Chapter 5

Organizing Vertical Boundaries: Vertical Integration and Its Alternatives

Chapter Contents
1) Introduction
2) Technical Efficiency versus Agency Efficiency
 The Technical Efficiency/Agency Efficiency Tradeoff and Vertical Integration
 Real-World Evidence
Example 5.1:    The Virtual Corporation
3) Vertical Integration and Asset Ownership
Example 5.2:    Vertical Integration of the Sales Force in the Insurance Industry
4) Process Issues in Vertical Mergers
5) Alternatives to Vertical Integration
 Tapered Integration: Make and Buy
Example 5.3:    Tapered Integration in Gasoline Retailing
 Strategic Alliances and Joint Ventures
 Collaborative Relationships:    Japanese Subcontracting Networks and Keiretsu
Example 5.4:    Inter-Firm Business Networks in the United States:    Women's Dress Industry in New York
City
 Implicit Contracts and Long-Term Relationships
6) Chapter Summary
7) Questions
                Chapter Summary

This chapter examines why vertical integration differs across industries, across firms within the same
industry, and across different transactions within the same firm.    The first part of this chapter assesses the
merits of vertical integration as a function of the industry, firm, and transactions characteristics.    It then
presents evidence on vertical integration from studies of a number of specific industries including
automobiles, aerospace, and electric utilities.

The next section examines whether there might be other factors apart from those discussed in Chapter 3
and 4 that might drive a firm’s decision to vertically integrate.    This chapter focuses on two alternative
classes of explanation.    One is theory offered by Sanford Grossman and Oliver Hart that stresses the role
of asset management and its effect on investments in relationship specific assets as a key determinant of
vertical integration.    The second theory analyzes “market imperfections” motivations for vertical
integration.

The final purpose of this chapter is to explore other ways of organizing exchange besides arm’s length
market contracting and vertical integration.    The alternatives discussed include:    1)    tapered integration
(i.e. making and buying) 2)    joint ventures and strategic alliances 3)    Japanese keiretsu, and 4)    implicit
contracts supported by reputational considerations.
 Approaches to teaching this chapter

Definitions
Agency Efficiency: The extent to which a firm’s administration and/or production costs are raised due to transaction
and coordination costs of arm’s length market exchanges or agency and influence costs of internal organization.

Technical Efficiency:    The extent to which a firm uses least-cost production techniques.

Tapered Integration:    A combination of vertical integration and market exchanges. Under this process, the
firm makes some quantity of the input internally and buys the rest from independent firms.

Strategic Alliance:    Two or more firms agree to collaborate on a project or to share information or
productive resources.

Joint Venture:    A particular type of strategic alliance where two or more firms create and jointly own a
new independent organization.

Keiretsu:    Close semiformal relationships among buyers and suppliers, best embodied by the Japanese.

Implicit Contract:    Unstated understanding between parties in a business relationship.

Overview
This chapter covers different segments, which can be taught separately. The first section is an extension of
Chapters 3 and 4. This section discusses the trade-offs that go into a decision of whether or not to
vertically integrate. Figures 5.1 and 5.2 illustrate theses trade-offs and yield three powerful conclusions
about the relationship between the attractiveness of vertical integration and the conditions of input
production and the product market scale of the firm.

 Scale and scope economics: A firm gains less from vertical integration the greater the ability of
outside market specialists to take advantage of economics of scale and scope relative to the firm.

 Product market scale and growth: A firm benefits more from vertical integration the larger the scale of
its product market activities.

 Asset Specificity: A firm benefits more from vertical integration when production of inputs involves
investments in relationship-specific assets.

At the end of this section real world firms are examined to see if they behave according to this theory;
much evidence suggests that they do.

The next section introduces an alternative theory, developed by Sanford Grossman and Oliver Hart, for
comparing vertical integration with market exchange. The main point of this section is that in comparing
vertical integration with market exchange, it is important to understand who has the residual rights of
control. Residual rights of control are defined as the rights of control that are not explicitly stated in the
contract. If one thinks of vertical integration as the changing rights of control over key assets, it affects
the efficiency of exchange. The economic value created is quite different and it has an important effect on
the efficiency of transactions. The Appendix to this chapter presents an algebraic illustration of this
theory.    The third section discusses the instances where companies might want to vertically integrate due
to market imperfections. It examines situations where vertical integration is aimed at undoing the effects
of market imperfections or increasing market power. This section may be skipped without loss of
continuity.
 The last section discusses alternatives to vertical integration and arm’s length market exchange. It
introduces four alternatives, namely: tapered integration, joint ventures, the Japanese Keiretsu and long-
term implicit contracts. The main point is that some economic trade-offs are useful in thinking about
decisions and desirability of these organizations.

