Inconsistencies of the PLS-PM approach to structural equation models with formative-reflective schemes

The principal aim of the paper is to show the inconsistencies of PLS-PM as a statistical tool to estimate structural equation models with formative-reflective schemes. We show that PLS-PM is not suitable to this kind of models, since it may even fail in identifying existing causal links between the variables of concern. We bring evidences of these facts discussing a real-data example and comparing the results by PLS-PM with those obtained by a simpler least squares approach