Forum: help
Monitor Forum | Start New ThreadRE: Tobit with Panel Data [ Reply ] By: Arne Henningsen on 2019-11-03 13:32 | [forum:47113] |
It looks like you are applying a random-effects rather tha a fixed-effects estimator. The standard errors are indeed estimated but they are all infinity. This could be caused, e.g., by perfect multicollinearity. Could this be the case? What are the standard errors if you estimate the model by plm() using a random-effects or a fixed-effects estimator? |
Tobit with Panel Data [ Reply ] By: Patricia Byrnes on 2019-08-16 21:55 | [forum:46907] |
Hello! I am trying to run a Tobit panel with panel data (fixed effects) using censReg. My panel data is 48 states and 16 years. The coefficients are estimated but the standard errors are not. Please help-- Thanks! --Patty The code is: pData <- pdata.frame( NoNJMA, c( "STATE", "YEARFACT" ) ) panelResult4 <- censReg(DBTSHRAVG70 ~ POP + YOUNG + OLD + POOR + MOB + PCY + CGI + UNEMP + TTU + MFTBASE + FED + BDRDS + AAA + ELECT + DIVIDED + SDEM + HDEM + TELINDEX, data = pData,left=0, method = "BHHH", nGHQ=15, iterlim=10000 ) summary( panelResult4 ) The result is: Observations: Total Left-censored Uncensored Right-censored 768 293 475 0 Coefficients: Estimate Std. error t value Pr(> t) (Intercept) 1.698e+02 Inf 0 1 POP 3.457e-06 Inf 0 1 YOUNG -3.045e+00 Inf 0 1 OLD -2.798e+00 Inf 0 1 POOR 8.116e-02 Inf 0 1 MOB 2.860e+00 Inf 0 1 PCY 9.480e-04 Inf 0 1 CGI -2.804e+01 Inf 0 1 UNEMP 1.275e+00 Inf 0 1 TTU -5.540e+00 Inf 0 1 MFTBASE -4.988e-06 Inf 0 1 FED -4.220e-01 Inf 0 1 BDRDS -2.852e-01 Inf 0 1 AAA1 -8.285e-01 Inf 0 1 ELECT1 2.538e+00 Inf 0 1 DIVIDED1 -2.663e+00 Inf 0 1 SDEM1 -3.026e+00 Inf 0 1 HDEM1 1.498e+00 Inf 0 1 TELINDEX 9.137e-02 Inf 0 1 logSigmaMu 2.918e+00 Inf 0 1 logSigmaNu 3.052e+00 Inf 0 1 BHHH maximisation, 156 iterations Return code 2: successive function values within tolerance limit Log-likelihood: -2334.847 on 21 Df |