Final Report Summary - SPATIAL (Using Natural Experiments to Understand the Spatial Economy)
The research in this grant has used natural experiments to estimate the strength of agglomeration and dispersion forces in cities and regions. Work on this grant has fallen into three different areas. First, in “The Economics of Density: Evidence from the Berlin Wall” we develop a quantitative model of internal city structure that features agglomeration and dispersion forces and an arbitrary number of heterogeneous city blocks. To estimate agglomeration and dispersion forces, we use data on thousands of city blocks in Berlin for 1936, 1986, and 2006 and exogenous variation from the city’s division and reunification. We estimate substantial and highly localised production and residential externalities. Our estimates suggest that doubling the density of employment in a city increases productivity by about 7%. However, these productivity externalities decay rapidly in space and are zero after approximately 500 meters. We also estimate positive effects of increasing residential density on urban amenities, and these externalities also decay rapidly in space. We show how our quantitative framework can be used to undertake counterfactuals for changes in the organisation of economic activity within cities in response, for example, to changes in the transport network.
Second, in “The Impact of Public Employment: Evidence from Bonn" we evaluate the impact of public employment on private sector activity using the relocation of the German federal government from Berlin to Bonn in the wake of the Second World War as a source of exogenous variation. To guide the empirical analysis, the paper develops an economic geography model in which public sector employment in a city can crowd out private employment through higher wages and house prices, but also generates potential productivity and amenity spillovers. We show that after the relocation of the government, Bonn experienced a substantial increase in total employment and population relative to a control group of cities. However, this increase was almost exclusively driven by an increase in public employment, while private employment has only marginally increased. Our estimates suggest that an additional job in the public sector has generated 0.4 additional private sectors jobs. This overall effect consists of an expansion in private sector non-tradable employment and a slight decrease in private sector tradable employment. The paper interprets this finding through the lens of the theoretical model and provides evidence for the mechanisms emphasised by the model.
Third, in “Estimating Neighbourhood Effects: Evidence from War-time Destruction,” we use Second World War destruction in London as a natural experiment to provide evidence on neighbourhood effects. The paper uses a newly collected and remarkable dataset on the level of war related destruction of each house in London and combines this with census and house price data and other micro data sources both before and after the Second World War. These data are combined with a quantitative model of the sorting of heterogeneous groups of agents across locations that differ in productivity, amenities and transport infrastructure. Our key finding is that both own and neighbours’ destruction affects the pattern of spatial sorting and that the effects of neighbours’ destruction are highly localised to a few hundred meters. The results imply that residential externalities in cities are strong, but also highly localised. These findings provide evidence for spatial sorting as a mechanism through which neighbourhood effects can occur.
Second, in “The Impact of Public Employment: Evidence from Bonn" we evaluate the impact of public employment on private sector activity using the relocation of the German federal government from Berlin to Bonn in the wake of the Second World War as a source of exogenous variation. To guide the empirical analysis, the paper develops an economic geography model in which public sector employment in a city can crowd out private employment through higher wages and house prices, but also generates potential productivity and amenity spillovers. We show that after the relocation of the government, Bonn experienced a substantial increase in total employment and population relative to a control group of cities. However, this increase was almost exclusively driven by an increase in public employment, while private employment has only marginally increased. Our estimates suggest that an additional job in the public sector has generated 0.4 additional private sectors jobs. This overall effect consists of an expansion in private sector non-tradable employment and a slight decrease in private sector tradable employment. The paper interprets this finding through the lens of the theoretical model and provides evidence for the mechanisms emphasised by the model.
Third, in “Estimating Neighbourhood Effects: Evidence from War-time Destruction,” we use Second World War destruction in London as a natural experiment to provide evidence on neighbourhood effects. The paper uses a newly collected and remarkable dataset on the level of war related destruction of each house in London and combines this with census and house price data and other micro data sources both before and after the Second World War. These data are combined with a quantitative model of the sorting of heterogeneous groups of agents across locations that differ in productivity, amenities and transport infrastructure. Our key finding is that both own and neighbours’ destruction affects the pattern of spatial sorting and that the effects of neighbours’ destruction are highly localised to a few hundred meters. The results imply that residential externalities in cities are strong, but also highly localised. These findings provide evidence for spatial sorting as a mechanism through which neighbourhood effects can occur.