Suppliers are passive instruments in this model. They are indifferent among contracts, since their expected profits are the same (zero) whichever choice the buyer makes. What drives the argument is that buyers can secure better terms only by relieving producers of demand cancellation losses. Buyers cannot have their cake (product supplied by the efficient technology at a price of p) and eat it too (cancel without cost). Price reductions are not awarded gratuitously.
Inasmuch as optimality is realized if h = k and a = 1, the ideal hostage would appear to be an offer of generalized purchasing power: money. A security bond in amount h = k would serve this purpose. The reason the argument does not terminate here is that such an arrangement does not assuredly engage the interests and cooperation of the supplier. Three reasons can be adduced for this condition: contrived cancellation, uncertain valuation, and incomplete conttacting. All are a consequence of joining bounded rationality with opportunism.
1. Contrived Cancellation
The issue of contrived cancellation has been addressed by Kenneth Clarkson, Roger Miller, and Timothy Mûris in their discussion of refusal of the courts to enforce stipulated damage clauses where breach has been deliberately induced (1978, pp. 366-72). Induced breach could arise where a party intentionally withholds relevant information, yet complies with the letter of the contract. Or it might involve perfunctory fulfillment of obligations where more resourceful cooperation is needed (pp. 371-72). In either case, induced breach is costly to detect and/or prove (p. 371).
This explanation for selective enforcement of liquidated damage clauses has troubled other legal scholars (Posner, 1979, p. 290), but a more satisfactory explanation has yet to be advanced. At the very least, the Clarkson et al. treatment reflects a sensitivity to the subtleties of opportunism—on which account private ordering is more complicated than the bare bones hostage model would suggest. Among other things, the expropriation hazard to which they refer may explain the use of ugly princesses.
Thus suppose that demand uncertainties are negligible, whence order cancellation hazards can be disregarded. Suppose further, however, that buyers differ in credit risk respects and that producers would, if they could, refuse sales to poor risks. Assuming that the difference between good and poor risks-is sufficiently great that a separating equilibrium is feasible,” producers could demand hostages (or, put differently, good risks could offer hostages) as a way by which to screen. Given, moreover, that the only use to which hostages are put is as a screen, a value of a= 0 would accomplish that purpose without exposing the buyer to an expropriation hazard (based, say, on a legal technicality). Specifically, a king who i\ known to cherish two daughters equally and is asked, for screening purposes, to post a hostage is better advised to offer the ugly one.
2. Uncertain Valuation
The model assumes that the value of the specific investment (k) is well specified. This need not be the case. Indeed, it may be difficult for buyers to ascertain whether the investments made in response to first period orders are of the amount or the kind that producers claim. That is not a serious problem if the production side of the market is competitively organized and fly-by- night concerns can be disregarded. Where, however, this cannot be presumed, the possibility that buyers will be expropriated arises. Producers may feign delivery competence (claim to have invested in specific assets in amount k but only committed k’ < k) and expropriate bonds for which h = k by contriving breach or invoking a technicality.
The hazard is especially great if the producer, who retains possession of the assets for which specificity is claimed, can preserve asset values by integrating forward into the buyer’s market upon taking possession of the hostage. Even though the producer is poorly suited to performing successor stage functions, the possession of specialized stage I assets effectively reduces the costs that would otherwise attend de novo stage II entry.
To be sure, the buyer who offers a hostage and recognizes a risk of contrived expropriation will adjust the original terms to reflect the risk. Spe- ci ically, contracts supported by hostages for which expropriation risks are beheved to be great will command less than those where the same hazards are believed to be lower. But this is to concede that, absent additional safeguards, neit er the transfer of product on marginal cost terms nor the efficient level an m of investment will assuredly attend contracts of type III. Deeper governance issues than those comtemplated by the simple model are evidently posed.
3. Incomplete Contracts!Haggling
Complex contracts are invariably incomplete, and many are maladaptive. The reasons are two: Many contingencies are unforeseen (and even unforeseeable), and the adaptations to those contingencies that have been recognized for which adjustments have been agreed to are often mistaken—possibly because the parties acquire deeper knowledge of production and demand during contract execution than they possessed at the outset (Nelson and Winter, 1982, pp. 96-136). Instrumental gap filling thus is an important part of contract execution. Whether it is done easily and effectively or, instead, reaching successive agreements on adaptations and their implementation is costly makes a huge difference in evaluating the efficacy of contracts.
Thus even if contrived breach hazards could be disregarded, producers who are entirely open and candid about contract execution may nevertheless be in a position to haggle—thereby to expropriate sellers—because contracts are incomplete or maladaptive. Specialized governance structures that have the purpose and effect of promoting harmonious adaptations and preserving the continuity of exchange relations arise in response to that condition. Use of knowledgeable third parties (arbitration) and reciprocal exposure of specialized assets are two possibilities.
4. Safeguards in Kind
The above-described difficulties with pecuniary bonding do not alter the proposition that buyers can purchase product from suppliers on better terms if they offer assurances than if they do not. But it raises the possibility that assurances will take forms other than bonding.
Assume that the buyers in question are not merely conduits but incur production and distribution expenses before making deliveries to customers. Assume further that buyers have access to two technologies, one fully general purpose while the other requires investments in specific capital that has value only in conjunction with servicing final demands for the product in question. Assume finally that the redeployable costs (v) of the special purpose technology are lower than those of the general purpose technology.
It then follows that suppliers will sell product at a lower price to buyers whose investments in sales and service are more specific than it will to those whose specific investments are less—even if no hostage exchange is made upon cancellation of an order. That is because such buyers will thereafter confirm orders in more adverse demand states than those who do not. Put differently, buyers who drive their redeployable costs down by making trans-action-specific investments present the supplier with a more favorable de- mand scenario that those who do not. In this sense, forward specific invest- ments constitute a credible commitment. The resulting contract will not, however, yield the efficient marginal cost pricing resuk. That will come about only if compensation is paid to suppliers upon order/cancellation.
Such a defect does not imply that hostage transfers should always attend unilateral trading. Such a rule suffers from the aforementioned expropriation hazard. More generally, all feasible trading alternatives may be flawed. A comparative institutional assessment of the main organizational alternatives would presumably include consideration of the following:
- Full compensation upon order cancellation, in which event buyers are exposed to expropriation hazards.
- Buyers invest in specific assets but refuse compensation, which cre- ates a more favorable demand scenario but still exposes suppliers to a (reduced) risk of uncompensated losses.
- As a compromise, suppliers create credible commitments and make partial but incomplete hostage payments upon order cancellation.
- The contractual relation is expanded by developing suitable reciproci-ty arrangement.
- The transaction is consolidated under common ownership, which is the vertical integration alternative.
Whether options 4 and 5 are feasible will vary with the circumstances. Reciprocity is examined in Chapter 8, while the tradeoffs that attend vertical integration were examined in Chapter 4. Here as elsewhere, informed choice among complex alternatives requires detailed knowledge of the institutional realities of economic life (Koopmans, 1957, p. 145). The attributes of the trading parties, the technologies to which they have access, and the markets in which they operate all have to be assessed.
Source: Williamson Oliver E. (1998), The Economic Institutions of Capitalism, Free Press; Illustrated edition.