Health policy has shifted towards placing a greater emphasis on the role of lifestyle and life circumstances in improving health. The factors that are associated with poor health status are known, but the comparative effectiveness of specific policy interventions in improving health and reducing inequalities in health is unclear. For example, there is little evidence that specific policies aimed at providing income support or poverty eradication have any measurable impact on health. Two previous reviews have addressed the evidence in this area but in a fairly restrictive way. One considered only randomised trials and the other excluded non-cash benefits. This article builds on the previous reviews in three ways: a broader scope of study designs and types of intervention is considered; more recent literature is reviewed; and it considers the extent to which an economic evaluation framework has been applied. A systematic search of electronic databases was carried out for literature published since 1980 and in the English language. Each study was appraised in terms of its relevance to the question of interest, and the quality of the study design was appraised in terms of its capacity to provide robust answers. Few studies were found with health outcomes as their main focus. Most of the studies that used secondary data sources or survey data were of poor quality. Where economic evaluations were reported, these tended to be restricted to financial assessments. Different types of interventions were evaluated. In studies of cash benefits, there was limited evidence that they had a positive effect on some health domains, mainly psychosocial. Studies in welfare-to-work interventions produced mixed results in terms of impact on either income or health; there was no consistent relationship between income gains and health improvements. Five welfare-to-work studies included 'benefit-cost analysis', but these were essentially financial assessments. Studies of benefits in kind did not meet the quality criteria for inclusion in this article. Overall, we found no evidence of the potential cost effectiveness of income support or anti-poverty initiatives in improving health, nor is there a strong effectiveness literature on which to build such analysis. However, the hypothesis that increased income may improve health cannot be said to have been properly tested. Studies generally analyse the incremental effect of changes to the welfare system and do not estimate the health effects of current provisions. The production function for good health is complex. Increasing income may be a necessary, but not a sufficient, condition for the creation of better health in those with low incomes.