Transaction Cost Theory

Transaction cost theory examines whether firms should make or buy something instead (Coase,
1937; Williamson, 1998). If a company can source the resources and manufacture its own
products, there is no need to contract with other companies. However, this is usually not the case,
and it is often advantageous for businesses to enter into commercial or other types of agreements
with other businesses.
The theory assumes that two trading companies are risk neutral; deal with each other basically as
equals; have extensive business experience; and employ specialized managerial, legal, technical,
and financial experts. With these assumptions, rather than focus on differences in the trading
partners (such as experienced versus naïve), the theory focuses on differences in contracting
issues between the trading partners and on the costs involved in those contracts (Williamson,
1998).
Focusing on transaction costs was a major shift in economics research, as prior to that time the
costs of running a business were mainly for production, and the costs of contracts and
transactions were assumed to be zero (Coase, 1937). The costs of transactions are usually not
measured directly, but are estimated using proxies of critical dimensions of transactions (Jobin,
2008). Transaction costs include negotiating, monitoring, and enforcing contracts (Hill, 1990)
and the costs of planning, adapting, and monitoring task completion (Williamson, 1985).
According to the theory, transactions differ in a number of ways, such as the degree to which
each party’s relationship-specific assets are involved, the amount of uncertainty about the other
party’s actions and about the future in general, the complexity of the trading agreement, and the
frequency with which transactions occur. These differences help a firm decide which governance
structures are preferable (Williamson, 1985).
Williamson (1985) identified four types of asset specificity: site, physical, human, and dedicated.
Site specificity refers to highly immobile assets that remain in place to save transportation and
inventory costs. Physical specificity refers to equipment and machinery that are specific to this
relationship. Human specificity refers to human capital or other employee education and
training that is specific to this relationship. Dedicated specificity refers to substantial
investments that were made only for this transaction and have no value outside this transaction.
When one party has specific assets, then it triggers opportunistic behaviors, which puts a firm at
risk and requires costly contractual safeguards to deter those negative behaviors (Poppo &
Zenger, 1998).
The central idea in transaction cost theory is that organizations change because managers seek to
economize their transaction costs (Williamson, 1985). The theory examines the costs of
transactions in one governance structure versus another. Organizations will perform the best
when they use governance structures that are the least expensive possible.
If parties to transactions are bilaterally dependent on each other, meaning that neither party can
easily make alternative arrangements, then the parties are vulnerable. In response, the parties
create value-preserving governance structures among the parties. These structures infuse order,
which helps mitigate conflict and allows the firms to realize mutual gain.
Transaction cost theory examines how trading companies protect themselves from the hazards
associated with exchange relationships with other companies (Williamson, 1975, 1985).
According to the theory, trading partners choose the most costeffective arrangement or
agreement that offers the best protection for their relationship-specific investments. Firms with
lower transaction costs are higher performers than are firms with higher transaction costs
(Williamson, 1985).
Transaction cost theory originally focused on the dichotomy between “make it” and “buy it.”
However, more recent research has focused on collaborative arrangements, called relational
governance, or alliances (Dyer, 1997).
Relational or alliance governance exchanges may be more beneficial and practicable than other
types of governance, such as when the market fails. However, relational governance exchanges
may be hard to legally enforce as they are often open ended and require such mechanisms as
trust, mutual dependence, parallel expectations, and fairness to sustain them. Geyskens,
Steenkamp, and Kumar (2006) found strong support for transaction cost theory for both make-
versusbuy and ally-versus-buy decisions.
Transaction cost research has begun to examine how firms align their transactions so as to pursue
multiple goals. Firms often assign different goals for different organizational units. These various
goals can lead to complicated decisions about whether to make, buy, or use allies. Multiple
decisions that are made by multiple entities can result in organizational transaction
misalignment, which can impede organizational performance (Bidwell, 2010).
Criticisms and Critiques of the Theory
Despite a large body of work and widespread support for transaction cost theory, there are
several gaps in the literature. For example, there is a lack of agreement on key term and concept
definitions. Inconsistencies and difficulties in measuring terms like opportunism, asset
specificity, and uncertainty make interpreting results from various studies difficult (Macher &
Richman, 2008)..
A second criticism of the theory is of its assumption that humans always behave
opportunistically, meaning with self-interest and without morality. Therefore, transactions must
always be governed by harshly specific contracts, because trading partners will always be out to
harm each other, and discovering avenues of harm after the fact will always be extremely costly
(Williamson, 1975).
Ghoshal and Moran (1996) argue that there are self-fulfilling prophecy problems with transaction
cost theory. As opportunism is always possible and often cannot be predicted, then it must
always be expected. As a result, trading partners must always distrust each other and create
contracts with as many controls as possible so that there will be enhanced behavior from both
trading partners.
Finally, critics have argued that transaction cost theory is biased toward the benefits of
integration and explicit contract safeguards (Poppo & Zenger, 2002). The theory has not been
able to explain anomalies, or situations where organizations can exist quite successfully without
typical governance structures (Chiles & McMackin, 1996).
Measuring Variables in the Theory

