Sudden rises in migration across the borders of the Global North have persistently attracted substantial media attention and fuelled hostility towards “irregular migrants” and “bogus refugees.” While existing qualitative studies have extensively criticised the migrant-refugee distinction, we offer unique quantitative evidence of how migration numbers and labels construct impressions of increased irregular migration while in fact creating “fake illegals.” We conduct a two-stage mixed-method analysis, demonstrating first that data on “irregular/illegal border crossings” (IBCs) published by Frontex have become an authoritative source of information on irregular migration flows cited in a corpus of mainstream news media articles. We then posit that while persecutions and violence in countries of origin may trigger migration, it is policies in destination states that determine who “is” and “isn’t” a refugee. We therefore develop a novel method to divide IBCs into those who would likely obtain asylum in 31 European destination states (“likely refugees”) and those who would not (“likely irregular migrants”) across time given asylum acceptance rates by nationality. We estimate that between 2009-2021 most border crossers labelled as “irregular/illegal” (55.4%) were actually “likely refugees,” a proportion we estimate to be 75.5% at the peak of arrivals in 2015. We thus find that sudden and large increases in border crossings concentrated in space likely concern forced rather than irregular migrants. Our constructivist approach thus unveils how migration data and categories both influence and are influenced by securitized border policies and reveals that, in this respect, borders start with numbers.