avatar

Floyd Jiuyun Zhang

Economist at Instacart
PhD, Stanford University
floydzhang93 (at) gmail.com


About Me

I am an economist based in the San Francisco Bay Area. I work on micro-economics and algorithms at Instacart, where I am an economist and machine learning engineer. Prior to Instacart, I received my PhD from Stanford University’s Graduate School of Business in 2022, where I researched the economics of political institutions and behavior. I have also studied at the Vancouver School of Economics at the University of British Columbia, earning a bachelor’s and a master’s degree in 2016 and 2017 respectively.

Publications and Working Papers

  1. Nature Human Behaviour
    Political Endorsement by Nature and Trust in Scientific Expertise During COVID-19
    Nature Human Behaviour (2023)
    In a large-sample pre-registered online experiment, participants were randomly assigned to receive information about the journal Nature's endorsement of Joe Biden. The informational treatment substantially reduced and polarized confidence in Nature and the scientific community among participants, and the behavioral impacts of this belief update include 38.3% lower demand for public health information attributed to Nature among Trump supporters. I found no evidence that the endorsement changed participants' views about Biden or Trump.
    Select coverage: Nature Editorial (official response), Nature News & Views, Politico, The Washington Post, The Wall Street Journal, Reason Magazine, Morning Brew, The Atlantic, Phys.org, Not Another Politics Podcast (University of Chicago), Copernicus (Jagiellonian University, Poland), Nature Asia (Japan), The Lancet (UK), i Newspaper (UK), La Nación (Argentina), Le Point (France), Süddeutsche Zeitung (Germany)
    Selective Exposure and Electoral Competition (with Avidit Acharya, Peter Buisseret, and Adam Meirowitz)
    Revise and resubmit, Journal of Politics
    We theoretically study how selective exposure to information affects electoral competition and policy polarization by analyzing a Calvert-Wittman probabilistic voting model in which voters only receive valence information slanted to confirm their predispositions. Contrary to predictions of existing work, we show that it reduces policy polarization relative to a benchmark where voters all consume information from a neutral source. The result is robust to non-Bayesian voters who are systematically manipulated by biased signals.

Work in Progress

  1. Rage Against the Merchant: The Effect of Job Automation on Trade Protection
    Using a shift-share instrumental variable exploiting differential penatration of industrial robots across industries, I find suggestive causal evidence that U.S. Congress members representing labor markets more exposed to automation vote for more restrictive trade policies. The magnitude of the impact of robots on protectionism is quantitatively similar to experiencing an import shock of comparably large labor market impact. This effect is driven by incumbent response instead of electoral selection and is more pronounced in low-information media markets. I interpret the results as systematic misattribution of labor market disruptions but also discuss caveats concerning the interpretation and the IV strategy.
    Nationalism and Preferences for Redistribution in Autocracies: The Opioid of the Masses?
    Why do autocrats promote nationalism? I present a simple model of endogenous identity formation in a stylized autocratic political economy where distributive conflicts are resolved by the threat of revolution. I show that the salience of national identity reduces citizen demand for income redistribution and political power, and that the structural parameters can be identified econometrically with a random utility model and a factorial survey experiment. Equilibrium implications of these behavioral predictions for growth and inequality are derived and discussed in relations to existing work on redistribution, authoritarian growth and control.

Referee Services

Select Grants and Awards

Teaching Assisant and Staff


Powered by Jekyll and Minimal Light theme.