Technical vs. Agency Efficiency

The relationship between Technical and Agency efficiency is one of the main themes of this chapter.
Figure 5.1 illustrates the trade off between Technical and Agency efficiency and is related to the
discussion in chapters 3 and 4 about the benefits and costs of using the market. Agency vs. trade-off
concepts should be explained by means of graphs because MBA students find these concepts difficult.
 
T = cost of producing the activity in-house – cost of producing the activity by a market specialist,
assuming that both produce as efficiently as possible.

 "Transactions" cost when activity is performed by a market specialist is defined broadly to include:

 direct costs of negotiating contracts costs of safeguards against hold-up

 costs of safeguards against hold-up.
 inefficiencies due to under-investment in relationship-specific assets and lost opportunities for cost
savings due to mistrust.
 costs associated with breakdowns in coordination and synchronization when activity is purchased.

A = "transactions" cost when activity is produced in-house – "transactions" cost when activity is
provided by a market specialist.

"Transactions" cost when activity is produced in-house include:

 agency and influence costs that arise when hard-edged market incentives are replaced by soft-edged
incentives of internal organization.

C = T + A = the full cost difference between vertical integration and reliance on market specialist.
When  C is positive, in-house production is more costly than reliance on market specialists; when  C is
negative, in-house production is less costly than reliance on market specialists.

Managerial Implications
Below are four general rules for managers of when to rely on the market and when to perform tasks in-
house.

1. Rely on market for routine items; produce in-house items that require specific investments in
design, engineering or production know-how.
 Asset specificity is high      ---      vertical integration is best.   
 Asset specificity is low        ---      reliance on market is best

2. Rely on market for items that require large up front investments in physical capital or
organizational capabilities that outside firms already have.

 When outside specialist benefit from significant economies of scale, reliance on the market is
best.
 When in-house division can capture nearly all economies of scale in activity, vertical integration
is best.

3. Vertical integration is usually more efficient for bigger firms than for smaller.

 Larger scale of product market activities      ---    vertical integration is best.

 Smaller scale of product market activities    ---    reliance on market is best.

4. Technological advances in telecommunications and data processing have tended to lower

coordination costs, making reliance on the market more attractive.

 Low coordination costs      ---      reliance on market is best.

 High coordination costs      ---      vertical integration is best.
Suggested Harvard Case Studies

Hudepohl Brewing Company HBS 9-381-092 (see chapter 2)

Pepsi-Cola Beverages HBS Case 9-390-034 A (see chapter 3)
Philips Compact Disc Introduction, HBS 9-792-035 (see chapter 3)
Tombow Pencil Co. Ltd. HBS 9-692-011 (see chapter 3)
Extra Readings
The sources below provide additional resources concerning the theories and examples of the chapter.

Anderson, E. and D.C. Schmittlein, "Integration of Sales Force: An Empirical Examination," RAND
Journal of Economics, 15, Autumn 1984, pp. 385-395.

Arrow, K., "Vertical Integration and Communication," Bell Journal of Economics, 6, Spring1975,
pp.173-182.

Borenstein, S. and R. Gilbert, "Uncle Sam at the Gas Pump: Causes and Consequences of Regulating
Gasoline Distribution," Regulation, 1993, pp. 63-75.

Caves, R. and D. Barton, Efficiency in US Manufacturing Industries, Cambridge: MIT Press, 1990.
Chandler, A.D., Jr., Scale and Scope: the Dynamics of Industrial Capitalism, Cambridge,
MA: Belknap,1990.

Clark, R., The Japanese Company, New Haven: Yale University, 1979.

Davidow, W.H. and M.S. Malone, The Virtual Corporation, New York: Harper Business, 1992.   

Grossman, S. and 0. Hart, "The Costs and Benefits of Ownership: A Theory of Vertical and Lateral
Integration," Journal of Political Economy, 94, 1986, pp.691-719.

