A Laboratory Investigation of Verification and Reputation Formation in a Repeated Joint Investment Setting

Two managers invest funds in either a division-specific or common project. Managers are privately informed of their own division-specific return ratio and issue forecasts to their partner. Experimental parameter values are chosen such that: (1) investment in the division-specific project constitutes a strictly dominated strategy and (2) socially efficient investment requires truthful forecasts and dictates that (only) the manager with the lower rate of return invest in the common project. This creates a role for ex post verification, the main manipulation in the experiment. In addition, random and continuous pairing of subjects are used. With randomly pairing settings, verification significantly increases honest forecasting and efficiency. With continuous pairings, verification improves coordination of investment strategies, but has only a muted effect on efficiency. Even without verification, continuous pairings increases efficiency. These results indicate that verification and repeated pairings act as substitutes in improving forecast quality and efficiency. Key words: Experimental economics; Verification; Reputation; Cooperation