Over the time a large number of reciprocal preferential trade agreements (RTAs) have been concluded among countries. Recently many studies have used gravity equations in order to estimate the effect of RTAs on trade flows between partners. These studies report very different estimates since they differ greatly in data sets, sample sizes and independent variables used in the analysis. So, what is the "true" impact of RTAs? This paper combines, explains and summarizes a large number of results (1827 estimates included in 85 papers), using a meta-analysis (MA) approach. Notwithstanding quite a high variability, studies consistently find a positive RTAs impact on bilateral trade: the hypothesis that there is no effect of trade agreements on trade is easily and robustly rejected at standard significance levels. We provide pooled estimates, obtained from fixed and random effects models, of the increase in bilateral trade due to RTAs. Finally, information collected on each estimate allows us to test the sensitivity of the results to alternative specifications and difference in the control variables considered.