Journal of Retailing 86 (3, 2010) 257–269

Transaction Cost Theory and International Business

Jean-François Hennart
CentER and Department of Organization and Strategy, Tilburg University, P.O. Box 90153, 5000LE Tilburg, The Netherlands

Abstract
I discuss the cross-fertilization between transaction cost theory (TCT) and international business (IB), showing how TCT provides a powerful
lens to study the institutions that organize international interdependencies, and especially multinational enterprises (MNEs). I then discuss some
of the insights that IB can provide for the further development of TCT. I argue that a full explanation of why MNEs exist must rely on information
asymmetry as well as asset specificity, and that the study of modes of foreign market entry leads to alternative viewpoints on equity joint ventures
and hybrids. I conclude by stressing the need for a simultaneous consideration of market transaction costs and internal organization costs when
examining governance choices.
© 2010 New York University. Published by Elsevier Inc. All rights reserved.

Keywords: Transaction cost theory; International business; Multinational enterprises; Global value chains; Hybrids; Equity joint ventures

It is fair to say that neoclassical economics proceeds on
the assumption that the ‘organization problem’, how individ-
uals find ways to cooperate so as to benefit from pooling their
joint efforts and from exploiting their differences, has already
been solved. Oliver Williamson’s major achievement has been
to transform this assumption into a research question by ask-
ing what mechanisms are and should be used by individuals to
solve the ‘organization problem’, that is to communicate their
needs and capabilities, to reach agreement on the terms of trade,
and to make sure their bargains are respected ex-post. By asking
the question and advancing testable hypotheses, Williamson’s
writings have spawned exciting research in a wide range of dis-
ciplines, from sociology to finance. It is only fitting that this
tremendous achievement be crowned with a Nobel Memorial
Prize.
As the next pages show, the assumption of zero transaction
costs that is implicit in neoclassical economics is particularly
constraining in international business (IB). IB scholars study the
governance of interdependencies between individuals located in
different countries, and hence separated by geographic, institu-
tional, and cultural distance. Organizing such interdependencies
is difficult given the lack of information caused by physical sep-
aration, communication and cultural difficulties which make
reaching agreement problematic, and uncertain enforcement
given that transactors are subject to different political jurisdic-
tions.
E-mail address: j.f.hennart@uvt.nl.
In this piece I discuss the cross-fertilization between transac-
tion cost theory (TCT) and IB. The article has two parts. In the
first I sketch how TCT has provided a powerful lens to look at
the IB core topic, the institutional arrangements used to organize
interdependencies between agents located in different countries,
and in particular the multinational enterprise (MNE). In the sec-
ond part I discuss some of the insights that IB might provide for
the further development of TCT.
The first section of the first part presents a brief history of
how TCT gained influence among IB scholars because it could
provide a persuasive explanation for the existence and the scope
of MNEs, which are central actors in IB. In the next section
I show how TCT is able to provide a unifying paradigm that
accounts for the many diverse forms taken by MNE activities.
The second part follows up on a few of the research oppor-
tunities that are suggested in the first part. In the first section
of the second part I argue that a persuasive account of the role
played by MNEs in the international economy cannot only rely
on asset specificity but often requires recourse to information
asymmetry and imperfect measurement arguments. The second
section looks at equity joint ventures and various contractual
arrangements which have played an important part in the coor-
dination of international interdependencies. While these two
forms have often been lumped together and labeled ‘hybrids’,
I show that they are very different. Equity joint ventures are
joint hierarchy, not hybrids of market and hierarchy. Hybrids,
on the other hand, are institutions that can be seen as simultane-
ously using the two basic organizing methods of output control
and behavior control. In the last section I stress the importance

0022-4359/$ – see front matter © 2010 New York University. Published by Elsevier Inc. All rights reserved.
doi:10.1016/j.jretai.2010.07.009
258 J.-F. Hennart / Journal of Retailing 86 (3, 2010) 257–269
of simultaneous considering market transaction costs and inter-
nal organization costs when explaining international governance
choices.
The contribution of transaction cost theory to international
business: the genesis
The institutional arrangements used to organize the interna-
tional interdependencies which IB scholars study can be purely
market-based, for example arm’s length exports using indepen-
dent distributors. They can use hierarchical processes, when
exchange takes place within MNEs, firms that coordinate activ-
ities across borders through employment contracts. Or they can
be hybrids of both market and hierarchy, and take the form of
contracts, as happens in licensing, franchising and in the orga-
nization of many vertical value chains.
MNEs are a relatively recent phenomenon. In the 1850s
and 1860s firms such as Singer Sewing Machines and Siemens
started to establish foreign factories. The trend picked up con-
siderable speed in the 1880s, slowed down during the Great
Depression, but then started again after the end of WW2, led
initially by US MNEs, and then by those of Europe and Asia
(Jones 2005).
The existence and growth of MNEs was left unexplained
until the late 1970s because it fit very awkwardly with extant
economic theories. International trade theories assumed perfect
mobility of products but perfect immobility of factors, yet the
international expansion of firms involved the international trans-
fer of factors of production such as knowledge and managerial
talent. More fundamentally, the economist’s world was one of
zero transaction costs. In such a world, MNEs are a paradox.
It is generally acknowledged that doing business in a foreign
country involves additional costs over those incurred at home
(Hymer 1976). In the absence of market transaction costs, the
costs of transferring factors of production such as knowledge
and reputation to a foreign buyer through market processes are
negligible. Then why would a firm ever prefer integrating into
the foreign production of goods and services incorporating its
know-how, with all its attendant difficulties, over licensing or
franchising foreign manufacturers?
If you cannot explain what you see, see what you can explain.
When they looked at MNEs, trade economists of the time
saw long-term capital exports, which they called foreign direct
investment (FDI) and which in their view only differed from
portfolio investments insofar as it carried control. They argued
that FDI, along with portfolio investment, flowed between coun-
tries in response to difference in real interest rates. This view
was not questioned until Stephen Hymer’s 1960 PhD thesis,
published in 1976 (Hymer 1976). Hymer looked at the foreign
activities of US MNEs and asked why, if as alleged by FDI the-
orists they were undertaken in response to higher interest rates
abroad than in the US, they were to a significant degree financed
in the foreign countries where US MNEs did business. He also
noted that US flows of FDI and portfolio investment often moved
in opposite direction. He asked why US FDI was undertaken by
industrial firms and not by banks and financial intermediaries,
and why it was concentrated in a small number of industrial
sectors, such as oil, motor vehicles, business machinery, and
tires. This led him to argue that FDI should be understood as an
attempt by homegrown monopolists to safeguard their monopoly
profits. When a reduction in international trade barriers brought
domestic and foreign monopolists in competition, domestic
monopolists attempted to take over or merge with their foreign
rivals, or to set up production in foreign countries to pre-empt
the emergence of such rivals. Through these strategies, domestic
monopolists expanded across borders and became MNEs. Hence
MNEs arose to reduce competition in final product markets in
industries where large barriers to entry had created and were
sustaining local monopolies. MNEs were institutions devised to
internalize the pecuniary externalities caused by imperfections
in the market for final products. These imperfections resulted
from barriers to entry and government intervention, i.e. they
were what Dunning and Rugman (1985) later called structural
market imperfections. Internalizing pecuniary externalities due
to structural market imperfections is a zero-sum game in which
firms gain at the expense of consumers. Hymer thus concluded
that MNEs reduced competition, and hence that their activities
should be severely restricted by governments. Hymer’s view of
MNEs as exploiting abroad their monopolistic advantages was
popularized by Kindleberger (1969) and further developed by
Caves (1971).
There is a subtle, but important difference between Hymer’s
view of MNEs internalizing pecuniary externalities due to struc-
tural market imperfections in the market for final products
and the transaction cost view that sees firms as internalizing
non-pecuniary externalities due to natural imperfections in the
market for intermediate products (Dunning and Rugman 1985;
Hennart 1982). Natural market imperfections arise because
agents are boundedly rational and opportunistic (Williamson
1975), and this means that exchange always entails posi-
tive transaction costs. When the rents available from market
exchange are lower than those available from organizing the
transaction within a firm, the MNE will expand abroad to inter-
nalize these non-pecuniary externalities (Hennart 1977, 1982).
This is a positive sum game which generates rents to be shared
by producers and consumers.
