The Latest NBER Working Papers http://www.nber.org/new.html We examine how state policies requiring firms to facilitate workplace retirement saving affect household balance sheets. Using data from the Survey of Income and Program Participation (SIPP), we compare private-sector workers likely exposed to Oregon’s Automatic-Enrollment Individual Retirement Account (Auto-IRA) policy with similar workers in not-yet-adopting states. We find that the Auto-IRA policy is associated with increases in IRA and employer-sponsored retirement plan ownership and assets. We also find increases in checking or savings account ownership and balances, as well as higher credit card debt. The results suggest that Auto-IRA policies spill over to household liquidity management and borrowing. https://www.nber.org/papers/w35373#fromrss https://www.nber.org/papers/w35373#fromrss This study provides pre-registered, experimental evidence on the use of non-generative artificial intelligence (AI) chatbots to support students in large-enrollment undergraduate courses. We find the chatbot messaging increased students’ final grades and engagement with academic supports, such as tutoring. Treatment effects were generally consistent across student demographics, with the exception of treated women in a Microeconomics course, who earned final grades that were seven percentage points higher than women in the control group. This study provides evidence that integrating AI-enabled outreach and communication to students in their college courses can enhance student engagement, learning, and course performance. https://www.nber.org/papers/w35397#fromrss https://www.nber.org/papers/w35397#fromrss Individuals may fail to recall and use information they already know when making decisions. We empirically investigate whether such “retrieval failures” distort consumption smoothing behavior among Zambian farmers, who derive their income from one annual harvest and then spend it down over the course of the year. We document that individuals underestimate upcoming spending by 50%, creating scope for under-saving. In order to improve recall, we randomize an intervention that prompts individuals to think through their future expenses associatively in categories—without providing any external information or guidance. Treated individuals increase “remembered” expenses by 36-60%; as predicted by the memory literature, effects are concentrated among small, irregular, and stochastic items. Immediate spending drops and, six weeks after the intervention, treated households hold 15% higher savings. They subsequently enter the hungry season—the final months of the year when consumption typically declines sharply—with one additional month of savings, leading to a flatter spending profile over the year. Households use the increased savings to self-finance additional farm investment, resulting in a 9% increase in the next year’s crop revenue. We replicate the intervention’s impact on beliefs among low-income Americans, suggesting that retrieval failures generalize across settings and populations. https://www.nber.org/papers/w35430#fromrss https://www.nber.org/papers/w35430#fromrss Evaluating optimized AI systems for asset pricing is fundamentally difficult for two reasons. First, models are trained on all data, implying that any backtest or analysis using historical data suffers from look-ahead bias. In addition, markets are reflexive — as investors adopt AI, prices adjust — which may erode the very patterns the AI system was trained to exploit. We introduce a real-time, out-of-sample benchmark designed to sidestep both problems. The benchmark measures how well AI systems can explain contemporaneous stock returns around earnings announcements using only information available at announcement time, including the text of the announcement itself. Applying this benchmark to a range of agentic AI systems — which extract structured signals from earnings call transcripts and optimize over those signals — we find that the best-optimized systems more than double the explained variation in returns relative to standard benchmarks (R2 increasing from 8% to close to 20%). We show that AI-based optimization can deliver efficiency gains relative to traditional machine learning methods while also improving interpretability as our approach produces human-readable economic mechanisms that explain price movements. These learned rules can be compared to the drivers of realized returns in existing asset pricing models to identify missing sources of variation in a data-driven, self-evolving way that integrates empirical learning with economic structure. We release an SDK for researchers to improve on our results. Saturating this benchmark would represent fundamental progress in understanding how capital markets process firm-level information. https://www.nber.org/papers/w35431#fromrss https://www.nber.org/papers/w35431#fromrss This study explores the effects of COVID-19 school closures on labor market outcomes of Canadian parents of school-aged children. Using newly collected data on grade-level-specific school closures across 141 Canadian cities along with individual-level panel data from the Labour Force Survey, difference-in-differences estimates provide robust evidence that restrictions on in-person schooling reduced employment among partnered mothers of school-aged children by approximately 2 percentage points. For fathers, labor supply adjustments are much