Panel nonlinear autoregressive stochastic frontier models with dynamic technical inefficiency

In this paper, we consider the relationship of an autoregressive panel data to exogenous variable‎. ‎In standard regression for finance data‎, ‎it has been assumed the errors are uncorrelated‎, ‎but this assumption fails in practical data. ‎So‎, ‎we introduce nonlinear autoregressive panel stochastic frontier model that error consists of statistical error and technical inefficiency‎. A semiparametric approach is proposed to estimate nonlinear autoregressive function for the model and apply Taylor series expansion‎, ‎which has a parametric framework as a representation of the nonlinear function. Expectation maximization estimation method is applied for parametric estimation and the obtained nonlinear autoregressive function is adjusted by a nonparametric factor‎. ‎