Sixteen years after Hymer came the 1976 publication of
Buckley and Casson’s The Future of Multinational Enterprise,
and its further development by Rugman in Inside the Multina-
tionals (1981). Buckley and Casson’s book, which was inspired
by Coase (1937) and by the interwar debate on the feasibil-
ity of using prices for central planning (Lange 1936), makes
no reference to Williamson. Buckley and Casson argue that
MNEs are more efficient than markets because they use inter-
nal prices. Using internal prices has particular advantages for
transferring knowledge because it makes it possible to apply
discriminatory pricing across markets, to sidestep government
restrictions through transfer pricing, and to replace failing future
markets. Hence Buckley and Casson see MNEs as internaliz-
ing both pecuniary externalities through price discrimination
and the avoidance of taxes, and non-pecuniary externalities
through the establishment of internal prices. These benefits have
to be offset by the costs of setting up internal markets, which
are a loss of scale economies due to their smaller size and a
 J.-F. Hennart / Journal of Retailing 86 (3, 2010) 257–269
need to do on-the-spot inspections (Buckley and Casson 1976,
pp. 36–45).
The first application of TCT to the MNE was made by Hen-
nart in his 1977 PhD thesis at the University of Maryland, later
published in book form as A Theory of Multinational Enterprise
(Hennart 1982). Hennart was inspired by Williamson (1975),
McManus (1972), and by property rights and agency theories
(e.g. Alchian and Demsetz 1972; Arrow 1974; Cheung 1969).
Buckley and Casson’s, Rugman’s, and Hennart’s books, fur-
ther work by these authors and by others such as Teece (1981),
and the incorporation of this thinking in Dunning’s eclectic
theory, all combined to persuade the IB field that MNEs could be
explained by looking at the comparative efficiency of firms and
markets in conducting international transactions. TCT was also
used by Anderson and Gatignon (1986) (Gatignon and Anderson
1988), Gomes-Casseres (1989), Hennart (1982, 1988a, 1991a,
2009), Hennart and Park (1993), and many others to explain the
modes chosen by firms to enter foreign markets, that is the choice
between fully or partly owned affiliates (i.e. equity joint ven-
tures) and that between greenfield entry and acquisitions. These
approaches, which all take a comparative institutional perspec-
tive but diverge significantly in how they explain why MNEs
exist, were given the generic name of ‘internalization theory’.
By 2009 John Dunning could write that “for much of the last two
decades. . .the theory of internalization. . .has been the dominant
explanation of the existence and growth of the MNE qua MNE”
(Dunning 2009, p. 44).1
Can transaction cost theory explain the existence and scope
of MNEs?
How does one explain the existence of MNEs and the scope
of their activities? This is the question Hennart tackled in his
PhD thesis (1977). He argued that TCT could provide a pow-
erful answer. While he acknowledged that MNEs sometimes
internalized pecuniary externalities and hence were an alterna-
tive to cartels or tacit collusion, he found it more interesting to
focus on how they could internalize non-pecuniary externalities.
This section, based on Hennart (1977, 1982, 2000), attempts to
show that TCT provides a unifying paradigm that can explain
the most common types of MNEs.
A firm may own manufacturing capacity and a distribution
system in its home country, but may be looking for foreign tech-
nology to manufacture products that will complete its product
line. A foreign manufacturer may have already developed such
products and may be willing to license his technology to the local
firm. For this potentially fruitful cooperation to take place, both
parties must be aware of each other and of the potential gains
of cooperating, they must be able to quickly settle on a price
for the technology, and they must be assured that the terms of
the transaction will be enforced. Because of bounded rationality
and opportunism, organizing this potentially lucrative interde-
1 As Dunning note on the same page, there has been lately an attempt by the
resource-based view to provide an alternative explanation for the existence and
development of MNEs.
259
pendence will incur information, bargaining, and enforcement
costs, i.e. transaction costs. There are two basic organizing meth-
ods that can be used to perform that task, the price system and
hierarchy, and each yields different levels of rents for a given
transaction because they use fundamentally different methods.
Consider first the need for enforcement. The price system
does it by rewarding agents for their output at market prices. With
bounded rationality and opportunism, however, measuring out-
put will be both necessary and costly. If measurement is costly,
it may not pay to do it perfectly, and hence it will be possible for
some agents to overprice or undersupply output, a behavior I call
‘cheating’ (Hennart 1982, 1993). Switching to hierarchy reduces
these incentives because under hierarchy agents (employees) are
no longer rewarded on the basis of their output but on the basis
of their behavior, i.e. they are rewarded for obeying the boss’
directives (Hennart 1982, 1993; Masten 1988).2 Hennart (1986)
argues that it is this switch from output-based to behavior-based
rewards, and not the setting up of an internal market in firms
as argued by Buckley and Casson (1976) and Rugman (1981),
that explains why firms can in some cases be more efficient than
markets in enforcing transactions. Replacing output by behavior
constraints reduces the incentives that employees have to cheat
but the uncoupling of output and reward has the unfortunate, but
unavoidable, consequence of lowering their incentives to work,
if work is less desirable than other pursuits. In other words,
employees will exert less effort and initiative than if they were
self-employed, i.e. they will ‘shirk’.
There are two basic ways to constrain behavior. One, which
can be called ‘internal control’, clan (Ouchi 1979), or socializa-
tion (Bartlett and Ghoshal 1989), consists in choosing employees
who have the same goals as the boss, or in persuading individu-
als with different goals to change them and internalize those of
the boss. The second, ‘external control’, consists in constrain-
ing the agent’s behavior, either through direct observation or
through bureaucratic rules and procedures. Because constrain-
ing behavior through either internal or external control is costly
and experiences diminishing returns, it will not generally pay to
fully do it. The cost of constraining behavior will vary with
the characteristics of transactions. External control is easier
when employees are gathered on location and behavior is a
good guide to performance, for example in machine-paced pro-
cesses, but harder when this is not the case, for example when
employees are spread out over space, as in extensive agricul-
ture, or in creative industries. Internal control is easier when
employees share a similar culture and have other social ties
(Ouchi 1979).
The analysis up to now has focused on enforcement costs,
but hierarchy and the price system also differ in the way they
solve the information and bargaining problems. Prices fulfill
that role in markets. When the cost of measuring output in all
2 Note that I use the term ‘hierarchy’ to describe a method of control, not
the managers who implement it in firms. As I argue below ‘firm’ is not the
same thing as ‘hierarchy’, and neither is ‘market’ equivalent to ‘price’. Prices
(output control) and hierarchy (behavior control) are organizing methods, while
markets and firms are institutions. Both institutions make use of both methods
simultaneously, but markets use predominantly prices, and firms hierarchy.
260 J.-F. Hennart / Journal of Retailing 86 (3, 2010) 257–269
of its relevant dimensions is high, prices will give inaccurate
signals, and it will pay to use a centralized system of managerial
directives rather than a decentralized system under which agents
react autonomously to price signals. Employees will be asked
to collect a small part of the needed information which will be
aggregated and synthesized by the boss and sent back to employ-
ees for execution. Because employees do not directly benefit
from the information they pass on, they will have fewer incen-
tives than market participants to collect and transfer accurate
information (Hennart 1993; Williamson 1975). The price sys-
tem solves the bargaining problem by making prices exogenous,
while hierarchy does it through fiat.
An MNE will be the chosen method to internalize non-
pecuniary externalities when organizing the interdependence of
agents located in different countries through hierarchy will result
in higher rents than using prices and when the benefits of orga-
nizing interdependencies net of their costs are positive.3 When
is this likely to be the case? Hennart argues that TCT can predict
which types of international interdependencies are likely to be
internalized by MNEs and that this explains the types of MNEs
observed in practice.
Knowledge and reputation
Internalization scholars have advanced that firms set up man-
ufacturing facilities abroad because this is the most efficient way
to exploit their knowledge (Buckley and Casson 1976; Cantwell
1989; Magee 1977; Rugman 1981; Teece 1981). Following
Arrow (1962), Hennart (1977, 1982) argues that imperfect prop-
erty rights in knowledge can be traced down to a fundamental
problem of information asymmetry, insofar as sellers have better
information than buyers on the content of the technology they
have developed, but cannot divulge it pre-purchase, since by
doing this they would be transferring their know-how to buyers
free of charge. Patent systems are supposed to solve this problem,
because they theoretically make it possible for inventors to make
their knowledge public while keeping full property rights to it.
In practice patents suffer from significant limitations (Hennart
1982; Levin et al. 1987). Tacit knowledge cannot be written into
patents, and hence cannot be patented. Patent enforcement is
costly, and in many countries patent laws are weakly enforced.
Knowledge interdependencies will be organized by licensing
contracts when knowledge can be described in patents, when
patents are easy to defend in court, and when they are strongly
enforced by public authorities (Arora and Fosfuri 2000; Caves,
Crookell, and Killing 1982; Contractor 1984; Davidson and
McFetridge 1984; Davies 1977). In the opposite case, typically
that of management or marketing know-how, high informa-
tion and enforcement costs will preclude contractual transfer.
