*2.2. Answer*

Our reference to a profit maximization behavioral assumption is exclusively related to the calculation of the Lerner index per se. This should not be confused with the cost minimization behavioral assumption, which is behind the estimation of a cost function.

As it is well known in the literature, the calculation of the Lerner index is based on the assumption of static profit maximization, hence the mark-up should always have to be non-negative. However, as Kumbhakar et al. (2012, p. 113) note, when the marginal cost (*MC*) is calculated through the estimation of a cost function there is no guarantee that it would take a non-negative value for each observation. In contrast, the Kumbhakar et al. (2012) stochastic frontier methodology, which we follow in our research, rectifies this problem by always producing non-negative values of the Lerner index.

#### **3. The Reasons for Not Estimating the Translog Cost Function**
