Institutional graduation rates occupy a prominent place in institutional research and public policy. Graduation rates are used in the College Scorecard, state performance funding initiatives, and potentially affect a significant proportion of public institutions revenues. Despite their widespread use, research suggests that institutional graduation rates are most strongly related to students' entering characteristics and stable institutional characteristics, but are only weakly related to characteristics institutions can directly control. One set of institutional characteristics that appears to be related to graduation rates are expenditures for instruction, academic support, student services, and institutional support. However, inconsistencies in research findings raise the possibility that estimates of the effects of expenditures on graduation rates may be biased due to omitted variables (i.e., unobserved heterogeneity). The present research uses within-/between-effects panel data models with IPEDS panel data to account for omitted variable bias and examine the effects of institutional characteristics, cohort characteristics, and institutional expenditures on graduation rates.