Estimating Average Causal Effects with Incomplete Exposure and Confounders