Spatial Knowledge Spillovers in R&D and Aggregate Productivity: Evidence from the Reunification of Germany

I quantify the importance of spatial knowledge spillovers in Research and Development (R&D) for aggregate productivity. Using new administrative dataset from German inventors, I causally estimate these spillovers by isolating quasi-exogenous variation from the arrival of East German inventors across West Germany after the Reunification of Germany in 1990. Increasing the number of inventors in a location by 10% leads to average inventor productivity gains of around 4.09%. I then embed these estimated spillovers into a quantitative spatial model of innovation, and use it to quantify the productivity gains from implementing policies that promote R&D activities. First, reducing inventor migration costs by 25% increases aggregate productivity by 5.87%. Second, the 25% subsidy for firms' expenditures in R&D within the 2020 German R&D Tax Allowance Act would increase aggregate productivity by 4.27%. Finally, the productivity gains from these policies increase with the level of spatial knowledge spillovers in R&D.

Climate Hazards and Resilience in the Global Car Industry

Climate change will increase the frequency and severity of natural disasters. I study the impact of an increase in the likelihood of weather-related disruptions on the spatial organization of firms. Using data on the global car industry and an event-study design, I document that the occurrence of a flood close to an assembly plant significantly reduces car production, which is partially reallocated to other unaffected plants of the same firm. I develop a novel quantitative multiregion model in which firms choose the location and capacity of their production plants to maximize expected profits under uncertainty about potential weather disruptions. In the model, firms have incentives to invest in production capacity in multiple locations to hedge against potential production disruptions in one of their plants. I estimate the model for the global automotive industry and use it to compute plant locations and capacity choices under different probabilities of weather disruptions according to possible climate change scenarios. In response to the heightened risk, I show that firms build additional plants, but these sites are smaller and operate with a larger spare capacity. The spatial reorganization of the firms’ plants is costly as it entails sizable productivity losses, resulting in higher consumer prices.

Fragmented Markets and the Proliferation of Small Firms: Evidence from Mom-and-Pop Shops in Mexico

Developing countries are characterized by the prevalence of small firms in the retail sector. We explain this phenomenon through a spatial model in which high transport costs lead to small effective market sizes and, consequently, to the proliferation of smaller and lower quality firms. We show that low costs of entry are key for this result. By obtaining a new, confidential panel of firm-level data surveying the universe of mom-and-pop shops in Mexico, we test the implications of our model. We exploit the deregulation of the Mexican gasoline market in 2017 as an exogenous shock on consumer transport costs. Where gas prices increased, the number of mom-and-pop shops differentially increased while their average size and quality fell. We give evidence of fragmentation and localized demand as the mechanism behind these effects. With our estimated model, we evaluate the welfare consequences of a licensing program in Mexico City which increased costs of entry for mom-and-pop shops. We show there are modest efficiency gains from having less stores in the market. 

Consumer Cities: The Role of Housing Variety

This paper estimates theoretically-founded housing price indices that account for differences in housing variety availability, where some varieties are available in certain cities but not others. I develop a model of housing demand and city choice that incorporates differentiated housing varieties and the discrete-nature of housing decisions. For estimation, I develop a method to jointly estimate the nests that varieties belong to and the elasticity of substitution across varieties within each nest. I find that households in larger cities benefit from having access to varieties not available in smaller cities. These consumption amenities are quantitatively large: the elasticity with respect to population is a third of the elasticity of housing prices with respect to population. After accounting for consumption amenities, real income increases with city size.

The dynamic allocation of public housing: Policy and spillovers

We consider the design of a large-scale public housing program where consumers face dynamic tradeoffs over apartments rationed via lotteries and prices. We show, theoretically and empiri- cally, that changing rules complements increasing supply. First, we present a motivating example in which supplying more housing leads households to strategically delay their applications. By waiting for better developments arriving tomorrow, households forgo mediocre developments available today, resulting in more vacancies. Turning to the data from the mechanism, we for- mulate a dynamic choice model over housing lotteries and estimate it. Under the existing mech- anism, we find that increasing supply fails to lower wait times. However, when a strategyproof mechanism is implemented, vacancies and wait times fall, but prices on the secondary market rise. Under this new mechanism, building more apartments lowers wait times and reduces the upward pricing pressure on the secondary market.

Are There Too Many Minibuses in Cape Town? Privatized Provision of Public Transit

Workers in low- and middle-income countries waste significant time commuting, partly due to gaps in public transit. In many African cities, privately-operated minibuses provide 50–100% of urban transit, at the cost of long wait times and poor personal safety for riders. Which externalities does the private provision of transit generate, and can policy interventions improve the market allocation? I build a micro-founded model of privatized shared transit subject to externalities in matching between buses and passengers, security provision, and road congestion. I then estimate the model with newly-collected data on minibus operations in Cape Town and stated user preferences for exogenously-varied commute attributes. I find that an optimal subsidy on minibus entry corrects matching externalities and particularly benefits low-skill workers on long routes. Government actions to improve security bring even more substantial welfare gains.

House Size and Household Size: The Distributional Effects of the Minimum Lot Size Regulation

What are the distributional effects of the minimum lot size (MLS) regulation on household welfare? An overlooked channel is how the MLS regulation increases physical house size. Using synthetic control methods, I show Houston’s reduction of the MLS (enacted in 1998) led to a 12% decrease in the size of new housing and an increase in the marginal cost of house size of around 14%. To quantify the distributional welfare effects stemming from these incentives, I build a quantitative model with housing and demographics and show that the effects of observed price changes induced by MLS regulations disproportionately hurt lower income and smaller households. Specifically, I find that the bottom decile of households (in terms of household size and income) are hurt about $25,000 more (in 2010 dollars) than the top decile. Finally, I show that the model's predicted locational selection of households by household size and income is consistent with empirical observations in Houston before and after the change in regulation.

