SMU ECONOMICS & STATISTICS WORKING PAPER SERIES Paper No. 05-2009
A dynamic factor model is applied to a large panel dataset of Singapores macroeconomic variables and global economic indicators with the initial objective of analyzing business cycles in a small open economy. The empirical results suggest that four common factors are present in the quarterly time series, which can broadly be interpreted as world, regional, electronics and domestic economic cycles. The estimated factor model explains well the observed fluctuations in real economic activity and price inflation, leading us to use it in forecasting Singapores business cycles. We find that the forecasts generated by the factors are generally more accurate than the predictions of univariate models and vector autoregressions that employ leading indicators.