Roads, Transit, and the Denseness of São Paulo's Urban Development

Taking Sao Paulo as our case study, we present causal inference that the construction of avenues and arterial roads crossing the urbanized area and connecting suburban and peripheral neighborhoods have generated urban expansion, extending the city’s urban footprint of the fastest growing city in the world in the mid-20th century. Each kilometer of new avenues and arterial roads generated between 5% and 9% increase in the local urbanization rate between 1947 and 1997. On the other hand, investments in rail transit have promoted vertical neighborhoods, increasing the floor area ratio in settings nearby Sao Paulo's central area. And each additional kilometer of transit lines was responsible for increasing local FAR in between 2.5 and 4.75, meaning 10 to 19 square kilometers of additional built-up area. This is the first joint causal inference of urban roads and transit different impacts on horizontal and vertical urban development. Our results also confirm that transit investments have stimulated specialization of land uses in Sao Paulo by attracting more commercial buildings to central areas and stimulating residential real estate development in areas further away. In this paper we take advantage of the fact that several avenues and arterial roads in Sao Paulo were built using the free course left by urban rivers that were channeled as an exogenous source of variation for the application of an instrumental variable approach. Abandoned streetcar routes were used as an instrument for recent rail transit investments. Long difference regressions were used to estimate the causal effect of these different types of transport on the type of urbanization in Sao Paulo. Besides the contribution to the academic empirical literature on the interaction between transportation and land use, our findings have huge implications for “new urbanism” movements and among contemporary planning strategies that promote transit as a strategy to promote sustainable urbanization.