Despite being central to government education policy in many countries, there remains considerable debate about whether resources matter for pupil outcomes. In this paper we look at this question by considering an English education policy initiative--Excellence in Cities--which has been a flagship policy aimed at raising standards in inner-city secondary schools. We report results showing a positive impact of the extra resources on school attendance and performance in Mathematics (though not for English) but, interestingly, there is a marked heterogeneity in the effectiveness of the policy. Its greatest impact has been in more disadvantaged schools and on the performance of middle and high ability students within these schools. A simple cost-benefit calculation suggests the policy to be cost-effective. We conclude that additional resources can matter for children in the poorest secondary schools, particularly when building on a solid educational or ability background. However, small changes in resources have little or no effect on the "hard to reach" children who have not achieved a sufficiently strong prior level. Data Appendix is included. (Contains 15 tables, 3 figures and 27 footnotes.)