From rivers to roads: Spatial mismatch and inequality of opportunity in urban labor markets of a megacity

Highlights

• Workers with longer daily commuting are less productive.

• Workers with higher accessibility to jobs are more productive.

• Policy choices in the past may impose strong costs to spatial sustainability in the present.

• Geography and history reinforced the patterns of inequality of opportunity in the labor market in SPMR.

Introduction

Location decisions of firms and workers shape the spatial distribution of economic activity between and within cities. On one hand, the interaction between cities is widely investigated in the literature of regional and urban economics, which tries to assess the extent to which urban scale affects the local concentration of different skills, sectors, etc., apart from defining each city's role in the regional system. On the other hand, within-city dynamics and internal heterogeneity is the focus of urban labor economic theory, simultaneously analyzing the interaction of land/housing markets with the labor market (Zenou, 2009).

Furthermore, the spatial scale of analysis shapes the choice of tools and measures to investigate specific issues. For instance, agglomeration economies are usually assessed by aggregated measures such as total population or density. However, this type of analysis does not provide a complete understanding of the intra-urban distribution of agglomeration economies effects (Melo & Graham, 2009). An alternative approach requires the estimation of market potential measures over labor market indicators, with distance decaying effects varying over space. In a way, accessibility to jobs embeds the concept of market potential and a higher chance of interactions and matching in the job market. Thus, agglomeration economies may have heterogeneous effects within cities, affecting location decisions of workers and firms.

In this paper, we look at how economic agents benefit from such interactions within a specific urban labor market area. Any equilibrium in the labor market will be directly related to the housing market. Therefore, wages and housing prices will alternatively be the interest variables, being jointly affected by, among other characteristics, the accessibility to jobs in the city.

We use the São Paulo Metropolitan Region (SPMR), the main economic and financial center of Brazil, as our case study. SPMR is the fourth largest urban agglomeration of the world, and the largest urban agglomeration in the country, with about 10% of the national population and 19% of Brazilian GDP. The city of São Paulo is the core of the metropolitan area and accounts for 5.9% of the country's population and 12% of its GDP. In 2007, the reference year of our main database, SPMR housed 19.5 million individuals, while the population of the city of São Paulo (MSP) reached 10.9 million inhabitants. The SPMR population density in the same year amounted to 2,458 inhabitants per km² (in an area of 7,947 km²), while in the MSP it was of 17,155 km² (in an area of 1,523 km²).

Furthermore, the rate of car ownership in SPMR was of 184 cars per 1,000 inhabitants in 2007, similar to the rate observed in 1997 (according to Origin-Destination surveys), with 3.6 million private cars in total. From the 38.1 million daily trips that happened in the SPMR in 2007, motorized vehicles were responsible for 66.0% of them. Finally, private cars did 31.7% of the trips from home to work. As we will discuss, the spatial mismatch between residential locations and jobs can be particularly relevant for poorer individuals, who rely more heavily on public transportation.

In what follows, we explore the relationship between commuting time, accessibility to jobs and urban prices in the context of urban markets in a megacity in the developing world. We start by lining up the theoretical framework that permeates the empirical analysis. The key point is the simultaneous determination of land use and urban prices in the context of the family of spatial general equilibrium urban models we consider. That brings challenges for dealing with simultaneity bias in the empirical estimation of urban wages (and housing prices) models. Our goal is to measure how individual wages are affected by the potential access to job opportunities. We develop an accessibility index that accounts for production externalities by weighing the travel time of each location to the various job locations in SPMR, in a rich geographical setting. The simultaneity bias in this case may arise as both residents and firms' locations are determined at the same time. Thus, wages paid in a specific part of the city will depend on local supply and demand for jobs, which, in their turn, will also affect wages. Wages may also influence the number of jobs openings through their impact on production costs faced by firms. Therefore, accessibility to jobs, which is the explanatory variable of interest in our analysis, depends on some of the factors that affect wages and may also be indirectly affected by wages. Then, a positive shock to individual wages may change the number of job openings, so that the measurement of the causal relationship from accessibility to wages will require that we isolate variations in local job openings that can be reasonably assumed to be exogenous to potential shocks on wages. We address this issue by relying on an instrumental variables estimation that uses the river shore access to the city's founding location as instrument for accessibility to jobs.

The paper adds to the debate on spatial sustainability in large human settlements in developing economies (Fan et al., 2014, Lau and Chiu, 2004). Policy choices in the past may impose strong costs to spatial sustainability in the present. In the case of São Paulo, current patterns of accessibility to jobs, imposed by physical infrastructure constraints, were heavily influenced by the development of the urban road network. Geography helped to define the design of such network that, together with a historical strategy that privileged motorized vehicle, reinforced the observed patterns of inequality of opportunity in the labor market.