sfcross iterations not concave for translog production function

[Jianying Wang]

Dear StataListers, I am working on my dissertation paper and encounter one big problem. My data is cross-section data about rice farmers, my analysis is at the plot-level. I want to use the one-step method to estimate technical efficiency of the rice farmers. I use the code like this: xi: sfcross lntvalue lnplotaream lntolt lnthlt lntmach lnc1 lnc2 p116 p117 /// lnlandlnland lnollnol lnhllnhl lnc1lnc1 lnc2lnc2 lnmaclnmac /// lnlandlnol lnlandlnhl lnlandlnc1 lnlandlnc2 lnlandlnmac /// lnollnhl lnollnc1 lnollnc2 lnollnmac /// lnhllnc1 lnhllnc2 lnhllnmac lnc1lnc2 lnc1lnmac lnc2lnmac /// i.village,d(t) emean(rplot rprobit ab dmigra07 pmigra07 lnplot07 lnparea07 lna_area lndhome lndroad /// lnharea dtra11 dcatt11 hhsize kidratio oldratio femaleratio age edu cpc sci exper /// scale2 scale3 p099 p103 p113 p114 p109 p115 i.village) and I tried all the tech() option, but stata shows it can not concave. The total sample is 783 and 12 villages. Then I tried to estimate the function first, run the code below, it can concave. xi: sfcross lntvalue lnplotaream lntolt lnthlt lntmach lnc1 lnc2 p116 p117 /// lnlandlnland lnollnol lnhllnhl lnc1lnc1 lnc2lnc2 lnmaclnmac /// lnlandlnol lnlandlnhl lnlandlnc1 lnlandlnc2 lnlandlnmac /// lnollnhl lnollnc1 lnollnc2 lnollnmac /// lnhllnc1 lnhllnc2 lnhllnmac lnc1lnc2 lnc1lnmac lnc2lnmac /// i.village So, what is the problem, how can I solve it? could you pls give me some suggestions, Thanks so much! Best, Jianying Wang College of economics Zhejiang University, China