Please help out. I would like to collect the number of observations(N) in a regression for each quarter in addition to coefficients, se and r squared. My data is panel data. Here is my code
I use N as number of observations but it does not generate anything other than missing observations. However, am able to generate coefficients, se and r squared if i remove n=r(N)
Here is my sample data with dataex
I use N as number of observations but it does not generate anything other than missing observations. However, am able to generate coefficients, se and r squared if i remove n=r(N)
Code:
*drop unwanted variables keep qdate ticker delta_kt *sort data sort qdate ticker *Create lag variable egen id = group(ticker) xtset id qdate generate delta_kt_1 = L.delta_kt *Cross-sectional regressions sort qdate statsby beta=_b[delta_kt_1] beta_se=_se[delta_kt_1] rsq=e(r2) n=r(N), by(qdate) clear: reg delta_kt delta_kt_1, noconstant format qdate %tq drop if beta==.
Code:
*example generated by -dataex-. To install: ssc install dataex clear ticker qdate delta_kt ST 2008q2 1.088697 FM 2008q2 .0505587 NIT 2008q2 1.088697 PC 2008q2 -.7651212 MP 2008q2 -.2089759 ILL 2008q2 -.2089759 SUN 2008q2 1.088697 NB 2008q2 -.2089759 NIC 2008q2 -1.506648 PI 2008q2 -1.506648 NM 2008q2 1.088697 NIC 2008q3 1.147993 ILL 2008q3 -1.506082 NM 2008q3 -.8425636 FM 2008q3 .1527145 ST 2008q3 .1527145 MP 2008q3 1.147993 PC 2008q3 -.5581983 NB 2008q3 -.8425636 NIT 2008q3 1.147993 FM 2008q4 -.3804566 NB 2008q4 .6419926 ST 2008q4 -.7480035 NIC 2008q4 -.5127734 MP 2008q4 -.1599282 TM 2008q4 -1.509042 ILL 2008q4 -.3804566 NM 2008q4 -.1599282 NIT 2008q4 1.604298 PC 2008q4 1.604298 ILL 2009q1 .9427915 TM 2009q1 -.8208599 RE 2009q1 .9427915 ST 2009q1 -1.114802 FM 2009q1 -1.526321 NIC 2009q1 .9427915 NB 2009q1 -.4289374 PC 2009q1 .1197541 NM 2009q1 .9427915 NIC 2009q2 -.4951594 NM 2009q2 -.661042
Comment