In this article I test empirically whether intratemporal preferences between (1) nondurables and labour supply and (2) nondurables and housing are separable. In doing so, careful attention is paid to the potential endogeneity of the total budget, labour supply and housing. For housing I offer two separate solutions to the problem of correlation between house size, proxied by number of rooms, and unobserved heterogeneity: first, instrument with a children's sex composition dummy and second, select on households residing in public housing. The results strongly reject separability between nondurables and labour supply. For nondurables and housing the results are more mixed.