Forecasting industrial production using models with business cycle asymmetry
利用产出与利率利差在扩张和收缩期的动态关系不对称性，提出一种无需事先确定转折点的双区制模型，用于预测月度工业生产，并证明其优于传统模型。
This paper exploits business cycle asymmetry observed in data, namely, a systematic shift in the dynamic relationship between the output and the interest rate spread across expansionary and contractionary periods in forecasting monthly industrial production. A bivariate model of monthly industrial production and the spread between the 6-month commercial paper and the federal funds rates is used as an example to illustrate forecast exercise. This paper's method does not require a forecaster to make an exact ex-ante determination of turning points in the output series which is being forecasted. Comparison of the forecast performance of various two-regime based and conventional models suggests that a measurable gain can be made by considering models which explicitly incorporate asymmetry in data.