Little is known about why minority Medicare beneficiaries spend less on and use fewer prescription drugs than white Medicare beneficiaries. We explored whether population differences in demographic characteristics, socioeconomic status, and health status were associated with observed disparities by race and ethnicity in the prescription drug spending and use of noninstitutionalized elderly Medicare beneficiaries. We used a nationally representative sample of 8101 white, 816 black, and 642 Hispanic Medicare beneficiaries from the 1999 Medicare Current Beneficiary Survey Cost and Use files. For each of these groups, we calculated total prescription drug spending, out-of-pocket spending, and number of prescriptions. We then used the Oaxaca-Blinder decomposition method to separate the impact of race and ethnicity on disparities in spending and use from the impact of differences in population characteristics across racial and ethnic groups. Much of the disparity in spending between whites and blacks and some of the disparity between whites and Hispanics can be attributed to race/ethnicity. Because of race/ethnicity, total spending for whites was 8.9% more than for blacks and 5.4% more than for Hispanics. Similarly, total out-of pocket spending for whites was 28.8% more than for blacks and 10.7% more than for Hispanics. Race/ethnicity also influenced the amount of prescription drug use. Whites used 2.3 more prescriptions than blacks and 1.6 more than Hispanics. However, these differences in use were offset by the impact of differences in population characteristics. Differences in factors identified in the Andersen model of access to care do not fully explain observed disparities in prescription drug use and spending. The portion of the disparities due to race and ethnicity may reflect patients' skepticism about medicine and medical care in general, patients' adherence to medical advice, patient-physician communication, physicians' prescribing habits, and usual source of care. Future research should explore whether these and other unobserved factors associated with race and ethnicity are responsible for disparities in drug spending and use.