The Aggregate Effects of the Great Black Migration

In the United States, four million African Americans migrated from the South to the North between 1940 and 1970. How did it impact aggregate US output and the welfare of cohorts of African Americans and others? To answer this question, I quantify a dynamic general equilibrium model in which cohorts of African Americans and others migrate across states. I compare the baseline equilibrium matched with the US data from 1940 to 2010 with counterfactual equilibria in which African Americans or others cannot relocate across the North and the South between 1940 and 1970. The mobility of African Americans and the mobility of others increased aggregate US output in 1970 by 0.7 and 0.3 percent, respectively. Therefore, although African Americans accounted for about 10 percent of the US population, their migration had a larger impact on the aggregate US economy than the migration of the other 90 percent did. The mobility of African Americans increased the welfare of African Americans born in Mississippi in the 1930s by 3.5 percent, decreased the welfare of African Americans born in Illinois in the 1930s by 0.2 percent, and did not change the welfare of others substantially.

Dynamic Spatial Competition in Early Education: an Equilibrium Analysis of the Preschool Market in Pennsylvania

High-quality preschool is one of the most cost-effective educational interventions, yet the United States invests little in early childhood education. Recent policy discussions call for increasing preschool enrollment and raising the quality provided, especially for disadvantaged children, but equilibrium responses of private providers which make up most of the market generate trade-offs between these objectives. Supply expansion may lower incentives to invest in quality, and price responses to demand subsidies can increase the costs faced by non-subsidized parents. This paper develops a dynamic model of the preschool market to evaluate the effectiveness of policies at achieving these objectives. The model nests a static equilibrium model of spatial competition and preschool choice within a dynamic model of providers’ entry, exit and quality investments. I estimate this model using data on the universe of child-care centers in Pennsylvania. I use the model to simulate the aggregate and distributional consequences of proposed approaches to early education expansion. I find that policies focused on expanding supply raise access but decrease the quality children attend due to parents’ value for proximity. Demand subsidies generate market expansion, but on their own do not create sufficient incentives for providers to invest in quality. Among the simulated policies, the most cost-effective at expanding high-quality enrollment combine demand subsidies targeted to low-income families with financial support to providers serving disadvantaged children. These policies increase access by reducing exit of providers, and expand high-quality enrollment for low-income children through subsidies. In addition, these targeted policies generate spillovers to the educational quality of non-targeted families by creating incentives for centers to invest in quality.

Does Market Power in Local Agricultural Markets Hinder Farmer Climate Change Adaptation?

What role do government policies which distort market competition play in impeding farmers’ climate change adaptation? We study this question in the context of India, where longer-run adaptation to climate change has been inadequate — posing a considerable risk to its ∼250 million agricultural workers. We exploit spatial discontinuities in intermediary market power, created by state-level laws that restrict farmer-intermediary transactions to the same state, to determine how spatial competition affects farmers’ adaptation. We find that a farmer selling in the 75th percentile of the competition index compared to one that faces the 25th percentile of the competition index achieves a 4.9 percent higher output for each additional day of extreme heat. This effect is driven by increased input usage by farmers in anticipation of higher prices after climate shocks, an effect limited only to to high competition areas. We then propose and estimate a quantitative spatial trade model with intermediary market power to examine the welfare implications of higher competition for adaptation. Our structural estimates suggest that the farmer’s economic loss (i.e. their climate damage function) due to extreme weather could be mitigated by 13.8 percent if government regulation distorting market competition are dismantled. These results highlight the importance of understanding the political economy of reforming these competition-distorting laws to accelerate climate change adaptation.

The Spatial Consequences of Financial Frictions: Evidence from Brazil

What are the regional and distributional consequences of government subsidies in credit markets? We provide theory and evidence to answer this question using detailed administrative data from Brazil. We build a dynamic spatial general equilibrium model with working capital constraints in which a government can subsidize credit across sectors and locations. We show that spatial linkages through trade, migration, and input-output relationships are crucial to understand the long-run consequences of such policies. Guided by the model, we evaluate the long-run sectoral and skill composition effects of a credit place-based policy in Brazil. We exploit the fact that treatment intensity varied discontinuously across a geological border. Using a dynamic regression discontinuity design, we find that, after the credit shock, treated municipalities become more agricultural-oriented and less skill-intensive. We estimate the model by matching the reduced-form moments and perform counterfactual simulations to evaluate the regional and distributional effects of credit subsidies. These simulations indicate that credit subsidies differentially increase welfare in richer regions with smaller effects on poorer ones. An alternative policy that improves bankruptcy procedures through court reform, decreasing the cost of credit in local labor markets, differentially improves welfare in poorer regions.

Mobile Human Capital and Diffusion of Ideas Across Cities

I study how internal migration of inventors affects local and aggregate growth through technological diffusion across cities. I propose a quantitative spatial theory of growth and knowledge diffusion through internal migration. My model highlights two mechanisms in which productivity growth can be higher in one city than another: (1) agglomeration forces driven by city size and (2) knowledge inflows through internal migration. I estimate the model using data on U.S. cities and find that large cities have significantly benefited from better access to technology through migration inflows from other cities. This migration effect explains approximately 40% of the cross-sectional variation in local productivity changes.

Human and Nature: Economies of Density and Conservation in the Amazon Rainforest

Conserving tropical forests could impact the standard of living of local populations. Human adaptation through sectoral or spatial reallocation of economic activity may undermine conservation policy goals. To derive policies that balance human and ecological well-being, this paper estimates a multi-sector spatial model that formalizes human-nature interactions in rainforests using high-resolution georeferenced data from roadless river basins in the Peruvian Amazon. Identification comes from plausibly exogenous variation in the structure of river networks. We find that the agglomeration externality in agricultural production outweighs dispersion forces in access to land, implying that higher concentration leads to higher productivity with less deforestation per farmer. We also find a strong congestion externality with spatial spillovers in natural resource extraction. The estimated agglomeration externality, primarily driven by economies of scale in transport technology and agricultural intensification, has large impacts on improving welfare and reducing deforestation but leads to natural resource depletion through general equilibrium effects. Counterfactual simulations demonstrate that combining well-targeted transport infrastructure and place-based protection policies can simultaneously achieve higher welfare, lower deforestation, and less natural resource depletion.

The Intergenerational Wealth Effects of Local Labor Markets

Between 1999 and 2019, income and house prices have diverged across local areas in the United States as some cities have seen persistent growth in their labor markets while others have not. These divergent trends across labor and housing markets have an effect on wealth, especially housing wealth, which persists across generations. This paper asks how the local markets of parents shape their children’s wealth and affect wealth inequality. Using an event study style analysis, I find that children who grew up in better local labor markets have, on average, $45,000 higher net-worth as adults. This association is only true for the children of homeowner parents. To measure the aggregate effect of this divergence on wealth inequality, I build a parsimonious, multi-region modeling framework, and find that the dispersion in local labor market growth accounts for 40% of the rise in wealth inequality amongst the bottom 90% of households, primarily because parental concerns about bequests make households in growing areas save disproportionately more.

