Approximating grouped fixed effects estimation via fuzzy clustering regression

We propose a new, computationally efficient way to approximate the “grouped fixed effects” (GFE) estimator of Bonhomme and Manresa (2015), which estimates grouped patterns unobserved heterogeneity. To do so, we generalize fuzzy C-means objective regression settings. As clustering exponent m $$ approaches 1, converges GFE objective, recast as standard generalized method moments problem. replicate empirical results (2015) show that our delivers almost identical estimates. In simulations, approach offers improvements in terms bias, classification accuracy, computational speed.