Veterans of US military service are a valuable resource for epidemiologic studies, and the Department of Veterans Affairs (VA) files provide an effective way to gather mortality information on veterans, so long as these files provide reasonably complete death reporting. To determine the completeness of VA death reporting, we assembled an independent sample of known veteran deaths among males born between 1936 and 1955 and assessed the performance of VA death reporting in this sample. We also compared VA death ascertainment to Social Security Administration (SSA) ascertainment. Based on the more than 4300 deaths in our study, we found VA death reporting to be approximately 90% complete by itself and 96% complete when used in conjunction with SSA death reporting. In addition, we found no evidence that VA death reporting changed substantially after passage of the Omnibus Budget Reconciliation Act of 1981, which limited eligibility for VA death benefits. Because veterans make up a large segment of the US population, our findings have particular relevance for studies in which mortality is a primary end point.