Labor Mobility, Nonlocal Outside Options, and Wages

In labor markets with frictions, outside options are a key determinant of workers' wages and firms' rents. When outside options improve, workers can benefit by switching jobs, but even those that remain in the same job can realize gains through leveraging their improved bargaining position to renegotiate wages. In this paper, I study the geographic dimension of workers' outside options and effects on labor mobility and wages for job stayers. A large share of job-to-job (J2J) flows are across metropolitan areas (MSAs), and at the MSA level, a greater J2J outflow rate is associated with higher average earnings. To obtain causal estimates of the effect of nonlocal outside options on wages and mobility, I construct measures of exposure to labor market conditions in other markets and use a shift-share instrumental variables strategy to identify exogenous variation in non-local labor demand. I find that increases in labor demand in an MSA's network of labor markets are associated with increased job-to-job outflows with an elasticity of about .30, and greater wage growth for job stayers with an elasticity of .14. Labor mobility is much more responsive to demand in closely connected MSAs, which account for about 83% of the total mobility effect. I find similar mobility responses across education levels, but the effect of nonlocal outside options on wage growth is concentrated on workers without a 4-year college degree and in industries with lower average education levels.

How do Establishments Choose their Location? Taxes, Monopsony, and Productivity

To study the distribution of economic activity across space and the effects of place-based policies, I develop a model of the location choice of new establishments incorporating taxes, monopsonistic labor markets, and spillovers. Estimates using administrative data from Germany indicate that establishments generally have a preference for lower taxes, as well a preference for lower worker outside options which enable establishments to pay lower wages. The degree to which various types of productivity spillovers matter in the location decision of establishments varies greatly between industrial sectors. I also quantify the effects of a counterfactual place-based policy and find that commuting zones display highly heterogeneous wage and economic activity responses to the same policy due to differing degrees of labor market power across space.

The Dynamics of Urban Development: Evidence from New York

This paper studies the persistent impact of zoning and land use policies on urban development and welfare. It exploits an ambitious policy experiment led by then-mayor, Michael Bloomberg, in New York City during 2002-2013. The reform relaxed zoning regulations and increased the floorspace allowance, equivalent to 5% of the city's total housing stock. Using a boundary discontinuity design and high-resolution spatial data that tracks investment activity over time, the paper shows that the quantity of supply responded sluggishly to the relaxation of the zoning. The policy-induced new development generated positive spillover effects in surrounding areas but were delayed by the slow response in the housing supply. To quantify the reform's implications for welfare, the paper develops a dynamic spatial model. It features the interaction between the dynamic of housing supply and the spatial reallocation of heterogeneous workers. Estimating a dynamic spatial model with a large number of heterogeneities is computationally challenging. The paper uses Euler's method combined with a key property from competitive land market equilibrium to reduce the computation burden. Using the estimates, it solves the dynamic spatial model to study the equilibrium impact of the reform. It finds that the reform leads to a 0.7% increase in total housing supply in the long run and delivers welfare benefits for both high-skilled and low-skilled workers. Contrary to the common perception, it finds that the reform slows down the gentrification in the city. Counterfactual simulations further show that the fixed adjustment costs are pervasive in this market, eroding 67% of the policy's potential impact. The findings highlight the persistence of supply frictions thwarting the growth of productive cities.

Locally Optimal Place-Based Policies: Evidence from Opportunity Zones

The extent to which public policy can encourage new investment into areas that need it, and how those policies should be targeted, remain open questions. This paper evaluates the impact of Opportunity Zones on new residential and commercial development, and quantifies how policymakers could have achieved a more efficient response through alternative designations of the investment tax credit. Using a novel dataset on the location and timing of new development projects in large U.S. cities, I find that receiving the tax credit increases new development in census tracts by 2.9pp (20.5%). I also find positive spillovers on nearby development. Both effects are larger in neighborhoods with more available land to develop, more elastic housing supply, and lower home values. Through a model of new development that accounts for location-heterogeneities, dynamics, localized spillovers, and the equilibrium behavior of developers, I find that the policy as implemented had city-wide impacts on new development on the order of 2.7%. However, optimally chosen Opportunity Zones would have substantially increased the investment response. The results suggest that there is substantial scope for equity and efficiency improvements in how the program was implemented.

Displacing Congestion: Evidence from Paris

This paper shows that road-closing policies may have adverse short-run effects on pollution by reallocating traffic toward more congested roads. I study the impact of the 2016 closure of the Voie Georges Pompidou, a one-way expressway crossing downtown Paris, on traffic and pollution displacement. To do so, I rely on a difference-in-difference strategy based on the direction and the timing of traffic, which I implement on detailed road-sensor data. I show that the closure lowered average speed by over 15% on two sets of substitute roads: central streets nearby and the already congested southern ring road. Using air quality data, I show that NO2 emissions increased by 6% near the ring road and by 1.5% near local roads. The reduced-form results on traffic are quantitatively consistent with a calibrated model of shortest route choice, which allows me to recover the underlying rerouting patterns. Even though few displaced commuters diverted to the ring road, they triggered a massive pollution increase because of the U-shaped relationship between emissions and traffic speed. Overall, I estimate that up to 90% of the pollution cost was borne by lower-income residents around the ring road, who lived far away from the new amenity created by the closure and mostly outside the jurisdiction responsible for the closure decision. Finally, I study counterfactual closure scenarios to assess under which conditions those adverse effects could have been mitigated.

