possibility that

    one firm could exercise market power over

    others since many alternative buyers and

    sellers exist. Neoclassical economic analysis

    concentrates on equilibrium market out-

    comes. T here is no consideration of how

    business relationships arise. Instead, transac-

    tions are treated as though they occur in a

    frictionless economic environment, somewhat

    analogous to the physicists perfect vacuum.

    Somewhat ironically, the neoclassical theory

    of the firm has little to say about the firm; it

    does not provide a rationale for the existence

    of firms, an explanation of the growth of firms

    or an analysis of the internal organization of

    firms[1-4]. The firm is instead treated as a

    black box, a featureless production function

    which turns inputs into outputs; as such it is a

    component of the neoclassical explanation of

    the workings of a competitive economy, butone which is little understood. When one

    strips neoclassical theory down to its key

    assumptions, it is not surprising that analysis

    undertaken using its framework provides few,

    if any, insights for those interested in supply

    chain management.

    New institut ional economics

    Coase[2] identified some limitations to the

    neoclassical paradigm for understandingrelationships between firms. These ideas later

    became the foundation for new institutional

    economics. Coase argued that in order to

    understand what a firm does, one must first

    understand why a firm exists and, therefore,

    what forces govern the organization of eco-

    nomic activity. Unlike standard neoclassical

    economics, the Coasian approach recognized

    that there are costs to using the market mecha-

    nism. T hese include the costs of discovering

    what prices should be, the costs of negotiating

    individual contracts for each exchange trans-

    action and the costs of accurately specifying

    the details of a transaction in a long-term

    contract. These costs were later termed

    transaction costs[5].

    The costs of using the market can be avoid-

    ed if a firm becomes vertically integrated and

    assumes the burden of co-ordinating econom-

    ic activity internally through within-firm

    managerial direction. However, this means

    that a firm must assume the alternative costsof administering vertical flows of products and

    organizing factors of production. Provided

    that a firm can carry out these activities inter-

    nally at a lower cost than would be the case if

    the transaction were co-ordinated through an

    open market, then one would expect, all other

    things being equal, the organization of eco-

    nomic activities to be carried out by a vertical-

    ly integrated firm. Coase argued that:a firm will tend to expand until the costs of

    organising an extra transaction within the firm

    become equal to the costs of carrying out thesame transaction by means of exchange on the

    open market or the costs of organising in anoth-

    er firm[2, p. 395].

    Hence, he provided a rationale for the exis-

    tence of the firm which was based on the costs

    of carrying out a transaction. T hese insights,

    however, did not have a major impact on

    economic thought until more than 35 years

    later.

    In the 1970s, interest in transaction costs

    increased. Pioneering work in the develop-ment of a theory of transaction costs was

    carried out by Williamson[6,7]. Gradually, a

    body of theories based on the concept of

    transaction costs emerged; these include the

    transaction cost economics of Williamson[7],

    the property rights school[8-10], agency

    theory[11], the economics of the multination-

    al enterprise[12-14] and a transaction cost

    approach to economic history[4,15].

    Although focusing on separate economic

    problems, these approaches all have their

    roots in the original ideas of Coase[2] and use

    the concept of transaction costs to explain the

    organization of firms and the way in which

    they interact along a supply chain. (For an

    16

    A t ransact ion cost approach to supply chain m anagement

    Jill E. Hobbs

    Supply Chain M anagement

    Volume 1 Number 2 1996 1527

  • 7/27/2019 858117

    3/13

    excellent summary of theoretical develop-

    ments, empirical applications and a proposed

    framework for synthesizing the transaction

    and agency costs perspectives, see [16].)

    Tran saction cost sTransaction costs are simply the costs of carry-

    ing out any exchange, whether between firms

    in a marketplace or a transfer of resources

    between stages in a vertically integrated firm,

    when the neoclassical assumption of perfect

    and costless information is relaxed. They arise

    wherever there is any form of economic orga-

    nization, i.e. within a vertically integrated firm,

    in a market or in a command economy (in

    which market transactions are largely absent).

    It should be noted that the word transactionis used in a broader context than the normal

    English language usage which would not

    consider the movement of a finished product

    from a production line to the loading dock as a

    transaction because it is co-ordinated within

    the firm by managerial direction. The transac-

    tion cost approach treats this as a within-

    firm transaction. Unlike the frictionless

    economic system implied by neoclassical

    theory, T CA recognizes that transactions do

    not occur without friction and labels the costs

    which arise from the interaction between and

    within firms as transaction costs.

    It is useful to divide transaction costs into

    three main classifications: information costs,

    negotiation costs, and monitoring (or enforce-

    ment) costs. Firms and individuals face costs

    in the search for information about products,

    prices, inputs and buyers or sellers. Negotia-

    tion costs arise from the physical act of the

    transaction, such as negotiating and writing

    contracts (costs in terms of managerial exper-

    tise, the hiring of lawyers, etc.), or paying forthe services of an intermediary to the transac-

    tion (such as an auctioneer or a broker). M on-

    itoring or enforcement costs arise after an

    exchange has been negotiated. This may

    involve monitoring the quality of goods from a

    supplier or monitoring the behaviour of a

    supplier or buyer to ensure that all the pre-

    agreed terms of the transaction are met. Also

    included are the costs of legally enforcing a

    broken contract, should the need arise. I t

    should be noted that the relaxation of neoclas-sical assumption of perfect and costless infor-

    mation gives rise to all three types of transac-

    tion cost (information, negotiation and moni-

    toring costs). Information plays a key role in

    all three cases, however, the lack of informa-

    tion prior to a transaction is explicitly consid-

    ered to be an information cost.

    Transaction cost ana lysis ke y concept s

    A number of other disciplines including

    psychology, political science, economic histo-

    ry and law have all contributed to the theo-

    retical development of T CA. Drawn, in part,

    from these associated disciplines, four key

    concepts underpin TCA. T hey are bounded

    rationality, opportunism, asset specificity and

    informational asymmetry.

    Bounded rationality

    Bounded rationality means that although

    people may intend to make a rational decision,their capacity to evaluate accurately all possi-

    ble decision alternatives is physically limit-

    ed[17]. A useful analogy is that of a chess

    player who, although able to view the position

    of all playing pieces on the chessboard, cannot

    feasibly evaluate all potential outcomes of a

    move, given an opponents counter-moves,

    and given their own counter-counter moves,

    and so on[18]. Bounded rationality poses a

    problem only in situations of complexity or

    uncertainty where the ability of people to

    make a fully rational decision is impeded.

