Why do firms make large cash payouts, given the tax advantages of retention? The dividend puzzle is based on the premise that low or near-zero payouts are optimal (although not uniquely so) in frictionless markets, hence should be strictly optimal when payouts are taxed. This logic reflects misunderstandings about the nature of payout policy irrelevance in frictionless markets and the implications of adding personal taxes to the standard finance model. In frictionless markets, all optimal policies require substantial payouts. Those with low or near-zero distributions are strictly sub-optimal, and are infeasible given rational expectations. Payout policy irrelevance does not carry over to general equilibrium, so it is inappropriate to conclude that many or most firms can make low payouts for extended periods. Imposition of personal taxes perturbs the frictionless equilibrium on the margin, but does not alter the implication that all optimal policies require substantial payouts. In short, the standard finance model (with or without taxes) implies, requires, and predicts the large payouts observed in the world.