Estimating the Technology Coefficients in Linear Programming Models

    • preprint
    • Published in RePEc
Abstract
Linear constraints for mathematical programming models are . demonstrated to be random coefficient regression (RCR) models when estimating constraint coefficients from samples. Monte·carlo experiments show an RCR estimator preferable to least squares although least squares is also acceptable. Dependence between output levels and technical coefficients can lead to biased estimates.
All Related Versions