Nonresponse bias and business turnover rates: The case of the characteristics of business owners survey

This article considers the problem of nonresponse in the Characteristics of Business Owners survey. It presents a novel approach for assessing the extent to which nonresponse is nonignorable. The method exploits the fact that all of the owners of multiowner firms were surveyed. The article uses information received from a responding owner to derive inferences about a nonresponding owner of the same firm. It provides evidence that nonrespondents are more likely than respondents to discontinue ownership. The article estimates a variety of models of the joint process of ownership survival and survey response and quantifies the effects of ignoring nonresponse.