This assumption, when combined with the assumption of independent observations generated from a random sample, implies that ei is strictly exogenous. How do we check whether this is a reasonable assumption? We need to ask whether ei includes any variables that have an effect on SALES (are correlated with SALES), and are also correlated with PRICE or ADVERT. If the answer is yes, strict exogeneity is violated. This might happen, for example, if the pricing and advertising behavior of Andy’s competitors affects his sales, and is correlated with his own pricing and advertising policies. At the moment, it is convenient if we abstract from such a situation and continue with the strict exogeneity assumption.2