Influence of non-response in business tendency surveys on the properties of expectations

Expectations concerning key economic variables certainly influence decisions undertaken by economic agents. Since assumption of rationality forms the basis of neoclassical economic theory, question of whether expectations of industrial enterprises are indeed formed rationally deserves careful attention. Direct data on expectations are available mainly through business tendency surveys. Non-response problem is present in almost every survey, and much higher non-response rates are observed for expectations than for realizations. Weighting systems used to control for size of respondents may also introduce bias into expectations data derived from business tendency surveys. In this paper, we analyze two basic properties of expectations rational in sense introduced by J. F. Muth – that is, unbiasedness and orthogonality – taking into account issue of non-response and weighting schemes. We propose several sample balance statistics to correct for changing sample structure that results from non-response; it depends on time, and is different from general population. We find that rationality of expectations of Polish industrial firms is not sensitive to these factors: independently from non-response and weighting issues, expectations concerning relative changes in production remain unbiased but not efficient with respect to freely available information.