Transfer within a firm may then be more efficient because the
incentives facing sender and receiver of know-how are now to
collaborate in the transfer, not to cheat each other (Hennart
3 This is an important, but neglected, condition. At a given level of technology,
some interdependencies are just too expensive to organize, whether by firms or
by markets (Hennart 1982).
1982).4 Innovating firms will expand into foreign manufacturing
through the establishment of greenfield plants or the acquisition
of local firms. The same logic explains why many types of tacit
knowledge are difficult to obtain through in-licensing (Davies
1977, 1993). If the needed tacit knowledge is embedded in for-
eign firms, the knowledge-seeking firm will have to access it
through the full or partial takeover of these foreign firms or
through the establishment of greenfield joint ventures with them
(Hennart 1988a, 2009; Mowery, Oxley, and Silverman 1996).
While most IB scholars have focused on firms going abroad
to exploit their knowledge, historians have noticed that this is
not the only motivation for foreign expansion. Fieldhouse (1986,
p. 25), for example, notes that MNEs ventured abroad for many
different reasons and that “motives for FDI [MNEs] were there-
fore infinitely more complex than any unitary theory of the MNE
could possibly comprehend, and had no necessary connection
to the internalization concept”. As the following pages show,
however, TCT can account for the wide variety of forms taken
by MNE expansion.
Reputation earned in one country is often valuable in another.
Just like know-how can be embedded in a patent, reputation can
be embedded into a trademark which can be rented if its unautho-
rized use (counterfeiting) can be controlled and if free-riding by
trademark renters can be curbed. A trademark is a public good to
all those using it, and users can free-ride on the collective reputa-
tion of those using the trademark by selling trademarked output
at less than advertised quality. If quality differences are costly to
detect, and a significant fraction of the customers are non-repeat,
trademark users will have incentives to engage in this form of
cheating because only a small part of the reputation loss due to
cheating will be borne by them. If sophisticated measurement
methods were available that could instantly measure the impact
the output of cheaters has on the reputation capital of all those
sharing the trademark and immediately debit their bank account,
then a simple trademark rental would be possible. Given that this
is beyond present measurement capabilities, the next best thing is
a franchising contract by which franchisors specify some behav-
ioral rules to be followed by trademark renters (franchisees) so
as to limit their ability to reduce quality. If this is difficult to do,
and if the reputation rents at risk are large, trademark owners
will have to integrate into the production of the products using
their trademark. Employees, paid a salary less related to outlet
profits, have fewer incentives to cheat, but higher incentives to
shirk. Everything else constant, MNEs will be chosen over fran-
chising to organize international interdependencies involving
reputation when (a) counterfeiting can be curbed and the poten-
tial rents from reputation are high; (b) it is relatively difficult
to devise and enforce behavioral constraints that when followed
will assure a given level of output quality; and (c) it is rela-
tively easy to control shirking by employees. Hence reputation
is generally exploited in-house in banking, advertising, man-
agement consulting, and high-level legal advice, sectors where
4 This advantage will be lost if the internal transferor and transferee of know-
how are put in separate profit centers and granted bonuses based on their units’
profits, as this may recreate the level of cheating found in markets.
 J.-F. Hennart / Journal of Retailing 86 (3, 2010) 257–269
behavioral rules that ensure output quality are difficult to spec-
ify and distribution is selective, while it is generally exploited
through franchising in fast food, hotels, employment agencies,
car rental, and gasoline distribution, sectors where distribution
is intensive and such behavior rules are easier to contractually
specify (Fladmoe-Lindquist and Jacque 2005; Hennart 1982,
2000).5
Global value chains
Value chains, the successive tasks necessary for bringing a
product or service to market, become international when the
stages are located in different countries. Value chains can be
organized on spot markets, within MNEs, or by various types
of contracts (Gereffi, Humphrey, and Sturgeon 2005). Of the
three, contracts seem to be the most common (Campbell and
Parisotto 1995). Three conditions seem to lead to value chains
being organized within an MNE: (1) their efficient functioning
requires that some or all of the parties make transaction-specific
investments; (2) participants in the value chain can free ride on
each other because the separate impact of their output on the rep-
utation of the product is difficult to assess; (3) timely exchange
of tacit information between value chain participants is crucial
but difficult to negotiate. For expository convenience, I start
with the coordination of the upstream stages of value chains,
that between raw materials, parts, manufacturing, and assem-
bly stages, and then discuss the coordination of downstream
stages, that between manufacturing/assembly and wholesaling
and retailing.
The hierarchical coordination of flows of natural resources
through MNEs has been a persistent feature of the international
economy since the nineteenth century which saw the emergence
of vertically integrated MNEs in oil, aluminum, bananas, and
rubber. The efficient processing of some minerals requires facil-
ities which are optimized for a particular ore source (Hennart
1988b; Stuckey 1983). Once built, the facilities are thus depen-
dent on a single ore seller. Spot ore purchases would then expose
processors to sellers jacking up the price ex-post if dependence
is not bilateral (they would risk being held up). If this is the case,
processors will ask ore sellers to promise not to make unilateral
changes in the terms of trade. This, however, requires being able
to anticipate all potential contingencies, a difficult task when
facilities are long-lived, and there is considerable non-indexable
uncertainty. Then contracts provide only limited protection and
vertical integration may be more efficient (Stuckey 1983; Franz,
Sternberg, and Strongman 1986). The preceding considerations
explain why MNEs internalize the international trade of cer-
tain natural resources such as oil, bauxite, and certain types
of tin ores (Hennart 1988b; Stuckey 1983; Teece 1976) and
why assemblers integrate into the manufacture of some com-
5 For some empirical evidence in a domestic context see for example Caves
and Murphy (1976), Brickley and Dark (1987), Kidwell, Nygaard, and Silkoset
(2007) and Lafontaine and Shaw (2005). Franchisors can also be expected to
free-ride. They may collect money up-front by selling the right to franchise large
territories, and then leave the business. To reassure franchisees that they will not
act that way, most franchisors own some outlets.
261
ponents, for example why car assemblers vertically integrate
into engine production but purchase other parts (Monteverde
and Teece 1982).
The same considerations apply to the governance of interde-
pendencies between manufacturers, distributors, and retailers.
The effective distribution of products and services in for-
eign market usually requires physical investments (warehouses,
inventories, repair facilities, transportation equipment) as well as
intellectual ones (training salespeople to demonstrate and repair
the product). These tasks are usually performed by indepen-
dent agents and/or distributors and interdependences between
manufacturing and distribution organized through market sale,
for example by having manufacturers sell products to distrib-
utors which resell them to final consumers. In some specific
cases, however, coordination through market sale or distribu-
tion contracts will be relatively less efficient than coordination
through hierarchy and manufacturers will vertically integrate
into wholesaling and/or retailing (or wholesalers and/or retail-
ers will integrate into manufacturing). Such integration will
occur for the three reasons mentioned above. First, distrib-
utors may be reluctant to invest in physical or intellectual
assets with low resale value outside the specific trading rela-
tionship because such investments expose them to being held
up (Williamson 1985). As in the case of raw materials, dis-
tribution contracts can protect distributors, but only when the
environment is predictable. When this is not the case, integrating
manufacturing and distribution within an MNE is often the best
solution (Anderson and Coughlan 1987; Klein, Frazier, and Roth
1990). Singer Sewing Machine, for example, integrated into
both domestic and foreign retailing in the 1850s because inde-
pendent agents were unwilling to make the transaction-specific
investments that would maximize sales: they were unwilling to
learn how to demonstrate the machines, to carry stocks, and
to finance the sale of what was then a new and costly prod-
uct (Hennart 1982; Wilkins 1970). Second, the difficulty of
separating the impact that manufacturers, distributors, and retail-
ers have on product quality accounts for vertical integration
in value chains for perishable goods (Williamson 1985). The
quality of bananas, for example, depends on careful handling
at each stage, but it is difficult to know who is responsible
for bruised bananas by the time they are with the customer.
Banana distribution also requires important transaction-specific
investments. These two reasons explain why banana distribu-
tors integrated into banana transport and plantations in the last
years of the 19th century (Read 1986; Litvak and Maule 1977;
Wilkins 1970). Third, successful sales are sometimes crucially
dependent on timely marketing feedback from retailers, but it
may be difficult for manufacturers to craft contracts that ade-
quately reward them for providing this tacit information. In
that case vertical integration may be efficient. The clothing
retailer Zara, for example, has vertically integrated into gar-
ment manufacture to respond more quickly to fashion trends
(Tokatli 2008). Vertical integration between manufacturers, dis-
tributors and retailers seems, however, to be more the exception
than the rule. In most cases the coordination between manufac-
turers and wholesalers/retailers takes the form of contracts by
which the value chain participants agree to follow behavioral
262 J.-F. Hennart / Journal of Retailing 86 (3, 2010) 257–269
rules imposed by a channel leader (who can be a manufacturer
or a representative of a group of manufacturers, a wholesaler,
or a retailer) (Brown 1984; Menard 1996). To insure consistent
quality, British supermarkets, for instance, have signed contracts
with their African suppliers of fresh produce that specify how the
latter should grow their crops, process, and transport them. These
contractual stipulations are enforced through on the spot inspec-
tions, in some cases without prior notice (Dolan and Humphrey
2000).
