摘要：This paper explores the population distribution across U.S. states over time. We test for Zipf's Law on the size distribution of state populations and Gibrat's Law for the growth of state populations. State populations follow a lognormal distribution more closely than they do a Zipf or Pareto distribution. State population growth is negatively related to current state population in the 19th century but not in the 20th century, and is positively related to market potential in the 20th century but not in the 19th century.
关键词：Zipf's Law; Gibrat's Law; US states; dynamic panel data models
