Slow onset climate change has the potential to cause significant migration flows. Scientists have recently made considerable efforts to quantify these flows based on empirical methods. However, the literature on international migration has failed to come to a clear conclusion as many studies found no significant effects of climate, while others did. In this paper, we aim to uncover a factor which likely contributes to the mixed picture in the literature: how migration flow data is obtained from migrant stock data. Using the influential study of Cattaneo and Peri (2016) as a workhorse, we demonstrate that the derived empirical results depend heavily on the applied method to derive migration flows. Therefore, our study reveals the necessity for future research on international migration to test the sensitivity of estimated effects to changes in the construction of migration flows.