基于MATLAB实现的随机前沿模型代码【含固定效应】案例+实验数据

*********************************************************************************************** The codes have been tested in Matlab2010b and Matlab2012b. If you find any problem when using these codes. Pls contact me. Kerry Du [kerrydu1986@gmail.com] [caike22@126.com] ************************************************************************* 11 The matlab codes are written for estimating the fixed-effect SFA model using within-tranformation method proposed by Wang and Ho (2010) Reference: "Estimating Fixed-Effect Panel Stochastic Frontier Models by Model Transformation", Journal of Econometrics, 2010, 157 (2), 289-96. The zipped files contains: 1.PSFA2010.m - Main estimation function % USAGE: % [parameter,ste]=PSFA2010(dta,nx,para0) % [parameter,ste,robste,LL,BCindex] % =PSFA2010(dta,nx,para0,dis,flag,options) % INPUTS: % dta - a N*T by (K+L+3) matrix of data % The 1st column is id of DMU (firms/countries);the 2nd % is time variable; the 3rd is the dependent % variable(Y);the 4th -(3+K) are indepent variables (Xs)of % the frontier; the Kth-end are non-stochastic % inefficiency determinants (Zs). % That is to say, ID t Y X1...XK Z1...ZL % nx - The number of independent variables of the frontier (Xs) % para0 - starting values for the parameters % dis - The type of inefficiency distirbution, valid types are: % 0 if half normal distribution, as default % 1 if truncated normal distribution % flag - The type of the frontier, valid types are: % 0 if production function, as default % 1 if cost function % options - options for matlab function "fminunc" % OUTPUTS: % parameters - A K+L+2(+1 is his=1) vector of parameters. % the 1st-Kth of parameters are coefficients of Xs % the [K+1]th-[K+L]th of parameters are coefficients % of Zs; the [K+L+1]th is the log(sigma_u-square); the % [K+L+1]th is the log(sigma_v-square);(the last is mu % if dis=1). % ste - A K+L+2(+1 is his=1) vector of standard error. % robste - A K+L+2(+1 is his=1) vector of robust standard error. % robste - the log-likelihood function. % BCindex - BC estimator of efficiency index 2. robustvcv.m - utility function borrowed from Kevin Sheppard's MFE toolbox 3. hessian_2sided.m - utility function borrowed from Kevin Sheppard's MFE toolbox 4. demo.m - an example for estimation 5. Wang(2010)_data.csv - data for demo.m, which is from the empirical study of Wang and Ho (2010)