Our paper examines three related questions: Will directors who have friendship ties with the CEO manage earnings to benefit the CEO in the short term while potentially sacrificing the welfare of the company in the long term? Will public disclosure of friendship ties mitigate or exacerbate such behavior, and will disclosure of friendship ties influence investors' perceptions of director decisions? We conduct an experiment involving 56 active and experienced corporate directors from U.S. firms and a second experiment with M.B.A. students. We find that friendship ties caused directors to be more willing to approve reductions to research and development (R&D) expenses that cause earnings to rise enough to meet the CEO's minimum bonus target more often than when the directors and CEO were not friends. However, disclosing friendship ties resulted in even greater reductions in R&D expenses and higher CEO bonuses than not disclosing friendship ties. In a second experiment, we find that shareholders were more likely to agree with directors' decisions to approve cuts to R&D when friendship ties were disclosed. These findings have potentially important implications for corporate governance because they suggest that friendship ties between the CEO and board members can impair the directors' independence and objectivity, and that disclosure of the relationships can worsen this effect.