Job Market Paper
Abstract: This paper quantifies the cost of an additional dollar of fiscal stimulus in future tax increases (fiscal adjustment) in the U.S. quarter by quarter from 1962 to 2026. I build a medium-scale, two-agent New Keynesian model with downward nominal wage rigidity and estimate it using Bayesian methods. Downward nominal wage rigidity is an occasionally binding constraint that gives rise to unemployment in downturns, making the model’s dynamics nonlinear over the business cycle. To account for these nonlinearities, I derive a new present-value measure of fiscal adjustment that is conditional on the state of the economy and use it to quantify fiscal adjustment over time. I find that on average the government needs to raise about $1.02 in future taxes for each dollar of fiscal stimulus. Stimulus is least costly in an economy with idle resources and low inflation because it is more likely to raise output than prices, expanding the tax base and increasing tax revenues. An example is the Great Recession, when fiscal adjustment was about $0.98 for each dollar of stimulus in 2009Q1. In contrast, stimulus is most costly when it raises inflation more than output, prompting the monetary authority to raise interest rates and making debt more expensive to service. Examples include the 1970s oil crisis and the COVID recovery, when fiscal adjustment was $1.06 in 1974Q3 and $1.07 in 2021Q1, respectively.
Estimated self-financing share of an additional dollar of stimulus in the U.S. between 1962Q1-2026Q2. Fiscal adjustment is 1 minus the self-financing share.
Revise and Resubmit at Journal of Applied Econometrics
Abstract: We propose a clustered local projection (clustered LP) method to estimate impulse response functions in a class of time-varying models where parameter variation is linked to a low-dimensional matrix of observables. Our iterative method consists of three steps: it classifies the data using k-means, estimates impulse response functions via GMM, and evaluates differences across clustered LP estimates using an ad hoc Wald Test to choose the right number of clusters. We employ our technique to examine how uncertainty influences the transmission of a contractionary monetary policy shock to medium- and long-maturity U.S. nominal Treasury yields, showing that the two types of uncertainty operate through complementary but distinct channels.