    Opportunism

    Opportunism has been defined by

    Williamson[7, p. 234] as self-interest seeking

    with guile. In other words, it recognizes that

    businesses and individuals will sometimes

    seek to exploit a situation to their own advan-

    tage. T his does not imply that all those

    involved in transactions act opportunistically

    all of the time, rather, it recognizes that the

    risk of opportunism is often present. T his riskis greater when there exists a small numbers

    bargaining problem[7]. For example, the

    fewer the number of alternative suppliers

    available to a buyer, the more likely it is that an

    existing supplier will act opportunistically to

    alter the terms of the business relationship to

    their own advantage, such as by demanding a

    higher price than that previously agreed.

    Asset specificity

    Asset specificity arises when one partner to anexchange (firm A) has invested resources

    specific to that exchange which have little or no

    value in an alternative use. Examples might be

    the installation of specialized machinery in a

    17

    A t ransact ion cost approach to supply chain management

    Jill E. Hobb s

    Supply Chain M anagement

    Volume 1 Number 2 1996 1527

  • 7/27/2019 858117

    4/13

    production plant or the development or pro-

    motion of a product unique to one market.

    Firm A faces the risk that its trading partner

    (firm B) will act opportunistically by trying to

    appropriate some of the rent from this

    investment. (The term rent in this context

    refers to the additional amount over and above

    the minimum return firm A requires to induce

    it to make the specialized investment.) K now-

    ing that firm A has made a specialized invest-

    ment, and is therefore locked-into the ex-

    change, firm B could renege on the previous

    agreement by offering firm A a lower price for

    the product. Provided that this lower price

    covers As operating costs and makes a contri-

    bution towards the fixed cost of the investment,

    A will have little choice but to accept Bs dis-

    counted price offer[19]. Of course, in manycountries firm A would have recourse to con-

    tract law to enforce this contract and could sue

    for damages if the contract were reneged on.

    However, firm A may incur legal costs and

    faces uncertainty over the outcome of any

    litigious action. The decision to apply the law

    of contracts will probably depend on whether

    the benefits to be gained from forcing firm B to

    uphold its contractual obligations outweigh the

    costs of legal action, allowing for the uncertain-

    ty of the judicial process. T he economic rent

    captured by firm B is known as specialized

    appropriable quasi-rent. The opportunistic

    behaviour displayed by firm B is termed post-

    contractual opportunistic behaviour (PCOB)

    or opportunistic recontracting.

    Informational asymmetry

    Transaction cost analysis also allows for

    the relaxation of the full or perfect informa-

    tion assumptions of neoclassical theory.

    Drawing on the economics of information

    literature[20-22], TCA recognizes that manybusiness exchanges are characterized by

    incomplete, imperfect or asymmetrical

    information. (Information incompleteness

    and uncertainty refer to the situation where

    all parties to a transaction face the same, but

    incomplete, levels of information. Therefore,

    they all face the same uncertainty. Informa-

    tion asymmetry arises when there is public

    information available to all parties but also

    private information which is only available to

    selected parties, so that all parties to thetransaction no longer possess the same levels

    of information.) Informational asymmetries

    can lead to opportunistic behaviour in two

    ways. T he first involvesex anteopportunism

    where information is hidden prior to a trans-

    action. This is known as adverse selection

    and was first defined by Akerlof in his 1970

    seminal paper on the market for lemons.

    Akerlof suggested that in a situation of asym-

    metric information, a seller may possess

    information about defects in a product (e.g.

    a faulty second-hand car or lemon) that is

    not available to the potential buyer. As a

    result, the seller can act opportunistically by

    failing to reveal these defects to the buyer

    prior to the transaction. Buyers of second-

    hand cars always face the risk that the seller

    is acting opportunistically, trying to sell them

    a lemon. Since buyers cannot tell the

    difference between a good car and a lemon,

    both cars must sell at the same market price.

    At this price there will be a reduced incentivefor owners of good cars to sell them on the

    second-hand market, whereas selling lemons

    may still be a viable option. Lemons tend to

    self-select the used car market and there is a

    higher probability of purchasing a lemon in

    this market. Akerlof used this reasoning to

    explain why the price of new cars depreciates

    so rapidly once they have been sold. Hidden

    information can lead to adverse selection

    and problems of opportunistic behaviour. In

    some countries, contract law mitigates the

    problem of adverse selection, providing

    some protection to the buyer.

    Moral hazard also arises from informational

    asymmetry. This isex postopportunism which

    occurs after a transaction because of the hidden

    actions of individuals or firms. T hese parties

    may have the incentive to act opportunistically

    to increase their economic welfare because

    their actions are not directly observable by

    other parties. For example, insurance compa-

    nies cannot observe the actions of their clients.

    Once individuals have obtained fire insurancethey take less care in the prevention of fires,

    thus leading to an increase in the incidence of

    fires and an increase in insurance premiums.

    Alternatively, they may act opportunistically to

    damage items intentionally in order to collect

    insurance payments. In either case, informa-

    tion asymmetry exists because the actions (or

    inaction) of the individuals are not directly

    observable by the insurance provider.

    The relationship bet w een t ransactioncosts and ver tical co-ord inatio n

    Transaction costs are important because they

    affect the organization of economic activity or

    18

    A t ransact ion cost approach to supply chain m anagement

    Jill E. Hobbs

    Supply Chain M anagement

    Volume 1 Number 2 1996 1527

  • 7/27/2019 858117

    5/13

    vertical co-ordination. (Some writers refer

    to this as the governance structure of con-

    tractual relations, notably Williamson[6,7],

    who invented the term.) According to M ighell

    and Jones, vertical co-ordination:includes all the ways of harmonizing the

    successive vertical stages of production andmarketing. The market-price system, vertical

    integration, contracting, cooperation singly or in

    combination are some of the alternative means

    of coordination[23, p. 1].

    Thus, there is always some kind of vertical co-

    ordination if any production takes place. Of

    course, vertical co-ordination is central to the

    study of supply chain management.

    Vertical co-ordination can be viewed as a

    continuum. At one extreme lie spot markets

    where goods are exchanged between multiplebuyers and sellers in the current time period,

    with price as the sole determinant of the final

    transaction. In other words, other aspects of

    the transaction are non-negotiable the buyer

    either accepts the product in its current form,

    or does not purchase it. (Negotiation over

    product quality, delivery schedules, etc. would

    constitute a more formal exchange relation-

    ship often resulting in some form of contract.)

    Examples of spot markets are auction mar-

    kets, stock markets and most consumer good

    purchases (e.g. purchases of food in a super-

    market). In a spot market transaction, man-

    agement of the supply chain, in any formal

    sense, is entirely absent. At the other end of

    the vertical co-ordination spectrum lies full

    vertical integration, where products move

    between various stages of the production-

    processing-distribution chain as a result of

    within-firm managerial orders rather than at

    the direction of prices.

