Research in Experimental Economicss: Volume 8

In this study, 1 investigated the impact of pre-play, non-binding, face-to-face communication on contributions in continuous and provision point games. I find that such communication increases cooperation leading to greater contributions. This effect is found both in the short-term (i.e. periods with communication) as well as in the long-term (i.e. periods with no communication that follow periods with communication). Both the short-term and long-term effects are stronger in the provision point games. In addition, I find that the magnitude of communication has a negative relationship with contributions. Groups that meet more often to communicate and those that communicate more tend to contribute less. This is a preliminary result that must be investigated more fully in future research.

Many public goods may be characterized as having multiple provision points. These goods are provided in discrete rather than continuous quantities and only if specified minimal levels of funding are attained. This chapter describes an experiment that examines allocations to a multiple provision point public good. Subjects determine the level of the good provided using a voluntary contribution mechanism. We vary the number of provision points (from one to five), leading to increasing degrees of separation between the social optimum and the Nash equilibria. In the final rounds of single-provision-point sessions, efficient outcomes are frequently observed In contrast, efficient outcomes are only rarely observed in multiple-threshold sessions. In our first treatment, a second provision point is added, resulting in a decrease in the Nash equilibrium and no change in the efficient outcome. Contributions fall relative to the single provision point sessions, but remain above the lower provision point. We then add a third higher, provision point that increases the group optimum contribution but leaves the Nash equilibria unchanged. Surprisingly, contributions decline even more sharply in this treatment and are often below the lowest threshold. In all of our treatments, however, efficiency gains over full free-riding are observed in many rounds.

Alchian and Demsetz's (1972) classic paper models team production as a public good. They claim detection of individual effort levels, rather than aggregate effort levels, reduces shirking (free riding). This chapter experimentally tests this claim. Participants are informed either about the individual contributions of others on their team or only about their team's total contribution. Average group contributions in the two treatments are the same. However, group contributions under individual feedback have a significantly higher variance than those under total feedback. Implications of these results for team production are discussed.

Subjects in the role of multiproduct monopolists are observed in markets where demand is initially unknown. In each market period, the seller must choose between selling each good individually or selling a bundle containing one unit of each good. The fact that individual and collective reservation prices (demand) are initially unknown complicate this choice. As such information is paramount in determining the relative profitability of the two sales strategies, the multiproduct monopolist finds it profitable to search across various offer prices. A seller accumulates this information during a costless “practice” phase, then enters a ‘market” phase where profits or losses have pecuniary rewards. Results indicate that a multiproduct monopolist is often able to select the more profitable sales strategy.

This chapter compares four methods of rationing an input that is subject to supply capacity shocks. We describe the application of the mechanisms in terms of electricity. The mechanisms we test are random interruption, priority service, proportional service%all option and proportional service/put option. None of the mechanisms, as implemented, are first-best but can be ranked by their theoretical allocative efficiency. According to the theory, both versions of proportional service should produce the highest allocative efficiency whereas random interruption ranks lowest. In the laboratory, the two proportional mechanisms have radically different behavioral properties. The proportional/put option mechanism performed as predicted by theory. In spite of its theoretical superiority to random interruption and priority service, the proportional/call option mechanism performed no better than random interruption in the laboratory.