You are regressing stock B's returns on stock A's returns. Let y_hat be the vector of predicted returns and e be the vector of residuals. What is the dot product of y_hat and e? What if you run the regression without an intercept?

你在用股票 A 的收益率去线性回归预测股票 B 的收益率。设预测收益率向量为 y_hat，残差向量为 e。这两个向量的内积是多少？如果不加截距项呢？