Instrumental variables (IV) methods allow for consistent estimation of causal effects, but suffer from poor finite-sample properties and data availability constraints. IV estimates also tend to have relatively large standard errors, often inhibiting the interpretability of differences between IV and non-IV point estimates. Lastly, instrumental variables' idiosyncratic nature reduces their availability in data sets alongside outcome and other variables of interest. Most prior work on two-sample IV has exploited multiple data sets with the goal of attaining identification. Arellano and Meghir (1992) correspondingly propose a method equivalent to two-sample two-stage least squares (TS2SLS) to identify a model of labor supply. Under the assumption that the different samples utilized are drawn from the same population, these estimators identify parameters of interest consistently. Given its computational convenience and favorable asymptotic properties, the TS2SLS estimator is a natural choice for instrumental variables estimation under data combination. This paper aims to explore the properties and potential applications of data combination, specifically through the lens of the TS2SLS estimator. The paper, in its final form, will demonstrate the finite sample properties of the TS2SLS estimator and provide guidelines to empirical researchers to identify when using auxiliary data through the TS2SLS estimator results in preferable estimates. This will be done analytically in a basic framework where feasible, but more general propositions will be argued through simulation evidence. (Contains 1 table.)