Coordination, search, contracting, monitoring, and enforcement costs measures. Jobin, D.

(2008). A transaction-cost based approach to partnership performance evaluation. Evaluation, 14,
437–465.
Proposed Evaluation Methodology

TCE can fit into an evaluation as either an additional line of enquiry or as the key evaluation question.
The evaluation questions in the approach I propose address the following criteria: relevance,
effectiveness, economy, and efficiency. To address relevance, evaluators need to compare the TCs
associated with two or more governance structures that coordinate the same type of transaction. By
examining the relationship between transaction costs and outcomes, auditors and evaluators can
establish the effectiveness of the partnership from a transaction cost perspective. By reducing TCs,
resources dedicated to unproductive activities are reduced, affecting positively the economy criterion.
An increase in economy due to reduced TCs will likely improve the efficient use of partnership inputs,
since fewer resources will go to TCs and more to productive outputs, hence improving efficiency.

By examining a partnership from the TCE perspective, evaluators and auditors are able to understand
and examine its performance from a micro-analytic perspective – to open the partnership’s black box.
The proposed evaluation methodology involves four steps:

(1) overview of the partnership;

(2) dimensionalizing the transaction according to the critical dimensions and assumptions;
(3) devising a measurement strategy;
(4) (4) devising an analysis strategy and a reporting strategy

Table 2 provides a sample of TC evaluation questions that can be answered using the TCE framework,
while Box 2 provides an overview of the measurement strategy that deals with the critical dimensions
and factors relevant to partnership performance, including a productivity index. It also provides a
sample of indicators and methods used in a variety of studies. Of course, the contents of Table and Box
2 are only suggestions. Evaluators and auditors will need to adapt them, as each partnership
performance evaluation is unique.

Table 2. Transaction-Cost-Based Evaluation Questions

Trust and transaction cost measures. Dyer, J. H., & Chu, W. (2003). The role of
trustworthiness in reducing transaction costs and improving performance: Empirical evidence
from the United States, Japan, and Korea. Organization Science, 14, 57–68.

Extent of trust-based governance, assessing reputations, and other scales. Carson, S. J.,
Madhok, R., & John, G. (2003). Information processing moderators of the effectiveness of trust-
based governance in interfirm R&D collaboration. Organization Science, 14, 45–56.
Measure Development

The initial fieldwork consisted of open-ended interviews with engineers and scientists enrolled in a
management of technology class. The primary purpose of this phase was to assure us that the
conceptual issues were indeed material in our selected context. Our decision to sample from the
research-intensive sectors grew out of these discussions. Following this phase, we sent a draft
questionnaire to about a dozen participants to verify the suitability of the wording of items. This phase
led to some minor changes in wording.

The items used for each construct are described below and are reproduced in the Appendix.

Trust-Based Governance (TBG). An eight-item scale measures the extent to which norms of trust
existed, operationalized as expectations for the fulfillment of obligations, mutuality, flexibility, and
information exchange. Items were adapted from Zaheer et al. (1998) and Noordeweir et al. (1990). The
items query for expectations not due to contracts.

Task-Related Skills of Client (CLIENTSKILLS). An eight-item scale measures task-related client skills
present at the outset of the engagement. The scale consists of two dimensions within a second-order
factor structure that incorporates the experiences and competencies of the client organization and
those of individual managers. Each dimension is equally weighted for analysis.

