RESEARCH
Bubbles AND TEXT
Talking and bubbles: narratives in Social and News Media JOB MARKET PAPER, LINK (New version: March 2024)
Abstract: I examine talking and listening during bubble events. In a structural model, I show that bubbles can form where agents make optimal choices to listen and talk about narratives. Listening is a dynamic decision based on the future value of changing one's mind. I develop a framework to structurally analyze this model using textual data from Social and Traditional Media. This allows me to show (a) the model describes two real, Social Media era, bubbles well; (b) the novelty of narratives plays an important role in price volatility; and (c) that bubbles can form faster with Social Media. The latter occurs because Social Media, in contrast to Traditional Media, leads more narratives to receive substantial coverage. The framework may be useful for modelling other economic variables impacted by social interactions e.g. the emergence of political, social and environmental innovations.
A tale of two Bitcoin bubbles (2017,2020): the role of textual risk factors and how they differed across episodes
Abstract: Bitcoin has been one of the most prominent bubble assets during the social media age. Correspondingly the speed and magnitude of run-ups and crashes of Bitcoin's price are distinct when compared with many historically famous bubbles. In this paper, I use social media text data to examine its role in the formation of two distinct Bitcoin bubbles. I show that econometric methods typically applied to asset prices to identify bubbles also work on recovered textual data, constructed using cutting-edge text analysis methods for identifying key verbal content. These findings suggest potential new avenues for bubble identification beyond simply monitoring the price. I also show that the causes of the first Bitcoin bubble are semantically distinct from those that drove the second.
CLimate change and financial frictions
Do macro-climate models need financial frictions: empirical evidence and a structural model. LATEST DRAFT.
(Winner: third year best paper award)
Abstract: Little has been said in the literature to date on how financial markets will react to climate change. Given the importance of these markets in the economy, and how the economy responds to shocks, this is a significant omission. I contribute to this gap by providing empirical evidence that establishes the financial consequences of climate shocks and building a theoretical model to parameterize the role of financial markets in the climate problem. My empirical results show that routine climate shocks drag on firms' ability to raise financing via higher credit spreads; while larger disasters cause much larger spikes in credit spreads of 60-100 basis points for an average firm. I use this evidence to motivate a financial macroeconomy with climate model (a DSGE model with a climate externality and collateral constraint) and state important results and intuition. The marginal externality damage equation from the macroclimate literature still holds but reductions in economic output are amplified.
Pandemics and macroeconomics
Disease-economy trade-offs under alternative pandemic control strategies (with Antonio Bento, Daniel Kaffine, Akhil Rao and Ana Bento) | 2022, Nature Communications, 13, 3319. LINK
Abstract: Public policy and academic debates regarding pandemic control strategies note disease-economy trade-offs, often prioritizing one outcome over the other. Using a calibrated, coupled epi-economic model of individual behavior embedded within the broader economy during a novel epidemic, we show that targeted isolation strategies can avert up to 91% of economic losses relative to voluntary isolation strategies. Unlike widely-used blanket lockdowns, economic savings of targeted isolation do not impose additional disease burdens, avoiding disease-economy trade-offs. Targeted isolation achieves this by addressing the fundamental coordination failure between infectious and susceptible individuals that drives the recession. Importantly, we show testing and compliance frictions can erode some of the gains from targeted isolation, but improving test quality unlocks the majority of the benefits of targeted isolation.
MIGRATION
How asylum seekers in the United States respond to their judges: evidence and implications (with Emily Nix) LATEST DRAFT.
Abstract: Every year thousands of migrants seek asylum in the United States. Upon entry they encounter U.S. immigration judges who exhibit large variability in their decisions, with on average a 20 percentage point within court gap in grant rates between the least versus most lenient judges. We document one potential consequence of these large discrepancies across judges: asylum applicants may defect from the court prescribed behavior in response. Focusing on the years 2009-2015, preceding and during a major increase in asylum applicants, we find that asylum seekers quasi-randomly assigned to less lenient immigration judges are more likely to be absent for their immigration hearings. We show that this type of endogenous response to decision-maker leniency leads to bias in second-stage estimates when using randomly assigned judges and variation in judge leniency as an instrument. We conclude that the extreme variability in judicial decisions in United States immigration courts causes important distortions in the behavior of those subject to such caprice.