Work in Progress
Work in Progress
with Ruben Gaetani and Martí Mestieri
Previous version circulated as "Cities and Technological Waves"
Cities rise and fall
Technologies come and go
Select them wisely
We develop a spatial model of endogenous growth via frictional knowledge diffusion to examine the effect of technological waves---defined as long-term shifts in the importance of specific knowledge fields---on local population dynamics. We calibrate the model using a new dataset of historical geolocated patents spanning over one hundred years. We find that frictions to idea diffusion across locations and technological fields account for two-thirds of the empirical relationship between exposure to technological waves and local growth in the United States during the twentieth century. Counterfactual experiments suggest that future technological scenarios may have large geographical effects.
CEPR Working Paper DP16794
Knowledge Access: The Effects of Carnegie Libraries on Innovation (last version) (Conditionally Accepted, RESTAT)
with Peter Nencka
More innovation
Information is the key
Libraries matter
Between 1883 and 1919, Andrew Carnegie funded the construction of more than 1,500 public libraries across the United States, reducing the costs of accessing knowledge for millions of people. We study the effect of these libraries on innovation using new data on city-level patenting and a novel control group: cities that qualified to receive a library grant and applied to be part of the program, but ultimately did not build a library. Patenting in recipient towns increased by 7-11 percent in the 20 years following library construction. We show that access to scientific knowledge and opportunities to interact with fellow patrons are possible mechanisms.
Patents are commonly used as the main source of data for empirical studies related to innovation and technological change. The large amount of information about the underlying innovative process contained in each patent has certainly contributed to their popularity. Nevertheless, due to the lack of reliable data, historical analysis has focused on relatively small time frames or on specific dimensions of patents data. The goal of this paper is to fill this gap. I build and release a comprehensive time series of the universe of U.S. patents. The data set contains all the variables commonly used in the literature and, importantly, geolocates every inventor and assignee reported in each grant over the period 1836-2016.
Global Innovation Spillovers and Productivity: Evidence from 100 years of World Patent Data (last version)
with Kristina Manysheva and Martí Mestieri
We use a panel of historical patent data covering a large range of countries over the past century to study the evolution of innovation across time and space and its effect on productivity. We document a substantial rise of international knowledge spillovers as measured by patent citations since the 1990s. This rise is mostly accounted for by an increase in citations to US and Japanese patents in fields of knowledge related to computation, information processing, and medicine. We estimate the causal effect of innovation induced by international spillovers on sectoral output per worker and TFP growth in a panel of country-sectors from 2000 to 2014, and on aggregate income per capita since 1960. To assess causality, we develop a shift-share instrument that leverages pre-existing citation linkages across countries and fields of knowledge, and heterogeneous countries' exposure to technology waves. On average, an increase of one standard deviation in log-patenting activity increases sectoral output per worker growth by 1.1 percentage points. We find results of similar magnitude for sectoral TFP growth and long-run aggregate income per capita growth.
A Quantitative Model of Income Segregation in a Knowledge Economy (draft coming soon)
with Ruben Gaetani
Older Working Papers
We study the relationship between local corruption and income misreporting, in the context of the EITC program. Using a newly assembled dataset of corruption at an MSA level, we find that public officials’ corruption predicts more than 70% of the variation in bunching observed in the distribution of EITC eligible self-employed workers’ reported income. We employ a research design that exploits exogenous variation in institutional accountability and find that a raise in our corruption measure by one standard deviation causes sharp bunching among self-employed workers to increase by 0.60 to 0.83 standard deviations. We argue and show suggestive evidence that information via news coverage is a relevant channel through which corruption affects tax evasion. The results are consistent with a framework in which individuals weigh their tax evasion decision against the cost of social stigma which is decreasing in the (perceived) level of institutional corruption.
Can Skill Mismatch Explain Geographic and Time Variation in the Returns to College Majors? Evidence from Online Job Postings (last version)
with Paul Mohnen and Bledi Taska
Early-career returns to college majors vary substantially across cities and over time, even after accounting for differences in the cost of living. In this paper, we show that mismatch between college majors and the skills demanded by local employers in the years following graduation is an important contributing factor. Exploiting data on the near-universe of online job postings in the U.S. between 2010 and 2015, we build a novel measure of skill mismatch capturing the “distance” between individual college majors and local labor demand. Intuitively, skill mismatch is low for a particular college graduate when a large fraction of job postings in her city are suitable for her major. We find that a one standard deviation increase in our mismatch measure is associated with a 1.5% decline in hourly wages among recent college graduates, controlling for individual characteristics and two-way interactions of college major, city and year fixed effects. College graduates are also more likely to be unemployed and more likely to hold non-college jobs conditional on employment when their major is less in demand. In ongoing work, we quantify how much of the cross-sectional and time variation in the returns to college majors can be explained by skill mismatch.