Teachability of Task Skills (TEACHABLE). A five-item teachability scale was adapted from Kogut and
Zander (1993) and focuses on the lack of need for first-hand experience and/or face-to-face
communications to learn the skills involved in the work.

Colocation of Task Execution (COLOCATE). A single item asked whether a significant portion of the
supplier's R&D work on the project was conducted at the client's physical site. The responses are
entered as a dummy variable in the empirical analysis.

Parallel Task Execution (PARALLEL). A four-item scale addresses the extent to which R&D activities on
the project were performed by the client paralleling the activities of the supplier in terms of dollar
expenditures, value-added, hours of work, and numbers of workers.

Task Performance (PERFORMANCE). A 13-item scale measures performance on the outsourced R&D
task.

There are four dimensions, including (1) the (ex post) expected contribution of the technology to the
client's product performance, (2) a comparison of actual performance to goals and objectives, (3)
adherence to timelines, and (4) the creativity and innovativeness of the technology. Items are adapted
from Ancona and Caldwell (1992), Andrews and Smith (1996), and Gatignon and Xuereb (1997). Each
dimension is equally weighted in the final scale.

Control Variables. Items for multi-item control variables also appear in the Appendix. These include a
three- item scale of the expected continuity of future interactions with this supplier (EXPCONT) using
Heide's and Miner's (1992) items. The ease of assessing and communicating the reputation of this
supplier (REPUTATION) is measured with a four-item scale developed specifically for this study.

Finally, three control variables are measured with single items. The estimated past history of the
interfirm ties is measured in months (HISTORY), and the financial stake of the client is measured as the
estimated development budget in dollars (INVEST). The number of sup- pliers is measured as the current
number of independent suppliers engaged on the focal project (#SUPPLIERS).

No questionnaire

Volume uncertainty scale. John, G., & Weitz, B. A. (1988). Forward integration into
distribution: An empirical test of transaction cost analysis. Journal of Law, Economics, and
Organizations, 4, 337–355.
1. MEASURES

Each sales manager was asked to complete the questionnaire for a distinct self-selected product line; all
the questions were answered at the level of this product line. The specific questions used to measure
each of the variables involved in the propositions are described below.

2. CHANNEL

The use of resellers was captured by responses to the following question: What percentages of sales are
made to the following types of customers?

____________end users

____________ channel members (wholesalers, distributors, retailers)

100% total

The percentage sold to end-users (% DIRECT) constitutes the amount going through a direct channel.
Sales to channel members represent the use of resellers.

3. SPECIFIC ASSETS (ENVUNCT)

In order to measure this variable, we capitalize on the notion that products requiring a good deal of
training and experience specific to the line represent situations where specific assets are present. This
was measured with the following item:

How much time is required for a newly hired salesperson with experience in the industry to become
adequately familiar with your products and customers? ____________ months.

This measure (SPSKILLS) does not measure the overall level of skill needed; rather, only the
nontransferable component is measured. It specifies a person with prior experience in the industry. To
the extent that the skills are transferable, such an individual would not need much time to adapt to this
situation. This measure is similar to one used by Anderson.

4.ENVIRONMENTAL UNCERTAINTY

The inability to predict relevant contingencies is measured with the rating scale items described below.
These items capture the extent to which the downstream marketing environment is volatible and
turbulent.

How would you describe these products compared to other products in general?

The measure (ENVUNCT) was formed by computing the average response to these five items. Social
science researchers use this summated scale strategy to enhance the reliability of measures. Such a
summation is reasonable only if the items are unidimensional in a factor-analytic sense and are
internally consistent. Evidence of unidimensionality is provided by extracting all the principal
components in these data. The results showed that only the first principal component had an eigenvalue
greater than one, which indicates that one factor adequately describes the variance in the items.
Evidence of internal consistency was provided by computing Cronbach's alpha, and this estimate (0.73)
is well above the suggested 0.6 cutoff for basic research

5. BEHAVIORAL UNCERTAINTY

The second aspect of uncertainty addressed here is the difficulty of assessing performance. We measure
this variable by capitalizing on the notion that it is more difficult to ascertain adherence to contractual
agreements by downstream actors when critical points in the selling cycle are separated by relatively
longer periods of time. When transactions are performed instantaneously, the performance of each
party is more readily observable.

