Abstract

This paper examines the information content in Singapore and foreign electronics indicators with a view to forecasting domestic electronics output. Our aim is to develop a summary index of activity that can be used to predict turning points in the global electronics cycle, as well as to generate quantitative forecasts of the local electronics industry's performance. For this purpose, indicators that historically tended to presage the cyclical fluctuations in world electronics demand are identified. Three rigorous selection procedures - cross correlation analysis, causality tests and marginal predictive content tests - yielded four alternative subsets of leading indicators, each consisting of up to six component variables. These variables, which span a whole spectrum of indicators including orders, inventories, shipments and prices, are combined to form composite Electronics Leading Indices (ELIs). To evaluate the predictive power of the ELIs, they are incorporated into vector autoregressive (VAR) models to forecast domestic electronics production. The results of this exercise show that the use of the ELIs in forecasting leads to a reduction in the magnitude of prediction errors.