Finance
Williamson’s (1988) view of debt and equity as alternative
governance structures provides a powerful lens with which to
understand poorly understood MNE types. Lending is risky
because the two parts of the transaction, providing funds and
being repaid, are not simultaneous, and because money is fun-
gible. The main safeguards available to lenders include lending
(1) to borrowers they know; (2) to projects with which they are
familiar and (3) when there is good collateral. Because geo-
graphic distance creates barriers to the flow of information, we
can predict that international lending will be restricted to a spe-
cific category of projects. Indeed the evidence shows that banks
have limited their international lending to short-term loans on
receivables, while long-term lending has been typically to gov-
ernments, or to borrowers with good collateral such as land.
Whenever banks have deviated from this, as they did in the
1970s, the results have been disastrous. One alternative to the
contractual exchange of financial funds across countries through
international bank lending or international bonds is to organize
the interdependence between lenders and borrowers through
hierarchical methods. This means that individuals with projects
will self-finance and ask others to become co-owners in syndi-
cates, or that they will sell shares to the public at large on stock
exchanges. This solution gives lenders better and more flexi-
ble control on the use of their funds and alleviates the tendency
that borrowers have to take excessive risk (Jensen and Meckling
1976; Williamson 1988). Internationally, the hierarchical trans-
fer of capital has taken two main forms, an intermediated one
where MNEs based in capital-abundant countries borrow on
their good name and use the funds to finance activities in
capital-poor ones, and an non-intermediated form, the ‘free-
standing company’ (Hennart 1994a, 1994b, 1998; Wilkins and
Schroeter 1998). Free-standing firms were the dominant type
of MNEs before 1914 and are still being formed today. They
were single business operations, operating a single mine or a
single plantation in a foreign country. They were domiciled in
the main equity markets, but had only a skeletal head office
there. Their actual operations were in capital-poor countries,
and in activities not widely understood and offering poor col-
lateral, then in mining and plantations, today in mining but
also in R&D-intensive activities such as biotechnology (Hennart
1994a, 1994b, 1998). TCT explains why these free-standing
firms, often viewed by business historians and economists as
purely speculative or a by-product of colonial expansion (Casson
1994), were in fact viable and long-lived business institutions
(Hennart 1994b).
Conclusion
IB scholars study the institutions used to organize interna-
tional economic interdependencies. Hence it is not surprising
that they have found TCT particularly useful. As the preced-
ing pages show, TCT offers a persuasive explanation of why
and when MNEs are the most efficient institution to organize
international interdependencies. The arguments used have relied
on asset specificity, but also on information and measurement
costs. The analysis of the governance of international transac-
tions also reveals the relative importance of two institutional
arrangements, equity joint ventures and contracts (Dolan and
Humphrey 2000; Gereffi, Humphrey, and Sturgeon 2005), which
have been labeled ‘hybrids’ in the TCT literature (Boerner
and Macher 2003; Kreps 1990; Menard 1996; Oxley 1997;
Williamson 1996). This relative importance of hybrids in the
international context may be due to the high cost of extend-
ing hierarchical control over borders (Hennart 1991b). The next
sections address these issues.
Possible contribution of international business to
transaction cost theory
Applying a theory to a new context is usually helpful in fur-
thering its development, and IB is no exception. In the rest of
this article, I focus on a few of the many areas where I see IB
making useful contributions to TCT. I will argue that applying
TCT to IB suggests a broader conceptualization of transaction
costs, provides a better understanding of equity joint ventures
and hybrids, and calls for more attention to the costs of using
hierarchy to organize international interdependencies.
Beyond asset specificity
The analysis of the various types of international transactions
that have been organized by MNEs shows that the concept of
asset specificity, while useful to explain the hierarchical organi-
zation of vertical value chains, may be insufficient to provide a
full understanding of the internalization of capital flows, and of
that of knowledge and goodwill.
Williamson (1988) argues that the viability of lending hinges
on the extent to which collateral is asset specific. While it is
true that this an important factor, an explanation of the choice
between debt and equity that would entirely rely on this concept
would play down the role of information asymmetry between
lenders and borrowers and that played by reputation in enforc-
ing loan transaction. Lenders may lend if they have confidence
in the integrity and capabilities of borrowers, and a good under-
standing of their projects, even in the absence of good collateral.
Similarly, reputation effects may play an important role where
recourse to court enforcement is not feasible and it may be
difficult to repossess collateral. This suggests that information
asymmetry between lenders and borrowers may be as important
as asset specificity in fostering hierarchical links between saving
and investment, a point made by Williamson (1975, pp. 141–49)
when explaining the efficiency of internal capital markets.
 J.-F. Hennart / Journal of Retailing 86 (3, 2010) 257–269
Similarly, explaining the choice between licensing and verti-
cal integration in terms of asset specificity does not seem to tell
the whole story. While there is evidence that licensees and pur-
chasers of plant and equipment are often dependent on licensors
for help with troubleshooting the implementation of the licensed
technology, and hence are sometimes in a position to be held
up by licensors (Hennart 1989), arguing that asset specificity
drives the choice between licensing and integration implies that
licensing contracts involve technology that is less transactor-
specific than that used in internal transfers. Yet one would
expect all licensees to make transaction-specific investments
since licensing agreements usually involve patented innovations,
and patented products and processes are unique, since novelty
is a requirement for patenting. Arguments based on informa-
tion asymmetry seem more persuasive. IB scholars have shown
that newer and more tacit technologies, and those available from
only few sources, are transferred internally, whereas older and
better known technologies tend to be licensed (Davidson and
McFetridge 1984; Davies 1977, 1993; Kogut and Zander 1993).
The same argument seems also to apply to the international trans-
fer of reputation. There the choice between operating outlets
with employees vs. franchising them does not seem to be driven
by asset specificity, but, as we have seen, is better understood
as a tradeoff between the cost of curbing free-riding by fran-
chisees and that of monitoring employees (Brickley and Dark
1987; Fladmoe-Lindquist and Jacque 2005; Kidwell, Nygaard,
and Ragnhild 2007).
Equity joint ventures
The general view in TCT is that equity joint ventures (EJVs)
are hybrids, one of the three discrete governance structures
alongside market and hierarchy (Boerner and Macher 2003, p.
5; Kreps 1990; Williamson 1991, p. 292). Oxley writes that the
EJV is “the classic form of hybrid organization” because it sits
between markets and hierarchy in terms of incentive intensity,
administrative controls, and contract support. She argues that
“the intensity of incentives in a joint venture is not reduced to the
same extent as in a fully integrated structure (since parties to the
transaction retain a degree of autonomy)”. Administrative con-
trols are also in-between markets and hierarchy because “. . .in
contrast to directives from senior management or the board of
directors in a fully integrated firm, the directives from joint
venture companies are subject to negotiation and compromise”.
Lastly, because of the need for continued cooperation, the gov-
ernance of joint ventures is not characterized by classical control
law but “by the highly adjustable framework of ‘neoclassical’
or ‘relational contracting”’ (Oxley 1997, pp. 390–1). Consistent
with this view of EJVs as half-way between pure market and
pure hierarchy, Williamson (1991, p. 292) has argued that while
hierarchy is used when asset specificity is very high, and markets
when it is very low, EJVs and hybrids are used when it is at inter-
mediate levels. Consequently, some authors (e.g. Erramili and
Rao 1993) have modeled the choice between EJVs and wholly
owned affiliates as determined by the level of asset specificity,
with wholly owned operations chosen when asset specificity is
high and EJVs when it is lower.
263
The study of modes of foreign market entry suggests, how-
ever, that EJVs should not be seen as hybrids of market and
hierarchy, but instead as joint hierarchy. Successful operation in
a foreign market requires the bundling of non-location-bound
factors such as technological, marketing assets, and/or financial
resources, and location-bound factors such as land, labor, dis-
tribution networks, permits and licenses, and so forth. In Fig. 1
I assume for simplicity that all these location-bound factors are
held by a single local firm so that there are only two input suppli-
ers, an MNE contributing knowledge and a local firm supplying
all the other factors of production needed for local production.
