The organizational failures framework: Information Impactedness

Information impactedness is a derivative condition that arises mainly because of uncertainty and opportunism, though bounded rationality is involved as well. It exists when true underlying circumstances relevant to the transaction, or related set of transactions, are known to one or more parties but cannot be costlessly discerned by or displayed for others.

It will be useful, for the purposes of understanding the information impactedness condition, to distinguish between buyer, seller, and arbiter to the transaction. Also, ex ante information impactedness, which exists at the time of the original negotiations, should be distinguished from ex post information impactedness, which develops during the course of contract execution. The occasion to engage an arbiter appears only m conjunction with the ex post information impactedness condition.

The relation of information impactedness to first-mover conditions ought also to be emphasized. The reason why outsiders are not on a parity with insiders is usually because outsiders lack firm-specific, task-specific, or transaction-specific experience. Such experience is a valuable resource and can be used in strategic ways by those who, by being awarded initial contracts, have acquired it.

1. General

It is generally conceded that if information is asymmetrically distributed between the parties to an exchange, then the exchange is subject to hazards. As Arrow explains it:         . . the critical impact of information on the optimal allocation of risk bearing is not merely its presence or absence but its inequality among economic agents” ( 1969, p. 55). I submit, however that: (1) it is not merely asymmetry alone but asymmetry coupled wit (a) the high costs of achieving information parity and (b) the proclivity of parties to behave opportunistically that poses the problem; (2) information problems can develop even when parties have identical information and, a fortiori, if information differences exist; and (3) the distribution of information between the parties is of special concern in small-numbers bargam-ing contexts.

The last proposition is reasonably obvious. Consider therefore a small-numbers exchange relation in the context of (2) above, where both buyer and seller have identical information, and assume, furthermore that this informations is entirely sufficient for the transaction to be com-pleted. Such exchange might nevertheless experience dificulty if, despite identical information, one agent makes representations that the true state of the world is. The problem develops here because human agents are given to making opportunistic representations if the arbiter can discern whitch party is lying, and to what degree, only by incurring the expense associated with making independent observations. Consequently, only when buyer, seller, and this information is adequate, can one say with confidence tha the transaction will go through without difficulty.

Next, consider a variant of this situation where all parties, includin the arbiter, have identical information, but the information is incomplete, and the parites have  failed to fully specify, ex ante, both the rules for in-ferring, onthe basis of incomplete information, what state of the world obtains and the rules by which additional observations are to be taken. Contractual ambiguities can plainly develop since parties can make ex post representations in favor of mappings which favor their interests. Costly haggling then results.

Suppose, instes, that the parties have different information but that, were the information possessed by both parties to be candidly revealed, the resulting information pool would be sufficient for the exchange to be completed. Suppose, furthermore, that neither party has an information “advantage” at the outset. Nevertheless, problems can develop which even an agreed upon mapping fails to rectify. Opportunism here takes the form of selective disclosure or distortion of the data to which each party uniquely has access-which, provided again that it is costly for an arbiter to be inde-pendently apprised as to what true conditions exist, is to be anticipated.

The problems are compounded if an information asymmetry condi-tion exists. One of the agents to the contract has deeper knowledge than does the other, and it is costly for the party with less information to achieve information parity. Selective disclosures or distortions are then all the more hazardous for the party who is at an information disadvantage. I emphasize, however, that the problems posed by information asymmetry differ in degree but not in kind from those posed when information sets are identical yet incomplete or, while different, neither party can be said to enjoy a strategic information advantage.

In any event, information impactedness need not impair market exchange if ( 1) the parties are not opportunistic, (2) an unbounded rationality condition were to obtain, or (3) a large-numbers competition condition prevails — both presently and prospectively. If, however, all of these conditions are violated, a shift of the transaction from market to hierarchy may obtain an account of the above indicated advantages (see Section 2.3) of hierarchy in curbing opportunism.

