A Study On The Expenditure On National Debt And The Forecast Of Economic Cycle

Evidence from both theory and practice shows that, term spread is extremely useful in forecasting business cycles. Currently, as economic growth becomes unsustainable and vulnerable, to predict the level of economic activities in the near future is vital for monetary policy makers. This paper studies the performance of term spread in forecasting business cycles in China. Based on HP filter, the quarterly real GDP is de-compromised, and the resulting output gap estimates are used for deciding China’s business cycles. The estimated China’s business cycles are roughly in consist with economic facts and the lagging index. China’s Treasury bond markets and term spreads are then introduced, after which the10-,2-and1-year Treasury bonds are selected to construct the term spread. The main part is regressing the state of economy one to eight quarters ahead on the term spread using a "probit" model, with the sample period from the first quarter in2001to the fourth quarter in2012. The results show that the10-2year term spread has predictive power, while the10-1year term spread does not. A table showing the term spread and forecasted probability of recession is given. Similar regressions on leading index, purchasing manager index, stock index and bond index provides evidence of forecasting power for all of the variables. However, they differ in forecasting horizon, as leading index and purchasing manager index are useful in a short-term period of one year, term spread and stock index are useful in a mid-term period of half a year to1.5years, and bond index is useful in a long-term period of1.5years to2.5years. From perspective of monetary policy makers, the mid-term forecasting horizon is of significant importance, since most monetary policies will take effect one year later. In order to test whether term spread contains useful information that the leading indices does not contain, we put both term spread and the leading indices into regression. Results have accepted the hypothesis, since term spread increases the models’predictive power in mid-term forecasting horizon. Adding the term spreads of US and Germany will lead to very significant coefficients estimates, however the signs are all wrong, suggesting the probability of pseudo correlation. Therefore, we cannot tell whether foreign term spreads are useful or not.