11, Porte des Sciences
Using a finitely repeated game, we study whether fraudulent announcements of high returns of investment by project managers can be mitigated by reputation and by the introduction of standard market mechanisms when feedback is imperfect. In our laboratory experiment project managers announce to potential investors the likely return of their funds. Announcements are cheap talk and while some categories of lies can be detected ex post by investors, other remain deniable. We find that reputation (in the sense of fixed matching) reduces the relative frequency of extreme and detectable lies but cannot reduce the frequency of deniable lies. Instead of encouraging more honesty, market mechanisms lead project managers to make more risky lies to attract investors. Reputation and the associated threat of punishment in competitive markets reduce fraud but cannot eliminate the negative effect of competition on ethics.