weaker and concentrated on single fathers. Finally, auxiliary analyses using administrative tax records provide suggestive evidence that school closures increased public benefit receipt among single parents. https://www.nber.org/papers/w35432#fromrss https://www.nber.org/papers/w35432#fromrss This paper examines the labor market effects of overtime coverage in the United States, where salaried workers are covered for overtime if their base pay falls below a legislated salary threshold. Using an event-study design with administrative payroll data and state-level threshold changes from 2014-2021, I find evidence against conventional models of overtime. Contrary to the historical intent of policymakers, firms do not increase employment by substituting more workers for fewer hours. However, contrary to compensating differential models, firms also do not offset the costs of overtime by lowering workers' base pays. Instead, employers raised salaries above the threshold to keep workers exempt from overtime, indicating that monitoring and adjusting workers' hours is costly for firms. Taken together, these results suggest that expanding overtime coverage increases workers' earnings without negatively impacting employment. https://www.nber.org/papers/w35433#fromrss https://www.nber.org/papers/w35433#fromrss Earnings inequality among college graduates has grown, and returns to a bachelor's degree (BA) vary substantially across colleges. We experimentally evaluate a program that shifted high-achieving, low-income students into higher-quality colleges. Students randomly offered a guarantee of four years of tuition and fees at the University of Michigan were 21 percentage points more likely to enroll at a UM campus. The offer shifted students across colleges rather than into college: most compliers would otherwise have attended less-selective four-year colleges. HAIL increased four-year BA completion by 4 percentage points, but had no effect on attainment after six years. It substantially increased degree quality: students offered the scholarship were 12 percentage points more likely to earn a bachelor's degree from a highly or most selective college. IV estimates indicate that students induced to attend a University of Michigan campus were 56 percentage points more likely to earn such a degree. The increase in college quality predicts earnings gains of 18 to 25 percent. Faster degree completion adds about another 1 percent to the present discounted value of lifetime earnings. https://www.nber.org/papers/w35434#fromrss https://www.nber.org/papers/w35434#fromrss The persistent surge in U.S. inflation that began in 2021 caught forecasters and policymakers by surprise. The 2021 inflation shocks were viewed as transitory, not persistent, leading to large forecast errors in late 2021 and 2022. This paper asks whether time series models – using only data on current and past inflation, but incorporating stochastic volatility and exhibiting time-varying persistence – performed better. Univariate models, using real-time data, did not. Multivariate models, incorporating sectoral inflation measures, did. https://www.nber.org/papers/w35435#fromrss https://www.nber.org/papers/w35435#fromrss We study whether information frictions about the relative effectiveness of study methods distort students’ effort allocation. Standard models treat effort as a single input whose level, rather than composition, determines outcomes. We instead model effort as an allocation across methods with heterogeneous effectiveness. Combining panel surveys, administrative records, digital activity logs, and a randomized information intervention with over 2,000 undergraduates, we test whether imperfect information leads to effort misallocation and quantify consequences for performance. At baseline, students hold divergent beliefs, and many devote time to passive strategies such as rereading, despite evidence that active retrieval is more effective; these choices predict lower performance and larger self-assessment errors. We randomize treated students into general-feedback and personalized-feedback arms, benchmarking habits against peers in the latter. General feedback has little effect. Personalized feedback widens the perceived effectiveness gap between active and passive methods by 42 percent, shifts time toward effective practices, and raises exam scores by 0.05–0.08 standard deviations. A dynamic model with learning-by-doing and convex allocation costs shows that, for students with the weakest baseline beliefs, personalized feedback is equivalent to a 16.5 percent reduction in adoption costs. These results identify information frictions about multidimensional effort as a determinant of academic performance. https://www.nber.org/papers/w35436#fromrss https://www.nber.org/papers/w35436#fromrss Artificial intelligence will probably generate major changes in the US economy, although the nature, timing, and magnitude of those changes are highly uncertain. We analyze a set of long-term scenarios involving different combinations of faster productivity growth, greater income inequality, job displacement, and a higher capital share of income. For each scenario, we assess the implications for federal debt and potential policy responses related to faster economic growth, the distribution of income, support for workers who are laid off, and taxation and ownership of capital. Given the uncertainty surrounding AI’s economic effects, policies