Public support to families with pre-school children can be in the form of cash benefits (e.g. child allowances) or of "in-kind" support (e.g. care services such as kindergartens). The mix of these support measures varies greatly across OECD countries, from a cash/in-kind composition of 10%/90% to 80%/20%. This paper imputes the value of services into an "extended" household income and compares the resulting distributive patterns and the redistributive effect of these two strands of family policies. On average, cash and in-kind transfers each constitute 7-8% of the incomes of families with young children. Both instruments are redistributive. Cash transfers reduce child poverty by one third, with the estimated impacts in Austria, Ireland, Sweden, Hungary and Finland performing above average. When services are accounted for, child poverty falls by one quarter and poverty among children enrolled in childcare is more than halved. This reduction is highest in Belgium, France, Hungary, Iceland and Sweden. The paper also presents simulations in which cash transfers are replaced by services, and vice versa, to provide a better understanding of these effects. The results from these simulations do not allow us to draw "generalised" conclusions as to which of the two instruments fares "better". However, in a majority of countries, if all in-kind spending on childcare were transformed into cash benefits, a lump-sum approach (i.e. a basic income supplement to all children) would be more effective in reducing poverty than an up-rating of present child benefits. The analysis in this paper is exploratory in that it considers only the first-round distributive effects of the policy instruments and does not capture additional indirect and longer-term redistributive effects, in particular possible labour supply effects and their potential impact on household incomes. The hypothetical simulations constitute extreme cases in that the entire volume of early childhood education and care (ECEC) services is replaced by cash transfers, and vice versa. The simulations nevertheless provide useful benchmarks for estimating potential losses or gains in redistribution when key elements of the early childhood policy mix are to be changed. (author abstract)