Perfect Multicollinearity and Dummy Variable Trap: Explaining the Unexplained

Recently we have come across some confused references to ‘dummy variable trap’ (DVT) during an Econometrics workshop organized at a University in Kerala, India. A google search has generated a large number of so-called ‘machine learning’-based tutorials of the very same content. In addition to these internet sources of such insufficient/incorrect information, a number of (new generation) econometrics text books also have unfortunately been found to cater to such confusions. The confusion arises from the inadequate care in discussion by some textbook authors that spreads through the mass of new generation half-wit tutorial bloggers and other media, who further venture to simplify it, and finally grips the careless novices, who get lured by the ‘simple logic’ of it. Unfortunately, they choose to ignore the authoritative text books as well as the need for an approach of mathematical logic. Note that these text books are also often insufficient to bring to light the concepts clearly. Hence this paper seeks to explain this issue in the framework of its mathematical logic.