The Impact of Construction Constraints on Spatial Allocation during Economic Transition

I show that without persistent spatial planning and building permitting that does not reflect pent-up demand for new construction, large cities would be even larger as well as overall economic output would be higher.I propose a novel housing supply function explicitly depending on maximum allowed density of development, amount of zoned developable land and stringency of local permitting process. Additionally, the function incorporates the concept of kinked supply function resulting from durability of housing capital. I then embed this function within a quantitative spatial model framework based on Ahlfeldt et al. (2015). Analyzing data from the Czech Republic during its transition to a market economy between 1991 and 2011 I find substantial variation in housing supply price elasticities with inter quartile range spanning from 0.39 to 0.77. However, low supply price elasticities in many places are not caused by constraining planning, but rather low price levels of real estate.Counterfactual analyses show combination of relaxation of stringent policies by upzoning by 50%, increasing developable densities by 50% and keeping permitting stringency at first pentile of national level would increase indirect utility by up to 4% along with re-allocation of up to 2.6% workforce generally from rural areas to cities.

Location Choice, Commuting, and School Choice

We study the interaction between residential location choices and public school applications/assignments under centralized school choice. Residential locations determine commuting distances to schools and admissions probabilities, which households consider when deciding where to live—a neglected factor in the empirical school choice literature. We propose a framework that captures this interaction, whereby households’ observed and unobserved heterogeneity generates residential and school segregation. We estimate the model using New York City student-level data. Variation from a boundary discontinuity design separately identifies preferences for access to school from other location amenities. The decomposition exercise illustrates that this preference for access to school explains 30% of school segregation. In the counterfactual policy analysis, we study to what extent equalizing admissions probabilities to the highest-performing schools across locations can desegregate schools. Whereas the policy reduces school segregation by 7% when we take households’ location choices as given, the effect is dampened by half from households’ endogenous location choices. In contrast to a widespread concern, the choice to opt out of public schools to outside schooling options plays a smaller role.

Trade Costs, Supply Chains, and the Decline of the Heartland

This paper studies how changes in domestic trade costs can cause regions to decline. The agriculture-intensive states of the American Midwest (the heartland) lost population relative to the rest of the country over the postwar period. I document that the price of shipping agricultural relative to manufactured goods fell considerably over this same period. To show how these two facts may be linked, I outline a simple version of a trade model and derive comparative statics of the price, production, and population effects of a decline in agricultural shipping costs. The model highlights how the input-output structure of the economy contributes to these population effects. I validate the model's predictions by studying how a 1963 Supreme Court ruling that sharply reduced the cost of shipping wheat versus flour affected the flour milling industry. Finally, I calibrate a multi-sector, multi-location version of the model to the U.S. in 1950 and find that observed declines in agricultural trade costs can explain nearly 10% of the postwar population decline in the heartland.

How Mortgage Financing Costs Affect Rental Housing: Pass-Through & Pricing

This paper studies the effect of changes in financing costs on landlords and tenants using property-level variation across mortgages (from refinance & acquisitions) and within mortgages (from interest only periods).

Rural marketplaces and local development

Marketplaces are an age-old way to connect geographically separated producers and consumers,and they remain widespread in rural areas of low-income countries. How do these gatherings shapedevelopment around them? This question is not easily answered, since we typically lack comprehensivemarket maps and localized indicators of development. To address these long-standing data gaps, Icombine historical sources with novel satellite-based methods to map marketplaces and measure localpopulation density. I focus on Kenya over the last five decades and establish three stylized facts. First,while rural population quadrupled, two thirds of weekly markets operating in 1970 no longer do sotoday. Second and despite many markets no longer operating, population concentrated on averagearound markets that were active in 1970. Third, markets further from large cities saw the mostpopulation concentration relative to their surroundings. To rationalize these findings and deriveimplications for policy design, I extend a model of rural-urban trade with markets as population-independent locations that aggregate otherwise sparse supply and demand and enable economies ofscale in transportation. The model explains when new markets emerge, why some markets decline,and which complementary policies catalyze markets for local development.

The Sharing Economy as Disaster Remedy: The Case of Airbnb

Environmental disasters increasingly displace people without much notice worldwide. The massive growth in sharing economy has unlocked a new source of short-term housing supply for the displaced people. In this paper, I quantify the welfare impacts of home sharing on the short-term displacement in the context of wildfire evacuations for one of the world’s largest home-sharing markets - the Los Angeles area. I develop and estimate a structural model of the home-sharing market under informational asymmetry on customer type (disaster refugee versus regular traveler), which highlights two welfare channels, namely the increased choice set of housing as well as the altruistic sharing by hosts. I find that the displacement loss amounts to at least 31% of the property damages by wildfire. Airbnb can reduce this welfare loss by 52%, with a quarter of mitigation contributed by supplier generosity. I show that altruistic sharing is mostly conducted by hosts who are more well-off as indicated by the value of demographics and home characteristics. I also estimate the welfare spillovers as a result of the free riding of regular travelers. I recommend that a platform targeting on displaced people can improve the efficiency and equity of mitigation gains.

Moving Into Risky Floodplains: The Spatial Implication of Flood Relief Policies

This paper studies the welfare implications of post-flooding relief policies in a dynamic spatial economy with flood risk and mobile individuals. We argue that mobility leads to trade-off in the design of relief policies: on the one hand, migration frictions imply imperfect adjustment via moving after flooding, generating equity gains of relief policies; on the other hand, policies induce endogenous sorting into floodplains, potentially causing spatial misallocation of population depending on location characteristics. To confront the data, we develop a fully quantitative general equilibrium model with rich geographic linkages, industry structure,and flood risk. We develop a new solution method with neural network to overcome the course of dimensionality. We calibrate the model to the economy of Texas coastline and the event of Hurricane Harvey. We show that existing post-Harvey relief transfer improves U.S.welfare compared to a zero-relief economy when the transfer is financed equally by the rest of U.S.. However, a more welfare-enhancing policy, instead, is to reduce the transfer into the most affected locations while increase the amount into the Houston metropolitan area.This reshuffled policy induces less spatial misallocation of population within Texas coastal floodplains. Finally, a resilient policy of providing moving subsidy to incentivize individuals to move away from Texas coastal floodplains leads to lower U.S. welfare compared to a zero-subsidy economy.