What is the most efficient way to bundle them when their sale
is subject to positive transaction costs, and specifically, who,
of the MNE or the local firm, should hold title to the profits
that will be generated by their joint effort? It can be argued
that the party who should receive title to the profits, i.e. who
should be paid with what remains after the other factor owner
has received his ex ante contractually agreed payment, should be
the one whose contribution is the hardest to measure (Eswaran
and Kotwal 1985). The reason is that the residual claimant can
be expected to self-monitor to maximize the output of the bun-
dle. In Fig. 1, I assume that the knowledge supplied by the MNE
can be either explicit and patented knowledge whose transfer
incurs low transaction costs (it is ‘easy to transact’ – column
1), or tacit and poorly protected, in which case its transfer will
incur high transaction costs (‘difficult to transact’ – column 2).
Likewise, the package of factors of production owned by the
local firm can either be transferable at low transaction costs –
easy to transact (row 1) or at high transaction costs – difficult
to transact (row 2). This simple model predicts that when MNE
knowledge is easy to transfer but local factors are hard to transfer,
the solution that will maximize total rents will consist in having
the local firm hold the residual, and purchase MNE knowledge
on the market (through licensing contracts, the purchase of cap-
ital goods incorporating that knowledge, or the acquisition of
firms in which the knowledge is embedded). Likewise, when
MNE knowledge is hard to transfer and local factors are easy to
acquire, the optimal solution will be to have the MNE purchase
or rent the local factors held by the local firm (or purchase the
local firm) and keep the residual – the MNE will establish a
greenfield wholly owned affiliate or fully acquire the local firm.
Whenever the market exchange of both the knowledge held by
the MNE and the local factors held by the local firm would incur
high transaction costs, the optimal solution is to give both parties
a claim to the profits of the bundle. This results in a ‘residual
sharing agreements’ (Hennart 1988a, 2009) in which both input
suppliers share the profits and have joint decision rights over
the venture. Greenfield EJVs, partial acquisitions, partnerships
and sharecropping contracts are examples of residual sharing
agreements. Such arrangements will be efficient when giving
both parties medium incentives to both cheat and shirk is prefer-
able to giving strong incentives to one of the parties to cheat,
and the other to shirk, as would be the case if title to the prof-
its was fully allocated to one of the parties. Hence EJVs are
not chosen when transaction cost are at intermediate levels (for
example because asset specificity is at an intermediate level)
but when they are simultaneously high for two or more par-
264 J.-F. Hennart / Journal of Retailing 86 (3, 2010) 257–269

Fig. 1. Optimal assignment of residual rights in foreign market entry. Source: Hennart (2009).

ties who are contributing complementary inputs. In that sense hierarchy, as an organizing method, does not reward agents on
joint ventures are not hybrids, they are hierarchy, albeit joint the basis of their output, but on the basis of their behavior.
hierarchy. Breaking the connection between output and rewards reduces
the incentives that agents have to cheat but also those they have
to maximize output, and hence the behavior of employees will
Hybrids
have to be monitored by their superiors. If imposing behav-
ior constraints also experiences diminishing returns, it will not
If EJVs are not hybrids, then what are hybrids? Carter and
pay to monitor or socialize employees to the point where shirk-
Hodgson note that “compared with vertical integration, there is
ing is reduced to zero, so the total enforcement costs of using
less of a consensus over the nature and causes of hybrid rela-
hierarchy (total shirking costs) will be those of curbing shirk-
tionships. Many empirical researchers regard the understanding
ing through monitoring and/or socialization, and the residual
of hybrid relationships as a major challenge and propose that
amount of shirking that will not be worth curbing.
Williamson’s transaction cost economics framework is inade-
Fig. 2, adapted from Hennart (1993), shows how the optimal
quate for this task. Many different phenomena are described
institution chosen will depend on the level of total enforcement
as hybrids, and what exactly constitutes a hybrid relationship
costs. Institutions are arrayed on the x-axis, and are defined by
is open to dispute” (Carter and Hodgson 2006, p. 468). Yet,
their mix of output and behavior constraints. The y-axis mea-
because hybrids play such an important role in organizing inter-
sures the level of enforcement costs, totalcheating costs, total
national interdependencies, understanding their nature is both
theoretically and empirically important.
A starting point for understanding hybrids might be to ask
why replacing a market transaction by a hierarchical one (i.e.
having a task taken from a contractor and given to an employee)
may increase the rents of organizing the interdependence. To
simplify, I will concentrate on enforcement costs and ignore
information and bargaining costs. As argued above, the price sys-
tem rewards agents for their output. Given opportunism, some
market participants may be incited to both undersupply the pos-
itive dimensions of output and to oversupply the negative ones,
i.e. to cheat. Market participants will thus have to invest in mea-
suring output. If measurement experiences diminishing returns,
perfect measurement will not be worthwhile. Total cheating
costs, which are the enforcement costs of using the price system,
will be the sum of measurement costs and the unavoidable level
or residual cheating due to imperfect measurement. The costlier
it is to measure all the relevant positive and negative dimensions
of output, the higher total cheating costs.
When cheating costs are high, switching the transaction to
hierarchy may increase rents. This is because, as argued above, Fig. 2. Organizing methods and institutions.
 J.-F. Hennart / Journal of Retailing 86 (3, 2010) 257–269
shirking costs, and their sum, total enforcement costs. The level
of total cheating costs for a given transaction is given by line CC .
Consistent with the argument made above, CC falls as output
incentives are loosened because this loosening reduces incen-
tives to cheat. SS , the total level of shirking costs, rises as an
agent’s rewards is increasingly based on her behavior, since the
payoff to spend paid time to indulge in other pursuits increases
with the proportion of her income paid on a time-basis. The level
of the CC and SS curves for a given transaction reflect how
effective output measurement and monitoring of behavior are
at enforcing transactions. Fig. 2, for example, shows a transac-
tion for which using only output measurement incurs lower total
enforcement costs than using only behavior constraints, but the
opposite could be the case for other transactions. One assump-
tion of the graph is that behavior and output measurement are
substitutes. In other words, it is not efficient to both increase out-
put and behavior constraints at the same time. This assumption
is discussed below. The non-linearity of the CC and SS curves
reflects the assumption that each organizing method experiences
diminishing returns as it is applied more intensively. Point M
corresponds to the economist’s view of pure markets, that is full
output incentives and zero behavior constraints. At point M total
shirking costs (given by line SS ) are zero, while total cheating
costs (given by line CC ) are at their maximum. Point H cor-
responds to a purely hierarchical firm, where agents are fully
rewarded on the basis of their behavior, and hence where cheat-
ing costs are zero while shirking costs are at their maximum.
Point T corresponds to a mix of output and behavior constraints
that is biased towards output incentives. As I argue below, con-
tracts are examples of this type of institutions insofar as they
include hierarchical elements in what are still basically market
relationships (Stinchcombe 1985). Point B corresponds to a mix
of output and behavior constraints biased towards behavior con-
straints, for example an employment contract with significant
output-related rewards (piece work, bonuses, and commissions).
The institution chosen for a given transaction will be the one that
uses the mix of output and behavior constraints that yields the
lowest level of total enforcement costs, which are the sum of
total cheating and total shirking costs. Fig. 2 corresponds to
the situation where the most efficient institution is a contract
because the increase in total shirking costs that would accom-
pany the imposition of additional behavior constraints would
be higher than the reduction in cheating costs such a move
would accomplish.
Personal selling, for example, can be obtained from indepen-
dent agents called manufacturer representatives (reps) or from
salaried employees. Reps are independent contractors paid com-
missions on the sales they make. They have strong incentives to
maximize sales, and usually achieve higher sales than employ-
ees, but at the expense of those aspects of their output which are
not measured. Hence even though they may be asked to provide
demonstration, instruction, after-sales service, and to truthfully
represent product characteristics and the extent to which they
fit the needs of their customers, the model would predict that
reps rewarded on sales are more likely to provide less of these
ancillary services and to be less than fully honest in their repre-
sentations to customers (Anderson 1988; Anderson and Oliver
265
1987; Robertson and Anderson 1993; Wilkins 1970).6 In other
words, they are likely to cheat. A move to the right of M means
a decrease in output incentives and an increase in behavior con-
straints, and hence a decrease in cheating but an increase in
shirking costs. Manufacturers may want to keep the benefits of
output incentives and avoid turning reps into employees because
it is costly to monitor the behavior of salaried salespeople, espe-
cially if they do outside sales. On the other hand, they may want
to control the more costly of the negative side effects caused
by output incentives. A solution is to have reps agree to fol-
low some behavioral rules imposed by manufacturers. Pre-war
British manufacturers required their foreign reps to hold min-
imum levels of stocks, specified a minimum level of showing
they had to do, and reserved the right to refuse orders, to moni-
tor stocks and to inspect sales records (Nicholas 1983). Further
increases in behavior constraints imposed on reps are likely,
however, to reduce their incentive to generate sales, as they
will balk at having their behavior constrained by manufactur-
ers while they alone bear the negative earnings consequences of
such constraints. At some point they will resist the imposition of
additional behavior constraints and will demand a base salary in
addition to their commission, i.e. they ask to become employees.
