Production analysis with asymmetric noise

Abstract Symmetric noise is the prevailing assumption in production analysis, but it often violated practice. Not only does asymmetric cause least-squares models to be inefficient, can hide important features of data which may useful firm/policymaker. Here, we outline how introduce into a or cost framework as well develop model inefficiency said models. We derive closed-form solutions for convolution and distributions, log-likelihood function, inefficiency, show determinants heteroskedasticity, efficiency skewness allow heterogenous results. perform Monte Carlo study profile analysis examine finite sample performance proposed estimators. R Stata packages that have developed apply three empirical applications our methods lead improved fit, explain hidden by assuming symmetry, approach still able estimate scores when exhibits well-known “wrong skewness” problem analysis. The are modeling risk linked outcome variable allowing error asymmetry with without inefficiency.