Coauthored

Describing and explaining global interdependence from trade require an approach that incorporates fundamental features of the modern global economy, such as trade in intermediate goods, substitutability across products, and alternative trading partners. We use theoretical models of structural gravity to construct measures of dyadic dependence that directly incorporate these features. The measures are valid under a wide variety of classical and new theories for trade. The measures describe (1) how much damage is done to a country's welfare when dyadic trade is interrupted and (2) how much additional trade it would take from their existing trade partners to compensate. We then show three important findings. First, the new measures differ from traditional measures based on aggregate trade statistics in meaningful ways. Second, they tell a new story of how dependence and interdependence have changed over time. Contrary to much common wisdom, the world has _not_ marched inexorably towards greater interdependence. Average levels of dependence and interdependence -- the degree of symmetry in dependence relationships -- have both increased over time. But this is driven by a large increase in the degree to which countries are reliant on one partner, especially China. For many countries, dependence on China has risen to such a degree as to overwhelm any secular increase in interdependence. These important dynamics are obscured by aggregate trade statistics. Third, we use China's foreign assistance, as exemplified by the Belt and Road Initiative, to assess the degree to which states can successfully manipulate their dependence relationships. We find that the Chinese assistance has increased partners' dependence on China over time, without a reciprocal increase in Chinese dependence. These important effects of Chinese assistance are not apparent in analysis of aggregate trade statistics.

When one country or international organization makes accusations about violations of international law, the intended audience is often third-party states who might support punishing the offender. When do these accusations persuade publics in those countries and when do they trigger backlash? We show that reactions to accusations about international law violations are consistent with a theoretical model that allows for both types of responses – persuasion and backlash – depending on the audience member’s prior beliefs and trust in the information source. We provide evidence from large survey experiments in four global swing states: India, South Africa, Turkey, and Indonesia. Swing states are where persuasion or backlash matter most, since allegations about international law could conceivably tilt their support toward the accuser or the accused. In our survey experiments, when the International Criminal Court makes accusations that Russia violated international law, this persuades certain subsets of the population. When the United States makes an identical accusation, it fails to persuade—and often backfires—because publics in global swing states hold lower trust in the United States as a credible source. We further show how accusations affect perceptions of the accuser, not just the accused. We show a feedback effect, where information sent today can increase or decrease views of the credibility of the information source. This can then magnify or mute the effect of future accusations. Accusations by the ICC improve respondents’ views of the Court’s credibility. Accusations by the United States further undermine its credibility.

Conventional wisdom holds that partisan polarization dooms the prospects of redistribution in the United States, particularly because conservatives resist increasing government spending. We argue that globalization weakens this opposition, but only for certain redistribution policies. The perceived harms of globalization increase bipartisan support for social investment, like education spending, which promises to strengthen the present and future workforce. By contrast, conservatives remain opposed to social assistance programs, such as unemployment benefits and cash aid. The contrasting views on social investment and assistance reflects how many Americans view globalization as undermining the meritocratic foundations of the “American Dream.” Social investment is seen as restoring the link between hard work and success, whereas social assistance is not. Using nationally representative survey experiments, we show that prompts about globalization boost support for social investment, especially among conservatives. But those prompts do not increase support for social assistance. We complement this with observational data on education spending at the local level from 1992-2022. The areas hardest hit by trade shocks -- \textit{both} Republican and Democratic counties -- increased education spending at higher rates, despite fiscal pressures from trade shocks. Our findings suggest that partisan divides over redistribution are surmountable, and that the United States need not remain confined to protectionism, piecemeal industrial policy, and lackluster support for workers.