If the market reacts with a lag to actions taken, however, it becomes more difficult to attribute output to
efforts. Each survey respondent answered the following question:

What is the typical time between an initial contact concerning the product and the ultimate placement
of an order? _______ months.

This measure (BEHUNCT) measures the length of the selling cycle and indexes the difficulty of assessing
downstream performance.

6. SCALE VARIABLES

There are two scale variables included in the study. In the first of these, we asked the informant
to report the sales volume of the product line in question (SALES).
What is the annual sales volume of this product line? $________. The second scale variable
described the density of sales territories. The density of a sales territory refers to the geographic
concentration of customers. Denser territories make it possible to assign a company employee to
that territory and to use that person's time efficiently. In relatively sparser territories, it may not
be possible to exhaust the available time of company personnel assigned to that area. By
contrast, independent resellers are better able to exhaust scale economies by aggregating the
products of different firms. We measure territory sparseness as follows:
What percent of the typical field salesperson's time is spent on the following activities?
__________face-to-face selling
__________service, after sales support
__________paper work
__________travel time
__________other
100%
Our measure of (lack of) density (SPARSE) is the fraction of time reported for travel. Obviously,
more time is spent on the road when the density of the territory is lower. This is the same
indicator of density used by Anderson.

Asset specificity and measurement difficulty scales. Poppo, L., & Zenger, T. (2002). Do
formal contracts and relational governance function as substitutes or complements? Strategic
Management Journal, 23, 707–725.
Measurement
Questionnaire items, unless stated otherwise, were measured using a 7-point scale in which '1'
rep- resented 'low degree' and '7' represented 'high degree.' Table 1 presents the means and
correlations for each of the measures in the study.
Performance
When measuring exchange performance, most work in transaction cost economics focuses on
governance efficiency (Masten, Meehan and Snyder, 1991; Walker and Poppo, 1991; Uzzi,
1997; Artz and Brush, 2000). Yet, to incorporate both production and governance efficiency, we
examine overall satisfaction with exchange performance, rather than governance costs (see
Poppo and Zenger, 1998). This composite measure is consistent with previous measurements of
alliance performance found in the strategy literature (Mohr and Spekman, 1994; Saxton, 1997).
The underlying logic of our composite measure is that satisfaction is a focal consequence of a
working partnership. It is not only a close proxy for concepts such as perceived effectiveness, but
is also predictive of future actions by partner firm managers (Gladstein, 1984). At the same time,
however, by not assessing governance costs directly, we are constrained in our capacity to assess
optimal levels of governance. Thus, absent costs, more relational governance or contractual
complexity should be preferred.
1997; Artz and Brush, 2000). Yet, to incorporate both production and governance efficiency, we
examine overall satisfaction with exchange performance, rather than governance costs (see
Poppo and Zenger, 1998). This composite measure is consistent with previous measurements of
alliance performance found in the strategy literature (Mohr and Spekman, 1994; Saxton, 1997).
The underlying logic of our composite measure is that satisfaction is a focal consequence of a
working partnership. It is not only a close proxy for concepts such as perceived effectiveness, but
is also predictive of future actions by partner firm managers (Gladstein, 1984). At the same time,
however, by not assessing governance costs directly, we are constrained in our capacity to assess
optimal levels of governance. Thus, absent costs, more relational governance or contractual
complexity should be preferred.
Relational governance
In this study, we view relational governance as a composite factor with the following underlying
norms and dimensions: open communication and sharing of information, trust, dependence, and
cooperation. This specification is consistent with previous measurement (Macneil, 1978;
Anderson and Narus, 1990). We modified three indicators from these empirical studies and
asked the key informant to indicate their degree of agreement with each of the following
statements using a 7-point scale (1 = strongly disagree, 7 = strongly agree): (1) the buyer has an
extremely collaborative relationship with the vendor (RG1); (2) both parties share long- and
short-term goals and plans (RG2); and (3) the buyer can rely on the vendor to keep promises
(RG3) (Cronbach alpha= 0.78).
Contractual complexity
Following Macneil (1978), we measured the degree to which the parties created a complex con-
tract to deal with future contingencies by asking key informants to indicate their level of
agreement (Contract, 1 = strongly disagree, 7 = strongly agree) with the following statement: the
formal contract is highly customized and required considerable legal work. The measurement of
