We investigate the long-run stock-bond correlation using a novel model that combines the dynamic conditional correlation model with the mixed-data sampling approach. The long-run correlation is affected by both macro-finance variables (historical and forecasts) and the lagged realized correlation itself. Macro-finance variables and the lagged realized correlation are simultaneously significant in forecasting the long-run stock-bond correlation. The behavior of the long-run stock-bond correlation is very different when estimated taking the macro-finance variables into account. Supporting the flight-to-quality phenomenon for the total stock-bond correlation, the long-run correlation tends to be small/negative when the economy is weak.
|Research areas and keywords
- Business Administration
- DCC-MIDAS model, Long-run correlation, Macro-finance variables, Stock-bond correlation
|Publisher||Department of Economics, Lund Universtiy|
|Number of pages||39|
|Publication status||Published - 2014|
|Name||Working Paper / Department of Economics, School of Economics and Management, Lund University|