Understanding Social Interaction Effects in the Workplace: An Experimental Approach

Field evidence suggests that people’s work morale depends on those of their co-workers. We are interested in understanding the behavioural logic of such “social interaction effects”. We design a three-person gift-exchange experiment where workers can observe each others’ effort levels and adjust their own effort if they so wish. Our design avoids the “reflection problem” (Manski 1993) which plagues identification of social interaction effects. The one-shot nature of our experiment also excludes any strategic incentives for changing efforts. We nevertheless observe many adjustments, most of them down to the lower effort level. We show that efforts are strategic complements rather than strategic substitutes as predicted by various models of social preferences. Conformism seems to be the dominant motive behind the social interaction effects we observe. * Acknowledgments: This paper is part of a research project “Soziale Interaktionen, Unternehmenskultur and Anreizgestaltung”, financed by the Grundlagenforschungsfonds of the University of St. Gallen. We are grateful for very helpful comments by Martin Kolmar, Bentley MacLeod, Rupert Sausgruber, Karl Schlag, Jean-Robert Tyran, as well as by the participants of the Verein für Socialpolitik conference in Bonn, the meeting of the Swiss Society of Economics and Statistics, the IZA Bonn Workshop on Behavioural Labour Economics, the ESA meeting in Rome, and research seminar participants in Innsbruck and Nottingham. a University of St. Gallen, FEW-HSG, Varnbüelstrasse 14, CH-9000 St. Gallen; [email protected]. b IZA, CESifo and University of Nottingham, School of Economics, Sir Clive Granger Building, University Park, Nottingham, NG7 2RD, United Kingdom; [email protected].