Enforcement without Courts? Combining Formal and Relational Contract Governance

Abstract

Research into relational contracting – its attributes, its advantages over orthodox, “formal” methods of contracting, and its limitations – has long been constrained by the lack of scalable methods for measuring relational practices in applied settings. Using conventional social science methods, it is difficult, time-consuming and expensive to reliably discern nuanced differences in contracting behaviour for anything other than a small handful of cases. However, rapid advances in computational techniques, coupled with the growing availability of big data on government purchasing, raises the prospect of new, larger-scale research into relational contracting practices and outcomes. We begin to explore that potential by developing an automated approach to distinguishing formal and relational enforcement clauses in 10,174 contracts issued by Danish local and regional governments during 2021-2025. Formal enforcement uses the (threat of) sanction by third parties to reduce opportunism; relational enforcement rests upon norms, reputation, and the promise of future business. We develop, trial and refine a promptbook that enables an off-the-shelf large language model to identify formal and relational enforcement mechanisms in a large corpus of government contracts. Initial results suggest that formal enforcement clauses strongly dominate public contracts, and that enforcement design varies little with transaction hazards such as quality measurement problems and asset-specific investments.