Outliers and Model Selection: Discussion of the Paper by Søren Johansen and Bent Nielsen
本文讨论模型选择与稳健估计应联合处理，指出脉冲指示变量饱和法虽能实现这一点，但会导致变量数多于观测数，并以Tobin食品数据为例说明。
Abstract It is argued that model selection and robust estimation should be handled jointly. Impulse indicator saturation makes that possible, but leads to the situation where there are more variables than observations. This is illustrated by revisiting the analysis of Tobin's food data.