When No Bad Deed Goes Punished: Relational Contracting in Ghana and the UK

Experimental evidence to date supports the double theoretical prediction that parties transacting repeatedly punish bad contractual performance by reducing future o ffers, and that the threat of punishment disciplines opportunistic breach. We conduct a repeated gift-exchange experiment with university students in Ghana and the UK. The experiment is framed as an employment contract. Each period the employer makes an irrevocable wage o ffer to the worker who then chooses an e ffort level. UK subjects behave in line with theoretical predictions and previous experiments: wage o ffers reward high e ffort and punish low eff ort; this induces workers to choose high eff ort; and gains from trade are shared between workers and employers. We do not find such evidence among Ghanaian subjects: employers do not reduce wage o ffers after low e ffort; workers often choose low e ffort; and employers earn zero payo ffs on average. These results also hold if we use a strategy method to elicit wage o ffers. Introducing competition or reputation does not signi ficantly improve workers' e ffort. Using a structural bounds approach, we find that the share of sel fish workers in Ghana is not substantially di fferent from the UK or earlier experiments. We conclude that strategic punishment in repeated labor transactions is not a universally shared heuristic.