that are robust to different scenarios would be especially valuable. https://www.nber.org/papers/w35437#fromrss https://www.nber.org/papers/w35437#fromrss The central trade-off for designing Disability Insurance (DI) is between providing insurance to those in need while maintaining incentives to work. This paper develops a novel revealed-preference approach to identify the insurance value of DI benefits. We show that comparing the DI take-up response to a change in benefits versus a change in wages identifies the insurance value. Implementing our framework in Canada, we estimate that increasing DI benefits by $1 creates an additional disincentive cost of $0.60 but creates an insurance value of $2.20. Thus, our approach suggests that DI benefits are not overly generous in the Canadian context. https://www.nber.org/papers/w35438#fromrss https://www.nber.org/papers/w35438#fromrss We harmonize survey data on interest rates paid by approximately 15,000 small and medium enterprises across 125 firm surveys and 285,000 households across 83 household surveys spanning developing and rich countries to study the relationship between monetary policy rates and borrowing costs faced by SMEs and households. Using within-country variation in policy rates over time, we find that pass-through to firm and household borrowing rates is stronger in richer countries than in poorer ones. https://www.nber.org/papers/w35439#fromrss https://www.nber.org/papers/w35439#fromrss This paper examines the choice between public and private healthcare systems among elderly veterans, who are dually eligible to obtain care through the publicly operated Veterans Health Administration (VA) and through private providers financed by Medicare. We analyze health system choice among veterans who move across areas with differing rates of VA utilization to quantify the relative importance of individual-specific factors (e.g., preferences, income, health) and place-specific factors (e.g., local access, quality, and convenience). Our estimates indicate that 50–60% of geographic variation in VA use is attributable to demand-side individual factors, with the remainder explained by place-based factors. We also document important heterogeneity across types of care, with place-based factors playing a larger role for inpatient and emergency care than for outpatient and primary care. Additional analysis suggests that the supply-side features emphasized in recent legislation—distance to VA facilities and wait times—explain only a small share of estimated place effects. These findings highlight the importance of individual factors in health system choice and have implications for policies aimed at reducing geographic disparities in VA utilization. https://www.nber.org/papers/w35440#fromrss https://www.nber.org/papers/w35440#fromrss Japan has the world’s oldest population and a universal Long-Term Care Insurance (LTCI) system, yet how care use and costs differ between cognitive and physical impairment remains underexamined. Using the nationally representative Comprehensive Survey of Living Conditions (2016 and 2019 waves; 324,466 adults aged 65 and older), we compare the utilization, intensity, and cost of long-term care across three groups: those with a dementia diagnosis (defined as regularly receiving medical treatment for dementia), those with physical (ADL) limitations but no dementia, and those with neither. A dementia diagnosis is associated with dramatically higher care use—75.3% receive some care, versus 5.3% of the unimpaired reference group. In models adjusting for demographic and socioeconomic characteristics, dementia raises the probability of receiving formal care by 55 percentage points and informal care by 52 percentage points—roughly double the effects of ADL limitations—and, conditional on use, is associated with about 113 additional hours of formal care per month (some 60% more than physical limitations alone). We estimate formal-care costs at 1.07–4.03% of GDP (depending on the valuation method) and co-residing informal-care costs at 0.79% of GDP. Per-capita formal-care costs are substantially higher for those with dementia (about 3.5 versus 2.0 million JPY annually under the self-reported approach), whereas informal-care costs are nearly uniform across impairment types—underscoring the intensive, and largely invisible, contribution of family caregivers. With the dementia and mild-cognitive-impairment population projected to reach roughly 12 million by 2040, these findings point to mounting fiscal and family-care pressures from cognitive impairment in Japan. https://www.nber.org/papers/w35442#fromrss https://www.nber.org/papers/w35442#fromrss We develop a model of aberrant behavior by Bitcoin miners and test it with a new 2017-2025 dataset. Miners’ rewards, comprised partly of user fees, exhibit variability across blocks of transactions. When large reward disparities exist between adjacent blocks, miners have incentives to attempt alternative versions of prior blocks and claim other miners’ rewards for themselves. Regression analysis shows that fee differentials are associated with these attacks and longer waiting times between blocks. These patterns imply potential destabilization of the Bitcoin blockchain as future mining rewards become more volatile due to gradual withdrawal of fixed block subsidies. https://www.nber.org/papers/w35443#fromrss https://www.nber.org/papers/w35443#fromrss I study insurance