Impact of Product Characteristics and Market Conditions on Contract Type: Use of Fixed-Price Versus Cost-Reimbursement Contracts in the U.S. Department of Defense

ABSTRACT:

Transaction cost economics is used to produce a conceptual framework that helps explain public-sector contract decisions. When a product is easy to specify, easy to produce, and there is a thick market of buyers and sellers, fixed-price contracts are more likely; when a product is difficult to produce, difficult to specify, and the market has few buyers and sellers, cost-reimbursement contracts are more likely. These arguments were tested with five years of data (FY 2004–2008) drawn from the Federal Procurement Data System (FPDS), the most comprehensive and largely untapped database on federal contracting practices. Over 2,000 Defense Department contracts were examined, charting contract type (i.e., fixed-price vs. cost-reimbursement) across simple and complex products. The results confirm conventional wisdom about public sector procurement practice, at least within the Defense Department: product characteristics and market conditions drive the use of fixed-price and cost-reimbursement contracts. This research lays the foundation for research on contract outcomes by identifying factors that drive the use of different contract types and producing unique measures of product characteristics that are not currently available in the literature.

KEYWORDS:

• cost-reimbursement contracts
• fixed-price
• market conditions
• product characteristics

Notes

We use the generic term “product” to refer to both goods and services.

See https://www.usaspending.gov. Over the last decade, federal contract spending rose from $346 billion in FY 2004 to a high of $541 billion in FY 2008. Contract spending has averaged around $450 billion over the last two fiscal years.

For example, the 2007 Report of the Acquisition Advisory Panel to the Office of Federal Procurement Policy and the United States Congress finds: “The federal acquisition workforce is an essential key to success in achieving the government’s missions. Procurement is an increasingly central part of the government’s activities” (U.S. Acquisition Advisory Panel, 2007, p. 352).

The FAR is a component of the Code of Federal Regulations; specifically, Title 48: Federal Acquisition Regulation System.

Obviously there will be circumstances when a buyer prefers a cost-reimbursement contract, and a seller, a fixed-cost contract, but all things being equal, we assume that buyers are likely to select contract types that keep costs low, and sellers are likely to select contract types that augment prices.

Resource dependency theory makes powerful arguments about how the power of one organization over another in a relationship influences the structure of the relationship (Pfeffer & Salancik, 2003; Provan, 1993). We acknowledge the theory’s insights but elect not to draw on it in crafting our framework, given that resource dependency theory focuses on ongoing interdependent interactions between parties, whereas the contractual-type decisions we examine typically occur before the exchange occurs and a relationship commences. In instances where contractual exchanges become interdependent, resource dependency theory offers a potential complement to transaction cost theory (Burt, 1983).

Federal contracting practice requires the purchasing government to determine the contract type before going to market (Brown, 2013).

We provided random numbers only to initial contracts, not to all contract actions in our sample, so that the probability of being selected is the same across all contracts in the population. If we had provided random numbers to all contract actions on a list, contracts with more modifications would have been more likely to be selected.

Because contract managers can modify contracts through post-award negotiations with vendors, the initial contract does not always reflect the contract that ultimately governs the exchange. For this reason, in creating our sample we tracked both initial contract decisions and modified contract decisions. While contract managers in our sample often changed contract elements, like contract duration, we found no instances in which the type of contract, whether cost-reimbursement or fixed-price, was changed after the initial agreement.

The percentage of incomplete FPDS records is 15.5% (269 of 1,734). The incomplete rates for the DoD are on par with the overall rate of incompletion at 15.2%.

The survey was conducted with a protocol approved through a university’s institutional review board.

We pretested the survey instrument on a sample of 38 federal procurement personnel from a different professional association, the National Institute of Governmental Purchasing (NIGP). The pretest provided useful feedback on ways to improve the validity of the survey items for the constructs we are interested in.

About 20% of respondents neglected to provide biographical information. We are unable to determine whether they differed from the respondents who did provide information, but have no reason to think that they would differ in any systematic way.

Non-DoD respondents tended to rate products higher on the scales we describe below, but the differences were not statistically significant.

“Requirements definition” is a commonly accepted term in procurement, referring to the process of writing down a product’s attributes and capabilities.

Appendix 2 reports the wording of the survey prompt for each of the two measures.

There is a growing body of research on complex contracting (e.g., Malatesta & Smith, 2014) and product complexity in public sector procurement (e.g., Brown et al., 2010, 2013).

In separate analyses, we also include a variety of ex post measures of market competitiveness and concentration. These variables include the total number of bids, whether there was a single offer or agencies went out for competitive bids, and the percentage of the vendor’s total revenue for a given year represented by the contract (this last measure is drawn from Dun & Bradstreet’s Million Dollar Database). None of these measures is statistically significant or influenced the results for the variables described above. We suspect that this is because all of the variables measure information revealed after contract type is set.

Appendix 3 reports the mean product complexity rating and distribution of fixed-price and cost-reimbursement contracts for each product in our sample.

See http://www.defense.gov/contracts/contract.aspx?contractid=3206 and http://www.defense.gov/contracts/contract.aspx?contractid=5018 (accessed October 6, 2013)

This analysis is based on a logistic regression that added the production stage dummy variable to our original model. The dummy variable is statistically significant and positive, and the results for the other independent variables remain largely unchanged—the magnitude of the coefficients decreases slightly, and the competition variable is no longer significant at the 0.05 level. We opt to incorporate this analysis in the discussion rather than in our core theoretical argument because our proxy for stage of development—the dummy variable—is crude. The coding of the dummy variable is based on NAISC product categories rather than detailed contractual information on whether the purchase is for development of the product or routine production of the product.

Semi-structured interviews were conducted with seven current and former senior procurement personnel across a variety of federal agencies. All of those interviewed had served in a variety of acquisition positions in different agencies. Each interview followed the same semi-structured protocol, was recorded by hand, and then coded and scanned for key terms. Each interviewee was promised anonymity under a university institutional review board process.

Additional information

Notes on contributors

Yong Woon Kim

Yong Woon Kim is an Assistant Professor at Sang-Hur College, Konkuk University.

Alex Roberts

Alex Roberts is a Ph.D. candidate at the John Glenn College of Public Affairs, Ohio State University, Columbus, Ohio.

Trevor Brown

Trevor Brown is a Dean and Professor at the John Glenn College of Public Affairs, Ohio State University, Columbus, Ohio.