Green financial policies (GFPs) have emerged as an option to align financial flows with the objectives of the Paris Agreement. While conceptual studies suggest that GFPs that target the financial sector can contribute to emissions reductions in real economy sectors, empirical evidence on their effectiveness remains unclear. To fill this gap, here we analyse GFP policy output in G20 countries and conduct the first meta-analysis and assessment of empirical evidence in this area. We present a matrix that maps evidence (gaps) concerning the adopted policies. We find that more direct intervention instruments, such as credit allocation policies or green investment by public banks, are more widely studied, and the evidence appears mostly positive. Information policies, including taxonomies and standards, have received some empirical attention, but evidence for emission reductions remains weak. Finally, novel policy approaches such as green monetary or prudential approaches remain significantly under-evaluated. Based on the available evidence, we discuss the key impact mechanisms and respective outcome variables. Overall, our analysis highlights that the empirical evidence is still limited and derives priorities for future GFP research.