The paper presents a meta-analysis of the empirical literature on the Rank-Size Rule for the size distribution of cities. In its pure form, the Rank-Size Rule – also known as Zipf’s law after the seminal work of George Zipf (1949) – states that the second-largest city in a given territory has on average about one-half of the population of the largest city, the number 3 city one-third of that population, and the number n city 1/n of that population. The urban economics literature provides many empirical tests of Zipf’s law as well as competing theoretical hypotheses as regards the reasons why the hierarchy of cities should follow such a pattern, or might deviate from it. Our meta-analysis considers available estimates in the literature of Zipf’s law, covering 1669 estimates derived from 59 studies. We find that in general and on average, the rank-size rule does hold remarkably well, but only rarely it holds perfectly. Interestingly, Zipf’s law holds particularly well in the USA, while the opposite is true for city-size distributions in Europe. Deep economic factors such as GDP/capita, sectoral composition, trade, infrastructure, geography and institutional quality do not explain systematic variation in Zipf coefficients. In contrast, time and continent variables explain the variation in Zipf coefficients remarkably well. We conclude that history matters in the Zipf literature: more attention in the Zipf literature is needed for the role of arrival (birth) of new cities and volatility of city growth across the city distribution. Finally, our analysis strongly suggests that explaining the deviation from the rank-size rule is a more interesting question than whether or not it holds perfectly.