During the past 3 decades, research has identified several psychosocial evidence-based practices (EBPs) for people with severe mental illness (SMI). Starting from a different origin, the recovery movement has influenced perceptions of how EBPs and other services should be delivered, and also emphasized the value of peer supports. We now know much more than 30 years ago about the kinds of services that help people with SMI live satisfying lives in the community. Evidence-based and recovery-oriented services require additional resources but use them sparingly: they are highly individualized, often result in reductions in costs of other mental health services, such as hospitalizations, and favour reliance on and integration into community settings rather than mental health services. Nevertheless, access to such services remains very limited. During the same period, the place of medications in the services system has become a source of growing concern, and there are several reasons to believe that current spending on medications is excessive. Inadequate housing and community supports that increase lengths of stay unnecessarily and spending on ineffective, nonrecovery-oriented vocational services are only 2 additional forms of misallocation of resources. Devolving control over medication budgets to regional or local health authorities, introducing program budgeting and marginal analysis, and implementing individual budgets to give more control to service users (in addition to promoting shared decision making) merit further investigation as potential strategies to improve outcomes for people with SMI in Canada in the context of limited budgets.