Subcontracting and Competitive Bidding on Incomplete Procurement Contracts

[Job Market Paper]

abstract

This paper investigates the cost implications of contractual incompleteness and its effect on subcontracting decisions in the bridge construction industry. Construction contracts are incomplete because the original blueprints and specifications may require modifications during the course of construction. According to the transactions cost theory of the firm -- Coase (1937), Williamson (1985) -- such contract revisions can lead to significant bargaining and renegotiation costs. Furthermore, theory predicts these costs are larger if a subcontractor performs the work. Forward looking contractors anticipate these costs and incorporate them in their bids. I develop an empirical framework to quantify the impact of incompleteness on cost for both integrated and non-integrated transactions and apply it to 32 bridge contracts procured by the California Department of Transportation. Contracts contain many work items (e.g. casting concrete, drilling, traffic striping). For each item, contractors decide whether to perform work themselves or hire a subcontractor and submit a bid. The difference between the work item quantity specified in the original contract and the quantity installed per final contract specifications proxies for incompleteness. In estimation, I account for the strategic aspects of bidding to recover cost from bids and exploit the panel data structure to account for the endogeneity of subcontracting decisions. On average, incompleteness explains a small portion of cost, 2%, for integrated transactions and a large portion, 13%, for non-integrated transactions. The results provide quantitative evidence in support of incomplete contracting theories of the firm and have practical significance for evaluating procurement practices.