OBJECTIVE We sought to test the assumption that an emergency department observation unit can be funded through the reallocation of resources made available through the unit's impact in reducing inpatient admissions and facilitating bed closures. METHODS We conducted our study in a tertiary care center ED with 46,000 visits annually. For a 3-month period, all patients admitted to the hospital through the ED were screened by an emergency physician for suitability for admission to an observation unit. Any patient in the hospital for 3 days or less who did not undergo surgery or other inpatient procedure, and who was admitted through the ED, was considered a candidate for the observation unit. RESULTS Of 1,840 admissions, 147 patients met the admission criteria. Only 48 (32.2%) could have been treated in an observation unit, and these patients were not admitted to any single unit in high frequency. The potential savings from inpatient bed closures would only have amounted to 1.68 full-time equivalents-not enough to staff a 4-bed observation unit, which would require 5 full-time equivalents. CONCLUSIONS Because of the diffuse and inconsistent effect such a unit had on inpatient bed use, funding for an ED observation unit at our institution could not be justified on the basis of the closure of inpatient beds and transfer of resources.