[Limdep Nlogit List] SFA based on Distance functions

William Greene wgreene at stern.nyu.edu
Mon Mar 16 00:02:32 EST 2009


Pavlos.
CREATE ; MY = -Y $ then ;LHS=MY...
takes care of the sign on the LHS variable.
Adding ;COST to the frontier command takes care of the
sign on u(i).
/B. Greene


----- Original Message -----
From: "Pavlos C. Symeou" <p.symeou at gmail.com>
To: limdep at limdep.itls.usyd.edu.au
Sent: Sunday, March 15, 2009 7:52:38 AM GMT -05:00 Columbia
Subject: [Limdep Nlogit List] SFA based on Distance functions

Dear Limdepers,

I want to conduct an efficiency analysis by estimating a translog 
output-oriented production function for the sector of 
telecommunications. Since, the production function does not support 
multi-output specifications, I transform the production function into a 
distance function, according to Coelli and Perelman (2001). 
Specifically, the original production function would have the form "y  = 
βx + v - u", whilst the output oriented distance function has the form 
"-y  = βx + v + u"; that is, the output now appears with a negative sign 
and the inefficiency term "u" is not deducted from the RHS but instead 
is added. How do I model this in Limdep 9?

Namely, if I intended to estimate "y  = βx + v - u" I would invoke the 
following commands in Limdep 9:

FRONTIER    ; LHS = Y;    RHS = ONE, X $

Now, for the estimation of the distance function, should I have the 
following commands?

FRONTIER    ; LHS = -Y;    RHS = ONE, X;
                       COST $

where Y appears in a negative form on the LHS and COST is invoked to 
account for the positive sign of "u" on the RHS.

I look forward to hearing from you.

Best regards,

Pavlos

Reference:
Coelli, J. T. and Perelman, S. (2000). Technical Efficiency of European 
Railways: A distance Function Approach. Applied Economics 32, 1967-1976.


_______________________________________________
Limdep site list
Limdep at limdep.itls.usyd.edu.au
http://limdep.itls.usyd.edu.au



More information about the Limdep mailing list