Despite the many papers analyzing the causal effect of retirement on health, there is no consensus in the literature as to whether retirement deteriorates or improves health. This study is the first attempt to a large-scale quantitative meta-analysis. Using 576 results from 61 causal studies, we find that 15% of the results indicate that retirement deteriorates health, while 36% of the results indicate that retirement improves health. Our results indicate that the variation in effects found in the literature is primarily due to variation in health measures and most prominently due to the use of mortality versus other health measures.