Guest:
sorry, but I'm still not clear with what you're after.
sorry, but I'm still not clear with what you're after.
-
Login or Register
- Log in with
- You are not logged in. You can browse but not post. Login or Register by clicking 'Login or Register' at the top-right of this page. For more information on Statalist, see the FAQ.
-
Kind regards,
Carlo
(Stata 18.0 SE) -
Carlo Lazzaro I am sorry that I did not make myself clear. Well, to put it in another way: giving the model :Where the β0 is the intercept (constant) scaled by the lag of the value of Total assests (Ai,t-1), how can I run the regression in STATA? Thank you so much.Code:TACit/Ai,t-1 = β0 (1/Ai,t-1) + β1 (ΔREVit/Ai,t-1) + β2 (PPEit/Ai,t-1)+ ԑit
Comment
-
Guest:
if all the variables are scaled for -Ai,t-1-, the constant will be automatically calculated with that constraint..
Put differently, constant is something out of your control.
That said, I think you should do something along the following lines:
Code:<depvar>=<indepvars> <controlsifany>
Kind regards,
Carlo
(Stata 18.0 SE)Comment
-
Carlo Lazzaro Hi carlo, Thank you so much for your reply. This is what I have done and I want you please to tell me your opinion. Like I already point in my post, I need to generate errors the "epsilon" as a proxy for my variable "discretionnary accruals" using one of the models developped in the literature. So, I decide to stick with the Kothari model:
1- I run the -xtreg, fe- to check whether there are indeed individual effects in my data. H0 is rejected (0.0112< 10%)===> panel dataCode:TACit/Ai,t-1 = β0 (1/Ai,t-1) + β1 ((ΔSalesit-ΔARit)/Ai,t-1) + β2 (PPEit/Ai,t-1)+ β3 ROA (i, t-1)+ ԑit
Code:xtreg ACCTT varCA_CClts IMMOB Lag_ROA, fe Fixed-effects (within) regression Number of obs = 228 Group variable: i Number of groups = 38 R-sq: Obs per group: within = 0.0418 min = 6 between = 0.2677 avg = 6.0 overall = 0.0531 max = 6 F(3,187) = 2.72 corr(u_i, Xb) = -0.8667 Prob > F = 0.0457 ------------------------------------------------------------------------------ ACCTT | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+---------------------------------------------------------------- varCA_CClts | -.0782775 .0568273 -1.38 0.170 -.1903824 .0338274 IMMOB | .1373683 .0729324 1.88 0.061 -.0065077 .2812444 Lag_ROA | .2490675 .1757287 1.42 0.158 -.0975979 .5957329 _cons | -.1148862 .0420693 -2.73 0.007 -.1978776 -.0318948 -------------+---------------------------------------------------------------- sigma_u | .11087529 sigma_e | .09420287 rho | .58076406 (fraction of variance due to u_i) ------------------------------------------------------------------------------ F test that all u_i=0: F(37, 187) = 1.71 Prob > F = 0.0112
2- I test for time-fixed effects===>Prob>10%, so we failed to reject the null, therfore no time fixed effects are needed in our case.
Code:xtreg ACCTT varCA_CClts IMMOB i.année, fe
Code:testparm i.année ( 1) 2013.année = 0 ( 2) 2014.année = 0 ( 3) 2015.année = 0 ( 4) 2016.année = 0 ( 5) 2017.année = 0 F( 5, 183) = 1.25 Prob > F = 0.2898
3-I run the Breush-pagan LM test for random effects versus OLS===> Prob<10%, so we reject the null and conclude that random effects is appropriate;
Code:xtreg ACCTT varCA_CClts IMMOB Lag_ROA, re
Code:xttest0 Breusch and Pagan Lagrangian multiplier test for random effects ACCTT[i,t] = Xb + u[i] + e[i,t] Estimated results: | Var sd = sqrt(Var) ---------+----------------------------- ACCTT | .0108795 .104305 e | .0088742 .0942029 u | .0009049 .0300817 Test: Var(u) = 0 chibar2(01) = 2.35 Prob > chibar2 = 0.0628
4-Hausman test===> Prob<10%, so we reject the null, we have to use fixed effects
5- White test for heteroskedasticity (We reject H0=====> Existency of heteroskesticity): I use xttest3 afterCode:hausman fixe ---- Coefficients ---- | (b) (B) (b-B) sqrt(diag(V_b-V_B)) | fixe . Difference S.E. -------------+---------------------------------------------------------------- varCA_CClts | -.0782775 -.0894111 .0111337 .0217968 IMMOB | .1373683 -.0653583 .2027266 .0707456 Lag_ROA | .2490675 .1397222 .1093453 .1441856 ------------------------------------------------------------------------------ b = consistent under Ho and Ha; obtained from xtreg B = inconsistent under Ha, efficient under Ho; obtained from xtreg Test: Ho: difference in coefficients not systematic chi2(3) = (b-B)'[(V_b-V_B)^(-1)](b-B) = 8.28 Prob>chi2 = 0.0405Code:xtreg ACCTT varCA_CClts IMMOB Lag_ROA, fe
Code:xttest3 Modified Wald test for groupwise heteroskedasticity in fixed effect regression model H0: sigma(i)^2 = sigma^2 for all i chi2 (38) = 15736.78 Prob>chi2 = 0.0000
6- Wooldridge test for autocorrelation (We fail to reject H0==> No serial correlation)Code:xtserial ACCTT varCA_CClts IMMOB Lag_ROA Wooldridge test for autocorrelation in panel data H0: no first-order autocorrelation F( 1, 37) = 0.151 Prob > F = 0.6995
Becuase I have heteroskedasticity, I need to runright? and How Can I have my resid values? like I said, I'm interested in the resid because it's a proxy for my variable "discretionnary accruals"? And is there other tests I need to run before having the values of my epsilon? Thank you so much.Code:xtreg ACCTT varCA_CClts IMMOB Lag_ROA, fe robust
Comment
-
Guest:
congrats' on your really comprehensive pre and post estimate assessments.
