A Quantitative Analysis of the Yield Curve: Forecasting the Length and End of a Recession

Economists, finance professionals, businesses, and governments analyze economic activity to forecast recessions to put in place plans that will minimize the impacts of a recession. A tool that could identify how long a recession might last could reduce the unwanted effects of a recession. While researchers have identified that the interest rate term spread, a component of the yield curve, can effectively predict if a recession is probable and approximately when it might begin, it is not known whether the components of the interest rate yield curve might also be able to predict how long a recession might last. The purpose of this quantitative correlational study was to analyze the components of the yield curve to determine if they could be used to forecast how long a recession might last or when a recession might end. Data for the study was gathered for the U.S. economy for the period from 1942 to 2012. The data set for the entire time period and for each of the twelve recessionary periods was analyzed using the autodistributive lag and vector autoregression models. It was found that there was a statistically significant correlational relationship for two yield curve components, the interest rate term spread and the three-month T-bill rate, that could be used to forecast when a recession might end. Prior to these recessions the interest rate term spread drops rapidly and typically turns negative. The term spread began to increase during the recession and for the twelve recessions analyzed rose to the level of 100 basis points on average about seven months before the end of the recession. It was also found that the three-month T-bill rate rose significantly before the recessions, peaked, and then fell throughout the recessionary time period. For the recessionary periods analyzed the peak in the T-bill rate occurred on average about eleven months before the end of the recession.