Joskow, P., "Vertical Integration and Long-Term Contracts: The Case of Coal-Burning Electric
Generating Plants," Journal of Law, Economics and Organization, 33, Fall 1985, pp. 32-
80.

Klein, B., "Vertical Integration as Organizational Ownership," Journal of Law, Economics, and
Organization, 1989 pp. 199-213.

Machlup, F. and M. Taber, "Bilateral Monopoly, Successive Monopoly, and Vertical Integration,"
Economist, 27, 1950, pp. 101 -119.

Masten, S., "The Organization of Production: Evidence from the Aerospace Industry," Journal of
Law and Economics, 27, October 1984, pp. 403-417.

Masten, S., J.W. Meehan and E.A. Snyder, "Vertical Integration in the U.S. Auto Industry: A Note on
the Influence of Transactions Specific Assets," Journal of Economic Behavior and
Organization, 12, 1989, pp. 265-273.

McKenzie L.W., "Ideal Output and the Interpendence of Firms," Economic Journal, 61, 1951,
pp.785-803.

Monteverde, K. and D. Teece, "Supplier Switching Costs and Vertical Integration in the Automobile
Industry," Bell Journal of Economics, 13, Spring 1982, pp. 206-213.
Ohmae, K., "The Global Logic of Strategic Alliances," Harvard Business Review, March-April 1989,
pp.143-154.

Perry, M.K., "Forward Integration by Alcoa: 1888-1930," Journal of Industrial Economics, 29, 1980.
pp.37-53.

Perry, M., "Price Discrimination by Forward Integration", Bell Journal of Economics, 9, 1978,
pp.209-217.

Peters, T., Liberation Management. Necessary Disorganization for the Nanosecond Nineties, New
York:Knopf, 1992.

Spengler, J.J., "Vertical Integration and Antitrust Policy," Journal of Political Economy, 53,
1950,pp.347-352.

Wallace, D.H., Market Control in the Aluminum Industry, Cambridge: Harvard University Press,
1937.

Williamson, O.E., Antitrust Economics, Oxford, UK: Basil Blackwell, 1987, chaps. 1 through 4, for a
discussion of the role of transactions cost considerations in antitrust policy.

Williamson, O., "Strategizing, Economizing and Economic Organization," Strategic Management

Journal, 12, Winter 1991: 75-94.
Answers to End of Chapter Questions

1. Why is the “technical efficiency” line in Figure 5.1 above the x-axis?    Why does the
“agency efficiency” line cross the x-axis?

The technical efficiency line represents the minimum cost of production under vertical integration minus
the minimum cost of production under arm’s-length market exchange.    The technical efficiency line is
above the x-axis indicating that the minimum cost of production under vertical integration is higher than
the minimum cost of production under arm’s length market exchange, irrespective of the level of asset
specificity.    The reason costs are higher under internal organization is that outside suppliers can aggregate
demands from other buyers and thus can take better advantage of economies of scale and scope to lower
production costs.

The agency efficiency line represents the transactions costs when production is vertically integrated
minus the transactions costs when it is organized through an arm’s-length market exchange.    The agency
efficiency line is positive for low levels of asset specificity and negative for high levels of asset
specificity.    When asset specificity is low, the probability of hold up is low.    In the absence of a
significant risk of hold up, market exchange is likely to be more agency efficient than vertical integration
because independent firms often face stronger incentives to innovate and to control production costs than
divisions of vertically integrated firms (see chapter 3).    Agency costs of market exchange increase with
the level of asset specificity—above some threshold of asset specificity, using the market entails higher
transactions costs than using an internal organization.

2.    Explain why the following patterns seem to hold in many industries:                                                         

a. Small firms are more likely to outsource production of inputs than are large firms.

Figure 5.2 illustrates that a firm gains less from vertical integration the greater the ability is of outside
market specialists to take advantage of economies of scale and scope relative to the firm itself.    A small
firm might not be able to take advantage of the economies of scale and scope because its level of
production would not offset the significant, up-front setup costs or meet the demands of a large market
outside the firm.    A large firm might be able to produce a sufficient level of output and achieve the same
economies of scale and scope that an outside firm would have.

b. “Standard” inputs (such as a simple transistor that could be used by several electronics
manufacturers) are more likely to be outsourced than “tailor-made” inputs (such as circuit
board designed for a single manufacturer’s specific needs).

