AbstractMuch has been written about fishery‐induced evolution (FIE) in exploited species, but relatively little attention has been paid to the consequences for one of the most important parameters in evolutionary biology—effective population size (Ne). We use a combination of simulations of Atlantic cod populations experiencing harvest, artificial manipulation of cod life tables, and analytical methods to explore how adding harvest to natural mortality affectsNe, census size (N), and the ratioNe/N. We show that harvest‐mediated reductions inNeare due entirely to reductions in recruitment, because increasing adult mortality actually increases theNe/Nratio. This means that proportional reductions in abundance caused by harvest represent an upper limit to the proportional reductions inNe, and that in some casesNecan even increase with increased harvest. This result is a quite general consequence of increased adult mortality and does not depend on harvest selectivity orFIE, although both of these influence the results in a quantitative way. In scenarios that allowed evolution,Nerecovered quickly after harvest ended and remained higher than in the preharvest population for well over a century, which indicates that evolution can help provide a long‐term buffer against loss of genetic variability.