Research

WORKING PAPERS

The Global Race for Talent: Brain Drain, Knowledge Transfer, and Economic Growth [SSRN Link]

Conditionally Accepted at The Quarterly Journal of Economics

Awarded Best Job Market Paper by the European Economic Association and Unicredit Foundation

Abstract: How does inventors' migration affect talent allocation, knowledge diffusion, and productivity growth across countries? To answer this question, I use a novel micro-level dataset of migrant inventors from patent data, and I trace the network of migrants' co-inventors in the countries of origin and destination. I focus on the US-EU corridor, where a fourth of US patents are produced by immigrants, of whom 27% come from the EU. I document four new empirical results. (i) Gross migration flows are asymmetric, creating brain drain (net emigration) from the EU to the US. (ii) Migrants increase their patenting by 42% after migration. (iii) Migrants continue working with inventors at origin after moving, although less frequently. (iv) Migrants' productivity gains spill over to their collaborators at origin, who increase patenting by 15% when a co-inventor emigrates. To assess the implications of these results for the economy's innovative capacity and policy, I develop a novel two-country model of innovation-based endogenous growth. Inventors are born with heterogeneous talent, which increases endogenously over time by interacting and learning from others. In addition, they produce innovations and solve forward-looking, dynamic problems of migrating abroad or returning to the home country. Inventors who move abroad interact with individuals at both origin and destination, creating a network that diffuses knowledge within and across countries. I calibrate the model to match the empirical results, and I study the impact of innovation and migration policy. A 10 percentage-point decrease in the tax rate for foreigners and return migrants in the EU eliminates the brain drain in the short run but reduces knowledge spillovers in the long run. On net, after 25 years, EU innovation increases by 9%, but US innovation declines by 6%, which reduces technology diffusion to the EU. The former effect dominates in the short run, increasing EU productivity growth by 5% in the first 25 years. However, the latter effect dominates in the long run, reducing EU productivity growth by 6% in the new long-run equilibrium.  On the migration policy side, doubling the size of the US H1B visa program increases productivity by 9% in the US and the EU in the long-run, because it sorts inventors to where they are most productive and can learn most, increasing knowledge spillovers to other countries. 

Tapping into Talent: Coupling Education and Innovation Policies for Economic Growth, with Ufuk Akcigit and Jeremy Pearce. 

NBER Working Paper # 27862.  Forthcoming at The Review of Economic Studies

Abstract:  How do innovation and education policy affect individual career choice and aggregate productivity? This paper analyzes the various layers that connect R&D subsidies and higher education policy to productivity growth. We put the development of scarce talent and career choice at the center of a new endogenous growth framework with individual-level heterogeneity in talent, frictions, and preferences. We link the model to micro-level data from Denmark and uncover a host of facts about the links between talent, higher education, and innovation. We use these facts to calibrate the model and study counterfactual policy exercises. We find that R&D subsidies, while less effective than standard models, can be strengthened when combined with higher education policy that alleviates financial frictions for talented youth. Education and innovation policies not only alleviate different frictions, but also impact innovation at different time horizons. Education policy is also more effective in societies with high income inequality. 

Read more at: VoxEu.

The End of the American Dream? Inequality and Segregation in US Cities, with Alessandra Fogli, Veronica Guerrieri, and Mark Ponder

R&R at the Journal of Political Economy

Abstract:  Since the 1980s the US has experienced not only a steady increase in income inequality, but also a contemporaneous rise in residential segregation by income. What is the relationship between inequality and residential segregation? How does it affect intergenerational mobility? We first document a positive correlation between inequality and segregation, both over time and across metro areas. We then develop a general equilibrium model where parents choose the neighborhood where they raise their children and invest in their children’s education. In the model, segregation and inequality amplify each other because of a local spillover that affects the return to education. We calibrate the model to a representative US metro in 1980 and use the micro estimates of neighborhood exposure effects in Chetty and Hendren (2018) to discipline the strength of the local spillover. We first use the calibrated version of the model to explore the economy’s response to an unexpected skill premium shock. We find that segregation dynamics played a significant role in amplifying the increase in inequality and in dampening intergenerational mobility. We then use the model to explore the effects of policies designed to move poor people to better neighborhoods, like the Moving To Opportunity program. We show that scaling up MTO policies induces general equilibrium effects that limit their efficacy.

