An Agent advises a Decision-maker (DM) on a project choice. There are n ≥ 2 alternatives and a status quo or default option. Alternative i yields benefit bi ≥0 to both the Agent and the DM, whereas the status quo yields value b0 ≥0 to the DM but no value to the agent. Hence, the agent prefers any alternative over the status quo, but has congruent preferences with the DM among the set of alternatives. The Agent knows the vector (b1; : : : ; bn) privately, while the DM has only limited information based on hard or observable characteristics of each alternative. Specifically, it is common knowledge that each bi is drawn from an interval in R+ with cdf Fi, where different observable components for alternatives i and j are captured by Fi ≠ Fj .
Applications include a Hiring Committee advising a Dean on a potential hire in their academic field, a Divison Manager requesting resources from an Executive Committee for a new project, a Consultant pitching a consultancy project to a potential Client, or a Supplier advising a potential Buyer on an experience or credence good.
Communication, modeled as cheap talk, is essentially equivalent to the Agent making a recommendation that the DM choose a particular project and the DM then deciding whether to follow this recommendation or stick with the status quo. We show that whenever the status quo value is not too low, equilibrium communication is generally characterized by inefficient pandering: the Agent distorts his recommendation towards “stronger-looking” projects, and these projects are more likely to be accepted by the DM. At the same time, “weaker-looking” projects become more sellable when favorably pitched in comparison to a stronger-looking slate of alternatives.
Various organizational responses to the above pandering are studied, all of which require some commitment power from the DM: (i) At the interim stage -having observed the hard information about all alternatives- the DM is better off reducing the value of her outside option (the statusquo) or committing to follow the Agent’s recommendation (rubberstamping/complete delegation), so long as she would not always implement the status quo in the communication game.
(ii) Optimal interim commitment does not involve rubberstamping/complete delegation. Sometimes, optimal commitment can be implemented by delegating authority to a third party, who puts a positive but lower value on the status quo than the DM.
(iii) Ex-ante ignorance-not observing the hard information about alternatives-can be optimal when interim commitment is not possible. It is suboptimal, however, if the DM can wait to commit until after she has observed the hard information.