arXiv subjects

Paul Goldsmith-Pinkham

Publications and source records attributed to Paul Goldsmith-Pinkham.

14 recordsLinked to original sources

arXiv2026

Leniency Designs: An Operator's Manual

We develop a step-by-step guide to leniency (a.k.a. judge or examiner instrument) designs, drawing on recent econometric literatures. The unbiased jackknife instrumental variables estimator (UJIVE) is purpose-built for leveraging exogenous leniency variation, avoiding subtle biases even in the presence of many decision-makers or controls. We show how UJIVE can also be used to assess key assumptions underlying leniency designs, including quasi-random assignment and average first-stage monotonicity, and to probe the external validity of treatment effect estimates. We further discuss statistical inference, arguing that non-clustered standard errors are often appropriate. A reanalysis of Farre-Mensa et al. (2020), using quasi-random examiner assignment to estimate the value of patents to startups, illustrates our checklist.

econ.EM↗

arXiv2026

Non-robustness of diffusion estimates on networks with measurement error

Network diffusion models are used to study disease transmission, information spread, technology adoption, and other socio-economic processes. We show that estimates of these diffusions are highly non-robust to mismeasurement. First, even when the network is measured perfectly, small and local mismeasurement in the initial seed generates a large shift in the locations of the expected diffusion. Second, if instead the initial seed is known, even a vanishingly small share of missed links causes diffusion forecasts to be significant under-estimates. Forecast failure depends critically on the geometry of measurement error: we provide sufficient conditions for catastrophic failure when missing links bridge distant network regions (acting as shortcuts), and sufficient conditions for robustness when missing links are a uniformly, randomly thinned subset of the full network (preserving network structure). Such failures exist even when the basic reproductive number is consistently estimable. We explore difficulties implementing possible solutions and conduct simulations on synthetic and real networks.

econ.EM↗

arXiv2026

Human-AI Collaboration in Radiology: The Case of Pulmonary Embolism

We study how radiologists use AI to diagnose pulmonary embolism (PE), tracking over 100,000 scans interpreted by nearly 400 radiologists during the staggered rollout of a real-world FDA-approved diagnostic platform in a hospital system. When AI flags PE, radiologists agree 84% of the time; when AI predicts no PE, they agree 97%. Disagreement evolves substantially: radiologists initially reject AI-positive PEs in 30% of cases, dropping to 12% by year two. Despite a 16% increase in scan volume, diagnostic speed remains stable while per-radiologist monthly volumes nearly double, with no change in patient mortality -- suggesting AI improves workflow without compromising outcomes. We document significant heterogeneity in AI collaboration: some radiologists reject AI-flagged PEs half the time while others accept nearly always; female radiologists are 6 percentage points less likely to override AI than male radiologists. Moderate AI engagement is associated with the highest agreement, whereas both low and high engagement show more disagreement. Follow-up imaging reveals that when radiologists override AI to diagnose PE, 54% of subsequent scans show both agreeing on no PE within 30 days.

econ.GN↗

arXiv2025

Causal Inference in Financial Event Studies

Financial event studies, ubiquitous in finance research, typically use linear factor models with known factors to estimate abnormal returns and identify causal effects of information events. This paper demonstrates that when factor models are misspecified -- an almost certain reality -- traditional event study estimators produce inconsistent estimates of treatment effects. The bias is particularly severe during volatile periods, over long horizons, and when event timing correlates with market conditions. We derive precise conditions for identification and expressions for asymptotic bias. As an alternative, we propose synthetic control methods that construct replicating portfolios from control securities without imposing specific factor structures. Revisiting four empirical applications, we show that some established findings may reflect model misspecification rather than true treatment effects. While traditional methods remain reliable for short-horizon studies with random event timing, our results suggest caution when interpreting long-horizon or volatile-period event studies and highlight the importance of quasi-experimental designs when available.

econ.EM↗

arXiv2025

Anonymous Attention and Abuse

We analyze the content of the anonymous online discussion forum Economics Job Market Rumors (EJMR) and document its evolving interactions with external information sources. We focus on three key aspects: the prevalence and impact of links to external domains, the surge in discussions driven by Twitter posts since 2018, and the categorization of individuals whose tweets are most frequently discussed on EJMR. Using data on linked domains, we show how these trends reflect broader changes in the economics profession's digital footprint. Our analysis sheds light on EJMR's informational role but also raises questions about inclusivity and professional ethics in economics.

