编者按：美国全国经济研究所（NBER）是美国最大的经济学研究组织，其发布的工作论文代表着经济学研究最新的成果，每周一发布新论文。本周（1.20-1.26）共发布 20 篇新工作论文，本期将会推送最后 7 篇论文的题目与摘要，供读者学习。
学术财经研究团队翻译。
幸存者的心理健康与收入稳定的保护作用
Itzik Fadlon, Astrid Sophie Fugleholm, and Torben Heien Nielsen #33359
Abstract: We use administrative records on the healthcare utilization and economic outcomes of the universe of Danish households to characterize survivors' mental health following their spouse's death. We provide visually clear evidence for the inevitable immediate, large, and lingering adverse impacts and focus on studying the role of potential mediators: economic conditions and the presence of children. We find no evidence of heterogeneity in family composition. As for economic outcomes, baseline levels of income and net wealth play only a modest role: there is no meaningful cross-household inequality gradient in mental health declines, so that spousal death is devastating for both the rich and the poor. Rather, a key source of heterogeneity in the decline in mental health is the household's degree of income insurance, that is, the within-household income variation. Specifically, the least-insured households experience an immediate decline in mental health that is 80 percent larger. Our findings suggest that the consumption smoothing welfare gains from income protection policies can have important spillovers to improved mental health in the context of severe household events.
摘要：我们利用丹麦所有家庭的医疗保健利用和经济结果的行政记录，来描述配偶去世后幸存者的心理健康状况。我们提供了直观清晰的证据，显示配偶去世带来不可避免的、巨大且持续的负面影响，并重点研究潜在中介因素的作用：经济状况和子女的存在。我们未发现家庭组成的异质性证据。至于经济结果，收入和净财富的基准水平只起到了有限作用：在心理健康下降中没有显著的跨家庭不平等梯度，意味着配偶去世对富人和穷人都是毁灭性的。相反，心理健康下降的一个主要异质性来源是家庭的收入保险程度，也就是家庭内收入的变动。具体而言，收入保障最差的家庭在心理健康方面的立即下降幅度大约是其他家庭的80%。我们的研究结果表明，收入保护政策所带来的消费平滑福利增益，在严重家庭事件的背景下，可能对改善心理健康产生重要的溢出效应。
增长红利与超额利息
Danny Yagan #33375
Abstract: Standard deficit accounting neglects the growth dividend: the amount by which annual GDP growth shrinks the debt-GDP ratio. America's growth dividend has more than doubled since the Great Recession because the debt ratio has more than doubled, leading to headline deficits that far exceed changes in the debt ratio. Each year's change in the debt ratio can be decomposed into three components: the primary deficit (non-interest spending minus tax revenue), interest payments, and the growth dividend. The sum of the latter two is excess interest: the impact of past debt on the debt ratio, roughly equal to last year's debt ratio times the excess interest rate (r-g)/(1+g) where r is the average nominal interest rate on federal debt and g is the nominal GDP growth rate. Excess interest remains slightly negative in CBO's baseline ten-year projection. Hence, current debt is sustainable in the CBO baseline despite high interest payments, and primary deficits entirely drive the unsustainable projected debt ratio path and provide a good guide for how the debt ratio is projected to change. However, America's higher debt ratio implies higher vulnerability to the risk that the excess interest rate turns persistently positive.
摘要：标准的赤字会计忽略了增长红利：即年度GDP增长减少债务-GDP比率的金额。自大衰退以来，美国的增长红利已翻倍，因为债务比率已经翻倍，导致名义赤字远远超过债务比率的变化。每年的债务比率变化可以分解为三个组成部分：初级赤字（非利息支出减去税收收入）、利息支付和增长红利。后两者之和为超额利息：即过去债务对债务比率的影响，大致等于去年的债务比率乘以超额利率(r-g)/(1+g)，其中r是联邦债务的平均名义利率，g是名义GDP增长率。在CBO的基准十年预测中，超额利息仍然略为负值。因此，尽管利息支付较高，CBO基准预测中的当前债务仍然是可持续的，初级赤字完全推动了不可持续的债务比率路径，并为债务比率的变化提供了良好的预测指导。然而，美国较高的债务比率意味着在超额利率持续为正的风险面前，其脆弱性更大。
数字平台在塑造科技创业创新中的作用
Ginger Zhe Jin, Mario Leccese, and Liad Wagman #33370
Abstract: This chapter examines the multifaceted interactions between top digital platforms and technology ventures across capital, labor, innovation, and product markets. Exploring how venture investments, talent flows, strategic alliances, and competitive behaviors can shape the innovation ecosystem, the chapter highlights both the complementary and competitive dynamics between large incumbents and smaller entrants, and the benefits and potential inefficiencies that may arise from them, as demonstrated by the empirical and theoretical literatures. Throughout, the chapter identifies key areas for research that can support a rigorous evaluation of policy proposals concerning evolving market structures in the digital economy.
