Many hospitals have introduced formularies to reduce hospital pharmacy expense, among other reasons. This study provides empirical evidence of the influence of hospital formulary restrictions on pharmacy charges, all other hospital charges, and on length of stay, using a survey of hospital drug policies and hospital discharge data from Washington State in 1989. Limiting the number of drugs in particular therapeutic categories reduced total charges incurred for gastrointestinal disease and asthma patients, increased total charges for cardiovascular disease patients, and had no effect on charges for infectious diseases patients. Restricting availability of drugs reduced pharmacy charges, but these savings tended to be offset by increases in other charges. Combining the categories, we found that restricting availability of drugs did not affect charges. We conclude that across-the-board restrictions do not result in cost savings, although savings may be realized for particular drug categories.