markets in which scalable prediction, like AI, designs residual risk rather than merely classifies fixed risk. A complete-contracting benchmark shows that if prevention is observable, contractible, competitively supplied, and fully priced, it does not matter whether consumers, insurers, or vendors supply it. Adverse selection breaks such irrelevance. When high-risk consumers are more "AI-treatable," efficient prevention makes low-risk contracts attractive to them. A contract intended for low-risk consumers faces a risk-design trilemma: separate, prevent efficiently, or avoid cross-subsidy, but not all three. The result extends Rothschild-Stiglitz from distorted coverage to distorted risk-control technology and offers market design insights of AI in insurance markets. https://www.nber.org/papers/w35444#fromrss https://www.nber.org/papers/w35444#fromrss Misaligned incentives within organizations may explain why firms fail to adopt or fully benefit from productive technologies. We conducted a randomized controlled trial in Indian garment factories in which units received an anonymous worker-management communication technology, this technology paired with incentives for HR managers to communicate effectively with workers, or neither (control). We find that the technology alone had no impacts relative to control. But pairing the technology with HR incentives increased productivity by 5%, reduced absenteeism by 13%, and raised worker earnings by 3%. Impacts were driven by greater HR responsiveness and increased worker reporting of production-related issues. https://www.nber.org/papers/w35445#fromrss https://www.nber.org/papers/w35445#fromrss Industrial zones are a widely used industrial policy tool, yet their effects vary enormously. Matching the timing and location of zone establishment to firm-level data in Vietnam, we estimate that a new zone raises formal employment by 2,000 workers within seven years. This average masks extreme skewness: a handful of zones generate the gains while the rest produce almost none. Zones near major trade hubs account for most of the effect, driven by export-oriented foreign manufacturers. Accounting for spillovers, domestic firms face agglomeration shadows nearby but gain farther out, while foreign manufacturers operate as enclaves, generating few local linkages. https://www.nber.org/papers/w35446#fromrss https://www.nber.org/papers/w35446#fromrss The federal government appropriated $189.5 billion in pandemic aid to school districts through the Elementary and Secondary School Emergency Relief (ESSER) fund. We evaluate their impact on schools with district poverty shares near 5%, where a threshold-driven increase in funds enables us to implement a difference-in-discontinuities design. Federal funds were passed on to residents through reductions in local revenue collections, including property taxes, and did not result in increased per-pupil expenditures by school districts. We find no evidence that additional funds mitigated the declines in test scores. We provide suggestive evidence that political engagement, especially from parents, increased in districts that qualified for additional funds. This combination of engagement and reduced taxation may explain the greater enrollment and faster reopening we observe in such districts. https://www.nber.org/papers/w35447#fromrss https://www.nber.org/papers/w35447#fromrss We show that incorporating uninsurable countercyclical income risk into a standard international RBC model can qualitatively and quantitatively account for the quantity puzzles in open-economy macro, namely (i) the Backus-Smith puzzle, (ii) the Backus-Kehoe-Kydland puzzle and (iii) the weak correlation between the trade balance and the exchange rate. We also show that our model can simultaneously account for the Fama puzzle and the evidence that high interest rate countries have stronger currencies—which representative-agents models that rely only on financial or demand shocks cannot jointly account for. Crucially, our model resolves all these puzzles while relying solely on productivity shocks, and thus generates the observed domestic and cross-country macroeconomic comovement. https://www.nber.org/papers/w35448#fromrss https://www.nber.org/papers/w35448#fromrss Maritime transport remains the backbone of global trade, yet the port and shipping network that carries it has been transformed by containerization and related technological advances. Drawing on newly available granular data---digitized historical shipping records, georeferenced ship movements, and shipment-level routing information---we present five stylized facts on the structure and evolution of the maritime network. Global shipping activity is highly concentrated among a changing lineup of dominant top ports even as lower-ranked ports disperse, while state-owned Chinese port terminal operators increasingly account for these global volumes, boosting overall port operations while delivering efficiency gains mostly to Chinese vessels. We use these facts to organize a synthesis of a fast-growing literature: containerization reshaped which port cities could expand, reinforced hub-and-spoke concentration that yields large but localized welfare gains, embedded ports in multimodal networks that amplify the returns to infrastructure, and generated market power, congestion, and environmental costs. Together, this evidence shows how evolving maritime technologies simultaneously deepen