The Welfare Effects of Congestion Pricing: Evidence from High-Occupancy Toll Lanes

This paper estimates the effects of highway congestion pricing on traffic, economic activity, and welfare. Using spatial panel data on real-time traffic speed and flow in California, I first provide reduced-form evidence that changes in traffic patterns differ in the short- and medium-run when non-toll lanes are converted to toll-lanes with dynamic pricing. This implies that individual responses may involve not just changing where they drive, but also where they live or work. Therefore,to estimate the full equilibrium effects, I develop a quantitative urban model with endogenous commuting costs in which residential and commercial locations, driving routes, travel times, and toll costs are simultaneously determined. Based on model estimates, I estimate both the partial and general equilibrium effects of congestion pricing. In the partial equilibrium analysis, which holds the locations of residences and workplaces fixed, congestion pricing induces a spatial leakage of traffic externality as people divert from toll lanes to non-toll lanes; this reduces annual aggregate welfare by $1.8-$11.0 million. However, in the general equilibrium analysis, which allows for adjustments in residences, workplaces, and driving routes, congestion in the overall road network decreases because people re-sort to reduce commuting distances. In aggregate, when net toll revenues are redistributed, annual welfare increases by $2.4-$11.6 million.

Commuting and the value of marriage

Over time, as metro-areas sprawled to the suburbs, long commutes became common. In this paper I combine motivating evidence with a structural model to show how housing policies resulting in long commutes affect singles differently than couples (and men differently than women), if evaluated in a joint housing and marriage market equilibrium. First, I show that the gender gap in commuting among singles is negligible. Second, men in couples (not women) have much longer commutes than single men, and residential choice cannot explain this difference. This suggests that commuting features gains from specialization harnessed within couples, allowing men to take better jobs. I embed this feature in a quantitative spatial model with endogenous marriage and location choices that successfully captures the commuting and location patterns by marital status. In equilibrium, gains from specialization in commuting have the following implications: as metro areas expand in the model, commuting increases most for men in couples and employment falls most for women in couples, widening gender gaps in both outcomes. However, in terms of welfare it is singles who lose the most. Couples are able to partially evade commuting costs through specialization, lower housing costs and redistributing resources within the household.

 The Effects of Foreclosure on Households

US homeowning households tend to be less geographically mobile than households that do not own a home. However, it is unknown whether the inducement to move that foreclosure provides can lead households to make beneficial moves even under adverse circumstances. This paper documents how foreclosure affects the outcomes of US households through location. I link mortgage and foreclosure deeds and property characteristics with residential moves of US households and their corresponding neighborhood characteristics. Variation in the foreclosure process used across US states puts observably similar households at greater risk of foreclosure in those that follow a nonjudicial (faster) rather than judicial (slower) process. I use the resultant variation in foreclosure rates across borders in a spatial regression discontinuity design (RDD) to estimate local average treatment effects of nonjudicial foreclosure process and, incorporating two-stage least squares, of being foreclosed. Under a nonjudicial foreclosure process, households are more likely to be foreclosed (2 pp), to move out of their home (3 pp), and to undertake more distant moves (2 pp). Moves come not only from completed foreclosures but also from preventative moves by households at risk of foreclosure. Furthermore, households experience relatively greater increases in destination relative to origin tract median income ($1,000) and household income rank at age 29 of children who grew up there (0.5 percentile) under a nonjudicial rather than judicial foreclosure process. Together these results suggest that any potential long-term negative effects of foreclosure on household members are not driven by adverse changes in neighborhood characteristics.

Right Idea, Wrong Place? Knowledge Spillovers and Spatial Misallocation in R&D

A few cities perform a high share of research and development (R&D) in the United States. To the extent that R&D generates local knowledge spillovers which are not internalized by private researchers, the geographic distribution of R&D may be inefficient. Should policymakers favor R&D in certain cities, and if so which ones?

I unpack the rationale for place-based R&D policy through the lens of a spatial growth model. Researchers create new ideas, sell them, and enjoy spillovers from other researchers’ ideas. Geography matters if knowledge diffusion across space is imperfect. Imperfect diffusion changes the size of the spillover—the pool of ideas available for learning—and so generates variation in the social returns to R&D. I emphasize a complementary implication: imperfect diffusion introduces variation in the private returns to R&D because barriers to diffusion limit the size of the market for a new idea. The optimal policy subsidizes research in regions with high spillovers relative to profits, rather than simply high spillovers. Even if an idea does not create much profit, it can benefit future researchers. If knowledge diffusion were perfect, an idea’s place of origin would be irrelevant and there would be no efficiency motive for place-based R&D policy.

I turn to patents data to estimate the size of diffusion barriers and back out the social and private returns to R&D across cities. I consider two exogenous barriers: one from researcher to researcher, and another from researcher to firm. Prior literature has documented that inventors tend to cite nearby inventors. I put forth new evidence that inventors tend to sell their patents to nearby firms. Patent sales decline more steeply with distance than do patent citations. All else equal, an inventor in Denver is only 20% as likely to sell a patent to a firm in faraway Boston compared to a firm in nearby Colorado Springs. The same inventor is 55% as likely to receive a citation from Boston as to receive a citation from Colorado Springs. My interpretation is that researchers in remote markets earn lower profits for their ideas because they struggle to sell at distance. However, the same ideas are relatively visible to other researchers, who can learn effectively at distance.

The optimal R&D policy aligns the social and private returns to research across cities. I calibrate a quantitative version of the model to match the observed distribution of wages, production employment, and patenting. The quantitative model matches untargeted R&D employment well. I then implement the optimal R&D allocation with a budget-neutral subsidy and tax on research employment. In practice, the policy reallocates researchers to remote regions which under-innovate in equilibrium, consistent with the reduced-form intuition from the patent regressions. The policy increases patenting by 2.8% and aggregate consumption by 0.8% in the long run. The gains are purely from reducing misallocation, as I deliberately fix aggregate resources devoted to R&D. The policy has negligible effects on inequality between regions or workers.

House Prices and Rents in a Dynamic Spatial Equilibrium

House prices and rents do not always comove across locations and over time. To study the causes and welfare consequences of rent and price variation, I set up a quantitative dynamic spatial equilibrium model of housing demand and supply. In the model, price-to-rent ratios can vary because of differences in expected rental growth or discounting, and data on prices, rents, migration and construction contain identifying power to separate the two. I take the model to data in the case of Finland, where house prices have been diverging across regions, even if rents have not. Through the lens of the model, the rapid price divergence between big and small cities can be rationalized as the combination of an increase in the value of living in cities and regionally diverging discount rates. These changes have led to an important regional divergence in both renter welfare and housing wealth across smaller and larger cities.