    In between the two extremes of spot market

    transactions and vertically integrated firms liea myriad alternative ways of co-ordinating

    economic activity, from strategic alliances and

    formal written contracts, to vertical integra-

    tion. These represent different degrees of

    supply chain management some more for-

    mal than others. A strategic alliance is an

    agreement mutually entered into by two

    independent firms to serve a common strate-

    gic objective. I t is often more flexible than a

    contract or full vertical integration. Central to

    the success of a strategic alliance are trust

    between firms and a strategy which is to the

    mutual benefit of all the participants; some-

    times the alliance may also place legal obliga-

    tions on the parties. For example, a meat

    processor might reach an agreement with a

    group of pig producers to obtain finished pigs

    of a certain quality, providing producers with

    a list of acceptable breeders. A meat processor

    might also introduce a high-quality packaged

    pork product jointly developed with a major

    retailer under a strategic alliance[24].

    Under a contract, a firm usually devolves

    control over various aspects of the supply

    chain i.e. marketing and/or production of its

    product to a buyer. Contracts can be classi-

    fied into three broad groups[23]:

    (1) Market specification contracts represent

    an agreement by a buyer to provide a

    market for a sellers output. T he seller

    transfers some risk and the decisions over

    when the product is sold and how it is

    marketed to the buyer. Control over theproduction process, however, remains

    with the seller.

    (2) A production-management contract gives

    more control to the buyer than a market-

    specification contract. The buyer partici-

    pates in production management through

    inspecting production processes and

    specifying input usage.

    (3) Even more control rests with the buyer in

    the case of resource-providing contracts in

    which the buyer provides a market outletfor the product, supervises its production

    and supplies key inputs. Often, the buyer

    may own the product, with the seller paid

    according to the volume of output. This is

    the closest contractual arrangement to full

    vertical integration. For example, a feed-

    stuffs manufacturer might contract with

    pig producers, supplying feedstuffs, over-

    seeing production methods and market-

    ing the finished pigs.

    Quasi-vertical integration refers to a relation-ship between buyers and sellers that involves a

    long-term contractual obligation where both

    parties invest resources in the relationship. It

    differs from full vertical integration because

    the arrangement ceases at the end of an agreed

    period of time and the firms remain indepen-

    dent of one another. A joint venture is one

    example of quasi-integration. Participants

    share the costs, risks, profits and losses of the

    venture. Franchises and licences are other

    examples of quasi-vertical integration.

    Tapered vertical integration occurs when a

    firm obtains a proportion of its inputs through

    backward integration with a supplier. For

    example, a beef processing firm integrated

    19

    A t ransact ion cost approach to supply chain management

    Jill E. Hobb s

    Supply Chain M anagement

    Volume 1 Number 2 1996 1527

  • 7/27/2019 858117

    6/13

    backwards into beef production could obtain a

    proportion of its beef supplies from its own

    farms with the remainder procured from

    auction markets or direct from beef produc-

    ers. Alternatively, a firm could transfer a

    proportion of output forward through its own

    distribution network with the remainder sold

    on the open market.

    Full vertical integration occurs when one

    firm carries out two or more consecutive

    stages of the production-distribution chain. A

    firm can be integrated forwards (downstream)

    into distribution or retail functions or back-

    wards (upstream) into supply functions.

    According to TCA, one of the determi-

    nants of vertical co-ordination is the nature

    and level of transaction costs, wherein a

    change in the transaction costs arising fromthe exchange of a product may lead to a

    change in the management of that supply

    chain[25]. In this context, the key characteris-

    tics of transactions are:

    the degree of uncertainty surrounding the

    transaction;

    the degree of asset specificity; and

    the frequency of the transactions[7].

    A low level of uncertainty lends itself to spot

    market transactions. When aspects of the

    transaction (such as quality characteristics)are highly uncertain, a more formal type of

    vertical co-ordination where one party has

    more control over the outcome of the transac-

    tion may result, e.g. a strategic alliance, a

    contract or some form of vertical integration.

    Goods which are non-specific in nature, or

    produced with non-specific assets, have many

    alternative uses and would tend to be sold in a

    spot market. As asset specificity increases, we

    move along the spectrum of vertical co-ordi-

    nation towards a more formal type of supplychain management such as vertical integra-

    tion. Whether the result is a strategic alliance,

    a long-term contract or full vertical integra-

    tion may depend on whether one party, or

    both, make an asset specific investment[26].

    When transactions are carried out fre-

    quently, both buyer and seller will probably

    value repeat business and will not wish to

    tarnish their reputations by acting opportunis-

    tically. Frequent transactions also provide

    buyers and sellers with information about one

    another. For these reasons, transactions

    repeated frequently tend to be carried out in

    the spot market. As transactions become more

    infrequent, however, the incentive to act

    opportunistically and to exploit any informa-

    tional asymmetries that may be present

    increases and we move further along the

    continuum of vertical co-ordination towards

    the extreme of vertical integration.

    The economic theory underlying TCA

    provides considerable insights for supply

    chain management. A large number of

    testable hypotheses regarding supply chain

    management can be devised using the theoret-

    ical framework of TCA. Testable hypotheses,

    however, require information about transac-

    tion costs. L imited information has inhibited

    the use of T CA in the study of supply chains.

    M et hodo logical issues in measuring

    transaction costsOne of the problems with TCA is that suc-

    cessful measurement of transaction costs has

    not kept pace with theoretical developments.

    This is perhaps not surprising, since, unlike

    production costs, transaction costs the costs

    of economic organization are neither easy to

    separate from other managerial costs nor

    readily measurable. (T he problem is analo-

    gous to the difficulty accountants have in

    assigning costs of jointly used assets to indi-

    vidual enterprises in a multiple enterprisefirm.) T he complex nature of companies and

    market institutions means that the costs of

    their operation are not easy to quantify and

    the data which one might use to measure

    transaction costs are not usually collected by

    governments or by the standard accountancy

    practices of firms. Although one can recognize

    that there are indeed costs involved in valuing

    a good or in monitoring the actions of a buyer

    or seller, it is difficult to measure these costs in

    financial terms. Until, and indeed if, the datanecessary for this type of estimation are ever

    recorded by firms, an accounting approach to

    empiricizing the transaction cost approach is

    impractical.

    Economists have turned instead to other

    ways of measuring transaction costs. These

    applications can be divided into three broad

    types, each representing different methodolo-

    gies with different data requirements and each

    varying in the type of information which they

    provide about transactions:

    (1) those that evaluate the effect of transac-

    tion costs on vertical co-ordination (pri-

    marily vertical integration) across indus-

    tries using secondary data sources;

    20

    A t ransact ion cost approach to supply chain m anagement

    Jill E. Hobbs

    Supply Chain M anagement

    Volume 1 Number 2 1996 1527

  • 7/27/2019 858117

    7/13

    (2) industry-specific investigations of the

    impact of transaction costs on vertical co-

    ordination using secondary data; and

    (3) industry-specific investigations of the

    impact of transaction costs on vertical co-

    ordination using primary data.

