Periodic Reporting for period 2 - COMMUTE (Quantitative Models of Cities)
Berichtszeitraum: 2023-05-01 bis 2024-10-31
One of the most striking empirical regularities is the huge divergence in economic activity both across and within countries. Within countries this is most clearly illustrated by the vast differences in economic activity between cities and rural areas. In the US, for example, 17% of GDP is produced in just three metropolitan statistical areas, which occupy only 0.6% of the land area of the US (Allen and Donaldson 2020). Also within cities, there are enormous differences in economic activity and land use between peripheral locations and dense central locations. Despite these first-order facts, our quantitative understanding of the agglomeration and dispersion forces that shape the spatial distribution of economic activity is still imperfect.
Understanding the forces that shape the spatial distribution of economic activity is not only important in itself, but is also of key policy interest. Our limited ability to model these forces makes it difficult to predict the effects of policy interventions in cities, such as the construction of new transport infrastructure, changes in land use regulations or height restrictions on buildings. As a result, decisions on expensive transport infrastructure projects use simple rules of thumb which we know are far from perfect. The Covid pandemic and the move to more working from home have also increased the interest in understanding the forces that make some locations more attractive than others.
This project improves our ability to model the forces that shape cities by combining very detailed data for the entire population of Denmark over a more than 30-year period with advances in our ability to model cities. The project produces several new stylised facts. We document, for example, how people change locations over their life cycle and how choices are shaped by life events such as marriage and children. The project combines rigorous reduced form evidence with a quantitative spatial model, building on a literature that the PI has contributed to over the last 10 years. We develop the modelling in this literature to deal more realistically with worker heterogeneity and the dynamic impact of interventions in cities.
Understanding the forces that shape the spatial distribution of economic activity is not only important in itself, but is also of key policy interest. Our limited ability to model these forces makes it difficult to predict the effects of policy interventions in cities, such as the construction of new transport infrastructure, changes in land use regulations or height restrictions on buildings. As a result, decisions on expensive transport infrastructure projects use simple rules of thumb which we know are far from perfect. The Covid pandemic and the move to more working from home have also increased the interest in understanding the forces that make some locations more attractive than others.
This project improves our ability to model the forces that shape cities by combining very detailed data for the entire population of Denmark over a more than 30-year period with advances in our ability to model cities. The project produces several new stylised facts. We document, for example, how people change locations over their life cycle and how choices are shaped by life events such as marriage and children. The project combines rigorous reduced form evidence with a quantitative spatial model, building on a literature that the PI has contributed to over the last 10 years. We develop the modelling in this literature to deal more realistically with worker heterogeneity and the dynamic impact of interventions in cities.
During the first half of the project, we have concentrated on developing some of the sub-projects that form the overall work program. In particular, the paper “Neighborhood Effects: Evidence from Wartime Destruction in London” which is joint with Steve Redding, appeared as a NBER Working Paper in April 2024 and was presented at the NBER Summer Institute Session on Urban Economics in July 2024. The paper shows how plausibly random wartime destruction in London during WWII has long-lasting effects on the built environment and this has in turn changed the spatial distribution of population in London. We use these random changes to estimate how much different people care about the socio-economic composition of their neighbourhoods with the help of a quantitative spatial model.
Another key output of the first half of the project is the paper “The Geography of Life: Evidence from Copenhagen,” which is joint work with Gabriel Ahlfeldt, Ismir Mullalic and Caterina Soto Vieira. The paper presents a set of new stylised facts on how age and life events, such as marriage and childbirth, affect location choices in cities. This evidence is combined with a quantitative spatial model to explore how demographic changes, such as changes in the fertility rate or population ageing, will affect cities.
We are also making progress on other parts of the project, including work on the dynamic impact of policy interventions and the optimal use of space in cities. These working papers will appear in the near future.
Another key output of the first half of the project is the paper “The Geography of Life: Evidence from Copenhagen,” which is joint work with Gabriel Ahlfeldt, Ismir Mullalic and Caterina Soto Vieira. The paper presents a set of new stylised facts on how age and life events, such as marriage and childbirth, affect location choices in cities. This evidence is combined with a quantitative spatial model to explore how demographic changes, such as changes in the fertility rate or population ageing, will affect cities.
We are also making progress on other parts of the project, including work on the dynamic impact of policy interventions and the optimal use of space in cities. These working papers will appear in the near future.
In terms of advancing research significantly beyond the state of the art and expected results, the project innovates along several dimensions. First, the project combines quantitative spatial models that build on the work in Ahlfeldt, Redding, Sturm and Wolf (2015) with extraordinarily detailed register data for every person in Denmark. This data makes it possible to estimate the key structural parameters of quantitative spatial models using an unprecedented amount of information, and to document new stylised facts about how workers and firms make location decisions.
Second, the project extends the modelling of cities in two ways. On the one hand, the very detailed data allow us to develop models that have a much richer heterogeneity of workers than the current literature. On the other hand, we are extending the static quantitative spatial models that currently dominate the literature to a dynamic setting that can be estimated using the detailed data for Denmark. Third, while much of the literature has focussed on estimating the key parameters of quantitative spatial models, several of the papers in the second half of the project focus on using quantitative models to evaluate the effects of policy interventions in cities.
Second, the project extends the modelling of cities in two ways. On the one hand, the very detailed data allow us to develop models that have a much richer heterogeneity of workers than the current literature. On the other hand, we are extending the static quantitative spatial models that currently dominate the literature to a dynamic setting that can be estimated using the detailed data for Denmark. Third, while much of the literature has focussed on estimating the key parameters of quantitative spatial models, several of the papers in the second half of the project focus on using quantitative models to evaluate the effects of policy interventions in cities.