Racial Segmentation in the US Housing Marke

This paper studies racial segmentation in the US housing market since 1960. I document large differences in housing outcomes for Black and White households. In 1960, relative to a White household with the same income, a Black household at the median of the national income distribution was 13 percentage points less likely to own a house; 12 percentiles behind in the distribution of house values, if they owned; and even when renting they were 6 percentiles behind in the distribution of rental expenditures. By 2019, the rent and price gaps have declined by about half, whereas the gap in ownership rates has not changed. To interpret these facts, I use a dynamic housing assignment model with a choice to buy or rent housing. I estimate the degree of market segmentation by inferring differences in the quality of housing available to Black and White households, and the resulting differences in rents, prices, and the cost of owning a home. The model infers that Black households pay higher quality-adjusted rents and prices, especially at the high end, and thus sort into lower quality homes. In terms of lifetime consumption-equivalent welfare, relative to an integrated market, the average Black household is five percent worse off in 1960 and remains one percent worse off in 2019. 

Consumer Search and Firm Location: Theory and Evidence from the Garment Sector in Uganda

In low-income cities economic activity is spatially concentrated, with firms in the same sector operating next to one another in informal clusters. This paper studies the role of consumer information frictions in driving this behavior. I develop a quantitative equilibrium model in which imperfectly informed consumers favor searching in high-density locations that let them minimize the cost of gathering information. When choosing location, firms trade-off consumers’ preferences for agglomeration, fiercer competition induced by spatial proximity, and lower production costs from supply-side externalities. To estimate the model, I collect a new dataset from garment firms in Kampala. This combines transaction data- which allows estimating demand, customer data- which provides evidence on consumer search, and a mystery shoppers exercise- which yields accurate measures of product quality. I find that information frictions contribute to 8.2% of the observed agglomeration. Removing them leads high-quality firms to gain market share from lower-quality competitors, inducing an 11% increase in consumer welfare. Counterfactual scenarios show that the introduction of an e-commerce platform would induce a large share of firms to disperse, while also causing customers to shift to high-quality businesses. By contrast, urban decongestion policies would have an adverse impact of disproportionately harming high-quality firms by increasing consumers’ search costs.

Urban Policy and Spatial Exposure to Environmental Risk

In the past two decades, about half of the new homes in the United States were built in environmentally risky areas. Why is new residential development being exposed to such risk? I posit that land-use regulations restricting development in safer areas contribute to this pattern. I study this question in the context of exposure to wildfire risk in the San Diego, California, metropolitan area, where I show there is a positive correlation between built-out areas and fire safety. I estimate a quantitative urban model using detailed spatial data on zoning, density limits, lot size restrictions, wildfire risk, and insurance. In the model, the regulations benefit landowners and reallocate the population to unregulated at-risk areas. These effects depend on estimated disamenities from wildfire risk, insurance access, and the spatial correlations between regulations, wildfire risk, and location amenities. I find that land-use regulations raise city-level rents by an average 28% and explain 7% of the residents living in fire-prone areas. The estimated present-discounted cost of wildfire risk is $14,149 per person, with existing regulations accounting for 10% of that cost. Over the next 40 years, as wildfire risk intensifies, the population grows, and the current land restrictions become more binding, the number of exposed residents will grow by 12%. The results show that institutions that restrain relocating out of harm’s way, such as land-use regulations, can limit adaptation to climate change.

Wage Inequality and the Spatial Expansion of Firms

Multi-region firms increasingly dominate the U.S. economy. I study the implications of this trend for labor market inequalities. I document that multi-region service firms account for most of the rise in wage inequality since the 1980s, and provide evidence on the uneven nature of their spatial expansion: larger firms operate establishments in more locations, while hiring more skilled labor and paying higher wages in spatially-concentrated headquarters. I integrate this structure into a general equilibrium model, in which (a) firms open branches to serve local markets; (b) the output of headquarters workers is non-rival across branches; (c) firms have wage-setting power. The resulting wage distribution depends on the full network of firm spatial activity, and inequality rises with firms' geographical scope. The model admits tractable aggregation despite its complex micro-structure. I estimate it for 391 U.S. labor market areas and infer frictions to spatial expansion from the universe of HQ-branch linkages. Quantitatively, the decline in these frictions since the 1980s can account for multiple trends in U.S. labor markets, including rising inequality across establishments – between and within firms – and higher inequality and segregation across space.

The Impact of Place-Based Housing Subsidy on Academic Performance

Housing affordability and accessibility is an ongoing policy debate in the UnitedStates. An understudied issue is the extent to which access to affordable housing affectschildren’s human capital formation. In this paper, I study the short- and medium-runeffects of students in grades 2 to 11 moving into developments funded by the Low-Income Housing Tax Credit (LIHTC), a place-based rental housing subsidy policy. Igeocode and link the San Diego Unified School District administrative data to theCalifornia LIHTC database. Combining propensity score matching and difference-in-differences, I find an 8.73 percent decrease in the absenteeism rate, a 0.049 standarddeviation increase in standardized English scores, and a 0.048 standard deviation in-crease in standardized math scores for students who moved into LIHTC during thestudy period. I also find positive effects on high school completion, college enrollment,and college completion.

Option Value and Storefront Vacancy in New York City

Why do storefronts remain empty for more than a year in some of the world’s highest-rent retail districts? We construct and estimate a dynamic, two-sided model of storefront leasing to investigate possible explanations using data from New York City. The model incorporates salient features of the retail leasing market: heterogeneous tenant quality, high move-in costs, search frictions, asymmetric contract dissolution costs for landlords and tenants, and aggregate uncertainty in downstream retail demand. We estimate the model parameters by matching quarterly vacancy rates, lease-up rates, and tenant exit rates from a comprehensive, high-frequency storefront tracking service, combined with micro data on commercial leases. We find that tenant heterogeneity and move-in costs jointly explain long-run vacancy by generating dispersion in match surplus and therefore option value for landlords. In a counterfactual exercise, eliminating either feature results in vacancy rates of close to zero. Search frictions and aggregate uncertainty play much smaller roles. Finally, we use the estimated model to quantify the impact of a retail vacancy tax on long-run vacancy rates, average rents, and social welfare. Vacancies would have to generate negative externalities of $29.68 per square foot per quarter (about half of average rents) to justify a 1% vacancy tax on assessed property values.

