I am running a fe regression model on US listed companies to study the effect of lagged Relative Geographic Diversification (L1.Relative_GSD) on Firm Performance (ROA.) I am using a linear and quadratic term for L1.Relative_GSD and am interacting it with a dummy variable Crisis (Crisis=0 is the period 2001-06 and Crisis=1 is the period 2010-19.)
1. cL.Relative_GSD#cL.Relative_GSD has a p>|t| = 10.5%, but Crisis#cL.Relative_GSD has a p>|t| = 73.7% (not significant at all.)
2. Crisis#cL.Relative_GSD#cL.Relative_GSD is significant but Crisis#cL.Relative_GSD is not. (In such a situation, would it be correct for me to interpret the interaction results for Crisis#cL.Relative_GSD#cL.Relative_GSD?)
3. Should I be considered the adjusted R-squared before interpreting the results in #2? If so, where can I find the adjusted R-squared for xtreg? Thank you so much.
I am attaching my results screen for your reference.
Code:
. xtreg ROA_win05 l1.LnRev l1.LnLTDTA l1.CoAge l1.RnDIntensity l1.AdvExpIntensity TotalPSD GDPGrowth_Wg > t GDP_pc_Wgt CPI_Wgt c.l1.Relative_GSD##c.l1.Relative_GSD##i.Crisis if SIC_2Digit_Division=="Manufact > uring" & l1.Relative_GSD !=0 & l1.Relative_GSD !=. & HeadquartersCountryRegion=="United States" & FST > S>10, fe cluster(compid) Fixed-effects (within) regression Number of obs = 6,433 Group variable: compid Number of groups = 805 R-sq: Obs per group: within = 0.0575 min = 1 between = 0.1898 avg = 8.0 overall = 0.0981 max = 18 F(14,804) = 8.81 corr(u_i, Xb) = -0.1639 Prob > F = 0.0000 (Std. Err. adjusted for 805 clusters in compid) ------------------------------------------------------------------------------------------------- | Robust ROA_win05 | Coef. Std. Err. t P>|t| [95% Conf. Interval] --------------------------------+---------------------------------------------------------------- LnRev | L1. | -1.407216 .3320157 -4.24 0.000 -2.058935 -.7554957 | LnLTDTA | L1. | -.521211 .1351684 -3.86 0.000 -.7865356 -.2558863 | CoAge | L1. | .0137111 .0283719 0.48 0.629 -.0419806 .0694027 | RnDIntensity | L1. | .0026586 .0026161 1.02 0.310 -.0024766 .0077938 | AdvExpIntensity | L1. | -11.87018 9.741451 -1.22 0.223 -30.99186 7.251499 | TotalPSD | -.2000879 .3119765 -0.64 0.521 -.8124725 .4122966 GDPGrowth_Wgt | .2464468 .040049 6.15 0.000 .1678338 .3250599 GDP_pc_Wgt | .0000154 .0000294 0.53 0.599 -.0000422 .0000731 CPI_Wgt | .1231483 .0824054 1.49 0.135 -.0386069 .2849035 | Relative_GSD | L1. | .0011202 .9366034 0.00 0.999 -1.837356 1.839597 | cL.Relative_GSD#cL.Relative_GSD | 2.913843 1.793011 1.63 0.105 -.6056914 6.433377 | 1.Crisis | .8772725 .2731693 3.21 0.001 .3410634 1.413482 | Crisis#cL.Relative_GSD | 1 | .3113933 .9286133 0.34 0.737 -1.511399 2.134186 | Crisis#cL.Relative_GSD#| cL.Relative_GSD | 1 | -3.953864 1.851397 -2.14 0.033 -7.588007 -.3197211 | _cons | 16.47904 2.224425 7.41 0.000 12.11268 20.84541 --------------------------------+---------------------------------------------------------------- sigma_u | 8.7188974 sigma_e | 3.7498225 rho | .84390413 (fraction of variance due to u_i) -------------------------------------------------------------------------------------------------
0 Response to Interaction Effect between a Dummy Variable and Quadratic Term
Post a Comment