We investigate the comparative technical efficiency of producer cooperatives (PCs) and conventional firms (CFs) by looking at the performance of a mixed sample of Sardinian wine producing companies over the period 2004-2009. Thanks to the similarity of the habitats in which the firms operate, the peculiarities of the production environment, and the careful measurement of some key inputs through suitable aggregation of accounting data, the observed units are “twins” in all non- organizational respects, providing one natural setting for comparative work. The analysis is carried out in two steps: in the first, technical efficiency indicators for each firm in each year are calculated using Data Envelopment Analysis (DEA) with reference to a common production frontier. Subsequently, the measured efficiency scores become the dependent variables of a pooled truncated maximum likelihood regression in which we control for external covariates and firm type. To assess the procedure’s appropriateness, we test whether the separability condition that the support of the output variables does not depend on the set of external variables is satisfied. Moreover, a double bootstrap algorithm is run to compute valid standard errors and confidence intervals of the coefficients estimates. According to our findings cooperatives are less technically efficient than their capitalist counterparts and displays decreasing returns to scale. Both results are particularly worrying in light of the main challenges (liberalization of EU planting rights and climate changes) facing the wine industry in the near future.