    The discussion that follows is intended to be

    indicative of the types of analyses carried out

    rather than a comprehensive review of all

    applications of TCA. (M ahoney[16] discusses

    several examples of empirical research into the

    vertical integration decision and summarizes

    the methodologies used to measure vertical

    integration, uncertainty and asset specificity.)

    M ulti- industry evaluations of vert ical

    co-ordina tion using secondary dat a

    In an assessment of the transaction cost

    approach to vertical integration, L evy[27]

    presents a multi-industry study. Using a

    sample of 69 firms from 37 different industries

    over a four-year period, the author hypothe-

    sizes that vertical integration, as measured by

    the ratio of value-added to sales, is determined

    by the costs of transacting through markets as

    well as the costs of management. Levy is

    forced to use proxy measures of transaction

    costs; for example, a variable measuring theintensity of research and development expen-

    ditures is used to proxy asset specificity. T he

    minimum efficient plant scale relative to

    industry sales, a four-firm concentration ratio

    and the degree of diversification of the firm

    are other examples of variables used to explain

    vertical integration. M any of the variables

    used by L evy only indirectly measure transac-

    tion costs. Although the results indicate a high

    degree of explanatory power for the model,

    because many of the explanatory variables aremeasures of the degree of imperfect competi-

    tion rather than measures of transaction costs,

    the only conclusion that can be drawn is that

    transaction costs are one of a number of fac-

    tors that can lead to vertical integration.

    An empirical analysis of the effect of

    transaction costs on vertical co-ordination in

    US food industries is provided by Frank and

    Henderson[28]. (T here is a growing litera-

    ture discussing the concepts of vertical co-

    ordination and transaction costs in relation

    to agriculture and food markets see [24,29-

    31] however, there have been few attempts

    to empiricize these insights.) Working at a

    fairly disaggregated level across several food

    industries, they develop a vertical co-ordi-

    nation index to measure the extent of verti-

    cal co-ordination in an industry. T he index

    has two parts. F irst, it includes an input-

    output transactions matrix to calculate inter-

    dependencies between firms in an industry.

    Second, it includes a measure of the degree

    of administrative control over transactions.

    This is measured by the percentages of farm

    commodities procured through spot mar-

    kets, through market specification, produc-

    tion management and resource providing

    contracts as defined by M ighell and

    Jones[23] or through vertical integration.

    The resulting vertical co-ordination index

    ranges from zero (spot markets) to one (ver-

    tical integration).

    The index is regressed against proxy mea-surements of transaction costs divided into

    four categories: uncertainty, industry concen-

    tration, asset specificity and the costs of

    administering vertical co-ordination:

    (1) As uncertainty rises we expect vertical

    integration to increase because it is no

    longer possible to specify fully all contin-

    gencies in a contract. T he uncertainty

    surrounding food manufacturers input

    supply is measured by the percentage

    change in farm output supply between1981 and 1982.

    (2) Increasing concentration leads to a small

    numbers bargaining problem. A four

    firm concentration ratio measures concen-

    tration in food manufacturing industries.

    (3) As asset specificity rises, an increase in

    non-market vertical co-ordination is

    expected. Variables used to measure this

    include an advertising to sales ratio and a

    research and development expenditure to

    sales ratio for food industries.

    (4) As the costs of administering within-firm

    vertical co-ordination rise, a reduction in

    vertical integration is expected. The

    authors identify a number of characteris-

    tics of firms which affect the internaliza-

    tion of transactions (e.g. capital intensity)

    and use proxy measures of these charac-

    teristics (e.g. a capital to sales ratio) to

    measure the costs of administering verti-

    cal co-ordination.

    Although not all of the variables used were

    significant, the authors conclude that transac-

    tion costs are a primary determinant of verti-

    cal co-ordination. Due to the lack of adequate

    data for measuring transaction costs, Frank

    21

    A t ransact ion cost approach to supply chain management

    Jill E. Hobb s

    Supply Chain M anagement

    Volume 1 Number 2 1996 1527

  • 7/27/2019 858117

    8/13

    and Henderson[28] resorted to the use of

    industry characteristics as indirect proxy

    measurements of transaction costs. Conse-

    quently, this approach was not able to mea-

    sure directly the effects of specific transaction

    costs on vertical co-ordination.

    Industry-specific investigat ions oft ransaction cost s using seconda ry dat a

    Lieberman[32] assesses demand variability

    and transaction costs as factors jointly influ-

    encing vertical integration. (When the fluctu-

    ating demand of other buyers affects the price

    of an input, firms with stable input require-

    ments may integrate backwards to avoid

    paying a premium for the input.) Although the

    example used for the analysis is drawn fromthe chemical industry, the authors primary

    concern is to provide insights regarding these

    two causes of vertical integration which have a

    wider applicability. Using regression analysis,

    the dependent variable is a dummy variable

    equal to one if the firm manufactures its pri-

    mary input and zero if it does not. A number

    of proxy measures of transaction costs and

    demand variability are used as explanatory

    variables. For example, the reciprocal of the

    number of upstream firms is used to measure

    supplier concentration (an indicator of the

    small-numbers-bargaining problem outlined

    by Williamson[7]); the total fixed investment

    cost of the firms downstream plant is used to

    proxy asset specificity, together with a dummy

    variable to measure whether the upstream

    input is a gas (since this would necessitate a

    large fixed investment in a pipeline between

    plants). The results indicate that both transac-

    tion costs and demand variability can lead to

    vertical integration.

    Globerman and Schwindts[33] analysis ofthe organization of transactions in the Cana-

    dian forest products industries avoids the

    need to obtain actual measurements of trans-

    action costs. M any transactions along the

    chain from timber production to its various

    end-uses such as construction, paper produc-

    tion and newspaper production involve asset

    specific investments a characteristic of

    transactions which TCA indicates should lead

    to within-firm hierarchical, rather than market

    forms, of organization, i.e. to verticalintegration. Given the practical difficulties of

    obtaining reliable measurements of transac-

    tion costs, the authors adopt instead what

    they call a more pragmatic inductive

    approach[33, p. 201]. They examine the

    organization of the forest products industries

    to identify any anomalies between the

    observed industry structure and the predic-

    tions of TCA. All but one of Canadas 30

    largest forest products enterprises were inte-

    grated backwards into the ownership of

    timber rights. T imber mills are a highly

    dedicated asset, therefore mill operators are

    vulnerable to post-contractual opportunistic

    behaviour. Backward integration by mill

    owners (the purchase of licences or of the

    rights to log) is consistent with the TCA

    model. Similar inductive reasoning suggested

    that the vertical co-ordination arrangements

    between pulp/paper making activities and

    newsprint manufacturers were consistent with

    the predictions of T CA. T he authors con-clude that the major vertical linkages within

    the industry comply with the predictions of

    TCA and that transactional considerations, in

    particular asset specificity, are important

    determinants of vertical co-ordination.

