Cognitive impairment among residents has considerable resource implications for both individuals and those responsible for publicly funded care. Two linked surveys were carried out in England: (1) a longitudinal study followed 2500 admissions to publicly funded care up to 42 months after admission; and (2) a cross-sectional survey of 618 homes collected information about 11,900 residents. Information was collected about cognitive impairment using the Minimum Data Set Cognitive Performance Scale. Cognitive impairment was associated with source of funding and type of home. Although level of cognitive impairment has some effect, fees and costs were most influenced by type of home. At the same level of impairment, self-funded residents were more likely to be located in relatively low-cost settings than publicly funded residents. In independent homes fees were lower for publicly funded than for self-funded residents. Overall median length of stay of publicly funded admissions was 18 months. For the most part length of stay was not associated with level of cognitive impairment on admission. It is concluded that more information is needed about the effect of quality of care on people with cognitive impairment in different settings. If the same quality of care can be achieved in residential and nursing homes, the evidence would suggest that changes in placement policies could result in potential savings to the public purse.