Productivity is often computed by approximating the weighted sum of the inputs from the estimation of the Cobb–Douglas production function. Such estimates, however, may suffer from simultaneity and selection biases. Olley and Pakes (1996, Econometrica 64: 1263–1297) introduced a semiparametric method that allows us to estimate the production function parameters consistently and thus obtain reliable productivity measures by controlling for such biases. This study first reviews this method and then introduces a Stata command to implement it. We show that when simultaneity and selection biases are not controlled for, the coefficients for the variable inputs are biased upward and the coefficients for the fixed inputs are biased downward.
Exporting has always been thought of as one tool to improve productivity and, consequently, to spur economic growth in low‐ to middle‐income economies. However, empirical evidence of this so‐called ‘learning‐by‐exporting’ effect has been limited. This article determines whether learning‐by‐exporting is evident in two Turkish manufacturing sectors—the textile and apparel (T&A) and the motor vehicle and parts (MV&P) industries. A semi‐parametric estimator that controls for problems associated with simultaneity and unobserved plant heterogeneity is used to test the learning‐by‐exporting hypothesis. After controlling for these issues, our results suggest statistically stronger learning‐by‐exporting effects in the T&A than in the MV&P industry. The highly concentrated and capital‐intensive nature of the MV&P industry is the main reason for the lower learning‐by‐exporting effect in this sector. From a policy perspective, this implies that targeting export‐enhancing policies to industries with significant learning‐by‐exporting effects may lead to more productivity gains and would better stimulate an export‐led growth.