global integration and heighten the economic and geopolitical importance of critical nodes in the transport network---and of who controls them. https://www.nber.org/papers/w35449#fromrss https://www.nber.org/papers/w35449#fromrss In the canonical view of over-the-counter markets, dealers intermediate single-asset trades one at a time. We study a complementary role: when investors trade several assets at once, dealers absorb the joint position into inventory, insuring against execution risk. Using data on the near-universe of Canadian fixed-income trades, we find that bundled transactions account for 20 percent of investor volume. We develop a simple model of why, when, where, and at what prices investors bundle, and use it to explain three patterns. Switches—a purchase paired with a sale—transact at a discount, while uni-directional bundles—all buys or sells—transact at a premium, because a switch’s legs hedge within the dealer’s book while a uni-directional bundle’s compound. Bundling rises with volatility, when insurance against execution risk is most valuable, and the two types sort asymmetrically across bilateral and electronic venues. Together, these patterns reveal a dimension of intermediation the single-asset view overlooks, and one likely to grow in importance. https://www.nber.org/papers/w35450#fromrss https://www.nber.org/papers/w35450#fromrss Using 380 trillion tokens of realized AI consumption across more than four hundred large language models from the licensed proprietary OpenRouter dataset covering approximately 2 percent of current global monthly AI token consumption, we analyze how AI affects firms, markets, and workers. Leveraging the unprecedented size, scope and granularity of this data, we construct the AI Factor from growth in tokens, dollars, and users, estimate firm-level AI Betas from stock return comovement, and characterize the AI Premium. First, we build a high-frequency AI factor and decompose it into salient components. Second, we show that firms whose returns covary more positively with the AI factor—high AI beta firms—earn higher subsequent returns, and the AI premium is large and heterogeneous. A value-weighted longshort strategy earns 64.1 basis points per week, and the premium is large for loadings on the intensive, frontier-oriented margin of AI consumption—closed-source models, paying and seasoned users, and long prompts—but not on casual or open-weight use. Third, the premium reaches beyond technology firms into consumer-facing and capital-heavy parts of the economy, but is absent in emerging markets, including China. Fourth, the AI exposure is more positive in nonroutine interactive work and more negative in analytical, scientific, and operations-control skills—an occupation one standard deviation higher in interaction-and-communication content has 0.36-standard-deviation higher market-implied AI exposure. Additionally, we provide early evidence of the rise of the agentic economy. https://www.nber.org/papers/w35451#fromrss https://www.nber.org/papers/w35451#fromrss This scoping review curates empirical research on economic well-being among Latino households in peer reviewed journals from 1990 to 2025, following PRISMA guidelines and a preregistered search strategy. The search yielded more than 5,000 studies revealing a greater than five-fold increase since the early 2000s. The majority of studies examined Medicaid and health insurance and approximately half examined government benefit programs, particularly with respect to eligibility criteria and benefit receipt. Few studies examined the employment outcomes of Latinos in the context of government benefit availability and even fewer focus on Latinos and income support. A deeper content review revealed that studies examining Hispanics are typically through comparative analyses with other racial and ethnic groups. These studies find higher poverty and material hardship, and lower receipt of government benefits, among Latino families relative to non-Latino White households. Fewer studies consider heterogeneity within Hispanic populations. Research on economic well-being among Latino households, as distinct from more general investigations regarding immigrants, expanded in tandem with growth in the Latino population and shifts in social policy. Key policy-relevant open topics of research that affect Latinos include those related to tax credits, benefit generosity, employment, and considerations of mixed-citizen status households. https://www.nber.org/papers/w35452#fromrss https://www.nber.org/papers/w35452#fromrss Financial sanctions are widely viewed as a powerful tool of economic statecraft, yet direct evidence on their effects remains limited. We study how sanctions affect access to global payment networks using data on correspondent banking relationships, which link banks across countries and currencies to enable cross-border transactions. We first show that the dollar and euro networks offer the broadest global connectivity, supported by a small number of key hubs in the U.S. and Europe. Sanctions that restrict access to these hubs are therefore potentially powerful. We then show how financial sanctions imposed during 2021-2025 – a period of rapid growth in sanctions – reduced targeted banks' access to major-currency networks by severing correspondent relationships. Sanctioned institutions often retained access but only through longer, indirect, and more fragile chains of