Document Type


Publication Date



Credence-goods experiments have focused on stylized settings in which experts can perfectly identify the buyer’s best option and that option works without fail. However, in nature credence goods involve uncertainties that complicate assessing the quality of service and advice. We introduce two sources of uncertainty into a credence goods experiment. The first is diagnostic uncertainty; experts receive a noisy signal of buyer type so might make an ‘honest’ mistake when advising what is in buyers’ best interests. The second is service uncertainty; the services available to the buyer do not always work. Both sources of uncertainty make detection of expert dishonesty more difficult, so are expected to increase dishonesty by experts and decrease buyer trust (willingness to consult experts for advice and to follow expert advice) – decreasing efficiency of the interactions. We also analyze how buyers use ratings and whether ratings restrain both dishonesty and distrust by creating reliable reputations. In contrast to predictions, we find that uncertainty decreases dishonesty and increases trust. Also in contrast to predictions, ratings do not improve efficiency of the transactions under uncertainty – in part due to buyers’ tendency to ‘shoot the messenger’ (give low ratings) when they buy service that does not work due to bad luck, and to give experts the ‘benefit of the doubt’ (high ratings) when they buy service that may have been intentionally over-provided (not in the buyer’s best interest).


ESI Working Paper 21-15



To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.