    Industry-specific investiga tion s oftra nsaction costs using prim ary d at a

    Evaluating the determinants of vertical co-

    ordination at a general, multi-industry level isextremely difficult due to data limitations.

    The available published data sources often

    require the researcher to construct indirect

    and potentially confusing proxy measures of

    transaction costs. For this reason, most empir-

    ical work regarding transaction costs has been

    carried out on an individual industry level or

    at an even narrower level on a case study

    basis. However, this approach also faces data

    limitations. In addition, it has been criticized

    on the grounds that the results are not neces-

    sarily representative of the wider economicenvironment[27,34]. However, for the devel-

    opment of a credible empirical alternative to

    the measurement of transaction costs, the

    necessary transaction cost information would

    have to be collected on a routine basis, which

    it clearly is not. As Coase has lamented:just as important at the present stage would

    be the gathering in a systematic way of new data

    on the organization of industry so that we can

    be better aware of what it is that we must

    explain[35, p. 69).

    Given the problems of using secondary data to

    measure transaction costs, a more appropriate

    approach would be to use data which specifi-

    cally measure different transaction costs at the

    22

    A t ransact ion cost approach to supply chain m anagement

    Jill E. Hobbs

    Supply Chain M anagement

    Volume 1 Number 2 1996 1527

  • 7/27/2019 858117

    9/13

    level of the individual firm. Unfortunately, the

    accounting and management practices of

    most firms do not routinely collect this type of

    information. Thus, if researchers wish to

    apply TCA without recourse to general indus-

    try-level data to approximate transaction

    costs, it is likely that they will have to collect

    the data themselves through surveys of indus-

    try participants. C learly the researcher faces a

    trade-off: the use of surveys to collect primary

    data about transaction costs would be time

    consuming and resource intensive but should

    allow a more accurate measurement of these

    costs than the use of secondary data.

    Using a case-study approach, H allwood[36]

    presents a detailed study of the organization of

    offshore oil production. One of the objectives

    of this study was to determine whether theorganizational features of the industry are

    consistent with the predictions of the transac-

    tion cost paradigm. Data were collected

    through a series of interviews with oil compa-

    nies and offshore oil supply firms. Generally

    speaking, oil companies are not integrated

    backwards into input supply firms; instead, an

    invited tender bid auction system is used by oil

    companies to purchase most supply equip-

    ment. Under this system, oil companies invite a

    number of supply firms to submit sealed bids

    for the provision of an intermediate good or

    service; the bidder with the lowest price wins

    the supply contract. Using data from the survey

    of oil supply companies, Hallwood shows that

    the nature of transactions in this industry, i.e.

    their frequency, the potential for opportunistic

    behaviour and the extent of information asym-

    metry, affect vertical co-ordination in ways

    which conform to the predictions of TCA.

    The invited tender bid system is the preferred

    method of vertical co-ordination for many

    intermediate inputs because of lower transac-tion costs, which include lower information

    and monitoring costs. T he multinational

    nature of the offshore oil supply industry is

    also explained as an attempt by firms to mini-

    mize transaction costs, particularly to guard

    against the risks arising from a high level of

    asset specificity.

    An alternative to a descriptive case-study

    approach is to survey industry participants to

    identify individual transaction costs and then

    measure the effect of these costs on the organi-zation of supply chains. The first step would

    be to identify the industry participants and the

    vertical co-ordination relationships which

    exist. Are there separate firms carrying out

    separate functions along the supply chain, or

    are some stages conducted within a vertically

    integrated firm? Do suppliers/buyers along the

    chain transact through the use of open spot

    markets, contracts, strategic alliances, joint

    ventures, etc. or through a combination of

    these? What are the characteristics of these

    vertical co-ordination arrangements, for

    example, do spot markets take the form of

    auctions, electronic auctions or physical

    marketplaces? Do government, quasi-govern-

    ment institutions or marketing boards, etc.

    play a role? A thorough profile of the industry

    structure and the vertical co-ordination link-

    ages between industry participants must be

    developed before any analysis can be under-

    taken.

    Once an industry profile has been devel-oped, the potential transaction costs which

    industry participants may face can be identi-

    fied. T his process is often easier if the transac-

    tion costs are separated into information,

    negotiation and monitoring costs. A method-

    ology for analysing the transaction costs must

    then be chosen. Two such methodologies are

    proposed here. The first is the more resource

    intensive of the two but yields more detailed

    information regarding transaction costs. An

    in-depth survey of industry participants at one

    or more levels (e.g. producers, processors

    and/or retailers) can be carried out to identify

    and measure the important transaction costs

    facing those firms. T he challenge is to design

    survey questions which accurately describe

    the transaction costs, convey this information

    clearly to survey respondents and obtain

    responses which can be meaningfully

    analysed.

    This approach was used recently in an

    analysis of the marketing channel alternatives

    (live-ring auction sales, electronic auctionsales, direct sales to processor or sales through

    a group marketing scheme) available to Scot-

    tish farmers selling finished cattle[37]. For

    each of the marketing channel options, a list of

    possible information, negotiation and moni-

    toring costs was compiled. Examples of infor-

    mation costs incurred when selling cattle

    through a live-ring auction were the costs of

    obtaining information on likely auction prices

    and the price uncertainty associated in selling

    through an auction. Possible negotiation costsincluded the time and cost of transporting

    cattle to the auction and attending an auction

    sale. When selling cattle deadweight (direct to

    an abattoir) a potential monitoring cost would

    23

    A t ransact ion cost approach to supply chain management

    Jill E. Hobb s

    Supply Chain M anagement

    Volume 1 Number 2 1996 1527

  • 7/27/2019 858117

    10/13

    be the grade uncertainty which farmers face as

    they cannot know how their cattle will grade

    before they are slaughtered. Also, there is a

    degree of grade information asymmetry as

    farmers may not trust the deadweight grad-

    ing system. T hese, and the other transaction

    costs identified, will influence the farmers

    choice of marketing channel[38].

    What type of information should be collect-

    ed? How should this information be analysed?

