We estimate non-cash income from owner-occupied housing, subsidised rental housing, and free use of the main residence, and evaluate their impact on the household income distribution and selected inequality measures. We use a novel data set that provides inter-subjective information on housing quality, thus eschewing selection problems. We confirm the standard finding in the literature that imputed rents accruing to home owners have an equalising effect on the distribution of income and find similar evidence for non-cash income from subsidised rents. Whereas imputed rents equalise the upper part of the income distribution, subsidised housing has an equalising effect on the lower part of the income distribution. Overall, the effect of non-cash income from owner-occupied housing clearly dominates the distributional effects, which translates into a combined effect of around 13% higher equivalised household income for the bottom half and around 8% for the upper half of the distribution. Our data provide us with two rare opportunities: to distinguish between council and cooperative housing, and to apply all three approaches used to calculate imputed rents for owner-occupiers: the capital-, the self-assessment and the equivalent rent approach. We find that using the equivalent rent approach leads to the strongest reduction in income inequality. This suggests that, while the inequality reducing effects of imputed rents are robust, they might be overestimated in the literature.