Abstract
We use Bayesian dynamic factor models to disentangle the global and idiosyncratic country-specific factors of the stock market prices and other macroeconomic variables from a large group of countries. We find that the global factors account for significant portions of an individual country's stock market volatility and its macroeconomic fluctuations. The global macroeconomic shocks have strong effects on the stock price movement across countries in addition to domestic macroeconomic shocks. And a country's exposure to the global stock market risk to a large extent reflects that country's exposure to the global macroeconomic risks.
Notes
1 Some classical examples include Campbell and Shiller (Citation1988) and Cochrane (Citation1992) among many others. See Campbell (Citation1999) and Cochrane (Citation2011) for recent reviews of relevant literatures.
2 As in the standard latent factor models, the factor loading, , and the variance of
are not identified separately. We also impose the sign restriction such as
for all i. See Stock and Watson (Citation1989) for more on the identification of factor models.
3 Note that our Cholesky ordering gives a conservative estimate of the effect of global macroeconomic shocks on stock market returns relative to that of domestic macroeconomic shocks.