A Dynamic Nelson-Siegel Factor Model with Financial System Indicators.

Abstract

After the global financial crisis and the subsequent recession, given the nascent macro-finance literature, the interactions between macroeconomics and financial markets grow more important. A related question is how interest rates affect real GDP. I explore this interaction via a forecasting exercise. First, I use a vector autoregressive model with financial system indicators within the dynamic Nelson-Siegel framework to forecast interest rates. Then, I extend the analysis to credit spread modelling, after which I investigate whether credit spreads forecasted using the dynamic Nelson-Siegel factor model, in a contemporaneous regression with real GDP growth, beat the benchmark autoregressive model in real GDP growth and autoregressive distributed lag models in forecasting real GDP growth. I found that the proposed vector autoregressive model was able to generate good yield forecasts, and also produced credit spread forecasts ranked second to the dynamic Nelson-Siegel autoregressive factor model in forecasting real GDP growth.