intermediaries. Finally, in countries where sanctions expanded sharply, non-sanctioned banks also lost connectivity to Western networks, likely reflecting de-risking. These banks reoriented toward alternatives, particularly the Chinese yuan. Outside these heavily sanctioned economies, shifts away from the dollar and toward the yuan remain minimal on average. https://www.nber.org/papers/w35453#fromrss https://www.nber.org/papers/w35453#fromrss We examine a choice between bonus contracts offered to dealers of a U.S. auto manufacturer. In our data, dealers select the non-profit-maximizing option in 20 percent of observations, costing the mistaken dealers $18,453 per year on average. We examine how the propensity to make this mistake varies with competition, identified both cross-sectionally and within dealers over time. Both analyses show that greater competition substantially lowers the rate of mistakes. However, even in the most competitive markets, consequential mistakes persist. Our results suggest that competition disciplines mainly through within-dealer changes in behavior rather than entry and exit. https://www.nber.org/papers/w35454#fromrss https://www.nber.org/papers/w35454#fromrss We study how residential-commercial zoning affects the allocation of urban space. Using property-level data from Taipei and 34 U.S. metropolitan areas, we infer neighborhood-level zoning wedges from the allocation of residents, workers, and floor space. We find substantially greater dispersion in these wedges in U.S. cities than in Taipei, where mixed-use development is pervasive. The inferred wedges increase neighborhood specialization and reduce welfare. We then evaluate whether zoning is aligned with the neighborhood characteristics that would justify intervention. Although zoning is systematically related to comparative advantage, comparative advantage explains only a small fraction of the variation in zoning. The dominant effect of zoning in American cities is therefore not to promote efficient land use, but to increase the segregation of residential and commercial activity across neighborhoods. https://www.nber.org/papers/w35455#fromrss https://www.nber.org/papers/w35455#fromrss Both historically and today, support of aging parents has largely taken the form of in-kind transfers that require physical proximity, such as housing and caregiving. If Social Security substitutes for such support, it can relax constraints on where recipients' children live and work. We investigate the long-run intergenerational effects of the early Social Security program, exploiting within-occupation, cross-industry differences in coverage and a new dataset linking parents to their children's later-life outcomes. We find that sons whose parents had greater predicted coverage moved farther from their childhood homes, earned more, and lived in better neighborhoods late in life. We find no such effects for daughters, who tended to provide forms of support less easily replaced by Social Security. The gains considerably exceeded the associated Social Security benefits for the average family, with migration to better-matched labor markets a likely key driver. We propose that the early program enabled families to realize gains from migration that were back-loaded, uncertain, and difficult to contract on. https://www.nber.org/papers/w35456#fromrss https://www.nber.org/papers/w35456#fromrss Discrete choice experiments are widely used to estimate workers’ willingness to pay (WTP) for job amenities under the assumption that varying an attribute does not change workers’ beliefs about other job attributes. We test this assumption by embedding an amenity with a known market price—a popular monthly public transport pass—in a largescale discrete choice experiment with German workers. Many workers, including public transport users, overvalue the ticket by more than 100%, despite WTP for other attributes aligning with the literature. A complementary belief-elicitation experiment shows that advertising an amenity, such as the pass but also common amenities like work from home, causally shifts beliefs about unlisted attributes of the job. Posted wages similarly signal unlisted attributes so that wage variation, the money metric for WTP calculation, is itself contaminated by belief spillovers—such as higher pay signaling heightened stress. These spillovers imply that discrete choice estimates capture perceived bundles rather than isolated attributes, and distort current estimates of non-wage compensation and monopsony power. https://www.nber.org/papers/w35457#fromrss https://www.nber.org/papers/w35457#fromrss Firms often rebrand to counter negative shocks, but can it work? Using a historical natural experiment, we analyze a large sample of companies from the same industry (banking) that shared very similar names, but overwhelmingly decided to change them in response to a common, negative news shock. U.S. entry into World War I in 1917 created a sudden anti-German backlash against anything that invoked an association with the now enemy. The shock itself was thus orthogonal to bank fundamentals and pre-existing trends, but pushed banks to reconsider their brands. After 1917, the few German-named banks that kept their tainted names saw significant declines in assets, deposits, and market share relative to other banks. However, German-named banks that rebranded mitigated much of the negative shock. Specifically, German-named banks that adopted