Investment Forecasting with Business Survey Data

Business investment is a very important variable for short- and medium-term economic analysis, but it is volatile and difficult to predict. Qualitative business survey data are widely used to provide indicators of economic activity ahead of the publication of official data. Traditional indicators exploit only aggregate survey information, namely the proportions of respondents who report “up” and “down”. As a consequence, neither the heterogeneity of individual responses nor the panel dimension of microdata is used. We illustrate the use of a disaggregate panel-based indicator that exploits all information coming from two yearly industrial surveys carried out on the same sample of Italian manufacturing firms. Using the same sample allows us to match exactly investment plans and investment realisations for each firm and so estimate a panel data model linking individual investment realisations to investment intentions. The model generates a one-year-ahead forecast of investment variation that follows the aggregate dynamics with a limited bias