Contact Information
Department of Economics
Northwestern University
2211 Campus Drive
Evanston, IL 60208
Phone: 224-410-7955
jjnorris@u.northwestern.edu
Personal Website
Curriculum Vitae
Education
Ph.D., Economics, Northwestern University, 2019 (expected)
M.A., Economics, Northwestern University, 2014
M.Sci., Experimental and Theoretical Physics, University of Cambridge, 2013
B.A., Natural Sciences, University of Cambridge, 2012
Fields of Specialization
International Trade, Macroeconomics
Job Market Paper
“Fiscal Multipliers in Integrated Local Labor Markets”
Download (PDF)
I develop and analyze a spatial mechanism for generating the aggregate fiscal multiplier. Government budgets balance nationally. This allows the geographic distribution of government spending to differ from the geographic distribution of taxation. Given asymmetric economic geography, the combination can result in a net increase in aggregate GDP. This mechanism is orthogonal to canonical New Keynesian or Neoclassical mechanisms. Using models from International Trade, I construct a tractable general equilibrium representation of the fiscal multiplier in a spatially rich framework, accounting for geographic heterogeneity and interdependence due to trade. I develop an identification strategy and structurally estimate the model by combining a government spending shift-share instrument with the general equilibrium structure. I digitize a new historical dataset on interstate trade and apply my framework to US Federal defense procurement in the late 20th century. I estimate that the spatial channel generates variation in the multiplier of 50% relative to canonical mechanisms. Analogous to state-dependent fiscal multipliers popularized in the literature, my findings suggest an meaningful analogy with geography-dependent fiscal multipliers.
Other Working Papers
“Disproportionate Gains: A Home Market Effect in an Almost Arbitrary Geography”
Download (PDF)
The Home Market Effect (HME) provides a mechanism for trade driven by demand: locations with a relatively large consumer base for an industry specializes in production of that industry. The theory originates with Krugman in the 1980s, yet, despite the many iterations since, sharp theoretical predictions are only given in two-location models. As the world has more than two locations, empirical operationalization requires, therefore, assumptions outside of the model. By making only a single additional assumption relative to the canon — the matrix of interregional iceberg trade costs is positive semi-definite — I prove that the HME continues to hold in many location models on average. Furthermore, the empiricist needs to make little amendment: the HME hypothesis test valid in many-locations is almost identical to that already being ran using the two-location hypothesis.
Works in Progress
“Spillovers through Supply Chains”, with Caleb Kwon
Over the past decade, there has been a surge in using local labor market variation to inform macroeconomic questions. Although it offers the promise of sharper identification, only relative local treatment effects are typically identified, whereas it is the absolute local treatment effects that are relevant for policy. We demonstrate theoretically that the fundamental challenge in recovering absolute local treatment effects from relative local treatment effects is spatial interdependence. We present an approximate solution to this and develop a empirical framework to implement it. This is made feasible using a transactional level dataset on international trade. We construct exogenous shocks to a firm by exploiting a beneficial consequence of spatial interdependence: local shocks in economically distant firms can be used to instrument changes in economically close firms.
“Identification under Spatially Correlated Treatment Effects”
Draft coming soon
Increasing geographic disaggregation in a research design is appealing due to it increasing the sample size, N. The unfortunate inevitability is that this also increases the likelihood of spillovers as the distance between observations decreases (transitioning from states to counties to blocks, for example). The consequence is that, if treatments across space are correlated, then treatment in all other locations must be separately controlled for in the regression in order to consistently identify own-treatment effects. Yet a seeming parameter problem ensues: there are N observations but N2 coefficients (the effect in i of treatment in j, for all i,j pairs). Typical spatial econometric solutions leverage time asymptotics or impose strong functional form assumptions. I show that in fact neither are necessary: I develop an own-treatment effect estimator that is consistent with only N asymptotics and no functional form assumptions. I demonstrate it’s functionality using simulations and show that such bias is otherwise present in a remarkably common identification strategy: the Bartik Instrument.
Teaching
Department of Economics, Northwestern University, 2014-2018
Ph.D.: Introduction to Econometrics (Prof. Charles Manski)
Undergraduate: Econometrics (Prof. Jeff Lewis), Economics of Healthcare (Prof. Frank Limbrock), International Trade (Prof. Kiminori Matsuyama), International Finance (Prof. Richard Walker), PublicFinance (Prof. Lee Lockwood)
Download Teaching Evaluations
Kellogg School of Management, Northwestern University, 2014-2018
Executive: Strategic Marketing Decision (Prof. Tim Calkins, Prof. Julie Hennessy)
M.B.A.: International Finance (Prof. Sergio Rebelo), Public Finance for Business Leaders (Prof. David Besanko)
References
Prof. Martí Mestieri (Committee Chair)
Prof. Treb Allen
Prof. Martin Eichenbaum
Prof. Kiminori Matsuyama