I found my error terms to be normally distributed after running an OLS regression.
(You can check the distribution of the error terms in the png file attached).
How do I decide between a random-effects or fixed-effects model?
The Breusch Pagan test gives me message that I am violating one of its assumptions.
I assume that the Breusch Pagan test assumes non-normal distributed error terms, is that true?
My data is pooled cross-section over 26 regions, with many household observations in each region.
That is my data:
Code:
input float(lnexp_edu handling_edu0 ln_remittance ln_exp_total) double(oecd rem_freq femmigrant childsec childtert elderlyageabove62 womenabove15) float(hh_edu head_gender) byte(religion hagriland urban) float(employment foodpoverty country) -1.93268 .1875 7.539896 5.041759 1 6 0 0 2 0 3 2 0 1 1 1 .9138418 .10346045 6 -3.439349 .2327361 7.385745 6.037602 0 12 0 2 0 1 3 1 0 1 1 0 .6355928 .20125365 6 -5.225298 .1523096 6.469454 4.6046762 0 6 0 0 0 2 2 1 0 1 1 0 .3414017 .1891791 6 -4.6051702 .027922885 6.287133 4.790279 1 3 1 0 0 0 0 2 0 2 0 1 1.1116916 .4281095 6 -4.6051702 .15868984 6.287133 3.802226 1 6 0 0 0 0 1 1 0 3 0 1 .4122944 .288766 6 -2.2791886 .14560167 4.3900127 4.248833 0 3 0 1 0 0 1 1 1 1 1 0 .58518666 .624305 6 -2.579669 .14560167 4.3900127 3.1998334 0 6 0 3 0 2 2 1 0 4 1 0 .58518666 .624305 6 -1.7409674 .027922885 7.465788 6.102809 0 12 0 0 1 0 2 3 1 2 1 1 1.1116916 .4281095 6
In my case that is replacing a 0 dollar value by 1 dollar cent.
These are are the two regressions I run and the error message of the Breusch Pagan test.
Code:
xtset region eststo fe: xtreg lnexp_edu ib(last).phandling_edu0##c.ln_remittance ln_exp_total ln_remittance oecd rem_freq femmigrant childsec childtert elderlyageabove62 womenabove15 i.hh_edu head_gender i.religion hagriland i.urban employment foodpoverty i.country, fe eststo re: xtreg lnexp_edu ib(last).phandling_edu0##c.ln_remittance ln_exp_total ln_remittance oecd rem_freq femmigrant childsec childtert elderlyageabove62 womenabove15 i.hh_edu head_gender i.religion hagriland i.urban employment foodpoverty i.country, re hausman fe re Test: Ho: difference in coefficients not systematic chi2(29) = (b-B)'[(V_b-V_B)^(-1)](b-B) = -122.40 chi2<0 ==> model fitted on these data fails to meet the asymptotic assumptions of the Hausman test; see suest for a generalized test Array
Any help and comments are appreciated.
Kind regards,
Alexander
0 Response to Random and Fixed Effects with normally distributed error terms
Post a Comment