A solution to log of dependent variables with negative observations
We propose a spatial autoregressive stochastic frontier model, which allows for endogenous weighting matrix (i.e., the spatial weighting matrix is not independent of the two-sided error term). The parameters of the model are estimated via the maximum likelihood estimation method. Monte Carlo simulations illustrate that our model performs well in finite samples. As an example, we employed our methodology to the US banks and found evidence for endogenous spillovers. The empirical example suggested potential biases in the parameter estimates when endogeneity of spillovers is ignored.
Karakaplan, M.U., Kutlu, L. & Tsionas, M.G. A solution to log of dependent variables with negative observations. J Prod Anal 54, 107–119 (2020). https://doi.org/10.1007/s11123-020-00587-5
J Prod Anal