2. Some Examples


Problems can develop under the identical but incomplete information condition if the exchange is made contingent on which state of the world eventuates. Unless the parties have fully stipulated how ex post signals are to be mapped into state of the world descriptions, differences in opinion as to which state actually has obtained can be anticipated. That is, even though both parties have identical information with respect to the condition of the environment, they need not agree on what state of the world has actually been realized. To the contrary, if the consideration to be exchanged varies conditionally with which state of the world is agreed (or otherwise declared) to have eventuated, each party can be expected to make opportunistic mappings in support of outcomes favorable to himself.

Thus, suppose that party A agrees to supply party B with X¯, on date d, if the mean temperature on date d -1 is less than or equal to T0, and X¯ + Δ otherwise. Suppose also that both parties A and B have free access to temperature readings on date d—1 at 4:00 A.M., 12:00 noon, and 8:00 P.M. If on date d the unit-weighted average of the temperatures on the preceding day is well below or well above T0, the transaction goes through without difficulty. Suppose, however, that the unit- weighted average is slightly less than T0, while weights of 0.95, 1.10, and 0.95 would increase it above T0. Party B may now assert that everyone knows that the noon temperature deserves to be assigned a greater weight in computing the daily mean, and that + Δ should be delivered accordingly. Party A objects, thus haggling and the need to collect additional information ensue.14 Moreover, in circumstances where the state of the world is multidimensional, the occasion for such disputes to arise naturally increases.


Recall the insurance example in Section 3 of Chapter 1. Efficient riskbearing is impeded by information impactedness conditions in both ex ante and ex post respects. Ex ante information impactedness exists if. as is usually the case, insureds know better than insurers what their true risk characteristics are. Moreover, good risks cannot easily distinguish themselves from poor risks because their representations cannot be taken at face value: poor risks can make the same claims. Because it is costly to establish what the true risk attributes of the parties are, information asymmetries and opportunistic proclivities on the part of poor risks combine to yield the adverse selection problem.

The moral hazard problem is also the result of an information impactedness condition, in this instance of the ex post variety. Given that a party has been awarded insurance coverage, the question of efficient loss mitigation is posed: Will the insured take the appropriate steps to reduce his exposure to risk, or will he reallocate his assets away from loss mitigation in favor of other activities? The latter is to be expected if insurers are unable easily to determine whether losses are due to negligence or to environmental circumstances. Insureds then can simply disclaim negligence, whether it was a contributing factor or not. Again, information asymmetries and opportunism combine to yield the problem.


First-mover advantages have been referred to above and will appear in several of the chapters which follow. The basic phenomenon is this: Winners of initial contracts acquire, in a learning-by-doing fashion, nontrivial information advantages over nonwinners. Consequently, even though large-numbers competition may have been feasible at the time the initial award was made, parity no longer holds at the contract renewal interval. The information acquired through experience is impacted in the sense that (1) original winners may refuse to disclose it (which is a manifestation of opportunism) or (2) they may be unable, despite best efforts, to disclose it (because of bounded rationality of the language impeded variety). Small numbers bargaining situations thus evolve in this way. Markets frequently give way to hierarchies on this account.

What is sometimes referred to as “knowhow” is an interesting illustration of a first-mover advantage, where knowhow refers to an unpatented and probably unpatentable technology and is to be distinguished from trade secrets in that knowhow “suggests a continuing flow of information and data . . . [ whereas trade secret] suggests a single transfer of one secret or a package of secrets with no continuity of flow” (Eckstrom, 1963, p. 127). Knowhow implies a condition of information impactedness, as the following statement by the U.S. Justice Department, in asking relief in the United Shoe Machinery case, suggests (emphasis added):

Affirmative relief from United’s . . . control of the present technology and technicians in the Shoe Machinery field can only be accomplished by releasing to potential customers United’s complete knowhcrw. in as current a form as it is possible to do. This means putting United’s potential customers in possession of detailed models and blue-prints for the construction of those machines themselves. It means the release of all construction and operating manuals, rules and notices relating to machines, including the general layout plans that United has been in the habit of making available to shoe manufacturers. It envisages the making available of technical personnel to iron out kinks in the production and operating processes and to give to the potential competitors that intuitive knowledge based upon training and experience that is incapable of translation into written form. The employees of United’s potential competitors need to be given as much and as detailed help as United s own employees, if not more.