this construct by a single item is limiting. To further validate this measure, our survey also
requested respondents to indicate the length of the contract (in pages), which previous work has
shown as an indicator of contractual complexity (Joskow, 1988). While a lower response rate for
this question precluded its use as a second indicator of a complex contract, the correlation
between the two items was significant (p = 0.65).
Asset specificity
Firm-specific assets were defined by human assets, physical assets, and company-specific
routines and knowledge that were not redeployable to alternative uses (Williamson, 1985). Since
human capital is a critical component of information services, our measurement focused
primarily on specialized human assets, such as knowledge and skills. We used three items to
measure the degree to which the assets used to produce an information service were custom-
tailored to the buying firm: (1) To what degree must individuals acquire company-specific or
division-specific information to adequately per- form the IS function? (AS1); (2). To what
degree is your approach to this function (or set of applications) custom-tailored to the company?
(AS2); and (3) How costly would it be to switch outsourcing vendors? (Consider the time
required to locate, qualify, train, make investments, conduct testing, and develop a working
relationship) (SW) (Cronbach alpha = 0.83).
Measurement difficulty
Our measurement focused on the level of ease in measuring worker performance: To what
degree is it easy to measure the collective performance of those individuals who perform this
function? (Meas, 1 = very difficult, 7 = very easy). We reverse scored this item to create a
measure of measurement difficulty.
Technological change
In general, information services have had high levels of technological change. Still, there is
likely to be variation in the level of technological stability across industries, firms, and services.
We measured the degree of change in both skills and technology using two items: (1) To what
degree are the underlying skills associated with this IS function (or set of applications) rapidly
changing? (CH1); and (2) To what degree is the optimal configuration of hardware and software
required to perform this function (or set of applications) rapidly changing? (CH2) (Cronbach
alpha = 0.84).
Longevity of relationship
A key determinant of relational governance is the duration of time that the two parties have
worked with one another because a history of trade is necessary for the development of relational
norms (Macneil, 1978, 1980). We measured this construct using a 7-point scale (Long, 1 =
strongly disagree, 7 = strongly agree) and the following item: the buyer has worked with the
vendor for years and years. This measure is used as an identifying instrument in our equation
predicting relational governance.
Tenure and IS budget
We use measures of the firm's IS budget and the IS director's experience as identifying
instruments in the contractual complexity equation. We expect that firms with greater IS
resources will develop more customized contracts. In our study, top managers' experience in IS
proxies for firm knowledge about outsourcing. We expect man- agers will heighten the
complexity of contracts as experience increases. In addition, we expect firms with larger IS
budgets are also likely have in-house legal counsel with experience in drafting complex
contracts. We have no theoretical basis for believing either construct is related to relational-
governance. Our measure of resource scale is the natural log of size of the operating bud- get for
IS (Budget). Our measure of experience is the job tenure in years (ln) of the IS executive
(Tenure). These measures are used as identifying instruments in the equation predicting contract
complexity.
Controls
Because significant variance in industry exists for our sample of firms that outsource IS, we
control for potential differences that might exist in out- sourcing practices. We also use dummy
measures to control for the primary business in which the firm operates: banking/financial
services (Banking), manufacturing (Manu), and insurance (Insur).

Suggestions for Further Research

1. Examine the relative importance of various transaction dimensions and their impact on
organizational behaviors and governance choices.
2. Look at how firms align their transactions to accomplish multiple goals.
3. Explore how contracts are written that manage both governance and production problems
both internally and externally.
4. Explore situations and conditions where firms can transact successfully without the need
for governance structures.
Implications of the Theory for Managers
Transaction cost theory examines the importance of costs not directly related to the production of
goods and services. When you are budgeting for specific projects, be sure to include all of the
extra costs involved for you that are not directly related to producing your product or services,
such as searching, bargaining, monitoring, and enforcing your contracts with others. For
example, if you are buying a product, include the purchase price, but also include in your costs
the time you spent searching stores, comparing products, and making your purchasing decision.