“…However, when the individuals in the sample differ in technology and efficiency with differenced information, especially when such heterogeneity changes with time, the linear or parametric specification cannot fully describe the heterogeneity in the production function and may induce a bias in the measurement of technical efficiency. Conventional methods (either DEA or SFA) attribute the model misspecification errors to inefficiency (Fu, 2005;Balaguer-Coll et al, 2007;Grösche, 2009;Joseph et al, 2010;Coelli, 1992, 1995;Kumbhakar and Lovell, 2000;Wu 2003). Researchers have relaxed distributional assumptions in the error component and parametric assumptions in SFA to achieve a more reliable measurement of technical efficiency (Greene, 2005;Kneip and Simar, 1996;and Henderson and Simar, 2005).…”