That said, epsilon can be obtained just typing:
after -xtreg,fe rob-.Code:predict epsilon, e
Kind regards,
Carlo
(Stata 18.0 SE)Comment
-
Carlo Lazzaro So, do you think that I don't need to run other tests such as Ramsey test for misspecification or multicollinearity? Thank you so much for your feedback. Your responses to my post and to others as well in this plateform have been really helpful.Comment
-
Guest:
- usually, if standard errors do not look terribly weird (whatever that means in one's research field), there's no reson to worry about multicollinearity. See also wise comments on this topic by the always helpful Clyde Schechter and read, if available at your university library, Chapter 23 of https://www.hup.harvard.edu/catalog....40&content=toc.
- RESET test (-estat ovtest-) is not available after -xtreg-. Hoverewer, there's a work around, that you can find in the following toy-example:
In the previous toy-example the regression model is clearly misspecified, as -test- reject the null of a 0 coefficient for squared fitted values.Code:. use "http://www.stata-press.com/data/r15/nlswork.dta" (National Longitudinal Survey. Young Women 14-26 years of age in 1968) . xtreg ln_wage age, fe rob Fixed-effects (within) regression Number of obs = 28,510 Group variable: idcode Number of groups = 4,710 R-sq: Obs per group: within = 0.1026 min = 1 between = 0.0877 avg = 6.1 overall = 0.0774 max = 15 F(1,4709) = 884.05 corr(u_i, Xb) = 0.0314 Prob > F = 0.0000 (Std. Err. adjusted for 4,710 clusters in idcode) ------------------------------------------------------------------------------ | Robust ln_wage | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+---------------------------------------------------------------- age | .0181349 .0006099 29.73 0.000 .0169392 .0193306 _cons | 1.148214 .0177153 64.81 0.000 1.113483 1.182944 -------------+---------------------------------------------------------------- sigma_u | .40635023 sigma_e | .30349389 rho | .64192015 (fraction of variance due to u_i) ------------------------------------------------------------------------------ . predict fitted, xb (24 missing values generated) . g sq_fitted=fitted^2 (24 missing values generated) . xtreg ln_wage fitted sq_fitted , fe rob Fixed-effects (within) regression Number of obs = 28,510 Group variable: idcode Number of groups = 4,710 R-sq: Obs per group: within = 0.1087 min = 1 between = 0.1006 avg = 6.1 overall = 0.0865 max = 15 F(2,4709) = 507.42 corr(u_i, Xb) = 0.0440 Prob > F = 0.0000 (Std. Err. adjusted for 4,710 clusters in idcode) ------------------------------------------------------------------------------ | Robust ln_wage | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+---------------------------------------------------------------- fitted | 7.143466 .738485 9.67 0.000 5.69569 8.591242 sq_fitted | -1.816243 .2188485 -8.30 0.000 -2.245289 -1.387198 _cons | -5.167788 .6209677 -8.32 0.000 -6.385175 -3.950401 -------------+---------------------------------------------------------------- sigma_u | .4039153 sigma_e | .30245467 rho | .64073314 (fraction of variance due to u_i) ------------------------------------------------------------------------------ . test sq_fitted=0 ( 1) sq_fitted = 0 F( 1, 4709) = 68.87 Prob > F = 0.0000 .
To have an idea about how the auxiliary regression works, take a look at -linktest- entry in Stata .pdf manual.Kind regards,
Carlo
(Stata 18.0 SE)Comment
-
-
Hello, I want to run Kothari et al. (2005) model on Stata but I don't know how to measure the performance matched ROA can I ask how did you do that? Also you are assuming that this model doesn't have a constant but when I checked Kothari et al. (2005) article they have said they used a constant am a bit confusedComment
Comment