Research

Tax incentives are widely used to encourage the construction of rental housing, and reduce housing costs for low-income households. This paper examines an important source of local variation in the impact of these incentives: the home-biased supply of capital from potential landlords. Building upon administrative data covering the universe of dwellings in France, I document that private investors own undiversified and lumpy amounts of rental housing, face substantial return frictions across projects and locations, and strongly prefer properties close to their own home. A spatial equilibrium framework combined with a frictional portfolio choice rationalizes these features. It makes the distribution of landlords’ own residence location critical to housing supply in the long-run. Because spatial frictions reduce the private return to more remote rental investments, they lead mobile agents, who tend to be renters, to agglomerate near the residence of affluent and immobile owners. The model also predicts that home bias in the supply of capital regulates the heterogeneity in housing market responses to place-based subsidies. Exploiting quasi-experimental evidence from a location-specific investment tax credit targeted at individual landlords in France, the Pinel law, I evidence its substantial causal impact on dwelling sales and new construction. Out-of-town individual investor involvement rises in treated cities, and the policy has stronger effects in locations more open to outside capital. The incomplete capitalization of the incentive in new unit prices confirms that under imperfect capital mobility, landlords bear part of the incidence of local subsidies in the form of higher net returns. In the medium-run, subsidies expand the local housing stock with limited crowding-out, inducing substantial in-migration of renters and population growth. Therefore, landlord home bias is a key factor in the allocation of capital and people across space, and determines the efficiency and distributional effects of place-based policies.

This paper demonstrates and quantifies the effect of monetary campaign promises on electoral behavior. Using granular geographic variation in the expected benefit across 35,000 municipalities, and an instrumental variable strategy exploiting formulaic assessments established in the 1970s, I show that a promise to repeal a local housing tax causally contributed to Emmanuel Macron's success in the 2017 French presidential election, resulting in an aggregate boost of 1.4 million votes. High-frequency data show that information search, voting intentions, and market-based predicted chances of victory all experience a quantitatively consistent increase at the time of the promise.

I provide quasi-experimental evidence that housing quality regulations can raise costs and reduce quantities, but also foster endogenous market segmentation. Since 1977, French law imposes minimum standards on new home construction, only above a size threshold. Using exhaustive data on building permits, I evidence that the cost of new construction jumps at the discontinuity. In turn, quality standards distort the distribution of quantity choices: the size of new homes bunches below the time-varying regulatory notch. Nonetheless, a substantial number of households are either unaffected, or comply with the policy by distorting quality decisions, in line with the predictions of a structurally estimated model with two-dimensional preferences for quality and quantity. Homeowner demographics, dwelling features, and spatial location decisions are all discontinuous at the threshold, implying that size-dependent regulations induce heterogeneous households to sort across fragmented sub-markets.

This paper demonstrates that unpaid rent risk makes landlords reluctant to supply housing services to fragile tenants; and that insuring owners against it improves the access of renters to high-opportunity neighborhoods. We study the implementation of Visale, a publicly funded rent guarantee insurance policy in France, free of charge to eligible tenants and landlords. We exploit tax registry information on all French households, exhaustive data on Visale beneficiaries and claim payouts, and quasi-experimental eligibility variation across renters. We demonstrate that the non-payment guarantee increased access to private-sector rental housing for eligible tenants. The effects are stronger for immigrants and those with low or volatile incomes, who often do not satisfy standard screening criteria for landlords. The scheme eased the spatial mobility of low-income renters towards higher-wage, higher-rent locations. It led to new household formation, some reallocation of the vacant housing stock, and substitution out of public housing, but may have displaced ineligible households in tighter housing markets.

I assess the impact of capital gains taxes on real estate, exploiting novel quasi-experimental variation and detailed data on millions of housing sales in France. In the short-run, I show that investors delay sales and bunch transactions strategically to benefit from a more generous tax treatment for longer holding periods. The effects are stronger for smaller units held by sophisticated investors in high price-appreciation locations. After a 2012 reform substantially increased capital gain taxes on units held for intermediate holding durations, transaction volumes for more exposed dwellings rose sharply between the announcement and implementation of the tax hike ("anticipation"), and dropped durably afterwards ("lock-in"). I evidence the role of several avoidance mechanisms in driving the quantitative response. The shift in the composition of sales towards lower-price appreciation units and non-taxable dwellings had ripple effects on employment and activity in the most exposed localities. Overall, my estimates suggest that capital gains taxes in housing markets can have large fiscal externalities, despite the revealed presence of significant optimization frictions.