This salary plus commission package, represented by point B,
will therefore be optimal if both measuring output and constrain-
ing behavior experience diminishing returns. If price incentives
result in very high cheating costs, and if shirking is relatively
easy to control, then the manufacturing firm will use employees
paid a straight salary (point H). For example, a firm which wants
its salespeople to use low-pressure, expertise-based sales tactics
may find that obtaining this by contract is next to impossible
due to the difficulty of defining and measuring behavioral rules
on which both parties can agree, and that fully constraining and
rewarding behavior through an employment contract is more
efficient.7
While it is important to keep it mind that it only focuses
on enforcement, and neglects differences in how the price sys-
tem and hierarchy solve the information problem, the model
presented above has a number of implications that are sup-
ported by the empirical evidence. First, it predicts that the share
of salary (behavior-related pay) in a salesperson’s total com-
pensation package should be higher the more important the
hard-to-measure dimensions of output, such as demonstration
and service, and the easier it is to monitor behavior. John and
Weitz (1989) found that the share of salary in total compensation
was higher when what was expected of salespeople went beyond
making sales, specifically when salespeople spent more time on
non-selling activities, when customers had higher information
needs, and when output measures (sales and costs) were deemed
6 Unless they are selling to return customers and are in the business for the
long term.
7 The defining characteristic of the employment contract is that employees are
rewarded on the basis of appropriate behavior, and this also applies to salaried
salespersons. The most common criteria used by sales managers to evaluate their
subordinates were (in descending order): attitude, product knowledge, selling
skills, appearance and manners, and tied for the last place, communication skills
and sales volume (Jackson, Keith, and Schlacter 1983).
266 J.-F. Hennart / Journal of Retailing 86 (3, 2010) 257–269
by supervisors to be poor measures of performance. These find-
ings were supported by Krafft, Albers, and Lal (2004), who also
found that reps were used when travel requirement were high (i.e.
when their behavior was harder to monitor). Anderson (1985)
investigated a closely related variable, the choice between using
reps vs. company employees, and found that employees were
used when output was a bad measure of performance, a finding
confirmed by Krafft, Albers, and Lal (2004). Another implica-
tion is that because measuring output and monitoring behavior
experiences diminishing returns, most salespeople should be
rewarded by hybrid compensation packages, that is packages
that combine salary and commission. Such packages are in fact
common. In the US, for example, 70 percent of the manufactur-
ing firms surveyed reported rewarding salespeople with salary
and commission, 14 percent with straight salary, and 19 percent
with straight commission (Sales and Marketing Management
1989).
I have used the example of personal selling, but the model
is very general. As I have argued earlier, the high cost of mea-
suring and pricing the impact of free-riding on the goodwill
capital of a franchise chain prevents the arm’s length rental
of trademarks (point M). Given this, the next best thing is
a franchising contract (point T). These provide strong output
incentives which limit shirking, since franchisees get to keep
most of what they make, while cheating costs (free-riding on
quality) are kept in check through the imposition on franchisees
of behavioral rules devised to guarantee specific quality lev-
els (e.g. the QSC clauses in fast food). The attractiveness of
franchising is that it reduces shirking, so the higher the cost
of monitoring employees, the more likely franchising (Brickley
and Dark 1987; Fladmoe-Lindquist and Jacque 2005; Lafontaine
1992). When it is difficult to maintain standard quality by impos-
ing behavior constraints and franchising would result in too
much cheating (free-riding), and when distribution is selective
so the number of outlets is smaller and hence monitoring prob-
lems less serious (shirking costs are lower), the optimal solution
is be for the trademark owner to operate the outlets with employ-
ees (point H). If this solution results in too much shirking, outlet
managers may be paid bonuses based on the profitability of the
outlet (point B).
The preceding analysis also explains why most international
value added chains are neither vertically integrated within MNEs
nor handled in arm’s length markets. Take the case of agricultural
produce. As food safety has become more important, the costs of
measuring the relevant characteristics of agricultural output have
increased, thus reducing the efficiency of arm’s length trans-
actions. At the same time operating at all stages of the value
chain would force a firm to do business in unfamiliar industries
and countries, thus increasing monitoring costs. The problems
faced by buyers in buyer-driven value chains are similar to those
faced by manufacturers and franchisors, with the difference that
buyers need to assure that a large number of upstream produc-
ers supply output of consistent quality while manufacturers and
franchisors have the same problem with a large number of down-
stream retailers and producers. It is therefore not surprising that
the institutions chosen have been similar, with buyers entering
contracts with growers which, like in the case of franchising or
distribution contracts, stipulate detailed behavioral rules growers
must follow (Dolan and Humphrey 2000).
We have seen that firms can impose behavioral constraints
through both external means (setting up behavioral rules and
observing their compliance) or through internal means (having
employees who share the same goals or socializing them so that
they do). Likewise, setting up behavioral rules to complement
output monitoring in hybrid institutions need not necessarily
take the form of legal contracts. Behavioral compliance can
also be obtained through socialization, so that all parties to the
transaction will voluntarily agree to norms of behavior that will
maximize the global rents of the value chain. This results in what
has been called relational governance (Dyer 1996; Heide 1994)
and relational value chains (Gereffi, Humphrey, and Sturgeon
2005), a mode of governance that is facilitated by membership
in the same social group or by investments in socialization.
To sum up, contracts and relational governance should be
clearly distinguished from arm’s length transactions. In contrast
to EJVs, contracts and relational governance are true hybrids,
in the sense that they make use of both output incentives and
behavior constraints. It is also important to distinguish between
organizing methods and institutions. This is preferable to cat-
egorizing spot markets, hybrids, and firms into discrete types,
since these three institutions exhibit considerable heterogeneity,
even within a legal category. For instance, while they both are
called firms, small biotechnology research firms differ radically
from large pharmaceutical companies.
Benefits and costs of hierarchy
Lastly, the preceding examples of personal selling, fran-
chising, and global value chains make it clear that one cannot
understand governance choice by only looking at market trans-
action costs. For example, from the point of view of enforcement
costs, one cannot fully predict whether a particular outlet will be
franchised or company-owned based only on the likelihood of
free-riding (cheating). One must also simultaneously consider
the costs of monitoring employees (shirking). Keeping oppor-
tunities for free-riding constant, Brickley and Dark’s (1987)
empirical study of US franchising shows that distance from
HQ increases the probability an outlet will be franchised while
Fladmoe-Lindquist and Jacque (2005) find that an outlet is more
likely to be franchised the greater the geographic and cultural
distance between the country where it is located and the United
States. These are not isolated findings: in the empirical litera-
ture on franchising, the variables proxying for monitoring costs
receive consistently stronger empirical support than those prox-
ying for free-riding opportunities (Lafontaine and Slade 2002).
To understand governance, it is not sufficient to focus on mar-
ket failure, it is also necessary to develop a theory of the relative
benefits and costs of firm organization. From an enforcement
point of view, the cost and benefits of hierarchy are those that
derive from directly controlling behavior (as opposed to mea-
suring output) and hierarchy is an efficient organizing method
when a consistent output can be obtained by specifying behavior,
and when monitoring such behavior is relatively easy. The latter
can be proxied by geographic distance, and while this proxy is
 J.-F. Hennart / Journal of Retailing 86 (3, 2010) 257–269
very crude, it has shown surprisingly good explanatory power.
Naturally enforcement is not the only transaction cost. From an
information viewpoint, hierarchical coordination would seem
to have an advantage when the boss has better information than
the workers, with the price system being more efficient in the
opposite case (Hayek 1945; Hennart 1991b).
The need for a realistic theory of the benefit and cost of
hierarchy is particularly great in IB. If, as I have argued,
internalizing transactions means replacing a decentralized infor-
mation structure with a centralized one, and measurement
of output with control of behavior, then one can understand
why shirking and information loss will impose serious limits
on MNEs.8 The greater physical distance between headquar-
ters and employees in MNEs than in domestic firms makes
direct external control of behavior more difficult, interna-
tional accounting differences and differences in the environment
impair the efficiency of bureaucratic control, while the use
of socialization is made problematic by the MNE having
a mix of employees with different national cultures. Differ-
ences between domestic and foreign environments means that
employees located in foreign countries will have a substantial
information advantage over HQ, making centralized direction
less efficient and providing subsidiary managers with substantial
opportunities for suboptimization. As a result, one would expect
that the costs of organizing transactions within firms would
have a determining impact on the governance of international
transactions.
