The analyses at the U.S. state level indicates productivity measures estimated from the OMP programming approach with CRS technology is identical to the ideal Fisher index productivity measures for aggregate (single output and single input) technology. Divergence in productivity measures is observed not only due to choice of method –OMP and MTFP methods and various levels of commodity and input aggregation, but also between CRS and VRS technology. Due to the piecewise linear approximation of the nonparametric programming approach, the shadow share-weights are skewed leading to the difference in the productivity measures across methods, models and various levels of commodity aggregation.