Low levels of infrastructure quality and quantity in countries and regions can create trade impediments through increased transport or transaction costs. Since the late 1990s an increasing number of trade studies have been taking infrastructure into account, although many other studies continue to measure trade costs without considering infrastructure related factors. The purpose of the present paper is to quantify the importance of infrastructure for trade by means of meta-analysis and meta-regression techniques that synthesize various studies. The type of infrastructure that we focus on is mainly public infrastructure in transportation and communication. The null hypothesis that infrastructure does not have an impact on trade is examined by means of estimates obtained from thirty-five primary studies that yielded 517 infrastructure elasticities of trade. We explicitly take into account that infrastructure can be measured in various ways. Weighted average effect sizes are calculated by assuming random or fixed effects and taking account of variation across different methodologies, infrastructure definitions, geographical areas, model specifications, and publication characteristics. We estimate several meta-regression models (random effects, Hedges model of publication bias, weighted and ordinary least squares). Our results suggest that the estimated impact of infrastructure on trade varies especially according to the location of infrastructure, model specification, publication influence (journal rank), and the advocacy of the publisher. The meta-analysis suggests that at average study characteristics, and corrected for publication bias and unobserved heterogeneity, the infrastructure elasticity of exports is about 0.41 and the infrastructure elasticity of imports is about 0.16. These results suggest quite large benefit cost ratios of additional infrastructure investment in relatively open economies.