“Standard” inputs are more likely to be outsourced than “tailor-made” inputs for two reasons.    First, an
outside firm can reach economies of scale and scope if it is producing standard inputs for several different
electronic manufacturers.    An outside firm might not be able to produce a higher level of inputs to take
advantage of economies of scale and scope with tailor-made inputs because it is only producing for one
firm, not several firms.    By increasing the scale of production, outside firms will be able to take
advantage of decreased fixed and exchange costs.
 Second, when outsourcing tailor-made inputs, it is difficult to monitor performance and quality.   
Furthermore, specifying performance and quality expectations in a contract is usually not feasible.    The
outside firms would incur higher transactions costs.    In addition, the increase in design specifications and
complex components would accentuate the advantages of vertically integrating, according to the asset-
specificity hypothesis.    For example, Scott Masten’s study of the aerospace industry (chapter 5)
highlights the make-or-buy decision associated with producing complex components and encountering
potential hazards related to incomplete contracting.    See Chapter 5 for further details about incomplete
contracting.

3.      Use the arguments of Grossman and Hart to explain why stock brokers are permitted to keep
their client lists (i.e. continue to contact and do business with clients) if they are dismissed from
their jobs and find employment at another brokerage house.

Assuming incomplete contracts, Grossman and Hart’s argument is summed as follows:

 Ownership rights determine the resolution of the make-or-buy decision.    The owner of an asset has
residual rights of control not specified in the contractual agreement.
 When ownership is transferred, these residual rights are purchased.    They are lost by the selling
party, which fundamentally changes the legal rights of that party.
 There are two types of decisions: contractible (verifiable operating decisions) and noncontractible
(unverifiable up-front investments in relationship specific assets).      When negotiations break down,
control over operating decisions reverts to the party with residual right of control over relevant assets.
 The power in the relationship is determined by the nature of the vertical integration.
 Ownership should be given to the unit whose investments have the strongest impact on the total
profits of the venture.   
A brokerage firm’s assets are tied to each of its clients, and clients are the largest source of total profits of
the firm.    The nature of the relationship between a stock-broker and client is relationship-specific.    The
broker has ownership of each of his/her clients.    The broker is investing the time and information into
developing these relationships.    Therefore, the broker has the ownership of those assets.    Accordingly,
the broker has residual rights to the ownership and control of clients.    When ownership of the broker
changes, the broker’s relationship-specific assets, the selling party, the original brokerage firm, loses the
clients.    While ownership of specialized physical assets can be transferred, ownership of specialized
human capital can’t be.    The brokerage house simply provides products.    It is the broker who sources
and services their clientele.

The nature of the product (stock trades) dictates that brokerage houses allow the stock brokers to keep
client lists.    Customers often have an incentive to switch their brokerage business to obtain better value
for money.    This is particularly true for clients doing high volume of trades.    In these circumstances it
becomes important for the broker to generate extra effort to keep the client.    If the broker has the
ownership of the asset (client list), he is not subject to 'hold-up' by the brokerage house.    If he did not
own the client list, the brokerage house could force the broker to accept lower commissions.    This would
reduce the incentive for the broker to try hard to generate business for the brokerage house.

4. Analysts often array strategic alliances and joint ventures on a continuum that begins with
“using the market” and ends with “full integration”.    Do you agree that these fall along a
natural continuum?

No.    Alliances and joint ventures fall somewhere between arm’s-length market transactions and full
vertical integration.    As in arm’s length market transactions, the parties to the alliance remain
independent business organizations.    But a strategic alliance involves much more coordination,
cooperation, and information sharing than an arm’s length transaction.    A joint venture is a particular
type of strategic alliance in which two or more firms create, and jointly own, a new independent
organization.    The firms involved with a joint venture integrate a subset of their firms assets with the
assets of one or more other firms.

