Consider any regression model, linear or nonlinear, with dependent variable \(y\) and exogenous variables \(\mathbf{x}\), and iid errors \(\varepsilon\). Show that if the probability of missing data on \(\mathbf{x}\) is independent of \(y\), then the regression based on listwise deletion will provide a consistent estimate of the conditional mean function. [Show that the conditional distribution of \(y\) given \(\mathbf{x}\) is not affected by missing observations.]