Table 5: Regressions on Δ (log(S&P500)).

Sample 1973M02-2013M11 1973M02-2013M11 1973M02-2013M11
Sample size 490 490 490
Conditional Mean Equation
Constant 0.010900 (3.516905) 0.010654 (3.423197) 0.010340 (4.422293)
Coefficient on Δ (k) -52.13934 (5.081359) -52.41460 (5.118920) -52.77326 (5.125944)
Coefficient on Δ (log(RDISP)) 0.407951 (1.929993) 0.409207 (1.935600) 0.410721 (1.942396)
Coefficient on Δ (log( DOLLAR )) -0.188145 (1.180231) -0.188991 (1.185060) -0.189857 (1.191137)
Coefficient on Δ (log( CPI )) -0.173186 (0.223177)
Coefficient on absoluteΔ (log( CPI )) -0.095853 (0.123633)
Coefficient on the square of Δ (log( CPI )) 0.733436 (0.010172)
Coefficient on the average three laggedΔ (log( MS )) -0.904526 (3.583844) -0.904363 (3.583266) -0.905067 (3.608742)
Conditional Variance Equation
constant 0.0000595 (2.007861) 0.0000594 (2.009032) 0.0000592 (2.012091)
RESID(-1)^2 0.127511 (4.241274) 0.127515 (4.252679) 0.127473 (4.272594)
GARCH(-1) 0.849686 (23.71404) 0.849749 (23.79004) 0.849860 (23.92349)
Adjusted R-Square 0.076144 0.075953 0.075730
Log likelihood 863.0550 863.0352 863.0269
Durbin-Watson statistic 2.055479 2.055535 2.055796
Akaike Information criterion -3.485939 -3.485858 -3.485824
Schwarz information criterion -3.408899 -3.408818 -3.408784
Hannan-Quinn information criterion -3.455682 -3.455602 -3.455568

Notes: See notes under Tables 1-3. DOLLAR stands for the trade-weighted US dollar index. An increase in the DOLLAR variable is an appreciation of the US dollar. Bollerslev-Wooldridge robust standard errors and covariance are computed.