1. The test assets are excess returns to four carry trade portfolios based on the subset of the G10 countries. Panel A shows the factor risk prices ? obtained by FMB cross-sectional regression. I do not use a constant in the second-stage FMB regression. Robust t-statistics based on Shanken (1992) (SH) and Newey and West (1987) (NW) are reported in brackets. I also report the cross-sectional R 2 and ? 2 -test statistics with p-values in brackets that are based on both Shanken (1992) (SH) and Newey and West (1987) (NW) adjustments. Panel B shows the factor betas for time-series regressions of excess returns on a constant (a), the dollar risk factor (DOL) and the non-traded system-wide volatility connectedness risk factor (SV C) or the factor-mimicking portfolio of system-wide volatility connectedness innovations (SV C F M ). Heteroskedasticity and autocorrelation consistent (HAC) robust t-statistics based on Newey and West (1987) are reported in brackets. The sample period is from;The table reports factor prices and factor betas for a rolling window size of w = 250 days,1988
2. Panel A shows the factor risk prices ? obtained by FMB cross-sectional regression. I do not use a constant in the second-stage FMB regression. Robust t-statistics based on Shanken (1992) (SH) and Newey and West (1987) (NW) are reported in brackets. I also report the cross-sectional R 2 and ? 2 -test statistics with p-values in brackets that are based on both Shanken (1992) (SH) and Newey and West (1987) (NW) adjustments. Panel B shows the factor betas for time-series regressions of excess returns on a constant (a), the dollar risk factor (DOL) and the non-traded system-wide volatility connectedness risk factor (SV C) or the factor-mimicking portfolio of system-wide volatility connectedness innovations (SV C F M ). Heteroskedasticity and autocorrelation consistent (HAC) robust t-statistics based on Newey and West (1987) adjustments are reported in brackets. The sample period is from;The table reports factor prices and factor betas for a rolling window size w = 150 days. The test assets are excess returns to five carry trade portfolios based on the subset of developed countries,1988
3. based on a rolling window size of w = 150 days. Test assets are the excess returns to the five carry trade based on the subset of the 15 developed countries. Panel A shows results for non-traded factors of global FX volatility risk (V OL) and system-wide volatility connectedness risk (SV C), Panel B for the factor-mimicking portfolio of global;FX volatility innovations V OL F M and system-wide volatility connectedness innovations (SV C F M ), Panel C for SV C F M orthogonalised with respect to V OL F M,2012
4. The sample period is;F M;and robust t-statistics based on Newey and West (1987) and Shanken,1988