⚠️Automatic translation pending review by an economist.
« Heteroscedasticity » could almost be described as a complicated word used to scare students starting out in econometrics. It is actually used to describe a simple phenomenon: the fact that the variance of errors in a model (the famousɛi in the simplest possible modelyi=xiβ+ɛi) is not the same for all observations. The most basic model in econometrics, ordinary least squares (OLS), is based on a homoscedasticity assumption, according to which the variance of errors is the same regardless of the observation considered. Heteroscedasticity challenges this assumption and justifies the use of methods other than OLS for our estimation (GCS, GCSQ, etc.).
Julien P.