Conclusion
TCT has provided a remarkably useful lens to understand IB
phenomena. Applying TCT to the research questions asked by
IB researchers raises some challenging questions that call for
further theoretical developments. This is the mark of a powerful
theory.
Acknowledgement
I thank Arne Nygaard for his support and encouragement.
References
Alchian, Armen and Harold Demsetz (1972), “Production, Information Costs,
and Economic Organization,” American Economic Review, 62 (1), 777–95.
Anderson, Erin (1985), “The Salesperson as Outside Agent or Employee: A
Transaction Cost Analysis,” Marketing Science, 4, 234–5.
(1988), “Transaction Cost Determinants of Opportunism
in Integrated and Independent Sales Forces,” Journal of Law, Economics, and
Organization, 9, 247–64.
Anderson, Erin and Anne Coughlan (1987), “International Market Entry
Through Expansion via Independent or Integrated Channels of Distribution,”
Journal of Marketing, 51, 71–82.
Anderson, Erin and Hubert Gatignon (1986), “Modes of Foreign Entry: A Trans-
action Cost Analysis and Propositions,” Journal of International Business
Studies, 17 (3), 1–26.
8 See Tomassen and Benito (2009) for an interesting attempt to evaluate the
impact of these on subsidiary performance.
267
Anderson, E. and Richard Oliver (1987), “Perspectives on Behavior-Based Ver-
sus Outcome-Based Salesforce Control Systems,” Journal of Marketing, 51
(October), 76–88.
Arora, Ashish and Andrea Fosfuri (2000), “Wholly-Owned Subsidiary vs.
Technology Licensing,” Journal of International Business Studies, 31,
555–72.
Arrow, Kenneth (1962), “Economic Welfare and the Allocation of Resources for
Invention,” in The Rate and Direction of Inventive Activity, Kenneth Arrow
ed. Princeton: Princeton University Press, 609–25.
(1974), The Limits of Organization, New York: Norton.
Bartlett, Christopher and Sumantra Ghoshal (1989), Managing Across Borders:
The Transnational Solution, Boston: Harvard Business School Press.
Boerner, Christopher and Jeffery Macher (2003), “Transaction Cost Economics:
An Assessment of Empirical Work in the Social Sciences.” working paper,
Georgetown University.
Brickley, James and Frederick Dark (1987), “The Choice of Organizational form:
the Case of Franchising,” Journal of Financial Economics, 18, 401–20.
Brown, Wilson (1984), “Firm-Like Behavior on Markets: The Administered
Channel,” International Journal of Business Organization, 2, 236–7.
Buckley, Peter and Mark Casson (1976), The Future of the Multinational Enter-
prise, London: Macmillan.
Campbell, Duncan and Parisotto, Aurelio (1995), The Global Value Chain Con-
cept in Relation to the Institute’s Program of Work, Paper Presented at the
Workshop on the International Organization of Production, Geneva.
Cantwell, John (1989), Technological Innovations and Multinational Corpora-
tions, Oxford: Basil Blackwell.
Carter, Richard and Geoffrey Hodgson (2006), “The Impact of Empirical Tests
of Transaction Cost Economics on the Debate on the Nature of the Firm,”
Strategic Management Journal, 27, 461–76.
Casson, Mark (1994), “Institutional Diversity in Overseas Enterprise: Explain-
ing the Free-Standing Company,” Business History, 36 (4), 95–108.
Caves, Richard (1971), “International Corporations: The Industrial Economics
of Foreign Investment,” Economica, 38 (139), 1–27.
Caves, Richard and William Murphy (1976), “Franchising: Firms, Markets, and
Intangible Assets,” Southern Economic Journal, 42, 572–86.
Caves, Richard, Harry Crookell and Peter Killing (1982), “The Imperfect Market
for Technology Licenses,” Oxford Bulletin of Economics and Statistics, 45,
249–67.
Cheung, Steven (1969), “Transaction Costs, Risk Aversion, and the Choice of
Contractual Arrangements,” Journal of Law and Economics, 12, 23–42.
Coase, Ronald (1937), “The Nature of the Firm,” Economica, 4, 386–405.
Contractor, Farok (1984), “Choosing Between Foreign Direct Investment and
Licensing: Theoretical Considerations and Empirical Tests,” Journal of
International Business Studies, 15, 167–88.
Davidson, William and Donald McFetridge (1984), “International Technology
Transactions and the Theory of the Firm,” Journal of Industrial Economics,
32 (3), 253–64.
Davies, Howard (1977), “Technology Transfer Through Commercial Transac-
tions,” Journal of Industrial Economics, 26, 161–75.
(1993), “The Information Content of Technology
Transfers: A Transaction Cost Analysis of the Machine Tool Industry,”
Technovation, 13 (2), 93–100.
Dolan, Catherine and John Humphrey (2000), “Governance and Trade in Fresh
Vegetables: The Impact of UK Supermarkets on the African Horticulture
Industry,” Journal of Development Studies, 37 (2), 147–76.
Dunning, John (1977), “Trade, Location of Economic Activity, and the Multina-
tional Enterprise: A Search for an Eclectic Approach,” in The International
Allocation of Economic Activity, Bertil, Ohlin P. O. Hesselborn and P. M.
Wijkman, eds. New York: Holmes and Meier, 395–418.
(2009), “The Key Literature on IB Activities,
1960–2006,” in The Oxford Handbook of International Business, Alan Rug-
man ed (2nd ed.). Oxford: Oxford University Press, 39–71.
Dunning, John and Alan Rugman (1985), “The Influence of Hymer’s Disser-
tation on the Theory of Foreign Direct Investment,” American Economic
Review, 75, 228–32.
Dyer, Jeffrey (1996), “Specialized Supplier Networks as a Source of Compet-
itive Advantage: Evidence from the Auto Industry,” Strategic Management
Journal, 17, 271–92.
268 J.-F. Hennart / Journal of Retailing 86 (3, 2010) 257–269
Erramili, M. Krishnan and C.P. Rao (1993), “Service Firms’ International Entry
Mode Choice: A Modified Transaction Cost Analysis Approach,” Journal
of Marketing, 57, 19–38.
Eswaran, Mukesh and Ashok Kotwal (1985), “A Theory of Contractual Structure
in Agriculture,” American Economic Review, 75 (3), 352–67.
Fieldhouse, David Kenneth (1986), “The Multinational: A Critique of a Con-
cept,” in The Multinational Enterprise in Historical Perspective, Alice
Teichova ed. Cambridge: Cambridge University Press.
Fladmoe-Lindquist, Karin and Laurent Jacque (2005), “To Own or to Franchise:
International Control Decisions for Service Companies,” Journal of Applied
Corporate Finance, 9 (3), 98–109.
Franz, Juergen, Bo Sternberg and John Strongman (1986), Iron Ore: Global
Prospects for the Industry, Washington, DC: World Bank.
Gatignon, Hubert and Erin Anderson (1988), “The Multinational Corporation’s
Degree of Control over Foreign Subsidiaries: An Empirical Test of a Trans-
action Cost Explanation,” Journal of Law, Economics, and Organization, 4,
305–36.
Gereffi, Gary, John Humphrey and Timothy Sturgeon (2005), “The Governance
of Global Value Chains,” Review of International Political Economy, 12 (1),
78–104.
Gomes-Casseres, Benjamin (1989), “Ownership Structures of Foreign Sub-
sidiaries: Theory and Evidence,” Journal of Economic Behavior and
Organization, 11, 1–25.
Hayek, Frederic (1945), “The Use of Knowledge in Society,” American Eco-
nomic Review, 35, 519–30.
Heide, Jan (1994), “Interorganizational Governance in Marketing Channels,”
Journal of Marketing, 58 (1), 71–85.
Hennart, Jean-François (1977), “A Theory of Foreign Direct Investment.” Ph.D.
dissertation, University of Maryland.
(1982), A Theory of Multinational Enterprise, Ann
Arbor: University of Michigan Press.
Hennart, Jean-François (1986), “What is Internalization?,” Weltwirtschaftliches
Archiv, CXII (Winter), 791–804.
(1988a), “A Transaction Cost Theory of Equity Joint
Ventures,” Strategic Management Journal, 9 (4), 361–74.
(1988b), “Vertical Integration in the Aluminum and
Tin Industries,” Journal of Economic Behavior and Organization, 9 (3),
281–300.
(1989), “Can the New Forms of Investment Substitute for
the ‘Old Forms’. A Transaction Cost Perspective,” Journal of International
Business Studies, 20 (2), 211–34.
