Partial frontier efficiency analysis (Orderalpha)

[Neeraj Kumar]

Hi,
I am estimating efficiency score, for that I am using orderalpha command by Harald Tauchmann. I have read the article "Partial frontier efficiency analysis" by Harald Tauchmann. I Followed thy Commands used by him.

orderalpha X Y Z = avls, dmu(state) alpha(95) reps(500) table(full) generate(escore)

While using this commad it showing error "variable state does not uniquely identify dmus". However after that I generated new variable using "_n" then further run the command. After this it was not showing the same error. but some other errors has been appeared

 orderalpha LnNSDPpcC edu health = avgls , dmu(dmu) alpha(95) reps(500)  generate(escore)
matsize too small
    You have attempted to create a matrix with too many rows or columns or attempted to fit a model with too
    many variables.  You need to increase matsize; it is currently 400.  Use set matsize; see help matsize.

    If you are using factor variables and included an interaction that has lots of missing cells, either
    increase matsize or set emptycells drop to reduce the required matrix size; see help set emptycells.

    If you are using factor variables, you might have accidentally treated a continuous variable as a
    categorical, resulting in lots of categories.  Use the c. operator on such variables.

as per the error it is asking to set a metsize but metsize should be greater than 10. But as per Harald Tauchmann paper atleast one input and one output is required to run this command. In his paper as well I have notice, he used 4 input and 3 output variable. I don't know why it showing such error. I have read Milena Nikolova and Olga Popova article titled " Sometime your best just ain.'t good Enough..." , in which they also used the same method and they have 3 input variable and one output variable. However I have read Martin Binder and Tom Broekel paper " Happiness no Matter at cost?... " they have used orderm method but they also used one output and 3 input variable.
In My analysis X is Continuous variable, Y is categorical variable (3 categories) and Z also categorical variable (5 ) . I have noticed that in Brinder and Tom paper there input variable also having same nature as mine but they used orderm method.

I want to use orderalpha method kindly help to solve the issues.
Thank you so much

[Harald Tauchmann]

Dear Neeraj,
when using orderalpha, the number of DMUs (observations) is limited to the maximum number of rows allowed in a Stata matrix, which depends on the edition of Stata. So, the key questions are: how many DMUs enter your FDH analysis, and which edition and version of Stata you are using?
Best,
Harald