Are the well-known facts about urbanization in the United States also true for the developing world? We compare American metropolitan areas with comparable geographic units in Brazil, China and India. Both Gibrat's Law and Zipf's Law seem to hold as well in Brazil as in the U.S., but China and India look quite different. In Brazil and China, the implications of the spatial equilibrium hypothesis, the central organizing idea of urban economics, are not rejected. The India data, however, repeatedly rejects tests inspired by the spatial equilibrium assumption. One hypothesis is that the spatial equilibrium only emerges with economic development, as markets replace social relationships and as human capital spreads more widely. In all four countries there is strong evidence of agglomeration economies and human capital externalities. The correlation between density and earnings is stronger in both China and India than in the U.S., strongest in China. In India the gap between urban and rural wages is huge, but the correlation between city size and earnings is modest. The cross-sectional relationship between area-level skills and both earnings and area-level growth are also stronger in the developing world than in the U.S. The forces that drive urban success seem similar in the rich and poor world, even if limited migration and difficult housing markets make it harder for a spatial equilibrium to develop.

ABSTRACT. Much of the inequality literature has focused on national inequality, but local inequality is also important. Crime rates are higher in more unequal cities; people in unequal cities are more likely to say that they are unhappy. There is a negative association between local inequality and the growth of city-level income and population, once we control for the initial distribution of skills. High levels of mobility across cities mean that city-level inequality should not be studied with the same analytical tools used to understand national inequality, and policy approaches need to reflect the urban context. Urban inequality reflects the choices of more and less skilled people to live together in particular areas. City-level skill inequality can explain about one-third of the variation in city-level income inequality, while skill inequality is itself explained by historical schooling patterns and immigration. Local income also reflects the substantial differences in the returns to skill across, which are related to local industrial patterns.

We acknowledge support from the Taubman Center for State and Local Government. Chauvin acknowledges support from the Center for International Development at Harvard University. The views expressed herein are those of the authors and do not necessarily reflect the views of the National Bureau of Economic Research. At least one co-author has disclosed a financial relationship of potential relevance for this research. Further information is available online at http://www.nber.org/papers/w22002.ack NBER working papers are circulated for discussion and comment purposes. They have not been peerreviewed or been subject to the review by the NBER Board of Directors that accompanies official NBER publications.

In the last 50 years, population and incomes have increased steadily throughout much of the Sunbelt. This paper assesses the relative contributions of rising productivity, rising demand for Southern amenities and increases in housing supply to the growth of warm areas, using data on income, housing price and population growth. Before 1980, economic productivity increased significantly in warmer areas and drove the population growth in those places. Since 1980, productivity growth has been more modest, but housing supply growth has been enormous. We infer that new construction in warm regions represents a growth in supply, rather than demand, from the fact that prices are generally falling relative to the rest of the country. The relatively slow pace of housing price growth in the Sunbelt, relative to the rest of the country and relative to income growth, also implies that there has been no increase in the willingness to pay for sun-related amenities. As such, it seems that the growth of the Sunbelt has little to do with the sun.

One approach to urban areas emphasizes the existence of certain immutable relationships, such as Zipf's or Gibrat's Law. An alternative view is that urban change re ‡ects individual responses to changing tastes or technologies. This paper examines almost 200 years of regional change in the U.S. and …nds that few, if any, growth relationships remain constant, including Gibrat's Law. Education does a reasonable job of explaining urban resilience in recent decades, but does not seem to predict county growth a century ago. After reviewing this evidence, we present and estimate a simple model of regional change, where education increases the level of entrepreneurship.Human capital spillovers occur at the city level because skilled workers produce more product varieties and thereby increase labor demand. We …nd that skills are associated with growth in productivity or entrepreneurship, not with growth in quality of life, at least outside of the West. We also …nd that skills seem to have depressed housing supply growth in the West, but not in other regions, which supports the view that educated residents in that region have fought for tougher land-use controls. We also present evidence that skills have had a disproportionately large impact on unemployment during the current recession.Comments are appreciated and can be sent to eglaeser@harvard.edu, gponzetto@crei.cat, and kristina_tobio@ksg.harvard.edu.