Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
I show how improper conditioning of beliefs can reduce contribution in public goods environments with interdependent values. I consider a simple model of a binary, excludable public good. In equilibrium, provision of the public good is good news about its value. Naive players who condition expectations only on their private information contribute too little, despite the absence of free-riding incentives. In a laboratory experiment, contributions indeed fall short of the equilibrium prediction. Using modified games with different belief-conditioning effects, I verify that subjects fail to condition beliefs properly. However, improper belief conditioning cannot fully explain the results.