In an era where government is tasked to “do more with less”, it increasingly contracts out complex services even though evidence on cost savings is mixed.  As service complexity increases, the potential cost for managing contracts also increases.  The public management literature on contracting recommends developing time and resource intensive relational contracts to reduce uncertainty that contractors will perform and/or refrain from opportunistic behavior if unanticipated events occur.

Missing from the contract literature is a consideration of how structural properties may influence performance. I draw on exchange theory and network theory to develop arguments about how structural properties of ex post contract markets affect individual contractor performance. In exchange theory, power is relational and is the inverse of dependence.   When government contracts with more than one provider for the same service it creates an ex post contract market, where it has multiple provider options and providers must compete (either actively or passively) for clients.  Government may be able to rely on structural properties of contract markets (i.e. – competition among providers for clients) to induce performance; contractors must perform well in order to continue receiving clients. I predict that as government increases its number of contractors in the market, performance of all contractors will increase.

Data for this paper come from government funded residential programs for juvenile delinquents that are privately produced by both nonprofit and for-profit contractors. I analyze performance of over 100 residential programs nested in six distinct contract markets over a five-year period using hierarchical linear models (HLM). Performance measures are provided by the government funder and include cost, quality and effectiveness dimensions.  Cost and quality dimensions are related to contract maintenance and renewal while effectiveness is long-term measure that relates to the overall mission of the public organization to reduce juvenile recidivism.

Findings suggest that contractors perform better on measures related to contract maintenance and renewal when government has acceptable alternative contractors for service production. This finding is stronger for for-profit programs than nonprofit programs.  The finding is weaker for the long-term effectiveness measure, which is not directly linked to contract maintenance or renewal. These findings are important because they suggest that competitive pressures are important for performance in the ex post contract market.  Understanding how the structure of exchange in contract settings influences performance is important because it will help government understand where to strategically deploy resources for increased monitoring or relationship development with contractors to increase performance rather than applying the same approach to all contractors.  Additionally, I find a statistically insignificant correlation between measures related to contract maintenance and renewal with measures related to long term effectiveness.  This suggests that government measures- and rewards- outputs not related to its overall goals.  This finding warrants further investigation.