6Theory of Ownership
i. Theories of vertical integration are established to understand
capitalist institutions, such as the existence of firms with
ownership concentrated in the hands of one party or another,
alongside with market mechanism.
When applied to labor issues, these theories explain why
labor service is obtained through an employment relationship
(as in a factory system) rather than arms-length supply-
demand relationship, with the worker as an independent
contractor whose service, or, more precisely, product his
service, is purchased when it is needed (as in a putting-out
system). Nowadays, there are two ways for a computer
program writer to work for a firm: as an employee, or an
ii. Different approaches to understand capitalist institutions
a. Class interests (Marx)
b. Technology (neoclassic economics, e.g., Samuelson): the
firm is a production function. (Production cost economics.)
c. Transaction cost (in contrast to production cost)
economics sees the firm as a governance structure.
Capitalist institutions have the main purpose and effect of
economizing on transaction costs. The problem of
economic organization is contracting and contract
2. Early views on integrated organizations in free market economies. (Williamson, 1985, pp.2-4.)
a. Frank Knight (1922), Acknowledged the importance of studying human nature as we know it. He identifies “moral hazard” as an endemic condition with which economic organization must contend.
Problem: “moral hazard” was too technical and
misunderstood as a simple issue of purchasing and selling information. Coase (1952) argued that all advice or knowledge could be bought as required.
b. John R. Commons (1934), economic organization is not merely a response to technological features, e.g., economies of scale or scope, but often has the purpose of harmonizing relations between parties who are otherwise in actual or potential conflict.
Problem: The message made little headway against the prevailing view that the courts were the principal forum for conflict resolution. [Q: Why in the past economists tend to refer to conflict problems to the court and thus not discuss them, whereas nowadays they are more willing to discuss them?]
c. Ronald Coase (1937), whether transactions were organized within a firm (hierarchically) or between autonomous firms (in a market) is a decision variable. Which mode was adopted depended on the transaction costs that attended each.
Problem: Unless the factors responsible for transaction costs are identified, the reasons for organizing some transactions one way and others another would necessarily remain obscure.
d. Friedrich Hayek (1945), “the economic problem of
society is mainly one of rapid adaptation to changes in
particular circumstances of time and place.” This
emphasizes the importance of utilizing idiosyncratic
e. Kenneth Arrow (1959), “Traditional economic theory
stresses the sufficiency of the prices system as a source
of information, and this is correct enough at equilibrium.
In conditions of disequilibrium, a premium is paid for
the acquisition of information from sources other than
the prices and quantities” to which the firm has direct
Transaction cost is the cost of running the economic
3. Vertical integration in Oliver Williamson (1985).
i. The behavioral assumption
Bounded rationality: people satisfice, not optimize, due
to limited cognitive competence.
Self interest and opportunistic behavior.
iii. Critical dimensions for distinguishing among transactions
a. Behavioral assumptions
To understand opportunism, it is important to
distinguish between ex ante vs. ex post transaction costs.
Ex ante transaction costs: those incurred to draft, negotiate and safeguard (e.g., shared ownership, rule of law) an agreement.
Ex post transaction costs take several forms:
Maladaption costs incurred when transactions drift out of alignment in relation to the “shifting contract curve” (Aoki, 1983).
The haggling costs incurred if bilateral efforts are made to correct ex post misalignments.
The setup and running costs associated with the governance structure.
The bonding costs of effecting secure commitments.
b. General purpose vs. special purpose investments. While the latter is likely to be cost saving, the former is safer (more easily redeployed) should there be an interruption during the contract execution. “Do the prospective cost savings afforded by the special purpose technology justify the strategic hazards that arise as a consequence of their nonsalvageable character?” (Williamson, p.54.)
In accounting, a distinction is made between fixed and variable costs. For contracting purpose, it is important to distinct general purpose investment from special purpose investment, e.g., firm specific human capital.
4 types of asset specificity: Site, physical, human asset, and dedicated assets.
c. Uncertainty (informational problems)
Primary: State contingency.
Secondary (behavioral): What would the other party do?
iii. Contractual implications
Bounded Opportunism Asset Contract
Rationality specificity process
0 + + planning
+ 0 + Promise
+ + 0 competition
+ + + governance
0 = doesn’t exist; + = exists.
A cognitive map of contract can be drawn as a departure from competitive price mechanisms.
iv. The fundamental transformation.
a) Traditional thinking: The terms upon which an initial
bargain will be struck depend on whether
noncollusive bids can be elicited from more than one
qualified supplier. (Monopoly vs. competitive
b) Transaction cost economics fully accepts this
description of ex ante bidding. But it holds that a
condition of large numbers bidding at the outset does
not necessarily imply that a large number bidding
condition will prevail thereafter. Whether or not the
large number condition holds depends on specificity
of human or physical assets. If assets are relation
specific, “what was a large numbers biddings
condition at the outset is effectively transformed into
one of bilateral supply thereafter. This fundamental
transformation has pervasive contracting
v. Vertical integration as a solution.
The above analysis leads to the conclusion of vertical
integration. When market and contracting between
two independent parties cannot offer a satisfactory
and efficient solution to the moral hazard problem in
the bilateral relation between the two parties, the two
parties should merge and form a single firm.
What are some of the main problems in Williamson’s
theory of vertical integration?
Empirical challenge to the theory comes from Coase.