    Clearly, these questions are related. While it is

    possible to obtain monetary valuations of

    some transaction costs, e.g. transportation

    costs to/from an auction mart, survey respon-

    dents would find it difficult, if not impossible,

    to provide monetary estimates of other trans-

    action costs. Sometimes the analyst can imply

    a monetary value, for example, a respondentcan provide an estimate of the time taken to

    attend an auction sale and this can be assigned

    a monetary value using a shadow wage for the

    opportunity cost of the farmers time. Unfor-

    tunately, monetary values cannot be assigned

    to all transaction costs. The monitoring costs

    involved in grade information asymmetry

    would be extremely difficult to value in mone-

    tary terms without detailed (and accurate)

    information on the probability of an incorrect

    grade being applied to a carcase and an esti-

    mate of the average loss incurred by the

    farmer for an incorrect grade evaluation.

    Unless this type of information is routinely

    collected by those carrying out the grading,

    the data requirements become unrealistic.

    Instead, transaction cost data can be used

    in an econometric analysis of the choice of

    marketing channel, where the vertical co-

    ordination outcome is the dependent variable

    and the dependent variables are transaction

    costs. H obbs[37] collected data from a survey

    of 110 farmers in NE Scotland, includinginformation on the proportion of a farmers

    cattle sold through each marketing channel

    over an 18-month period. T his was the inde-

    pendent variable in the econometric analysis.

    The dependent variables were various infor-

    mation, negotiation and monitoring costs

    arising from the use of each marketing chan-

    nel, together with a number of socio-econom-

    ic and farm characteristic variables. Although

    it was possible to obtain direct approximations

    for some transaction costs (e.g. distance to theauction mart or abattoir as a measurement of

    transportation costs, percentage commission,

    etc.) most were categorical variables. Respon-

    dents were asked to rate whether different

    transaction cost aspects of using a marketing

    channel were a problem. A scale of 1 (not a

    problem) to 5 (major problem) was used. For

    example, farmers were asked whether it was a

    problem that cattle may not grade as expected

    when selling deadweight (a measure of grade

    uncertainty); they were asked whether not

    being present at the abattoir when cattle were

    graded was a problem (a measure of grade

    information asymmetry). T he time spent

    attending an auction sale or obtaining price

    information about auction and deadweight

    markets was measured in hours. Data were

    collected on 24 transaction cost variables.

    This enabled a regression of the proportion of

    cattle sold through live-ring auctions (versus

    direct to abattoirs) against the transaction cost

    and farm/farmer characteristic variables. Fivetransaction cost variables and three farm

    characteristic variables were found to be

    significant in determining the farmers choice

    of marketing channel. (For further informa-

    tion regarding these results or for copies of the

    questionnaire, see Hobbs[37].)

    The advantage of this methodology is that a

    comprehensive analysis of a large number of

    potential transaction costs is possible. The

    important transaction costs can be identified

    and the extent to which they influence vertical

    co-ordination is indicated by the size of the

    regression coefficient. The level of detail

    required in the questionnaire means that

    personal interview-style surveys may be the

    only practical way of collecting these data.

    This is relatively resource intensive. The

    success of this method depends crucially on

    the careful design of questions to elicit the

    appropriate information from respondents.

    A second analytical method which uses

    survey data provides less detailed information

    but is potentially less resource intensive. Thesupply channel (or marketing channel) options

    available to a firm can be described in terms of

    their key transaction cost attributes. Again, this

    requires the analyst to have developed, through

    consultation with industry representatives, a

    detailed profile of the industry structure and its

    vertical linkages. Having identified the key

    transaction cost attributes of the supply (mar-

    keting) channel, a series of hypothetical chan-

    nel scenarios can be developed using different

    combinations of these transaction cost attribut-es. Survey respondents are asked to rank or rate

    the channel scenarios in terms of preference.

    A regression analysis of the preference scores

    against the scenario attributes indicates the

    24

    A t ransact ion cost approach to supply chain m anagement

    Jill E. Hobbs

    Supply Chain M anagement

    Volume 1 Number 2 1996 1527

  • 7/27/2019 858117

    11/13

    extent to which respondents are willing to

    trade-off levels of one attribute for another and

    provides a measure of the relative importance

    of each transaction cost. Results are available at

    both an individual respondent level and for the

    whole sample. T his approach is known as

    conjoint analysis. (For more information about

    the use of conjoint analysis, see [39-41].)

    Hobbs[37] used conjoint analysis to

    analyse the importance of key transaction

    costs in determining the preferences of UK

    meat processors for different beef supply

    channels. Pre-survey interviews with industry

    representatives enabled a list of possible trans-

    action cost attributes to be reduced to four.

    These were: continuity of supply (long-term

    regular suppliers reduce a processors infor-

    mation and monitoring costs); the transporta-tion of cattle to the abattoir (direct transporta-

    tion from farm to processor reduces the

    buyers monitoring costs, imposes less stress

    on the animals and gives the processor greater

    control over the timing of delivery); basis of

    payment (if the processor pays on a per kilo

    liveweight basis, it incurs the uncertainty that

    cattle may not grade as expected, if payment is

    on a deadweight carcase grade basis, the grade

    uncertainty remains with the farmer); trace-

    ability of cattle (easy traceability lowers moni-

    toring costs for buyers). Scenarios were devel-

    oped using different combinations of these

    attribute levels and abattoirs were asked to

    rate each scenario on a preference scale.

    (Further details of the methodology, question-

    naire and results can be found in [37].)

    Similarly, conjoint analysis has been used

    to assess the importance of key transaction

    costs in determining the preferences of UK

    supermarket beef buyers[37]. Three key

    transaction costs were identified and com-

    bined in scenarios along with an attributerepresenting price. T hese attributes were

    chosen through pre-survey interviews with

    industry representatives.

    The results of both conjoint analyses pro-

    vided information on the relative importance

    of the transaction costs. Different supply

    channels exhibit these transaction costs to a

    greater or lesser degree. Predictions can be

    made regarding the consequences for supply

    chain management of changes in the market-

    ing or policy environment which alter thesetransaction costs.

    Conjoint analysis requires that a small

    number of transaction cost attributes be used.

    A large number of attributes would make the

    survey respondents task of evaluating the

    scenarios impossible. I t does not, therefore,

    generate as much information about individ-

    ual transaction costs as the econometric

    methodology described earlier. However, it

    enables an evaluation of the extent to which

    industry participants will trade-off transaction

    costs. Furthermore, it may be less resource

    intensive because the surveys can be carried

    out by mail or by telephone.

    Neither of the econometric or conjoint

    methodologies discussed above provides a

    monetary or percentage measurement of the

    size of individual transaction costs. Perhaps

    this is not necessary. Perceiving transaction

    costs as a measurable cost in the accounting

    sense of the word, or as a margin analogous

    to a marketing margin may be misguided. T henature of transaction costs is more subtle than

    either of these two interpretations allow.

