Trade and Geography

A conference on "Trade and Geography" took place in Cambridge on March 16-17. International Trade and Investment Program Director Stephen J. Redding and Research Associate Esteban Rossi-Hansberg, both of Princeton University organized the meeting. These researchers' papers were presented and discussed:

Enrico Berkes, Northwestern University, and Ruben Gaetani, University of Toronto

Income Segregation and Rise of the Knowledge Economy

Berkes and Gaetani analyze the effect of the rise of knowledge-based activities on spatial inequality within U.S. cities, exploiting the network of patent citations to instrument for local trends in innovation. They find that innovation intensity is responsible for 20% of the overall increase in urban segregation between 1990 and 2010. This effect is mainly driven by the clustering of employment and residence of workers in knowledge-based occupations. The researchers develop and estimate a spatial equilibrium model to quantify the contribution of productivity and residential externalities in explaining the observed patterns. Endogenous amenities account for two thirds of the overall effect. Berkes and Gaetani illustrate the relevance of the model for policy analysis by studying the impact of four proposed projects for Amazon's HQ2 on the structure of Chicago.

Fabian Eckert, Yale University, and Michael Peters, Yale University and NBER

Spatial Structural Change

Eckert and Peters study the spatial implications of structural change. The secular decline in spending on agricultural goods hurts workers in rural locations and increases the return to moving towards non-agricultural labor markets. They combine detailed spatial data for the U.S. between 1880 and 2000 with a novel quantitative theory to understand this process and to quantify its macroeconomic implications. They researchers find that spatial reallocation across labor markets accounts for almost none of the aggregate decline in agricultural employment. Despite ample migration, population net flows were only weakly correlated with agricultural specialization. Labor mobility nevertheless had important aggregate effects. Without migration income per capita would have been 15% lower and spatial welfare inequality would have been substantially higher, especially among low-skilled, agricultural workers.

Matthew J. Delventhal, Universitat Autonoma de Barcelona and Barcelona GSE

The Globe as a Network: Geography and the Origins of the World Income Distribution

How important are falling transport costs for patterns of population and income growth since 1000 CE? To answer this question, Delventhal builds a quantitative dynamic spatial model with an agricultural and non-agricultural sector, and endogenous fertility, migration, innovation and technology diffusion. In this model there exists an endogenous threshold for global transport costs, which is characterized by a simple network statistic. If transport costs are above this threshold, the world converges to a Malthusian steady state. If transport costs fall below this threshold, the world economy enters a process of sustained growth in population and income per capita. Taking this model to the data, Delventhal divides the globe into 2,249 3° by 3° quadrangles. He assigns each location an agricultural potential determined by exogenous climate and soil characteristics. He infers bilateral transport costs by calculating the cheapest route between each pair of locations given the natural placement of rivers, oceans and mountains. Delventhal calibrates the model so that in the year 1000 the world is in a Malthusian steady state. He then drops the cost of water and land transport exogenously in a way that is consistent with historical evidence and track the endogenous evolution of population and income until the year 2000. Qualitatively, this exercise generates slow but accelerating growth in both population and income per capita for the first 800 years, an abrupt takeoff in growth after 1800 CE with Europe in the lead, and a large increase in the dispersion of income per capita after 1800 CE. Quantitatively, the model accounts for 55% of the variation in population density across 10 major regions in 1000 CE, 44% of the variation in income per capita across regions in 1800 CE, and is able to generate 43% of the overall dispersion in income per capita in 2000 CE.

Bernt Bratsberg, Frisch Centre; Andreas Moxnes, University of Oslo; Oddbjorn Raaum, Frisch Centre; and Karen Helene Ulltveit-Moe, University of Oslo

Opening the Floodgates: Immigration and Structural Change

Bratsberg, Moxnes, Raaum, and Ulltveit-Moe investigate the impact of a large shock to labor supply on industry growth and structural change. The EU enlargement of 2004 and 2007 lead to an unprecedented migration wave to Norway. The country received the largest number of migrants relative to country size, compared to all other developed countries, over the ensuing decade. The researchers develop a simple factor-proportions theory and sufficient statistic approach that can be used to identify the aggregate impact of a labor supply shock across occupations on industry growth. Using detailed data on industry performance, immigration by occupation and occupational characteristics, the researchers introduce a new instrument that exploits the fact that language barriers in the Norwegian labor market are significant for foreign workers and that they vary across occupations and source countries. Their results point to migration leading to large adjustments in industry size, and in particular to sectors of the economy that are intensive in the use of immigrant occupations.