a new non-ethnic brand name avoided between 75% and 85% of the anti-German effect, and those that adopted a new patriotic brand name were able to neutralize it completely. Overall, we find that “crisis rebranding” paid off in our historical setting regardless of the chosen brand, with patriotic rebranding proving the most effective at offsetting the exogenous shock. https://www.nber.org/papers/w35458#fromrss https://www.nber.org/papers/w35458#fromrss We revisit fiscal sustainability through the lens of the government’s budget constraint. What constrains fiscal policy is not the stock of outstanding liabilities per se, but the fiscal cost of servicing debt. Using two centuries of U.S. fiscal data (1800–2023) and a long-run panel of advanced economies, we show that primary surpluses are systematically more closely associated with debt-service burdens, while debt ratios lose explanatory power when debt service is taken into account. Fiscal responses intensify when financing conditions deteriorate, specifically when the interest-growth differential is positive. We rationalize these findings using a simple flow-based framework in which debt stabilization depends on the responsiveness of fiscal surpluses to financing pressures. The results suggest that fiscal sustainability depends less on debt thresholds than on financing regimes and governments’ ability to absorb debt-service burdens. https://www.nber.org/papers/w35459#fromrss https://www.nber.org/papers/w35459#fromrss We analyze collaborative experimentation across multiple independent domains. Each domain contains infinitely many potential projects with asymmetric benefits. In each period and in each domain, two players can idle, jointly explore a new project, or jointly exploit a known one, with voluntary transfers. For intermediate discount factors, treating domains as independent during experimentation is suboptimal. The optimal experimentation policy exhibits common features of collaborative experimentation: lengthy exploration, temporary project exploitation, recall of past projects, and inefficient initial or terminal idling within certain domains. We connect these findings to research on buyer-supplier dynamics and persistent productivity differences. https://www.nber.org/papers/w35460#fromrss https://www.nber.org/papers/w35460#fromrss We develop a simple and intuitive Pigouvian perspective on optimal trade policy. Our approach unifies a wide range of rationales for taxing trade, from the classical optimal tariff argument to contemporary debates about global carbon emissions and geopolitics. We also clarify when trade policy intervention is warranted and when alternative domestic instruments should be used instead. https://www.nber.org/papers/w35461#fromrss https://www.nber.org/papers/w35461#fromrss We investigate the short- and long-run impacts of unexpected mortality on psychosocial health in the context of a large-scale high-mortality natural disaster, the 2004 Indian Ocean earthquake and tsunami. The effects of the tsunami-related mortality measured at the community level are contrasted with the effects of individual-specific loss of close kin using population-representative longitudinal survey data from the Study of the Tsunami Aftermath and Recovery. In the short-run, two years post-tsunami, both community and individual-level mortality exposures are significant predictors of elevated depression and post-traumatic stress symtoms. In the longer-run, ten years post-tsunami, individual-level exposures are slightly attenuated but remain significant predictors of both psychosocial health measures; in contrast, community-level mortality predicts post-traumatic stress but not depression symptoms. Mortality is linked to smaller household sizes in the short- and long-run and to less social support in the long-run, although the latter effect differs substantially for males and females. The estimates adjust for other tsunami exposure measures and, in the individual-level analyses, we compare people living in the same community at the time of the tsunami who were exposed to the same community-level mortality rate. https://www.nber.org/papers/w35462#fromrss https://www.nber.org/papers/w35462#fromrss Emerging market economies (EMEs) have historically been vulnerable to external shocks. Recently the global economy has experienced several major shocks yet EMEs have been remarkably resilient. Some of this improved performance can be attributed to prudent policies and stronger economic fundamentals, but the existing literature points to an overperformance mystery by some EMEs with weaker fundamentals. We evaluate an alternative measure of vulnerability based on the concept of a natural level of capital flows (KF*). Specifically, we hypothesize that EMEs are more vulnerable when prior to a large global shock capital flows—portfolio inflows, to be specific—exceed KF*. Results indicate that for major global shocks over the past two decades the pre-shock gap between actual portfolio and KF* has substantial predictive power for the post-shock performance of EMEs. https://www.nber.org/papers/w35463#fromrss https://www.nber.org/papers/w35463#fromrss The growing prevalence of cognitive impairment (CI) is one of the main drivers of age-related demand for health and long-term care (LTC). In Spain, CI is estimated to affect 18.5% of Spaniards over 65, and 45.3% in those aged 85 and above. This paper draws on a pooled pre-COVID data from