Theoretic challenge to the theory by
; Groosman and Hart (1986); and
; Hart and Moore (1988).
vi. Implication for employment relations.
a) Why is a stable relationship between labor and capital necessary? - It helps to alleviate various behavioral problems pervasive in arms-length transactions, e.g., pre-contract cheating (adverse selection), not making enough effort, not recognizing one’s specific investment, etc. This is efficiency enhancing.
b) Who should hire whom? Grossman and Hart (1986) studied this question.
4. A game theoretical model with relationship-specific investment (This example is from Gibbons, 1992)
Players: B wishes to purchase a unit of input from a seller S.
Timing of events:
； Date 0, B and S each make a specific investment, I & B
I, respectively. S
； Date 1, trade occurs.
； I , i = B, S, has only two possible values, H (for high) i
and L (for low). Specifically, assume H = $1.9 and L =
； The values of trade to B and S depend on the levels of
investment at date 0 and are summarized in the table
below. In the table, the first component refers to the
buyer’s benefit v, and the second to the seller’s cost c
(at date 1 – not to be confused with I at date 0). S
I = H I = L BB
I = H v=10 =9 vSb, 1b, 1
c= 6 c= 7 S, 1S, 1
I = L v= 9 =6 vSb, 1b, 1
c= 7 c= 10 S, 1S, 1
Exercise: Calculate net social surplus under each
SS(I = H, I = H) = ? Sb
SS(I = H, I = L) = ? Sb
SS(I = L, I = H) = ? Sb
SS(I = L, I = L) = ? Sb
； Note that the seller’s investment affects not only the
seller’s costs but also the buyer’s benefit, and vice versa.
For example, by developing a good seam, a coal mine
can provide better coal to an electric plant while
incurring a lower variable production cost itself.
Similarly, if the plant invests in a better boiler, it can
burn lower quality coal, thus reducing the seller’s costs.
； I is observable to B and S, but not verifiable and hence i
； v and c are private information.
； Whether B accepted or rejected the input at date 1 is
public (verifiable) information, and so is the price at
which the input is accepted (if it is accepted).
Exercise: Rank the involved variables by the quality of
； First best: I = I = H, which gives a total surplus of trade BS
10 – 6 – 3.8 = .2
； If I = I = L, no trade. BS
； If I = H and I = L, or vice versa, total surplus is BS
9 – 7 – 1.9 = .1
Contracts and first best outcome:
； If I is contractible, the first best can be achieved by
specifying I = H in the contract. Price for the input can be i
specified, too, and would be in the range
10 – 1.9 > p > 6 + 1.9. (Note: p=8 would work.) ； If quality and I have a one-to-one relationship, then the i
contract can also specify quality to achieve the first best.
Contracts and second best outcomes:
； The contract can specify an acceptance price p (if B 1
accepts the input, p is paid to S) and a rejection price p 10
(if B rejects the input, p is paid to the seller). 0
； Given p and p, the seller has 3 options: 10
1. I=L=0 (no investment), c=0 (no production of S
anything good), to receive p; 0
2. I=0, c=7, to receive p or p, depending on if B S01
accepts the input. (Note: total cost = 7 for S.)
3. I=1.9, c=6, to receive p or p, depending on if B S01
accepts the input. (Note: total cost = 7.9 for S.) ； S can always choose I = 0 and receive p (by making no S0
investment at date 0, and producing nothing good at date 1,
since neither I nor c is verifiable). For S to choose I=1.9 SS
and c=6, it’s necessary that
p – 6 – 1.9 > p, ？ p – p > 7.9. (1) 1010
； Similarly, B can always choose to 1) invest I = 0, and 2) B
pay p to reject the input and receive –p for payoff. For B 00
not to do so, it’s necessary that
10 – 1.9 – p > - p, ？ p – p < 8.1. (2) 1010
； But when inequalities (1) & (2) are both satisfied, S can
still choose I= 0, c = 7 if S expects B to set I= H = 1.9. S B
When I= 0 and I= H = 1.9, the value of the input to the S B
buyer is 9 > p– p. => B will accept the satisfactory input. 1 0
； When S is confident that B will accept the input, S will
choose I= 0, c = 7 because p – 7 > p – 6 – 1.9. S 11
； The equilibrium of I = I = H is disrupted, meaning SB
that the first best (socially optimal) solution is lost.
Exercise: Is there a Nash Equilibrium of B’s and S’s moves?
Find it, if your answer is “yes”.
； Note that, when it chooses I= H = 1.9, S incurs total cost s
6+1.9 > 7. So it always prefers I= L = 0, as long as he s
can be sure B will accept the product. B will indeed
accept the product since 9 > 8.1, which occurs when I = B
H = 1.9.
； Additional total cost=0.9, while the additional benefit=1.
This means there is a social loss due to underinvestment.
But, S is the one to incur the cost and B is the one to enjoy
the benefit. Hence the inefficiency.
； The result is sensitive to the v & c values, e.g., if without
I=H, c is very large (relative to that with I=H), then S SS
will choose I=H. S
； In this case, there is no informational asymmetry between
S and B. Informational asymmetry is found only between
the “insiders” and “outsiders”. However, it is adequate to
render court an ineffective means for contract enforcement.
If an agreement is court enforceable, which requires I to
be verifiable, the problem would go away.
； Without court enforcement, the problem is avoided if B
and S are under the same ownership.
5. Ownership allocation (Grossman and Hart, 1986; Hart and
i. Grossman and Hart had the following comments on Coase (1937), Klein et al. (1978) and Williamson (1979)