Peter Egger and Nicole Loumeau, ETH Zurich

The Economic Geography of Innovation

Egger and Loumeau outline a multi-region quantitative model to assess the importance of country-level investment incentives towards innovation at the level of 5,633 micro-regions of different size. While incentives vary across countries (and time) the responses are largely heterogeneous across regions within as well as across countries. The reason for this heterogeneity roots in average technology differences -- in terms of the production of both, output and innovation -- as well as in the geography (location) and amenities across regions. The cross-sectional unit of observation underlying the quantitative analysis are REGPAT regions, whose patenting output they measure and link to population as well as income statistics. The model and quantitative analysis take the tradability of output as well as the mobility of people across regions into account. In the counterfactual equilibrium analysis the researchers focus on the effects on three key variables - place-specific employment, productivity, and welfare - in a scenario where investment incentives towards innovation are abandoned. Egger and Loumeau find that the use of policy instruments which are designed to stimulate private R&D are globally beneficial in terms of productivity and welfare. Particularly, regions with high amenities and a low degree of transport remoteness tend to benefit from such policy instruments.

Jeffrey C. Brinkman and Jeffrey Lin, Federal Reserve Bank of Philadelphia

Freeway Revolts!

Freeway revolts -- mass protests by central-city residents concerned about local quality of life -- erupted across the U.S. following early urban Interstate construction in the mid-1950s. Brinkman and Lin present theory and evidence showing that the concerns of the freeway revolts (namely, disamenities from freeways) were important for the eventual allocation of freeways within cities and changes in city structure. Using panel data on U.S. cities and neighborhoods between 1950 and 2010, the researchers show that disamenity effects dominate access benefits in central cities: downtown neighborhoods closer to new highways declined more (or grew more slowly) in population, income, jobs, and land values compared with neighborhoods farther away. But in the suburbs, where access benefits are more important, proximity to a highway has the opposite effect. Further, actual freeway construction increasingly diverged from initial plans in the wake of the growing freeway revolts and subsequent policy responses, especially in central neighborhoods. Brinkman and Lin use a quantitative city structure model to quantify the effects of freeway disamenities and evaluate mitigation policies. The aggregate welfare benefits from burying or capping highways are large and concentrated downtown, explaining political opposition to urban freeways.

Nelson Lind, Emory University, and Natalia Ramondo, University of California at San Diego and NBER

Trade with Correlation (NBER Working Paper No. 24380)

Lind and Ramondo develop a trade model in which productivity -- the result of a country's ability to adopt global technologies -- presents an arbitrary pattern of spatial correlation. The model generates the full class of import demand systems consistent with Ricardian theory, and, hence, captures its full macroeconomic implications. In particular, their framework formalizes Ricardo's insight -- absent from the canonical Ricardian model -- that countries gain more from trade partners with relatively dissimilar technology. Incorporating this insight into the calculations of macro counterfactuals entails a simple correction to self-trade shares. Their framework enables general aggregation results which tie micro optimization to macro demand systems and guide counterfactual analysis based on micro estimates. Their quantitative application to a multi-sector trade model suggests that countries specialized in low correlation sectors have 40 percent higher gains from trade relative to countries specialized in high correlation sectors. After accounting for correlation, the model predicts that lower trade costs for imports from China, rather than Canada, have the largest impact on real wages in the United States.

Richard K. Mansfield, University of Colorado at Boulder and NBER

How Local Are U.S. Labor Markets? Using an Assignment Model to Forecast the Geographic Incidence of Local Labor Demand Shocks

Mansfield examines the welfare incidence of across locations and demographic groups for labor demand shocks featuring particular geographic and firm type compositions. LEHD data on the near universe of U.S. job transitions are used to estimate a rich two-sided assignment model of the labor market featuring thousands of parameters that is then used to generate simulated forecasts of many alternative local shocks. These forecasts suggest that existing local workers account for only 0.1% (2.7%) of total welfare (employment) gains, with 80% (56%) of welfare (employment) gains accruing to out-of-state workers. This is despite the fact that projected employment rate increases from a typical positive shock are 7 times larger for existing workers in the targeted Census tract than for workers from an adjacent tract, because workers in the target tract are a minuscule share of the national labor market. Further, the projected earnings incidence across local skill groups is highly sensitive to the shock's firm type composition.

Shushanik Hakobyan, International Monetary Fund, and John McLaren, University of Virginia and NBER

Local-Labor-Market Effects of NAFTA: The Other Shoe Drops

In previous work, Hakobyan and McLaren looked at the effect of US tariff reductions under NAFTA on US labor market outcomes, and found that blue-collar workers in industries and locations that lost tariff protection experienced slower wage growth compared to other workers. Here, they examine the corresponding reductions in Mexican tariffs on imports from the US. Surprisingly, the researchers find that blue-collar workers in industries or locations whose Mexico tariffs fell also experienced slower wage growth compared to other workers. Hakobyan and McLaren tentatively suggest that the most plausible explanation for this finding is that the tariff reductions made it easier for US manufacturers to use offshoring to Mexico to lower costs.

NBER Videos

National Bureau of Economic Research, 1050 Massachusetts Ave., Cambridge, MA 02138; 617-868-3900; email:

Contact Us