Scaling Up the American Dream: A Dynamic Analysis,  with Alessandra Fogli, Veronica Guerrieri and Mark Ponder

The Moving to Opportunity (MTO) program implemented in the ’90s offered vouchers to low-income people living in high-poverty neighborhoods to move to richer neighborhoods. We use a dynamic general equilibrium model with residential choice and endogenous local spillovers to explore the quantitative effects of scaling up this policy and compare them to those of alternative neighborhood-specific policies. We first contrast the MTO policy to a place-based policy that offers a transfer to all families living in a high-poverty neighborhood. The MTO program generates larger income gains for the children of the recipient families. However, as we scale up the policy, general equilibrium effects both dampen these gains and impose large welfare losses for the non-recipient families. The transfer place-based policy has overall higher average welfare gains, but it is less successful in reducing income inequality and residential segregation. We then introduce an alternative place-based policy that invests resources in local institutions, such as public schools, to directly improve the neighborhood spillover. This type of policy is less effective on impact, but realizes larger welfare gains over time, while also reducing both inequality and residential segregation. 

The Origins of Regional Specialization, with Josh Morris-Levenson.

Abstract: Why do US states specialize in different sectors? Using Census microdata from 1860-2020, we document that employment specialization is highly persistent, which suggests that specialization may deviate from natural advantage, and reallocation could increase aggregate output. Furthermore, growing sectors in a state disproportionately hire interstate migrants. Workers' mobility allows specialization to change; next, we show that mobility frictions contribute to persistence in specialization. We develop a quantitative spatial model in which workers move across state-by-sector labor markets in response to exogenous changes in local fundamentals, mobility frictions, and labor market-specific idiosyncratic skills. Persistence arises from mobility frictions and agglomeration, a reduced-form proxy for investment in physical capital. Removing mobility frictions reduces persistence by one third in the short run, even when investment follows its baseline path. Output increases substantially in the frictionless economy compared to the baseline due to improved matching of workers to labor markets rather than changes in the aggregate sectoral composition of states. 

PUBLICATIONS

Slow Household Deleveraging, with Veronica Guerrieri and Guido Lorenzoni. 

Journal of the European Economic Association, Vol. 18, Issue 6, Dec. 2020, pages 2755-2775. 

Abstract: We use a model of precautionary savings with housing and mortgages to study the effects of a deleveraging shock on consumer spending. We focus on deleveraging caused by a contraction in home values, and compute numerically the partial equilibrium effect of the shock. Our simulations show that household deleveraging is associated to a long and protracted weakness in consumption. These effects appear even if we assume, realistically, that housing wealth is illiquid and mortgage debt is long-term. We show that housing wealth matters for consumption decisions due to an insurance force: consumers know they can sell their house if they get hit by sufficiently negative shocks in the future. We also show that our slow deleveraging mechanism is amplified when incomes are affected by weak aggregate consumption demand through general equilibrium effects. 

WORK IN PROGRESS

Career Choice of Entrepreneurs, Inventors, and Economic Growth, with Ufuk Akcigit, Harun Alp, and Jeremy Pearce

Abstract: New technologies emerge and translate into economic growth through the team effort of inventors, entrepreneurs, and production workers. This paper provides a unified life-cycle framework that links sorting into these occupations, innovation in firms, and aggregate growth. We connect detailed microdata from Denmark on entrepreneurs, inventors, and firms to a novel quantitative endogenous growth model with occupational sorting, entrepreneurial startups, and innovation. Empirically, we find that while parental exposure is a key determinant of entrepreneurship, becoming an R&D worker is primarily determined by education and IQ. Entrepreneurs with higher IQs and higher educational attainment are more likely to be innovative: they hire more R&D workers, patent more, and grow faster. The quantified model highlights the importance of schooling for both the supply of R&D, through R&D workers, and the demand for R&D, from innovative entrepreneurs. The effect of family background on occupational sorting creates talent misallocation in the economy: individuals from lower-income families are less likely to access education, and R&D careers, and become innovative entrepreneurs. While R&D and entry subsidies can increase the appeal of R&D and entrepreneurship, education subsidies can improve access to education, sorting talented individuals from lower-income families into R&D and innovative entrepreneurship.

Neighborhood Segregation and Endogenous Racial Bias, with Alessandra Fogli, Martin Garcia Vazquez, and Veronica Guerrieri

The Geography of Innovative Firms, with Craig A. Chikis and Benny Kleinman