a longitudinal sample of individuals aged 65+ to examine the effect of CI and physical limitations on health and long-term care utilisation, estimates its costs, and financial burden. We report four sets of findings. First, we find that socioeconomic status at older age to be the strongest predictor of CI. Second, while both CI and physical limitations increase health and care adult care use, physical impairment is a stronger predictor of overall care utilisation (73% versus 55% for CI alone) and nursing home residence (2.0% versus 0.9%). Third, informal caregiving constitutes the overwhelming majority of dementia costs, accounting for 69–81% of the total. Finally, we estimate that the replacement cost of informal care would exhaust the full budget of Spain’s LTC system (SAAD). https://www.nber.org/papers/w35464#fromrss https://www.nber.org/papers/w35464#fromrss This paper studies social insurance with imperfect eligibility screening, focusing on Unemployment Insurance (UI) during its expansion in 2020 and 2021. We study the extent of imperfect screening by identifying anomalous payments using administrative tax data and UI policies, finding $214 billion in potentially-improper payments—concentrated in the Pandemic Unemployment Assistance (PUA) program—with approximately half detectable ex-ante through improved federal-state data sharing. There is substantial geographic variation, and a border design shows this is partly due to policy decisions made by states. To assess the implications for optimal policy, we first conduct simulations that replace PUA with means-tested, lump-sum transfers, finding that these transfers would have better insured against income losses at lower administrative cost. Second, we develop a model of opt-in versus automatic transfers that shows the targeting advantage of opt-in programs can reverse when ineligible recipients pass the benefit screen. Calibrated to 2020 UI, the model implies that shifting toward automatic transfers would have increased social welfare. https://www.nber.org/papers/w35465#fromrss https://www.nber.org/papers/w35465#fromrss We interpret exchange rate and yield curve responses to the April 2, 2025 U.S. tariff announcement through the lens of an equilibrium model capturing the portfolio balance mechanism. Disciplined by price elasticities from QE announcements, a 3-12% decline in the demand for dollar bonds (relative to annual U.S. GDP) accounts for the dollar depreciation and rise in dollar yields during this episode. News of a gradual rebalancing out of dollar bonds is consistent with immediate price impact since asset prices are forward-looking. Lower dollar bond demand was accompanied by higher risk aversion, which explains the cross-section of responses across currencies. https://www.nber.org/papers/w35466#fromrss https://www.nber.org/papers/w35466#fromrss We provide descriptive evidence on women's experiences in platform-based driving and delivery work using survey and administrative data from India and Indonesia. Partnering with major two-wheeler platforms, we study representative samples of drivers, oversampling women to examine gender differences (India: 404 women, 2,153 men; Indonesia: 892 women, 2,114 men). Female participation is extremely low -- 0.8% in India and 1.5% in Indonesia -- consistent with barriers related to safety concerns, cultural norms, and limited access to required assets. Among women who do enter, platform work is used differently than by men: women work fewer hours, are more likely to log in during free time, and place greater value on flexibility and supplemental earnings. Women earn less per hour than men -- by 6-8% in India and 8-22% in Indonesia -- differences that align with gendered choices over working times, locations, and service types, likely shaped by safety and domestic constraints. Women also face distinct challenges: higher reported harassment in both countries and, in Indonesia, widespread customer discrimination and higher accident rates. At the same time, a disproportionate share of women were unemployed before joining the platform, and women are more likely than men to report higher current earnings relative to prior work. Together, the evidence highlights how flexibility coexists with persistent gender-specific risks, constraints, and unequal returns in location-based platform work in low- and middle-income country contexts. https://www.nber.org/papers/w35467#fromrss https://www.nber.org/papers/w35467#fromrss We revisit the classic identification problem of separating supply and demand for a homogeneous good using data from multiple markets. We allow markets to be heterogeneous according to unobservables, a feature that arises if there are unobservable differences in consumer preferences or firm technology. We develop a new identification analysis based on hypothetical market types. We use this analysis to show how nonparametric, economically motivated assumptions carry empirical restrictions for a wide range of target parameters, including elasticities, but also welfare parameters, such as consumer surplus. Then, we develop computationally tractable methods for implementing partially identified linear random coefficients models in which the slopes of supply and demand are heterogeneous. We apply these methods to estimate the welfare impact and incidence of sales taxes in the United States. https://www.nber.org/papers/w35468#fromrss https://www.nber.org/papers/w35468#fromrss