Such first-mover advantages play an important role in several of the chapters which follow.

3. Internal Organization

Internal organization helps to overcome information impactedness in several respects. Most of the reasons are traceable to the advantages of internal organization over markets in opportunism respects — as developed in Section 2.3. above, Thus, internal organization (of the appropriate kind)17 serves to attenuate incentives to exploit information impactedness opportunistically. Also, the superior auditing powers of internal organization help to overcome information impactedness conditions.

The language advantages of internal organization also ought to be noted. This is significant in two respects. First, parties that are willing to disclose information to which they have selective access will find that this is easier if an efficient internal code has developed which permits idiosyncratic conditions to be communicated with little difficulty. The previously noted coding advantages of internal organization, in relation to markets, thus help to overcome information impactedness.

Second, internal organization is apt to be superior to market organization in experience-rating respects. Parties who might otherwise make opportunistic representations — with regard, for example, to their qualifications to perform a task or supply a component — will be discouraged from so doing if they face the prospect of being experience-rated. Since it is generally considered efficient that well-qualified and poorly qualified types be sorted out quickly and the appropriate discriminating wage or price paid to them, institutional modes that have better experience-rating characteristics are favored, ceteris paribus.

This is not to suggest that markets perform no experience-rating functions whatsoever. As Leff ( 1970, pp. 26-36) and others have urged, business reputation is a valuable resource and firms sometimes share contractual experiences by informally, and sometimes formally, pooling information. The extent to which such information-sharing is efficacious varies between markets: intermediate goods markets, in which transactions take place between firms, are generally more well developed in experiencerating respects than are final goods markets, in which transactions take place between households and firms ( Leff, 1970, pp. 29-33 ).

Although Leff restricts his comparisons to market-mediated exchange, experience-rating differences between market and internal organization also exist and are more germane to our interests here. The communication advantages that Leff imputes to intermediate product markets, in relation to final goods markets, are, I submit, even more well developed within the firm. For one thing, the experience rater and the decision maker will ordinarily be one and the same person — in which event there is no need for an interpersonal, much less interorganizational, communication to occur at all. The decision maker simply consults his own experience with regard to the transaction in question and decides accordingly. Second, because of the previously noted coding advantages of internal organization, subtle nuances within the firm may be impossible to achieve interorganizationally. Again, the volume of communication is reduced, this time on account of language economies.

Third, interfirm experience rating can be risky. Firm A may tell rival firm B that firm X is “possibly O.K.,” when in fact it knows firm X to be slightly substandard.” If it relies on these representations, firm B subsequently is put to a rivalry disadvantage. Moreover, inasmuch as the criteria for assignment to one performance category rather than another are not well defined, since firm A cannot be compelled candidly to document its rating, and as firm B’s use of X may vary slightly from firm A’s, firm B has little basis upon which to register a complaint, much less bring litigation.

Finally, even if firm A accurately reports firm X to be substandard, firm B will lack a depth account of why such an assignment was made. B may then conjecture that A’s problems were partly of its own making — for example, that a more carefully drawn contract would have avoided the difficulties. If B’s management considers itself to be more clever, the report may be discounted. Since, by contrast, depth accounts are more easily secured with respect to internal transactions, internal disbelief is less likely.

The upshot is that, however well-developed experience-rating is in interfirm contracting respects, intrafirm evaluations are apt to be even more refined. Both horizontal and vertical integration may occur for this reason.

Source: Williamson Oliver E. (1975), Markets and hierarchies: Analysis and antitrust implications, A Study in the Economics of Internal Organization, The Free Press.

Leave a Reply

Your email address will not be published. Required fields are marked *