    Instead, methodologies which identify the

    significant transaction costs and measure their

    importance (whether in absolute or relative

    terms) may suffice. Certainly, this improves

    our understanding of supply chain manage-

    ment and can indicate which transaction costs

    should be reduced in order to enhance the

    efficiency of vertical co-ordination within

    supply chains.

    There are many potential pitfalls when

    using primary data to measure transaction

    costs. The subtle nature of these costs means

    that great care must be taken in the design of

    questions. Some industry sectors are highly

    concentrated; this leads to additional prob-

    lems. First, the confidentiality of survey

    respondents must be protected; firms may be

    unwilling to provide information which they

    regard to be of a commercially sensitive

    nature. Any reluctance to take part in surveys

    could seriously damage the statistical reliabili-ty of the results, particularly if the industry

    sector comprises a small number of firms.

    Second, surveys of highly concentrated indus-

    tries are necessarily based on the responses of

    a very small number of firms, even if all firms

    participated. T his is the case for analyses of

    the UK supermarket retailing sector, for

    example. T hese problems notwithstanding,

    until the data which measure information,

    negotiation and monitoring costs are routinely

    collected by firms or by government agencies,analysts wishing to apply T CA to supply chain

    management will probably have to collect

    their own data through the use of industry

    surveys.

    25

    A t ransact ion cost approach to supply chain management

    Jill E. Hobb s

    Supply Chain M anagement

    Volume 1 Number 2 1996 1527

  • 7/27/2019 858117

    12/13

    Conclusions

    In order to further our understanding of the

    intricacies of supply chain management we

    need frameworks in which to test theories

    about supply chain management. Drawing

    from the new institutional economics litera-ture, transaction cost analysis may be one such

    framework. I t provides an explanation for the

    existence and structure of firms and for the

    nature of vertical co-ordination within a

    supply chain. Empirical measurement of the

    effect of transaction costs on vertical co-

    ordination appears to be rare in the main-

    stream economics literature but even rarer in

    the agricultural and food economics literature.

    The insights which TCA provides into the

    nature of vertical co-ordination need to be

    empirically verified through micro-analytical

    studies of supply chains. A convergence of

    interests is thus suggested for joint research.

    Transaction costs, and their reduction, lie

    at the heart of the interest in supply chain

    management. Proactive moves to enhance

    management of supply chains are fundamen-

    tally concerned with improving their efficiency

    to gain competitive advantage. Adversarial

    relationships along the supply chain increase

    transaction costs. Co-operation, teamwork

    and the rapid interchange of data amongcompanies in a supply chain will reduce trans-

    action costs. To gain a better understanding of

    how these new relationships reduce costs,

    those who wish to undertake analyses of

    transaction costs require information infor-

    mation only firms can provide. If firms do not

    collect information which can be used to

    analyse transaction costs, then their co-opera-

    tion in the collection of survey data is essen-

    tial.

    Here lies the challenge facing researchersand business practitioners alike to co-oper-

    ate in carrying out applied research into

    aspects of supply chain management. A few

    successful attempts have been made to empiri-

    cize the insights provided by TCA using data

    collected from individual firms. Requests for

    information from businesses need to be stated

    clearly and be capable of being satisfied by a

    reasonable amount of effort from a firms

    managers.

    The contribution of businesses to research

    efforts relating to supply chain management

    need not involve any proprietary or sensitive

    information. It will, however, involve taking

    time to answer questions, to provide input into

    the research agenda itself and to understand

    the requirements of those conducting the

    research. Researchers, on the other hand,

    must be sensitive to the concerns of those

    providing information relating to their activi-

    ties. In the past, this co-operation was not well

    developed, meaning that empirical analysis

    was confined to what was possible with pub-

    lished information largely within the con-

    straints of the neoclassical paradigm and

    companies were frustrated with the types of

    information available. Transaction cost analy-

    sis appears to provide a means to bridge the

    gap between those interested in the results of

    research and those who can provide the

    research.

    Notes and references

    1 Cheung, S.N.S., On the new inst i tut ional economics ,

    in W erin, L. and W ijkan der, H. (Eds), Contract Econom -

    ics, Blackwell, Oxford, 1992, pp. 48-71.

    2 Coase, R.H. , The natu re of the fi rm , Economica,

    Vol. 4,1937, pp. 386-405.

    3 Eggertsson, T. , Econom ic Behaviour and Institu tions,

    Cambridge Surveys of Economic Literature, Cambridge

    University Press, Cambridge,1990.

    4 North, D.C., Inst i tut ions, t ransact ion costs and

    economic growth , Economi c Inquiry, Vol. 25 No. 3,

    1987, pp. 419-28.5 Arrow, K.J. , The organizat ion of economic act iv i ty:

    issues pertinent to the choice of m arket versus non-

    market allocation , in Haveman, R.H. and M arglois, J.

    (Eds), Public Expenditure an d Policy Analysis,

    M arkham Publishing, Chicago, IL, 1970, pp. 59-71 .

    6 Wi l l iamson, O.E. , M arkets and Hierarchies: Analysis

    and Ant i-Trust Implications: A Study in th e Economics

    of Organization, Free Press, New York, N Y, 197 5.

    7 Wil l iamson, O.E., Transact ion cost economics: the

    governance of contractual relations , Journ al of Law

    and Economics, Vol. 22, 1979, pp. 233-62.

    8 Alchian, A.A. , Some economics of proper ty r ights ,

    Il Politico, Vol. 30 No. 4,1965, pp. 816-29.

    9 Demsetz, H. , Toward a theory of proper ty rights ,

    American Economic Review, Vol. 57 No. 2, 1967,

    pp. 347-59.

    10 Alchian, A.A. and Demsetz, H. , Product ion, in forma-

    tion costs, and economic organization , American

    Econo mic Review, Vol. 62, 1972, pp. 777-95.

    11 Jensen, M .C. and Meckl ing, W.H., Theory of the firm:

    managerial behavior, agency costs and ow nership

    structure , Journal of Financial Economics, Vol. 3

    No. 4, 1976, pp. 305-60.

    1 2 Ca sso n, M . , Enterprise and Competitiveness: A

    Systems View of Int ernational Business, ClarendonPress, Oxford , 1990 .

    1 3 Ca sso n, M . , The Econom ics of Business Culture: Gam e

    Theory, Transaction Costs and Econom ic Perform ance,

    Clarendon Press, Oxford, 199 1.

    26

    A t ransact ion cost approach to supply chain m anagement

    Jill E. Hobbs

    Supply Chain M anagement

    Volume 1 Number 2 1996 1527

  • 7/27/2019 858117

    13/13

    14 Teece, D.J., Transaction cost economics and the

    mult inational enterprise: an assessment , Journal of

    Economi c Behavior and Organization, Vol. 7, 1986,

    pp. 21-45.

