To assess whether newborns' insurance coverage was associated with differences in the allocation of hospital services. Retrospective analysis of computerized hospital discharge data, comparing resource allocation among newborns according to insurance status, controlling for race/ethnicity, diagnoses, hospital characteristics (ownership, teaching status, nursery level), and disposition. All California civilian acute-care hospitals. Population-based sample, excluding out-of-hospital and military hospital births. Resource allocation was studied among all newborns discharged in 1987 with evidence of serious problems (N = 29,751). Length of stay, total charges, and charges per day. Sick newborns without insurance received fewer inpatient services than comparable privately insured newborns with either indemnity or prepaid coverage. This pattern was observed across all hospital ownership types. Mean stay was 15.7 days for all privately insured newborns (15.6 days for those with indemnity and 15.7 days for those with prepaid coverage), 14.8 days for Medicaid-covered newborns, and 13.2 days for uninsured newborns (P less than .001). Length of stay, total charges, and charges per day were 16%, 28%, and 10% less, respectively, for the uninsured than for all privately insured newborns (P less than .001). Resources for newborns covered by Medicaid were generally greater than for the uninsured and less than for the privately insured. Both uninsured and Medicaid-covered newborns were found to have more severe medical problems than the privately insured. The findings cannot be explained by differences in medical need or by differences in non-medically indicated services; they constitute prima facie evidence of inequities that need to be addressed by policy changes.