5. What does the Keiretsu system have in common with traditional strategic alliances and
joint ventures?    What are some of the differences?
Keiretsu are closely related to subcontractor networks, but they involve a somewhat more formalized set
of institutional linkages.    Usually the key elements of the vertical chain are represented in a keiretsu.   
The firms in a keiretsu exchange equity shares, and place individuals on each other’s boards of directors
—U.S. firms are in many cases prohibited from these practices under the U.S. antitrust laws.   
6. The following is an excerpt from an actual strategic plan (company and product names
have been changed to protect the innocent):
Acme’s primary raw material is PVC sheet that is produced by three major vendors within the United
States.    Acme, a small consumer products manufacturer, is consolidating down to a single vendor.   
Continued growth by this vendor assures Acme that it will be able to meet its needs in the future.
Assume that Acme's chosen vendor will grow as forecast.    Offer a scenario to Acme management
that might convince them that they should rethink their decision to rely on a single vendor.    What
do you recommend that Acme do to minimize the risk(s) that you have identified?    Are there any
drawbacks with your recommendation?

Acme must have a valid reason for using a single supplier.    A key merit of Acme's approach is asset
specificity on the vendor's side, when the vendor might be encouraged to make specific investments.

However, there is the potential of hold-up problems created by decreasing the number of vendors down
from three to one.    The likelihood of hold-up problems would depend on the degree of asset specificity
and switching costs.    The amount of implicit and explicit costs related to hold-up problems would
depend on the level of inventory, customers' expectations and brand reputation.    For example, if Acme
can only produce its product with Company X's component and Company X ceases shipments, unless
Acme has sufficient amount in inventory of the components, Acme will incur significant losses
immediately due to lack of sales.
In order to minimize its risks, Acme could second source or backward integrate (partially or fully).   
Second sourcing or backward integration protects the firm from hold-up but may drive costs by reducing
purchasing discounts, increasing production coordination and requiring additional investments by the
second source.    An alternative solution would be for Acme to consider tapered integration.    Tapered
integration poses the usual make-or-buy problems, such as high costs, coordination problems and poor
incentives.
7.    Shaefer Electronics is a medium-size producer (about $18 million in sales in 1993) of electronic
products for the oil industry.    It makes two main products - capacitors and integrated circuits.   
Capacitors are standardized items.    Integrated circuits are more complex, highly customized items
made to individual customer specifications.    They are designed and made to order, they require
installation, and sometimes require post-sale servicing.    Shaefer's annual sales are shown in the
table below.
Shaefer relies entirely on manufacturers' representatives (MR's) located throughout the United
States to sell its products.    MR's are independent contractors who sell Shaefer's products in
exchange for a sales commission.    The company's representatives are not exclusive - they represent
manufacturers of related but non-competing products such as circuit breakers, small switches, or
semiconductors.    Often a customer will buy some of these related products along with integrated
circuits or an order of capacitors.    MR's have long experience within local markets, close ties to the
engineers within the firms that buy control systems, and deep knowledge of their needs.    In the
markets in which they operate, MR's develop their own client lists and call schedules.    They are
fully responsible for the expenses they incur in selling their products.
Once the MR takes an order for one of Shaefer’s products, Shaefer is then responsible for any
installation or post-sale servicing that is needed.
Shaefer recently hired two different marketing consultants to study its sales force strategy.    Their
reports contained the conclusions reported below.    Please comment on the soundness of each
conclusion.
a. "Shaefer should continue to sell through MR's.    Whether it uses MR's or an in-house sales
force, it has to pay sales commissions.    By relying on MR's, it avoids the variable selling
expenses (e.g., travel expenses for salespeople) it would incur if it had its own sales force.   
As a result, Shaefer's selling expenses are lower than they would be with an in-house sales
force of comparable size, talent, and know-how."
The conclusion of this marketing consultant is unsound - it is make-or-buy fallacy #1 - the fallacy states
that a firm should buy rather than make because by doing so it avoids setup and production costs
associated with making.    In this particular case, Shaefer does not avoid variable selling expenses by
relying on MR's.    Presumably the MR’s will negotiate a commission rate that will allow them (over some
reasonable period of time) to cover their variable selling expenses, such as travel costs.    This rate will be
higher than the rate that Shaefer would pay to otherwise equally talented, knowledgeable and productive
in-house sales people, whose selling expenses will be reimbursed directly by Shaefer, rather than
indirectly through the commission.
b. "Selling through MR's made sense for Shaefer when it was first getting started and
specialized in capacitors.    However, given its current product mix, it would not want to set
