This article investigates the impact of unions on the wages of hospital workers. Our OLS findings agree with previous OLS studies--unions increase registered nurses' (RNs) wages by five percent and by about eight to ten percent for other hospital workers. By contrast, we find (after correcting for selectivity bias in hospital unionization status) a direct union effect of about twenty percent on RN wages and in excess of thirty percent on wages of other hospital workers. While the results based on selectivity bias adjustments make us uneasy, we do not reject them out-of-hand. We also find indirect union effects (up to five percent) by other unionized occupations within a hospital and up to ten percent by other unionized hospitals in the local labor market. Prospective reimbursement programs have a negative impact on the wages of hospital workers but are only significant for non-unionized occupations. Our three empirical tests of monopsony all reject the view that monopsony is a factor in hospital wage-setting. Even considering the large union effects (based on selectivity bias adjustment), we conclude that unions have been a minor contributor to hospital cost inflation.