    15 North, D.C. and Thomas, R.P., The Rise of the W estern

    World: A New Econom ic History, Cambridge University

    Press, Cambri dge, 1973 .16 M ahoney, J.T. , The choice of organizat ional form:

    vertical financial ow nership versus other meth ods of

    ver t ica l in tegrat ion , Strategic M anagement Journal,

    Vol. 13, 1992, pp. 559-84.

    1 7 Si m on , H ., Adm inistrative Behavior, 2nd ed., M acmil-

    lan, New York, NY, 196 1.

    18 Douma, S. and Schreuder, H. , Economic A pproaches to

    Organizations, Prentice Hall International (UK), Hemel

    Hempstead, 1992.

    19 Klein, B. , Crawford, R.G. and Alchian, A.A. , Ver t ical

    integration , appropriable rents and the competit i ve

    contracting process , Journal of Law and Economics,Vol. 21 No. 2, 1978, pp. 297-326.

    20 St ig ler, G.J. , The economics of in format ion , Journal

    of Polit ical Economy, Vol. 69, June 1961, pp. 213-15.

    21 Akerlof , G.A. , The market for lemons : qual itat ive

    uncer ta inty and t he market mechanism , Quarterly

    Journal of Economics, Vol. 84, 1970, pp. 488-500.

    2 2 Ar ro w , K. J. , Informat ion and Econom ic Behavior,

    Federation of Sw edish Industries, Stockholm, 1 973.

    23 M ighel l, R.L. and Jones, L.A. , Vertical Coordinatio n in

    Agriculture, USDA ERS-19, Wa shingt on DC, 196 3.

    24 Spor leder, T.L. , M anageria l economics of ver t ica l ly

    coordinated agr icultura l fi rms , Am erican Journal ofAgricultural Economics, Vol. 74 No. 5, December 1992,

    pp. 1226-31.

    25 Note that t ransact ion costs are only one of a number

    of pot ential determinant s of vertical co-ordination . For

    example, the existence of econom ies and disec-

    onom ies of scale influence wh ether it is econom ically

    efficient for a fi rm to be vertically integrated. Also, the

    capital investment requirements and addit ional r isks

    involved in vertical integration could out w eigh any

    transaction cost advantages of integration. Some-

    tim es, firm s vertically integrate to reduce their taxa-

    tion bu rdens or to circumvent restr ictive import

    regulations. Although these other infl uences areimpo rtant, this paper deals w ith transaction cost

    mot ivations for vertical co-ordination.

    26 I f only one par ty (A) has made an asset specific

    investment, it is probably party A that w il l vertically

    integrate w ith th e other party (B) because of the r isk of

    B practising PCOB. Wh ere both parties have made an

    asset specific investment, either vertical int egration, a

    long-term contract or a strategic all iance may result.

    The outcome w ill depend on other features of the

    transaction, such as the level of uncertainty.

    27 Levy, D.T., The transactions cost approach to vertical

    integrat ion: an empir ical examinat ion , Review of

    Economi cs and Statist ics, Vol. 67, 1985, pp. 438-45.

    28 Frank, S.D. and Henderson, D., Transaction costs as

    determinant s of vertical coordinatio n in US food

    industr ies , American Journal of Agricultural Econom-

    ics, Vol. 74 No. 4, 1992, pp. 941-50.

    29 Barkema, A.D., New roles and al liances in the U.S.

    food system , presented at the spring meeting of the

    Federal Reserve System Committee on Agriculture andRural Development , Kansas City, M O, 20 M ay 1993.

    30 Sonka, S.T. , New Industries in Agriculture: Concepts,

    Issues and Oppo rtunit ies, mimeo, Department of

    Agricultural Econom ics, University of I l l inois at

    Urbana-Champaign, IL, 1992.

    31 M asten, S.E. , Transact ion-cost economics and the

    organization of agricultural transactions , presented

    at th e NC-194 Wo rld Food Systems Project Sympo-

    sium, Examin ing t he Econo mic Theory Base for Vertical

    Coordination, Chicago, IL, 17-18 October 19 91.

    32 Lieberman, M., Determinants of ver t ical in tegrat ion:

    an empir ical test , The Journal of Industrial

    Economics, Vol. 39 No. 5, 1991, p. 451-66.33 Globerman, S. and Schwindt , R., The organizat ion of

    vertically related transactions in t he Canadian forest

    products industries , Journal of Economic Behavior

    and Organization, Vol. 7, 1986, pp. 199-212.

    34 M atthews, R.C.O., The economics of inst i tut ions and

    the sources of growth , The Econom ic Journ al, Vol. 96

    No. 384, December 1986, pp. 903-18.

    35 Coase, R.H., Industr ia l organizat ion: a proposal for

    research , in Fuchs, V.R. (Ed.), Policy Issues and

    Research Opportun it ies in Industrial Organization,

    Nat ional Bu reau of Econom ic Research, New York, NY,

    1972, pp. 59-73.

    3 6 Ha ll w oo d, Transaction Costs and Trade between

    M ultinat ional Corporations: A Study of Offshore Oil

    Production, Unwin Hyman, Boston, M A, 1990.

    37 Hobbs, J.E., A t ransaction cost analysis of beef

    marketing in t he UK , unpublished PhD thesis,

    Department of Agriculture, University of A berdeen,

    Scotland, 199 5.

    38 Clear ly, other non- transact ion cost factors may

    infl uence a farmers choice of marketing channel.

    How ever, the purpose was to identify im portant

    transaction costs. Also, the transaction costs incurred

    by processors in purchasing fin ished beef catt le w il l

    affect their choice of supply channel, thereforeinfl uencing vertical co-ordination in t he supply chain.

    The same survey metho d could be used to id entify the

    import ant transaction costs facing p rocessors.

    39 Hair, J.F. Jr, And erson, R.E., Tatha m, R.L. and

    Black, W.C., M ultivariate Data Analysis w ith Readings,

    3rd ed., M acmillan, New York, NY, 1992.

    40 Steenkamp, J.B.E.M., Conjo int measurement in ham

    qual i ty evaluat ion , Journal of Agricultural Economics,

    Vol. 38 No. 3, 1987, pp. 473-80.

    41 Wit t ink, D.R., Vr iens, M. and Burhenne, W., Commer-

    cial uses of conj oint analysis in Europ e: result s and

    crit ical reflections , International Journal of Research

    in Market ing, Vol. 11 No. 1, 1994, pp. 41-52.

    27

    A t ransact ion cost approach to supply chain management

    Jill E. Hobb s

    Supply Chain M anagement

    Volume 1 Number 2 1996 1527