Over the last decade, there has been a notable surge in the movement of refugees across international borders, posing significant challenges for the international community. In response, various policy measures have been implemented, including the construction of border walls, with the aim of impeding refugee influx. However, scholars have expressed doubts regarding the effectiveness of these fortifications, suggesting that walls merely redirect migrants to alternative routes, discourage return migration, or alter migrants’ cost–benefit calculations. Despite these concerns, there has been a lack of rigorous testing to support or refute these claims beyond case-specific evidence. This article addresses this research gap by thoroughly examining the arguments surrounding the impact of border fencing on refugee flows. We conduct a systematic, cross-national test of these arguments with a two-way fixed-effects estimator, an equivalence test, and a recently developed matching estimator designed for use on time-series cross-sectional data. Our results strongly support those who are skeptical of the impact of walls. We consistently demonstrate either that border fencing has not had any causal impact on refugee flows between 1970 and 2017 or that the statistical state-of-the-art is incapable of discerning that true effect. In either scenario, the evidence suggests that border fences fail to deliver the anticipated outcomes. These findings hold significant implications as violence-driven refugee flows persist, underscoring that while walls may serve as politically attractive tools for populist leaders, their actual deterrent effects are highly questionable at best.