Forecasting the global electronics cycle with leading indicators: A Bayesian VAR approach
Developments in the global electronics industry are typically monitored by tracking indicators that span a whole spectrum of activities in the sector. However, these indicators invariably give mixed signals at each point in time, thereby hampering attempts at prediction. In this paper, we propose a unified framework for forecasting the global electronics cycle by constructing a VAR model that captures the economic interactions between putative leading indicators representing expectations, orders, inventories and prices. The ability of the indicators to presage world semiconductor sales is first examined by Granger causality tests. Subsequently, an impulse response analysis confirms the leading qualities of the selected indicators. Finally, out-of-sample forecasts of global chip sales are generated from two parsimonious variants of the VAR model, viz., the Bayesian VAR (BVAR) and Bayesian ECM (BECM), and compared with predictions from a bivariate model which uses a composite index of the leading indicators and a univariate autoregressive model. An evaluation of their relative accuracy suggests that the BVAR's forecasting performance is superior to the other models. The BVAR is also able to predict the turning points of the recent IT boom-and-bust cycle.