This paper assesses the empirical validity of Zipf's Law for cities, using new data on 73 countries and two estimation methods -OLS and the Hill estimator. With either estimator, we reject Zipf's Law far more often than we would expect based on random chance; for 53 out of 73 countries using OLS, and for 30 out of 73 countries using the Hill estimator. The OLS estimates of the Pareto exponent are roughly normally distributed, but those of the Hill estimator are bimodal. Variations in the value of the Pareto exponent are better explained by political economy variables than by economic geography variables.
This paper performs a test of Zipf's law (the size distribution of cities follows a Pareto distribution with shape parameter equal to 1) using data for Malaysian cities from five population censuses (1957, 1970, 1980, 1991 and 2000). For the full sample, Zipf's law is rejected for all periods except 1957, in favour of a city size distribution that is more unequal than would be predicted by Zipf's law. Results at the upper tail, where the distribution fits the Pareto distribution better, are more favourable to Zipf's law. Evidence is also found against Gibrat's law of proportional growth: smaller cities grow faster, as do state capitals and cities in the states of Sabah and Selangor.
a b s t r a c tThis paper explores the determinants of the choice of UK universities by overseas undergraduate applicants. We use data on overseas applicants in Business Studies and Engineering from 2002 to 2007, to 97 UK universities. Estimating using a Hausman-Taylor model to control for the possible correlation between our explanatory variables and unobservable university-level effects, we find that the fees charged may influence the application decision of some students, but that any relationship between levels of fees and applications is non-linear. The quality of education provided is positively and significantly related to the number of applications. Proximity to London and the existing popularity of a university among home applicants, are also significant predictors of university applications.
We investigate Zipf's Law on the size distribution and Gibrat's Law on the growth of sub-national populations in China, India and Brazil. We reject Zipf's Law for India, but not for China and Brazil; a log normal distribution also fits Brazil well, but not China and India. Gibrat's Law holds for Brazil; that is, lagged population is the best predictor of current population in Brazil. In China, market potential is an important predictor of population growth, while in India both crop area and market potential are important. Our results show that there is a diversity of experiences across countries, and we speculate that this diversity maybe caused by differences in the characteristics of the three countries.
JEL classification: R12
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.