The principle of inconsistent trinity in the selection of procurement systems

Abstract

This paper examines the economic characteristics of construction procurement systems partially in terms of transaction cost economics and partly those of project management. It proposes the theoretical principle of an inconsistent trinity, according to which, in selecting a procurement system, the client will inevitably face a trade‐off between the procurement system best able to deliver: (1) fastest completion of the project (and therefore higher discounted present value of revenue from the project); (2) least vulnerability when making changes (lower transaction hold‐up costs arising from temporal or process specificity); as well as (3) least vulnerability to non‐performance (lower transaction measurement costs from greater visibility to third parties for ordering of disputes). The model provides a potentially testable contribution to a transaction cost theory of construction procurement. It also has two implications: (1) the three main family types of procurement systems are held to be intrinsically different in economic terms, so selection of procurement system does matter; (2) no system can enjoy absolute advantage over others, so the proposed strategy of procurement system selection is to align the characteristics of procurement systems with attributes of projects.

Keywords:

• Transaction costs
• hold‐up
• measurement costs
• opportunism
• temporal specificity
• procurement system

Notes

1. In this paper, ‘specification’, where it appears without qualification, refers to a performance specification. The detailed written instructions as to method or descriptions of the work to be carried out that accompany designs as the basis of a contract to provide works are referred to as ‘input specifications’.

2. Rationality is treated as only bounded in relation to the uncertain future. It would however be possible to develop a version of the model in which design errors are made and construction begins with incomplete design.

3. This is a ‘conservative’ assumption as regards the severity of the hold‐up problem, because it weakens the power of the contractor to disrupt progress, as compared to the alternative feasible assumption. That would be to treat each change as requiring a voluntary agreed variation in the terms of the contract. By refusing so to agree, a contractor could proceed with the original works, ignoring the change. This would greatly disrupt the process of realization of value from the project (Uff, 2005, p. 255).

4. Walker and Chau (1999) only consider a measurement problem, and regard the hold‐up problem as negligible because in construction plant and other fixed‐asset types of specificity are very limited; they only give a separate role to client requirements as opposed to design in the case where DB route is chosen; they relate ‘requirements’ to project physical characteristics, rather than value characteristics; and make the usefulness of statements of requirements a matter of the degree of detail achieved, rather than contractibility and visibility to a judge.

5. In this paper loss of value (v) rather than cost of making good or reinstatement (m) is taken as the measure of damage. This is because that seems to fit with the ‘black box’ assumption. By the time the building is finished, it will often be too late economically to replace and reinstate defective work. That is, although earlier v>c (the cost saved by shading quality), now at completion m (cost of reinstatement) >v (loss of value). Courts will then apply v (Uff, 2005, pp. 207–9).

6. The proof is as follows:

	• The contractor faces a choice between delivering X and delivering less than X (X−v). If the client accepts their delivery of less than X, this yields the better payoff for the contractor, because they obtain some production cost saving, c, thereby. However, if the client ‘fights’ by taking them to court, the contractor faces a payoff reduced (in comparison with their payoff from the client not fighting) by the sum of both the amount of damages that will be set, D, and their share of legal costs, Lβ.
	• The client will fight if it is rational for them to do so. If they don't fight their payoff is X−v. If they do fight, their payoff is (X−v)+D−Lα. Thus it will be rational for the client to fight if D>Lα.
	• Now, suppose for the moment that the court is able to set D = v. Suppose further that v>c.
	• For the contractor (if there is any positive probability that the court will award damages to the client), it is only profitable to deliver X−v if the client won't fight. This will only be the case if D<Lα.
	• Since we have assumed D = v, it follows that the credible and profit‐maximizing strategy for the contractor will be to ‘shade’ the contracted value by an amount v that is just less than Lα.
	• Knowing this, the value the client expects to obtain, and will accept without fighting, is X−Lα.

7. A non‐credible threat is one that can be ignored, on the assumption everyone acts in their rationally calculated self‐interest, because implementation of the threat leaves the opportunistic party worse off than if they backed down. A's threat to put his fist through B's window is non‐credible. Availability of a brick may make it credible.

8. This is not to say this problem is necessarily trivial in practice. In particular, if construction costs are highly uncertain, ∏_O−∏_R may become larger (as ∏_R becomes negative and large) than the amount of retentions. If this is combined with ambiguity of allocation of risks and responsibilities under the contract, so that Lα rises, a credible hold‐up threat may emerge. See Chang and Ive (2006), for the Channel Tunnel case.

9. Masten et al. seem to argue as if the amount of loss faced by a contractor who causes delay will be limited to some proportion to the price of their contract. They suggest that if the contract is for a small work package only, this thereby strengthens the bargaining power of the contractor, by creating a disproportion between the value of the loss they can inflict on the client, and the loss delay causes to the contractor. This is no doubt true in cash flow terms, and would be relevant in the absence of court resolution of disputes. But if the real measure of that loss faced by the contractor is the value of damages the client can recover from them, and if the client can show the ‘line of cause’ by which the delay in one package caused the loss in value to the whole project, there will be no difference between the damages obtainable from a contractor with a small and a contractor with a large contract.

10. Thus, in contrast to Masten et al. (1991) a work package contractor (under MAN) has less bargaining power over the client than a main contractor has.