We tackle two limitations of standard instrumental variable regression in experimental and observational studies: restricted estimation to the conditional mean of the outcome and the assumption of a linear relationship between regressors and outcome. More flexible regression approaches that...