PUBLICATIONS Economic Shocks and Labour Market Flexibility (with Julien Labonne). 
Forthcoming, Journal of Human Resources. Online Appendix.


Anonymity or Distance? Job Search and Labour Market Exclusion in a Growing African City 
(with Girum Abebe, Stefano Caria, Marcel Fafchamps, Paolo Falco, and Simon Quinn)
R&R, Review of Economic Studies.

Coverage: voxdev,  voxeu

Can the state improve the lives of slum dwellers by supplying formal housing otherwise not provided by the market? Or will state-built housing, priced at the cost of production, be beyond the willingness to pay of poor households or built in the wrong location? To answer these questions, I study a lottery for large-scale government housing in Ethiopia. Winners of the lottery are sold apartments on the outskirts of the city. They then have the choice to move in or rent out these units. By moving in they pay a high implicit price in forgone rent determined by the market, which I find to exceed the cost to the state of providing the housing. I find that nearly half of lottery winners trade slum housing in the city centre for improved housing on the outskirts of cities. In addition, they make upgrades to their apartments, adding a number of amenities that they did not enjoy in their slum housing and did not come with government housing. I argue that this reveals unmet demand for improved housing and suggests that informal housing is a sub-optimal outcome for a large proportion of slum-dwellers in this setting. Moving to sites far from the city centre does not negatively affect labour supply or earnings. Although social lives are less vibrant in the new housing estates, lottery winners report significant reductions in conflict with neighbours and increased willingness to contribute to public goods.
Coverage: The Guardian. 

Good neighbours make good fences: randomly allocated neighbours and local public goods in urban Ethiopia. 
(Companion to job market paper) Draft available.
How does the composition of residents in urban neighbourhoods affect the provision of local public goods? In developing countries where the state does not always provide adequate local sanitation, security and maintenance goods, local communities’ well-being may depend on their ability to coordinate to provide these goods for themselves. Housing and urban policy influence how the inhabitants of diverse and unequal cities are distributed across neighbourhoods. I study a lottery for state-built housing in Ethiopia that randomly assigns groups of households to apartment blocks in newly built neighbourhoods. I collect data on the universe of randomly assigned households in 250 housing blocks at the time that they move in. I combine this with detailed data on the quality of local public goods at the time of moving in and two years later, from a specifically designed neighbourhood survey and household self-reports. I find clear evidence that the composition of neighbours affects local public good provision, in terms of building maintenance and beautification of public spaces, as well as self-reported community cohesion and willingness to contribute to public goods. Specifically, buildings with higher poverty rates and with lower rates of owner-occupation have significantly worse maintenance and community space improvements. By contrast, I find no evidence that ethnic diversity affects community and public goods outcomes.
(with Marcel Fafchamps, Girum Abebe, Simon Quinn, Stefano Caria, Forhad Shilpi, and Paolo Falco)

(with Stefano Caria and Marc Witte)

The long-term impacts of industrial and entrepreneurial work: Experimental Evidence from Ethiopia (with Chris Blattman and Stefan Dercon)
Draft available soon.

Can one-time interventions that aim to overcome barriers to entrepreneurship and wage employment place young people in developing countries on better long-term career trajectories? We study the effects of a cash grant of $300 plus business training intended to relieve capital constraints, and a one-time factory job offer designed to help low-skilled women with no work experience get their first opportunity in wage employment. We randomly assign young, mostly female, job applicants in Ethiopia to either one of these two interventions, or a control group. After one year, those who were given the grant were more likely to be self-employed, and those who were given the factory job were more likely to be working in factory employment. These effects have dissipated after five years: we find complete convergence in income, consumption, occupational choice, and hours of work between both grant and factory-job recipients and the control group. We find that the control group increases their participation in both self-employment and non-factory work over time. Recipients of the grant are not significantly more likely to be engaged in self-employment, nor more successful in the enterprises that they run. Individuals assigned to the factory job offer quit or lose these jobs relatively quickly, and are no more likely to find employment elsewhere

Private costs and social benefits of workfare: Evidence from urban Ethiopia 
(with Girum Abebe and Carolina Mejia-Mantilla)
Draft available soon.
Developing countries are investing in social safety net schemes. Public works are commonly used as an alternative to unconditional cash transfers.  But work requirements may significantly reduce welfare gains by limiting beneficiaries’ ability to earn income from non-public labour, particularly in urban areas where labour markets exhibit little seasonality. We study a large-scale urban public works program for poor households in Addis Ababa, Ethiopia. We worked with the state to randomize the program at the level of the urban district, and collect data on a panel of over 6,000 households. We find that the programme significantly reduces beneficiary households’ private labour supply, which reduces the total income transfer of the program by nearly 40%. We find evidence consistent with a small income effect and a large time-cost effect. Most of the reductions in private work come from household members who actively participated in the public works. We compare a regular phone survey to study a change in the public work requirement from 2 to 6 hours per day, half-way through the first year of the program. An increase in public work hours at the intensive margin, with no accompanying change in pay, leads to significantly larger reductions in private earnings. We find evidence that improved public goods may offset these costs: non-participants living in treated areas report dramatic improvements in neighbourhood cleanliness, greening and sanitation, one year after the program began. 


Urban density and labour markets: Evaluating slum redevelopment in Addis Ababa. (with Girum Abebe and Gharad Bryan)
All around the world, governments implement policies that influence where people live, particularly in dense urban neighbourhoods. These policies have important implications for the working of labour markets. Urban labour markets provide access to formal jobs without long commutes, access to customers and markets, and access to social networks that support labour markets, particularly in informal economies. Each additional person who lives in the city centre not only gains from these urban labour markets but also contributes to these agglomeration economies, imposing positive externalities on others. But increasing density in informal settlements can also have negative externalities in terms of congestion, health, and poor infrastructure, which could limit the potential of cities to create high productivity formal jobs. Estimating the scope and scale of these externalities is an empirical challenge. The government of Addis Ababa recently announced plans to redevelop large central areas of the city, which will lead to the relocation of 20,000 households out of informal settlements. Building on a sample of 30,000 geo-referenced surveys conducted in Addis before the announcement, we will estimate both the direct effect of being relocated out of a dense city centre, as well as the externalities of the relocations, on labour markets in the surrounding areas. We will draw on a variety of empirical techniques from labour and urban economics to estimate how these spillover effects extend over space.

Cost effective panel data collection for Kampala
(with Julia Bird, Gharad Bryan, and Astrid Haas)
The urban literature on developing country cities is still growing. One of the major constraints, however, is access to good quality data. Existing data sources are often costly or difficult to get, frequently do not measure prices, are not individual specific, do not allow for tracking of individuals, and are available infrequently and sporadically.  As a result of this, predictions about the impacts of projects, such as transportation infrastructure, or policy changes in urban contexts are often based on models that have to infer key prices from location choices and must be calibrated to data that is up to 10 years old.  Further, evaluations of projects are not able to make use of modern identification techniques such as synthetic controls which require long panels. In this project we test methods for collecting low-cost, high-frequency panel survey data that improves on existing data sources in three ways: 1) The data will provide information about prices, namely consumer prices, wages, and rents, at the individual and location level; 2) the data will allow for the creation of commuting flows at the individual or group level; 3) the data will allow for tracking of people when they move as well as places over time.