The effect of urban migration on educational attainment: evidence from Africa

Does rural to urban migration in developing countries improve the opportunities available to children? And does childhood urban exposure increase educational attainment? Using census data for 14 African countries combined with an age-at-move design, I show that childhood exposure to cities significantly raises primary school completion, school attendance, and literacy rates. The increase in educational attainment is robust to the inclusion of household fixed effects, visible in all subgroups and countries, and particularly large for girls. The paper hence provides evidence of a channel through which urbanization can stimulate economic growth in developing countries, even in the absence of structural transformation.

The Effect of Deindustrialization on Local Economies: Evidence from New England Textile Mill Towns

This paper documents the persistence of local shocks in the absence of social insurance from a key episode of deindustrialization in the US: the decline of the New England textile industry in the 1920s. Using a continuous difference-in-differences strategy, I show that exposure to the industrial decline did not significantly affect the town-level population, and individuals in textile-heavy towns did not migrate out of affected areas. Instead of out-migrating, young residents responded with increased educational attainment. Using a matched difference-indifferences design that exploits variation in timing and location of textile plant closures and variation within towns based on critical age of education, I find that young individuals in plant closure towns increased their eighth-grade completion but their labor market outcomes did not improve by 1940. Older workers switched to the agricultural sector and faced decreased occupational earnings. The localized effect of the shock suggests revisiting the debate on place- versus person-based policies, and highlights the importance of targeted education and training suited for the local economic needs and the role of a diverse local industrial base in local labor market resilience.

Local Attachment and Residential Mobility: Evidence from White Flight in Boston

I document and explain the persistence of household residential choices in Boston during the Great Migration. The large-scale in-migration of African Americans was a demographic shock to existing neighborhood composition. Using novel panel data that tracks individual residential choices over time, I document stark differences in out-migration responses to this shock among two religious groups. To understand these differences in neighborhood attachment, I estimate a dynamic panel data discrete choice model that disentangles state dependence in neighborhood choice from unobserved heterogeneity in preferences for neighborhood characteristics. I find that the differences in the two groups’ out-migration behavior are mostly explained by their differential state dependence. I show that this state dependence is strongly correlated with observable measures of investments in durable local infrastructure, such as the size of religious gathering places and the number of religious schools. My counterfactual policy analysis suggests that maintaining a mixed racial composition of urban neighborhoods in Boston during the relevant period would have required a voucher that reduces rents by 25 percent.

A Dynamic Spatial Knowledge Economy

Cities have long been thought to drive economic growth, but analyses of spatial policy have largely ignored the effects of such policies on growth. In this paper, I develop a spatial endogenous growth model in which heterogeneous agents make forward-looking migration decisions and human capital investments over the life cycle. Local externalities in human capital investment drive both agglomeration and growth. I show that the growth rate depends on the spatial distribution of human capital, making it sensitive to spatial policies. I calibrate the model to data on U.S. metropolitan areas. I then solve for the allocation that would be chosen by a national planner. I find that, relative to the decentralized allocation, the planner’s allocation features more concentration in the largest cities, higher levels of human capital overall, and faster growth. Commonly-proposed place-based policies fall far short of, or even oppose, the planner’s policies.

The Geography of Black Economic Progress After Slavery

This paper studies the long-run effects of slavery and Jim Crow on Black Ameri- cans’ economic outcomes. We trace each Black family’s linked census and adminis- trative records between 1850 and 2000 to measure how long they were enslaved and where they lived during Jim Crow. We show that Black families who were enslaved until the Civil War have considerably lower education, income, and wealth today than Black families who were free before the Civil War. The disparities between the two groups have persisted because most families enslaved until the Civil War lived in states with strict Jim Crow regimes after slavery. In a regression discontinuity de- sign based on ancestors’ enslavement location, we show that states’ Jim Crow regimes sharply reduced Black families’ economic progress in the long run, largely by limiting their access to education. Using quasi-experimental variation, we show that gaining school access closed 80 percent of the gap in human capital caused by exposure to strict Jim Crow regimes.

Moving Opportunity. Local Connectivity and Spatial Inequality

Within-country inequality across space is large. How does the connectivity of place determine underlying spatial inequality of opportunity? To answer this question, I derive a theoretical sufficient statistic result linking local opportunity to market access terms, developing a framework consistent with a broad class of spatial general equilibrium models. I empirically validate this result using a novel not-on-least-cost-path identification strategy and data from historical road maps in Benin, Cameroon, and Mali covering 1970 to 2020, that I digitize. Using these estimates to parameterize a structural case of the spatial model, I show that road building alters the spatial distribution of opportunity. By considering each possible road upgrade I show that although some roads decrease the standard deviation of opportunity by more than 2\%, others increase inequality by a similar amount. Policymakers also face an equity-efficiency trade-off: On average only 6 of the top 10 aggregate opportunity-increasing roads also decrease inequality of opportunity. A back-of-the-envelope calculation shows that between 13% and 44% of people living in low-opportunity locations would need to migrate to high-opportunity areas to achieve reductions in inequality similar to those I estimate to be possible through road upgrades.

A Penny for your Thoughts

How do communication costs affect the production of new ideas and inventions? To answer this question, we study the introduction of the Uniform Penny Post in Great Britain in 1840. This reform replaced the previous system of expensive distance-based postage fees with a uniform low rate of one penny for sending letters anywhere in the country. The result was a large spatially-varied reduction in the cost of communicating across locations. We study the impact of this reform on the production of scientific knowledge using citation links constructed from a leading academic journal, the Philosophical Transactions and the impact on the development of new technology using patent data. Our results provide quantitative causal estimates showing how a fall in communication costs can increase the rate at which scientific knowledge is exchanged and new ideas and technologies are developed. This evidence lends direct empirical support to an extensive theoretical literature in economic growth and urban economics positing that more ideas can emerge from communication between individuals.

