Economic activity is highly unevenly distributed across space, as reflected in the existence of cities and the concentration of economic functions in specific locations within cities, such as Manhattan in New York and the Square Mile in London. Understanding the strength of the agglomeration and dispersion forces that underlie these concentrations of economic activity is central to a range of economic and policy questions. These forces shape the size and internal structure of cities, with implications for the incomes of immobile factors, congestion costs, and city productivity. They also determine the impact of public policy interventions, such as transport infrastructure investments and urban development and taxation policies.
Although there is a long literature on economic geography and urban economics dating back to at least Marshall (1920), a central challenge remains distinguishing agglomeration and dispersion forces from variation in locational fundamentals. While high land prices and levels of economic activity in a group of neighbouring locations are consistent with strong agglomeration forces, they are also consistent with shared amenities that make these locations attractive places to live (e.g. leafy streets and scenic views) or common natural advantages that make these locations attractive for production (e.g. access to natural water).
A model of internal city structure and Berlin’s division and unification
In our paper (Ahlfeldt et al. 2014), we develop a quantitative theoretical model of internal city structure that incorporates both agglomeration and dispersion forces and also heterogeneity in locational fundamentals, while remaining tractable and amenable to empirical analysis. Locations differ in terms of their productivity, their amenities, their density of development (the ratio of floor area to ground area), and their access to transport infrastructure. Productivity depends on production externalities, which are increasing in the surrounding density of employment, and production fundamentals, such as topography and proximity to natural supplies of water. Amenities depend on residential externalities, which are increasing in the surrounding density of residents, and residential fundamentals, such as access to forests and lakes. Congestion forces take the form of an inelastic supply of land and commuting costs that are increasing in travel time, where travel time in turn depends on the transport network.
We combine this quantitative theoretical model with the natural experiment of Berlin's division in the aftermath of World War II and its reunification following the fall of the Iron Curtain. We make use of a remarkable and newly-collected dataset for Berlin, which includes land prices, employment by workplace, and employment by residence for thousands of blocks for the pre-war, division, and reunification periods.
Before structurally estimating the model, we first present reduced-form evidence showing that division leads to a reorientation of the gradient in land prices and employment in West Berlin away from the main pre-war concentration of economic activity in East Berlin, while reunification leads to a reemergence of this gradient. In contrast, there is little effect of division or reunification on land prices or employment along other more economically remote sections of the Berlin Wall.
As shown in Panel A of Figure 1 below, Berlin’s land price gradient was approximately monocentric prior to WWII, with the central business district (CBD) located in the district ‘Mitte’ -- just East of the future line of the Berlin Wall. Around this central point, there were concentric rings of progressively lower land prices, including two areas in the future West Berlin -- one around the Kurfuerstendamm and the other just West of the future line of the Berlin Wall, close to the Anhalter Bahnhof (see Panel B). Following the division, the second of these land price peaks was eliminated, as this area around the former Anhalter Bahnhof ceased to be an important centre of commercial and retail activity (see Panel C). Following the reunification, there was a re-emergence of high land price values in the pre-war central business district in Mitte (see Panel D), and in the areas of the former West Berlin close to the pre-war CBD (see Panel E).
Figure 1. Berlin land prices
We next structurally estimate the model and examine whether it can account quantitatively for the observed effects of division and reunification. The model yields a gravity equation for bilateral commuting flows, which we show is successful in accounting for observed commuting patterns, prior to WWII, during division, and after reunification. We find substantial and highly localised production and residential externalities.
- Our estimated elasticity of productivity with respect to the density of workplace employment for division is 0.07, which is towards the high end of the range of existing estimates using variation between cities.
- Our estimated elasticity of amenities with respect to the density of residence employment for division is 0.14, which is consistent with the view that consumption externalities are an important agglomeration force in addition to production externalities.
- Finally, we find that both types of externalities are highly localised within the city and after around 10 minutes of travel time, both production and residential externalities fall to close to zero.
We also use the model to undertake counterfactuals for division and reunification. We show that a special case of the model with exogenous productivity and amenities is unable to account quantitatively for observed treatment effects of division and reunification on the pattern of economic activity within West Berlin. In contrast, the model with the estimated agglomeration and dispersion forces generates predicted treatment effects close to the observed treatment effects.
Using our quantitative theoretical model and the natural experiment of Berlin’s division and reunification, we find important roles for both heterogeneity in locational fundamentals and agglomeration forces in shaping the internal structure of cities. We find strong but highly localised production and residential externalities that are important in enabling the model to account quantitatively for the observed effects of division and reunification. Our research highlights the importance of taking into account these agglomeration forces in evaluating public policy interventions.
Ahlfedlt, G, S Redding, D Sturm and N Wolf (2014) “The Economics of Density: Evidence from the Berlin Wall,” NBER Working Paper, 20354.
Marshall, A (1920) Principles of Economics, Macmillan, London.