Many surveys suffer from increasing unit and item nonresponse, which results in more reliance on imputation techniques. In this paper, we describe our experimental application of Sequential Regression Multivariate Imputation (SRMI) to impute income-related variables in the 2019 Consumer Expenditure Interview Survey. We incorporate administrative data from IRS W2s, 1040s, and 1099s to address potential biases from item nonresponse. We find that SRMI-imputed values from models with IRS records used as inputs perform better than models without IRS records when comparing overall distributions. Our findings suggest that SRMI imputation would informatively contribute to the distributions of several elements on the CE.