Stochastic Frontier Analysis (SFA) True Fixed Effects (TFE) model by Greene (2005)
Hi dear all;
When I run the Stochastic Frontier Analysis (SFA) True Fixed Effects (TFE) model by Greene (2005), oftentimes Standard Deviations, P-values, and Confidence Intervals are not reported for some of the variables, either in the frontier equation or in the inefficiency effects equation. Instead, a period (.) appears in their place. Why is this happening?
I have an unbalanced panel of 16 countries and 14 years, with around 185 observations in total.
My codes are as follows:
sfpanel lnY lnK lnL lnT Year, model(tfe) dist(tn) emean(Educ Institutions Year) ort(o)
As you can see, single output is lnY, inputs are the logarithms of capital (lnK), labor (lnL), and land (lnT), in a simple (log-linearized) Cobb-Douglas specification. Education and institutions are the inefficiency effects, with an output orientation, and a truncated normal distribution. Time trends to capture both technological change and inefficiency change are also included (the Year variable).
Thank you!
When I run the Stochastic Frontier Analysis (SFA) True Fixed Effects (TFE) model by Greene (2005), oftentimes Standard Deviations, P-values, and Confidence Intervals are not reported for some of the variables, either in the frontier equation or in the inefficiency effects equation. Instead, a period (.) appears in their place. Why is this happening?
I have an unbalanced panel of 16 countries and 14 years, with around 185 observations in total.
My codes are as follows:
sfpanel lnY lnK lnL lnT Year, model(tfe) dist(tn) emean(Educ Institutions Year) ort(o)
As you can see, single output is lnY, inputs are the logarithms of capital (lnK), labor (lnL), and land (lnT), in a simple (log-linearized) Cobb-Douglas specification. Education and institutions are the inefficiency effects, with an output orientation, and a truncated normal distribution. Time trends to capture both technological change and inefficiency change are also included (the Year variable).
Thank you!