itself up the way it is now if it were designing its sales force strategy from scratch.    But
with what it has got, Shaefer should be extremely cautious about changing."
Initially, Shaefer utilized outside sales representatives to market its standardized products in 1980.   
Historically, firms relied on independent marketing agents to sell and distribute fairly standardized
products.    This strategy made sense because the selling process did not require specialized know-how or
expertise to sell, according to the asset-specificity hypothesis.    However, there are limits to economies of
scale and scope in selling and distributing customized products.
If Shaefer was developing its sales force strategy in 1993, it would make sense for Shaefer to establish an
internal sales force from scratch.    Shaefer is selling standardized and customized products which do not
hold similar sales strategies.    MR's are probably not the best way to sell customized product, such as the
ICs.    First with the increase in capital intensity, independent wholesaling and marketing agents are no
longer able to benefit from the economies of scale and scope.    Consistent with the asset-specificity
hypothesis, Shaefer should begin to forward integrate into marketing and distribution, especially when
dealing with customized products requiring specialized investments in human capital or in equipment and
facilities.
Another factor against having the MR's selling the ICs is coordination costs.    Under the current
arrangement, an MR, who then has to communicate with Shaefer’s design and manufacturing people to
develop the specs for the customer’s order, makes the sale.    And after the sale has been completed, the
MR would have to arrange with Shaefer for installation and post-sale servicing.    Using MR’s sales staff
may compromise coordination between design, manufacturing, sales and installation and post-sale
customer service.    This situation leaves Shaefer vulnerable to competitors who can perform the order
taking, design manufacture, delivery and installation more quickly than Shaefer.
In addition, the salesperson's personal relationship with the customer is particularly important for
developing a long-term sales strategy.    In this "relationship-specific" situation, Shaefer will need to
provide post-sale servicing and potentially, sell future products.    If the sales process begins with an MR,
a customer's primary loyalty may belong to the MR, rather than to Shaefer.    The MR's "ownership" of the
client list would provide it with considerable bargaining power over Shaefer which would make it
difficult for Shaefer to: (1) keep costs and sales commission low; (2) enforce sales quotas; and (3) control
the time the MR's spend pushing Shaefer's products versus other products.    We don’t know for sure that
any of these problems are actually occurring, but one slight bit of evidence that may be suggestive of a
problem is that since 1990, Shaefer’s sales of integrated circuits have grown less rapidly than the industry
as a whole would predict.    Using contracts and/or commission schedules would assist in monitoring
performance; however, this arrangement would become relatively more costly.    Therefore, when
transactions involve relationship-specific assets, "making" is often preferred to "buying."
Overall, in the IC business, Shaefer is not just selling a physical product; it is selling its capabilities to
custom design, produce and install a device that will solve a particular need for a prospective customer.   
An in-house sales person is likely to have more specific knowledge of Shaefer's ability to adapt its
designs to solve particular customer needs than an MR who would be unable to provide similar and
extensive details due to its various products.    While it is true that MR’s know a lot about their clients'
needs, and probably know something about the capabilities of the manufacturers, the key question is: is
the MRs' knowledge about Shaefer's products sufficient enough to allow Shaefer to compete effectively
against rivals who are aggressively pushing their abilities to solve specific customer problems.
Of course, this line or argument begs the question of why Shaefer could not provide training for its MR's
to enhance their knowledge about Shaefer's products.    A possible answer relates to the earlier discussion
of asset specificity.    Providing an MR with training about the capabilities and technological
specifications of Shaefer's products creates relationship-specific assets that further cement the dependence
of Shaefer on its manufacturer reps. The possibility that its MR’s might engage in hold-up reduces the
marginal value of any investment Shaefer might make in its relationship with them.    Thus, while
providing training to the MR's may seem like a good idea in principle, it is far from clear that its benefits
outweigh the costs.
The above considerations would argue against using MR’s in the first place when starting a sales force
from scratch.    However, note that if Shaefer needed to decide whether to alter its current situation, then
other considerations would affect Shaefer's decision.    Primarily, the irony of relationship-specific assets
is that they can create dependency relationships that are difficult or costly to break.    Despite the potential
hold up problems for higher commissions and working less with MR's, the up front costs of developing
relationships with clients and learning how to sell Shaefer's products are less than starting an in-house
sales force from scratch at this time.    The level of attractiveness of MR’s is higher at this current point
than it would have been if Shaefer were designing its sales force strategy from scratch.