Local Human Capital and Firm Creation: Evidence from the Massification of Higher Education in France

Do college expansion policies promote local economic development? This paper exploits a massive construction policy of new colleges in France over the 1990s to assess the causal effect of higher education establishments. I start by studying their impact on education, firm dynamics, employment and wages at the city level. Leveraging the staggered implementation of the policy in an event-study design, I find a persistent rise in the level of education of the local workforce. Firm creation subsequently increases by 10% on average across all major industries. The rise in (i) tradable and (ii) skill-intensive industries indicates that the supply of educated workers played a major role in increasing firm entry. Incumbent firms experienced lower growth and a higher exit rate following the policy, suggesting displacement effects. Overall, the positive effects dominate, resulting in increased economic activity in treated cities. Employment stays constant but is subject to large composition effects: it increases for young skilled workers but decreases for older workers. In addition, province-level analysis suggests that new colleges had non-negative spillover effects on surrounding areas. Finally, I complement city-level results with findings on the long-run individual effects. Relying on differences between cohorts induced by the timing of the policy, I find that cohorts directly exposed to new colleges became more educated, more likely to be employed and hold more skilled positions.

Innovation Clusters and Spatial Inequality with a Local Brain Drain

Proximity to innovative counties makes low-educated counties experience slower growth in educational attainment in the US. Each innovative county, together with the other counties within 90km, is defined as an innovation cluster. The divergence inside innovation clusters can be explained by a local brain drain: talents are drawn to innovative counties by attractive high-skilled wages, given that migration costs are low. I document the fact that the local divergence of educational attainment increases the spatial inequality inside the innovation clusters, and contributes to a U-shape spatial inequality across the country during 1980-2010. I develop and calibrate a dynamic spatial model with two types of skill and endogenous education decisions, which reproduces the U-shape spatial inequality. Based on this model, I assess the place-based policies that some local governments subsidize tuition with income tax revenue. 50% tuition subsidies in 28 low-educated counties inside the innovation clusters increase 0.1% real GDP at the national level after 10 years. The place-based subsidies do not decrease spatial inequality, as people leave after getting higher education.

Returns to Political Contributions in Local Housing Markets

This paper investigates whether firms donate to political campaigns in order to influence supply in local housing markets. Using new data on campaign donors of U.S. mayoral candidates and a regression discontinuity design, I uncover three findings. Consistent with political favors, connection to the mayor causes residential development firms to sell more new housing units. Favors to donors shape local housing supply, since mayors attracting more donations from residential developers double permits for new housing construction. But differences in housing policy between mayors are empirically more important than favors for determining local housing supply.

Retail Demand Interdependence and Chain Store Closures

Why do retail stores agglomerate, and what happens when those patterns ofagglomeration change? I describe a model of interdependent retail demandbased on multi-stop shopping, and then estimate empirical interdependence bymeasuring how local store foot traffic responds to national chain bankruptciesand widespread closures. I find a 4%-5% average decline in nearby store foottraffic after a Sears or K-mart closure and a 5%-6% average decline in foot trafficafter a grocery closure, with considerable heterogeneity in observed declinesacross different nearby-store types. I fit a structural model of interdependentretail demand with parameters identified from these reduced-form estimates andthen use this model to simulate a decline in in-person grocery demand, such aswould occur under increased take-up of online grocery delivery. I project thatwhile the median nearby store would lose roughly 0.08% of traffic for every 1%decline in grocery traffic, up to a quarter of adjacent hardware stores wouldsee 0.4% or higher declines for every 1% drop in grocery traffic. The medianelasticity is projected to be 20% higher if in-person grocery demand decreasesonly for higher-income consumers. In a separate counterfactual, I simulate theeffect of remote work on retail demand and project that negatively-affectedstores would lose 0.15% of traffic on average for every 1% reduction in workerdensity, closely in line with other empirical studies.

The Emergence of Exclusionary Zoning Across American Cities

This paper identifies how changing postwar demographics in American cities caused their suburbs to adopt a form of land use control still widespread today — minimum lot sizes. I develop an algorithm detecting bunching on lot sizes, observable when governments' lot size controls bind developers from building denser housing. Applying the algorithm to national assessor records, I estimate which lot size controls first came into effect and their impact on new homes over time for 7,000 local governments. Most suburbs adopted lot size controls from 1945-1970, the same period when four million Black Americans left the South for opportunities in American cities. I then use the Second Great Migration as a natural experiment that shifted central cities' racial composition toward Black Americans. From 1940-1970, the rise in central city Black composition in non-Southern central cities modestly accelerated minimum lot size adoption while further explaining binding density controls applied to at least 600,000 housing units. Migration of lower-income whites into the same cities yields null effects on suburban lot size outcomes. In states that passed early legislation to desegregate public schools, Black migration had the largest effects on lot size restrictiveness. Together, the results indicate that local governments designed land use controls to exclude Black migrants from neighborhoods and public goods.

Cycling Toward Cleaner Cities? Evidence from New York City’s Bike Share Program

What is the impact of cycling on air quality in cities? This paper leverages the staggered rollout of New York City’s bike share program to estimate the effect of cycling on air pollution concentrations. I combine the universe of bike share trips with ground-level, high-resolution observational air pollution measures between 2010 and 2019. Through a routing algorithm, I use the location of bike share stations to map areas where road traffic is expected to decrease after the introduction of bike share. I compare these areas with others where traffic was likely unaffected using a staggered difference-in-differences strategy to retrieve causal estimates. I find that the deployment of bike share is associated with a 3% reduction in black carbon and 13% reduction in nitric oxide concentrations, both pollutants associated with road traffic. In addition, I investigate potential mechanisms and show that the introduction of bike share is associated with a decrease in short taxi trips in areas served by bike share, which I interpret as suggestive evidence that bike share substitutes road traffic.

Measuring Welfare Gains from Online Stores: Theory and Evidence from the Supreme Court’s Wayfair Decision

We study how the rise of e-commerce has reshaped consumer welfare and its distributional implications in the presence of retail oligopoly. Based on new data on shopping receipts, we document consumer heterogeneity in online retailing markets: households living in rural areas and with higher incomes are more likely to shop online. To quantify the welfare effects, we leverage an exogenous tax shock by the Supreme Court's Wayfair Decision to learn about online store substitutability and firm pricing responses. We then develop and estimate a structural demand and supply model focussing on the pet food retail market. The model allows us to decompose the consumer online welfare gains into gains from varieties (9 %) and convenience (5 %) and gains from pro-competitive effects (3 %). We further characterize the distributional effects of the rise of e-commerce and find it has reduced consumption inequality between rural and urban areas but increased consumption inequality between the rich and the poor.