econ.GN↗

arXiv2024

Anonymity and Identity Online

Economics Job Market Rumors (EJMR) is an online forum and clearinghouse for information on the academic job market for economists. It also includes content that is abusive, defamatory, racist, misogynistic, or otherwise "toxic." Almost all of this content is created anonymously by contributors who receive a four-character username when posting on EJMR. Using only publicly available data we show that the statistical properties of the scheme by which these usernames were generated allows the IP addresses from which most posts were made to be determined with high probability. We recover 47,630 distinct IP addresses of EJMR posters and attribute them to 66.1% of the roughly 7 million posts made over the past 12 years. We geolocate posts and describe aggregated cross-sectional variation -- particularly regarding toxic, misogynistic, and hate speech -- across sub-forums, geographies, institutions, and IP addresses. Our analysis suggests that content on EJMR comes from all echelons of the economics profession, including, but not limited to, its elite institutions.

econ.GN↗

arXiv2024

Contamination Bias in Linear Regressions

We study regressions with multiple treatments and a set of controls that is flexible enough to purge omitted variable bias. We show that these regressions generally fail to estimate convex averages of heterogeneous treatment effects -- instead, estimates of each treatment's effect are contaminated by non-convex averages of the effects of other treatments. We discuss three estimation approaches that avoid such contamination bias, including the targeting of easiest-to-estimate weighted average effects. A re-analysis of nine empirical applications finds economically and statistically meaningful contamination bias in observational studies; contamination bias in experimental studies is more limited due to smaller variability in propensity scores.

econ.EM↗

arXiv2024

Tracking the Credibility Revolution across Fields

This paper updates Currie, Kleven, and Zwiers (2020) by examining the credibility revolution across fields, including finance and macroeconomics, using NBER working papers up to May 2024. While the growth in terms related to identification and research designs have continued, finance and macroeconomics have lagged behind applied micro. Difference-in-differences and regression discontinuity designs have risen since 2002, but the growth in difference-in-difference has been larger, more persistent, and more ubiquitous. In contrast, instrumental variables have stayed flat over this period. Finance and macro, particularly corporate finance, has experienced significant growth in mentions of experimental and quasi-experimental methods and identification over this time period, but a large component of the credibility revolution in finance is due to difference-in-differences. Bartik and shift-share instruments have grown across all fields, with the most pronounced growth in international trade and investment, economic history, and labor studies. Synthetic control has not seen continued growth, and has fallen since 2020.

econ.GN↗

arXiv2024

Contagion Effects of the Silicon Valley Bank Run

This paper analyzes the contagion effects associated with the failure of Silicon Valley Bank (SVB) and identifies bank-specific vulnerabilities contributing to the subsequent declines in banks' stock returns. We find that uninsured deposits, unrealized losses in held-to-maturity securities, bank size, and cash holdings had a significant impact, while better-quality assets or holdings of liquid securities did not help mitigate the negative spillovers. Interestingly, banks whose stocks performed worse post-SVB also experienced lower returns in the previous year, following Federal Reserve interest rate hikes. Stock investors appeared to anticipate risks associated with uninsured deposit reliance, but did not foresee the realization of implied losses. While mid-sized banks experienced particular stress immediately after the SVB failure, over time negative spillovers became widespread except for the largest banks.

econ.GN↗

arXiv2022

Excess death rates for Republicans and Democrats during the COVID-19 pandemic

Political affiliation has emerged as a potential risk factor for COVID-19, amid evidence that Republican-leaning counties have had higher COVID-19 death rates than Democrat-leaning counties and evidence of a link between political party affiliation and vaccination views. This study constructs an individual-level dataset with political affiliation and excess death rates during the COVID-19 pandemic via a linkage of 2017 voter registration in Ohio and Florida to mortality data from 2018 to 2021. We estimate substantially higher excess death rates for registered Republicans when compared to registered Democrats, with almost all of the difference concentrated in the period after vaccines were widely available in our study states. Overall, the excess death rate for Republicans was 5.4 percentage points (pp), or 76%, higher than the excess death rate for Democrats. Post-vaccines, the excess death rate gap between Republicans and Democrats widened from 1.6 pp (22% of the Democrat excess death rate) to 10.4 pp (153% of the Democrat excess death rate). The gap in excess death rates between Republicans and Democrats is concentrated in counties with low vaccination rates and only materializes after vaccines became widely available.

