frontier

frontierProd

Description

The frontierProd action estimates the parameters of stochastic frontier production models using maximum likelihood methods. It supports various model types such as exponential, half-normal, and truncated-normal, and allows for the estimation of technical efficiency.

Settings
ParameterDescription
boundsImposes simple boundary constraints on the parameter estimates.
classSpecifies the classification variables.
displaySpecifies the list of display tables that you want the action to create.
freqSpecifies the observation frequency variable.
includeinternalnamesWhen set to True, adds an extra column to the parameter estimates table that shows the internal names for the parameters.
initialvaluesSpecifies the initial values for parameters.
modelSpecifies the model definition, including dependent variables, effects, and model options (like type: EXPONENTIAL, HALF, TRUNCATED).
optimizerSpecifies parameters that control various aspects of the parameter estimation process, such as the technique (algorithm) and convergence criteria.
outputSpecifies details for an output data table to contain scores, predicted values, residuals, and technical efficiency estimates.
outputTablesSpecifies the list of display tables that you want to output as CAS tables.
restrictionsSpecifies linear restrictions to be imposed on the parameter estimates.
tableSpecifies the table of data to be analyzed.
testsSpecifies linear hypotheses about the regression parameters that are specified in the model and tests to be performed on the hypotheses.
weightSpecifies details for the observation weight variable.
Data Preparation View data prep sheet
Creation of Production Data

Creates a CAS table with output (log_q) and inputs (log_l, log_k) for production function estimation.

Copied!
1DATA mycas.production_data; INPUT firm $ year log_q log_l log_k; DATALINES;
2F1 2020 5.2 2.1 3.5
3F1 2021 5.4 2.2 3.6
4F2 2020 4.9 1.9 3.2
5F2 2021 5.1 2.0 3.3
6F3 2020 6.0 2.5 4.0
7F3 2021 6.2 2.6 4.1
8; RUN;

Examples

Estimates a basic stochastic frontier production model using log inputs.

SAS® / CAS Code Code awaiting community validation
Copied!
1 
2PROC CAS;
3frontier.frontierProd TABLE={name='production_data'} model={depVars={{name='log_q'}}, effects={{vars={'log_l', 'log_k'}}}};
4 
5RUN;
6 
Result :
Estimates of the production frontier parameters.

Estimates an exponential frontier model, outputs technical efficiency statistics to a table, and uses a specific optimization algorithm.

SAS® / CAS Code Code awaiting community validation
Copied!
1 
2PROC CAS;
3frontier.frontierProd TABLE={name='production_data'} model={depVars={{name='log_q'}}, effects={{vars={'log_l', 'log_k'}}}, modelOptions={type='EXPONENTIAL'}} optimizer={algorithm='NEWTONRAPHSONWITHRIDGING'} OUTPUT={casOut={name='prod_efficiency', replace=true}, te1='TechEff', pred='PredOutput'};
4 
5RUN;
6 
Result :
An output table named 'prod_efficiency' containing the Technical Efficiency (TechEff) and Predicted Output (PredOutput) for each firm, along with the parameter estimates table in the report.