The response of wage growth to unemployment is a key parameter of interest for the conduct of monetary policy and the assessment of slack in the labor market. Using data on local labor market conditions, I investigate the existence, identify the magnitude, and explore the stability of the wage Phillips curve since the beginning of the European Monetary Union (EMU). After constructing measures of wage inflation in close to two hundred European NUTS-2 regions, I show that nominal wage growth responds strongly to local unemployment across labor markets in the monetary union. When unemployment is one percentage point higher in a region, nominal wage growth is reduced by 0.3 percentage points on average. The results are robust to the use of alternative proxies for slack in the labour market, including short-term unemployment, the employment-population ratio, and the output gap. I also report specifications taking into account various measures of inflation expectations, estimating a regional New Keynesian Phillips curve. Overall, my results suggest that the negative relationship between unemployment and nominal wage growth still holds in European local labor markets.

This paper tests whether a country's economic geography -- in particular, the distribution of population across space -- is an important source of comparative advantage. We develop a model in which sector-specific exports are determined by the distribution of productivity and factors of production within countries, and show how city-level data can be aggregated to measure determinants of country-level specialization. Countries with higher population-weighted population density specialize in sectors that benefit from agglomeration. Empirically, we estimate and find substantial variation within the US in the extent to which manufacturing sectors sort into dense locations. Combining this measure of "density affinity'' with satellite data on the distribution of population, we find that both US states and countries with higher population-weighted population density disproportionately export in sectors with high density affinity. Even conditional on aggregate endowments, the geographic distribution of population within states and countries affects Ricardian comparative advantage.

We document that migration flows upon retirement are predominantly from richer, more urban to poorer, more rural regions. In the context of rapidly aging societies, both local and national governments are increasingly implementing policies to attract mobile pensioners to their jurisdictions. In theory, the long-term local economic effects of senior immigration are ambiguous, while empirically there is little existing evidence on whether attracting mobile se- niors can be an effective tool to promote local economic development among lagging regions. We combine a unique collection of French microdata with a new empirical strategy to fill this gap and quantify the implications of senior migration patterns for spatial inequality over the past decades. We find that retiree in-migration has significant positive effects on the local economy over the course of a decade, including increases in the working-age population, total employment, GDP, average incomes, fiscal revenues and house prices, that are accompanied by a fall in the share of manufacturing activity. Combining these estimates with observed region-to-region net migration flows, we find that mobile seniors have been a significant force for reducing spatial disparities in economic activity over the past decades.

Is the unequal distribution of political influence across space — combined with the uneven geographic location of industries — an important determinant of economic performance and patterns of trade? Before the 1960s, US regions had vastly different amounts of representation per capita because districts were not re-drawn to reflect population movements. This changed in 1962-1964 following the Supreme Court’s one-person-one-vote decisions. We propose a model in which an industry’s political influence is determined by (i) its geographic distribution and (ii) the over/under-representation of the regions in which it operates. When re-districting is slow, non-comparative advantage sectors are over-represented in politics. Equalization of political influence leads to a reduction in trade barriers, and greater exports and output in industries that gain representation. We investigate the impact of representation using court-ordered re-districting as a natural experiment and combining newly constructed data on the distribution of industry employment in 1962, measures of industry-level trade and output, and historical maps of congressional districts and population distribution. Court-ordered re-districting led to an increase in exports and output in industries that gained representation. These industries also experienced abnormally high stock returns around the Supreme Court decision date. Our results suggest that the structure of electoral systems and distribution of political power across space shapes patterns of output and trade.

Unemployment risk varies substantially across local labor markets, but unemployment insurance (UI) policies are designed at the State or federal level. In a standard model, the redistribution-efficiency trade-offs faced when designing optimal UI can be substantially affected by spatial heterogeneity. On the one hand, consumption smoothing gains depend on compensating price differentials across places; on the other hand, local search externalities and directed mobility directly feed into the efficiency costs of UI benefits. In US data, I document strong co-movement of job opportunities, local price levels, and directed migration. Using variation in UI duration and maximum benefits allowance across states and time, I show that directed migration is dampened by more generous UI. Finally, I exploit discontinuities at state borders and the timing of sharp reductions in UI benefits to quantify the causal impact of UI on unemployment duration, local prices, and directed migration.

Landlords and tenants, on average, have opposite characteristics; but they display positive assortative matching within rental markets. In a nationwide data set containing administrative information on linked renter-occupiers and owners of investment properties for 4.5 million private rental dwellings in France, we document assortative matching by income level and composition, wealth, age, marital status and family structure, both across and within fine geographic segments. Consistent with a novel theory of rental housing assignment, the income correlation is only partially explained by observable characteristics such as location, size, or investment timing. It is mediated through high-wealth landlords listing higher-rent units occupied by high-income tenants, partly thanks to the filtering of the housing stock from past owner-occupiers to current renters. This pattern has quantitatively substantial implications for the measurement of returns to wealth, and the incidence of housing market policies, such as means-tested subsidies.