The determinants of unit non-response in the Ifo Business Survey

It is well-known that non-response affects the results of surveys and can even cause biases due to selectivities if it cannot be regarded as missing at random. In contrast to household surveys, response behaviour in business surveys is rarely examined in literature. This paper analyses a large business survey at a microdata level for unit non-response. The data base is the Ifo Business Survey, which was established in 1949 and is completed by more than 5000 firms every month. The panel structure makes it possible to use statistical modelling with the inclusion of different types of time dimensions, as well as firm-specific effects. The results show that there are strong time-dependent effects on the response rate and that non-response is more frequent in economic good times.