Abstract

This paper investigates the linkages between high frequency market volatility and macroeconomic variables through GARCH-MIDAS model. Engle, Ghysels and Sohn (2013) developed a new class of model to distinguish the short-term volatility and long-term components which are linked to the macroeconomic variable. MIDAS GARCH models incorporate weighted average of low frequency variables into the GARCH model for volatility changes. Volatilities observed in the daily data of NIKKEI index, USD/JPY foreign exchange rate and JGB 10 year government bonds are linked to the monthly data such as industrial production or consumer price index. Inclusion of macroeconomic variables improves the long-run forecast performance in the volatility. Macroeconomic variables play a significant role at short horizon.