(1991a), “The Transaction Costs Theory of Joint Ven-
tures: An Empirical Study of Japanese Subsidiaries in the United States,”
Management Science, 37 (4), 483–97.
(1991b), “Control in Multinational Firms: The Role of
Price and Hierarchy,” Management International Review, 31 (special issue),
71–96.
(1993), “Explaining the Swollen Middle: Why Most
Transactions are a Mix of Market and Hierarchy,” Organization Science,
4 (4), 529–47.
(1994a), “International Capital Transfers: A Transaction
Cost Framework,” Business History, 36, 51–70.
(1994b), “Free-Standing Firms and the Internalization of
Markets for Financial Capital: A Response to Casson,” Business History, 36
(4), 118–31.
(1998), “Transaction-Cost Theory and the Free-Standing
Firm,” in The Free Standing Company in the World Economy, Mira Wilkins
and Harm, Schroeter eds. London: Oxford University Press, 65–98.
(2000), “Transaction Costs Theory and the Multinational
Enterprise,” in The Nature of the Transnational Firm, Christos Pitelis and
Rodger Sugden , eds (2nd ed.). London: Routledge, 72–118.
(2009), “Down with MNE-Centric Theories! Market
Entry and Expansion as the Bundling of MNE and Local Assets,” Jour-
nal of International Business Studies, special 40th anniversary issue, 40 (9),
1432–54.
Hennart, Jean-François and Yeung Ryeol Park (1993), “Greenfield vs. Acquisi-
tion: The Strategy of Japanese Investors in the United States,” Management
Science, 39 (9), 1054–70.
Hymer, Stephen (1960). “The International Operations of National Firms.” Ph.D.
dissertation, Massachusetts Institute of Technology.
(1976), The International Operations of National Firms,
Boston: MIT Press.
Jackson, Donald, Janet Keith and John Schlacter (1983), “Evaluation of Selling
Performance: A Study of Current Practices,” Journal of Personal Selling and
Sales Management, 3 (November), 42–51.
Jensen, Michael and William Meckling (1976), “Theory of the Firm: Manage-
rial Behavior, Agency Costs, and Capital Structure,” Journal of Financial
Economics, 3, 305–60.
John, George and Barton Weitz (1989), “Salesforce Compensation: An Empir-
ical Investigation of Factors Related to Use of Salary vs. Incentive
Compensation,” Journal of Marketing Research, 26, 1–14.
Jones, Geoffrey (2005), Multinationals and Global Capitalism, Oxford: Oxford
University Press.
Kidwell, Roland, Arne Nygaard and Ragnhild Silkoset (2007), “Antecedents and
Effects of Free-Riding in the Franchisor–Franchisee Relationship,” Journal
of Business Venturing, 22, 522–44.
Kindleberger, Charles (1969), American Business Abroad, New Haven: Yale
University Press.
Klein, Saul, Gary Frazier and Victor Roth (1990), “A Transaction Cost Anal-
ysis Model of Channel Integration in International Markets,” Journal of
Marketing Research, 23, 196–208.
Kogut, Bruce and Udo Zander (1993), “Knowledge of the Firm and the Evolu-
tionary Theory of the Multinational Corporation,” Journal of International
Business Studies, 24, 625–4.
Krafft, Manfred, Sonke Albers and Rajiv Lal (2004), “Relative Explanatory
Power of Agency Theory and Transaction Cost Analysis in German Sales-
forces,” International Journal of Research in Marketing, 21, 265–83.
Kreps, David (1990), A Course in Microeconomics Theory, Princeton: Princeton
University Press.
Lafontaine, Francine (1992), “Agency Theory and Franchising: Some Empirical
Results,” Rand Journal of Economics, 23 (2), 263–8.
Lafontaine, Francine and Kathryn Shaw (2005), “Targeting Managerial Control:
Evidence from Franchising,” Rand Journal of Economics, 36 (1), 131–50.
Lafontaine, Francine and Margaret Slade (2002), “Incentives Contracts and
the Franchising Decision,” in Game Theory and Business Applications,
Kaylan Chatterjee and William, Samuelson eds. New York: Kluwer,
133–88.
Lange, Oskar (1936), “On the Economic Theory of Socialism,” Review of Eco-
nomic Studies, 4, 53–71.
Levin, Richard, Alan Klevorick, Richard Nelson and Sidney Winter (1987),
“Appropriating the Returns from Industrial Research and Development,”
Brookings Papers on Economic Activity, 3, 783–820.
Litvak, Isaiah and Christopher Maule (1977), “Transnational Corporations and
Vertical Integration: The Banana Case,” Journal of World Trade Law, 11/6,
537–49.
Magee, Stephen (1977), “Information and the Multinational Corporation: An
Appropriability Theory of Direct Foreign Investment,” in The New Inter-
national Economic Order, Jagdish Bhagwhati ed. Cambridge, MA: MIT
Press.
Masten, Scott (1988), “A Legal Basis for the Firm,” Journal of Law, Economics,
and Organization, 4, 181–97.
McManus, John (1972), “The Theory of the Multinational Firm,” in The Multi-
national Firm and the Nation State, Gilles Paquet ed. Don Mills, Ontario:
Macmillan.
Menard, Claude (1996), “On Clusters, Hybrids and Other Strange Forms: The
Case of the French Poultry Industry,” Review of Institutional and Theoretical
Economics, 152, 154–83.
Monteverde, Kirk and David Teece (1982), “Supplier Switching Costs and Ver-
tical Integration in the Automobile Industry,” Bell Journal of Economics, 19,
623–32.
Mowery, David, Joanne Oxley and Brian Silverman (1996), “Strategic Alliances
and Inter-Firm Knowledge Transfer,” Strategic Management Journal, 17,
77–91.
Nicholas, Stephen (1983), “Agency Contracts, Institutional Modes, and the Tran-
sition to Foreign Direct Investment by British Manufacturing Multinationals
Before 1935,” Journal of Economic History, 48, 675–86.
 J.-F. Hennart / Journal of Retailing 86 (3, 2010) 257–269
Ouchi, William (1979), “A Conceptual Framework for the Design
of Organizational Control Mechanisms,” Management Science, 25,
833–48.
Oxley, Joanne (1997), “Appropriability Hazards and Governance in Strategic
Alliances: A Transaction Cost Approach,” Journal of Law, Economics and
Organization, 13, 387–409.
Read, Robert (1986), “The Banana Industry: Oligopoly and Barriers to Entry,”
in Multinationals and World Trade, Mark Casson and Associates, London:
Allen & Unwin, 317–42.
Robertson, Diana and Erin Anderson (1993), “Control System and Task Envi-
ronment Effect on Ethical Judgment: An Exploration Study of Industrial
Salespeople,” Organization Science, 4 (4), 617–44.
Rugman, Alan (1981), Inside the Multinationals, New York: Columbia Univer-
sity Press.
Sales and Marketing Management (1989), 1989 Survey of Selling Costs, Febru-
ary 20, 20.
Stinchcombe, Arthur (1985), “Contracts as Hierarchical Documents,” in Orga-
nization Theory and Project Management., Arthur Stinchcombe and Carol,
Heimer eds. Oslo: Norwegian University Press, 121–7.
Stuckey, John (1983), Vertical Integration and Joint Ventures in the Aluminum
Industry, Cambridge, MA: Harvard University Press.
Teece, David (1976), Vertical Integration and Vertical Divestiture in the US
Petroleum Industry, Stanford: Institute for Energy Studies.
269
(1981), “The Multinational Enterprise: Market Failure
and Market Power Considerations,” Sloan Management Review, 22 (3), 3–17.
Tokatli, Nebahat (2008), “Global Sourcing: Insights from the Global Clothing
Industry—The Case of Zara, a Fast Fashion Retailer,” Journal of Economic
Geography, 8, 21–38.
Tomassen, Sverre and Gabriel Benito (2009), “The Costs of Governance in
International Companies,” International Business Review, 18, 292–304.
Wilkins, Mira (1970), The Emergence of Multinational Enterprise: American
Business Abroad from the Colonial Era to 1914, Cambridge, MA: Harvard
University Press.
Wilkins, Mira and Harm Schroeter (1998), The Free Standing Company in the
World Economy, New York: Oxford University Press.
Williamson, Oliver (1975), Markets and Hierarchies: Analysis and Antitrust
Implications, New York: Free Press.
(1985), The Economic Institutions of Capitalism, New
York: Free Press.
(1988), “Corporate Finance and Corporate Governance,”
Journal of Finance, 63/3, 567–98.
(1991), “Comparative Economic Organization: The
Analysis of Discrete Structural Alternatives,” Administrative Science Quar-
terly, 36, 269–96.
(1996), “Economic Organization: The Case for Candor,”
Academy of Management Review, 21, 48–57.