ABSTRACT
A problem of interest to econometricians is the specification and estimation of a frontier production function. The original formulation of the stochastic frontier model is due to Aigner, Lovell, and Schmidt (1977) and Meeusen and Van den Broeck (1977):
Y = f (x;β) + ε, (13.1)
where the error term ε = V − U , is composed by a symmetric component, V , representing measurement error, and by the non-negative technical inefficiency component U . The stochastic frontier f (x;β) + V allows for firms with same inputs, x, to have different frontiers due to unobservable shocks; indeed, (13.1) models the inefficiency of a company to attain its production frontier; see Parsons (2002) for a treatment of stochastic frontier analysis (SFA) in marketing science.
