Output Expectations and Forecasting of UK Manufacturing

Abstract

This paper analyses both quarterly data from the Confederation of British Industry (C.B.I.) Survey on respondents’ expectations of recent and forthcoming manufacturing output, and monthly Office of National Statistics (O.N.S.) figures on actual manufacturing output within the U.K. Quarterly output expectations of the C.B.I. manufacturers are explained from the monthly O.N.S. observations using a bounded rationality approach. The logistic formulation models the diffusion process across respondents. There is a backward-looking C.B.I. Survey perspective, explained by past O.N.S observations, and a forward-looking perspective, explained from future O.N.S. statistics. Also, the forecasting of monthly manufacturing output from earlier values along with the quarterly C.B.I. Survey information is examined, and tested against the alternative Pesaran/Thomas method. The study provides econometric evidence for the validity of the logistic model, and shows that bounded rationality can explain the formation of predictions among business managers in the U.K. manufacturing sector. The emerging consensus from the literature, supported by this paper, is that the logistic format is a superior ‘approximation’ to the true data generating process compared with the earlier standard Anderson/Pesaran/Thomas approach. An adjustment to the Survey is used, which achieves perfect symmetry with ‘up’ and ‘down’ versions of the data. The benefits of this adjustment are tested in the forecasting Section.