This multilevel study followed 4.5 million Swedish women and men from 1 January 1998 until 31 December 1999 in order to examine the association between neighbourhood income (defined as proportions of individuals with low income) and psychiatric hospital admissions. Individuals living in the poorest neighbourhoods exhibited a statistically significantly higher risk of being hospitalised for mental disorder than individuals living in the richest neighbourhoods, after adjustment for individual demographic and socioeconomic characteristics. The neighbourhood variance indicated statistically significant differences in psychiatric hospital admissions between neighbourhoods. Both individuals and neighbourhoods need to be targeted in order to enhance mental health in low-income neighbourhoods.