PREDICTIVE POWER OF YIELD CURVE – EVIDENCE FROM INDIA

In this paper, we explore the predictive ability of the slope of the sovereign (government) yield curve for the Indian market. Instead of showing the predictive power of the yield curve alone, we attempt to compare the predictive power of the yield curve slope with a composite index of lead indicators (CILI) constructed with potential high frequency lead indicators. The non-agricultural GDP (NAGDP) is used as the reference series and the quarterly data from 2005 to 2018 are considered for the study. The quarterly data of around 27 economic indicators are considered for the construction of CILI and these data are sourced from CSO, RBI, and CMIE for the period 2005 Q2 to 2018 Q1. The study suggests that the yield curve spread measured by the difference in the 10-year and 3-month yields has better predictive power to estimate the turns in economic activity compared to the CILI constructed. The yield curve has been able to predict the turning points of NAGDP about two quarters ahead, while the CILI leads NAGDP by one quarter. Thus, the yield curve can easily supplement the results from complicated econometric models for predicting economic activities. The paper also looks at, for the first time, at the predictive ability of both - the CILI and slope of the yield curve, during the period of the global financial crisis of 2008, domestic macro crisis of 2013, and unexpected and forced disruption of demonetisation at the end of 2016. It is observed that both, CILI and slope of yield curve, fail to capture or anticipate the slowdown caused by demonetisation in advance.