Quantitative inference from qualitative business survey panel data: a microeconometric approach

Business survey data are used widely as they offer timely information about the state of the economy. This paper addresses the problem of how best to infer a quantitative signal about economic growth from qualitative business survey data. A method drawing on the forecast combination literature is derived for producing quantitative best linear unbiased inference from qualitative panel survey data. This involves aggregating firm-level quantified estimates according to their reliability. We illustrate how the approach can be used to derive early estimates of official output growth data in the UK in an application to the panel of firm-level responses from the CBI's Industrial Trends Survey.