econ.GN↗

arXiv2022

Measuring Changes in Disparity Gaps: An Application to Health Insurance

We propose a method for reporting how program evaluations reduce gaps between groups, such as the gender or Black-white gap. We first show that the reduction in disparities between groups can be written as the difference in conditional average treatment effects (CATE) for each group. Then, using a Kitagawa-Oaxaca-Blinder-style decomposition, we highlight how these CATE can be decomposed into unexplained differences in CATE in other observables versus differences in composition across other observables (e.g. the "endowment"). Finally, we apply this approach to study the impact of Medicare on American's access to health insurance.

econ.EM↗

arXiv2021

Doctors and Nurses Social Media Ads Reduced Holiday Travel and COVID-19 infections: A cluster randomized controlled trial in 13 States

During the COVID-19 epidemic, many health professionals started using mass communication on social media to relay critical information and persuade individuals to adopt preventative health behaviors. Our group of clinicians and nurses developed and recorded short video messages to encourage viewers to stay home for the Thanksgiving and Christmas Holidays. We then conducted a two-stage clustered randomized controlled trial in 820 counties (covering 13 States) in the United States of a large-scale Facebook ad campaign disseminating these messages. In the first level of randomization, we randomly divided the counties into two groups: high intensity and low intensity. In the second level, we randomly assigned zip codes to either treatment or control such that 75% of zip codes in high intensity counties received the treatment, while 25% of zip codes in low intensity counties received the treatment. In each treated zip code, we sent the ad to as many Facebook subscribers as possible (11,954,109 users received at least one ad at Thanksgiving and 23,302,290 users received at least one ad at Christmas). The first primary outcome was aggregate holiday travel, measured using mobile phone location data, available at the county level: we find that average distance travelled in high-intensity counties decreased by -0.993 percentage points (95% CI -1.616, -0.371, p-value 0.002) the three days before each holiday. The second primary outcome was COVID-19 infection at the zip-code level: COVID-19 infections recorded in the two-week period starting five days post-holiday declined by 3.5 percent (adjusted 95% CI [-6.2 percent, -0.7 percent], p-value 0.013) in intervention zip codes compared to control zip codes.

econ.GN↗

arXiv2021

The Great Equalizer: Medicare and the Geography of Consumer Financial Strain

We use a five percent sample of Americans' credit bureau data, combined with a regression discontinuity approach, to estimate the effect of universal health insurance at age 65-when most Americans become eligible for Medicare-at the national, state, and local level. We find a 30 percent reduction in debt collections-and a two-thirds reduction in the geographic variation in collections-with limited effects on other financial outcomes. The areas that experienced larger reductions in collections debt at age 65 were concentrated in the Southern United States, and had higher shares of black residents, people with disabilities, and for-profit hospitals.

econ.GN↗

arXiv2021

Interacting Regional Policies in Containing a Disease

Regional quarantine policies, in which a portion of a population surrounding infections are locked down, are an important tool to contain disease. However, jurisdictional governments -- such as cities, counties, states, and countries -- act with minimal coordination across borders. We show that a regional quarantine policy's effectiveness depends upon whether (i) the network of interactions satisfies a balanced-growth condition, (ii) infections have a short delay in detection, and (iii) the government has control over and knowledge of the necessary parts of the network (no leakage of behaviors). As these conditions generally fail to be satisfied, especially when interactions cross borders, we show that substantial improvements are possible if governments are outward-looking and proactive: triggering quarantines in reaction to neighbors' infection rates, in some cases even before infections are detected internally. We also show that even a few lax governments -- those that wait for nontrivial internal infection rates before quarantining -- impose substantial costs on the whole system. Our results illustrate the importance of understanding contagion across policy borders and offer a starting point in designing proactive policies for decentralized jurisdictions.

physics.soc-ph↗