Results from the first order conditions for the least squares problem
Summary
- In this section \(e_i\) are the OLS residuals and \({\hat{y}}_i\) are the OLS fitted values
- Result for \(j=1, \ldots ,k\) :
\[\sum_{i=1}^{n}{ x_{i,j}e_i }=0\]
- Specifically, for \(j=1\) ( \(x_{i,1}=1\) ):
\[ \sum_{i=1}^{n}{ e_i }=0\]
- and \(\bar{e}=0\) where \(\bar{e}\) is the sample average of the OLS residuals.
- Result: The sample mean of the regressand data is equal to the sample mean of the OLS fitted values,
\[\bar{y}=\bar{\hat{y}}\]
- Result:
\[\sum_{i=1}^{n}{ e_i{\hat{y}}_i }=0\]
- Note that \(\sum_{i=1}^{n}{ e_i }=0\) and \(\bar{y}=\bar{\hat{y}}\) will fail to hold if we do not include “1” in our regressor data. \(\sum_{i=1}^{n}{ e_i{\hat{y}}_i }=0\) will still hold.