摘要：本章考察了顶级数字平台与科技创业公司在资本、劳动力、创新和产品市场方面的多维度互动。通过探索风险投资、人才流动、战略联盟和竞争行为如何塑造创新生态系统，本章强调了大公司与小型创业公司之间的互补性和竞争性动态，以及由此可能带来的利益和潜在的低效性，这一点在经验和理论文献中得到了证明。整章中，本章识别了可以支持对数字经济中不断变化的市场结构的政策提案进行严格评估的关键研究领域。
为年轻企业家现代化信贷获取：从FICO到现金流
Christopher M. Hair, Sabrina T. Howell, Mark J. Johnson, and Siena Matsumoto #33367
Abstract: Younger entrepreneurs are disadvantaged by traditional loan underwriting, which relies heavily on personal credit scores. With data from three fintech companies, we show that incorporating timely information about ability to repay from business checking account statements particularly improves default prediction performance for younger business owners. We develop a novel method to compare model predictions across subgroups—Tail Analysis for Comparative Outcomes (TACO)—which finds that switching from a Baseline (FICO-driven) model to a Cash Flow-enhanced model benefits younger entrepreneurs. We confirm this in causal analysis of approval decisions, showing that access to cash flow-intensive underwriting increases approval rates for younger vs. older entrepreneurs.
摘要：年轻企业家在传统贷款审查中处于不利地位，因为传统贷款审查通常严重依赖个人信用评分。通过来自三家金融科技公司的数据，我们展示了将来自商业支票账户对还款能力的及时信息纳入其中，特别能提高对年轻企业主的违约预测效果。我们开发了一种新方法——比较结果的尾部分析（TACO），用于比较不同子组的模型预测，发现从基准（基于FICO的）模型切换到现金流增强模型对年轻企业家有益。通过对审批决策的因果分析，我们证实了这一点，表明采用以现金流为主的审查方式能提高年轻企业家相对于年长企业家的审批通过率。
人工智能资产定价模型
Bryan T. Kelly, Boris Kuznetsov, Semyon Malamud, and Teng Andrea Xu #33351
Abstract: The core statistical technology in artificial intelligence is the large-scale transformer network. We propose a new asset pricing model that implants a transformer in the stochastic discount factor. This structure leverages conditional pricing information via cross-asset information sharing and nonlinearity. We also develop a linear transformer that serves as a simplified surrogate from which we derive an intuitive decomposition of the transformer's asset pricing mechanisms. We find large reductions in pricing errors from our artificial intelligence pricing model (AIPM) relative to previous machine learning models and dissect the sources of these gains.
摘要：人工智能中的核心统计技术是大规模变换器网络。我们提出了一种新的资产定价模型，将变换器嵌入到随机折现因子中。该结构通过跨资产信息共享和非线性，利用条件定价信息。我们还开发了一个线性变换器，作为简化的替代模型，从中我们推导出变换器资产定价机制的直观分解。与以前的机器学习模型相比，我们发现我们的人工智能定价模型（AIPM）在定价误差上有显著的减少，并剖析了这些收益的来源。
太阳能补贴的价值
Bryan Bollinger, Kenneth Gillingham, and A. Justin Kirkpatrick #33368
Abstract: Individuals trade present for future consumption across a range of economic behaviors, and this tradeoff may differ across socioeconomic groups. To assess these tradeoffs, we estimate a dynamic model of residential solar adoption and system sizing in California using household-level data on solar irradiance, electricity consumption, and electricity rates that offer plausibly exogenous variation in the future benefits from adopting relative to upfront costs. We find implicit discount rates of 15.3%, 13.8%, and 10.0% for low-, medium-, and high-wealth households. Counterfactual simulations demonstrate opportunities to reduce the regressivity of solar adoption, increase policy cost-effectiveness, and improve welfare for low-wealth households.
摘要：个人在一系列经济行为中将当前消费与未来消费进行权衡，而这种权衡可能在不同的社会经济群体之间有所不同。为了评估这些权衡，我们使用关于太阳能辐射、电力消费和电价的家庭级数据，估算了加利福尼亚州住宅太阳能采用和系统规模的动态模型，这些数据提供了在相对前期成本下，采用太阳能的未来收益的可合理外生变化。我们发现低财富、中财富和高财富家庭的隐含折现率分别为15.3%、13.8%和10.0%。反事实模拟表明，有机会减少太阳能采用的回归性，增加政策的成本效益，并改善低财富家庭的福利。
大规模移民时代的聚居区与同化：来自族裔天主教教堂的证据
Ran Abramitzky, Leah Platt Boustan, and Osea Giuntella #33362
Abstract: Immigrant enclaves offer valuable ethnic amenities but may delay assimilation. We study enclave formation in the Age of Mass Migration by using the centralized location decisions for “ethnic” Catholic churches. After a church opening, same-ethnicity residents of chosen neighborhoods experienced falling earnings but strengthened communal ties, as compared to residents of areas matched on baseline characteristics. Treated residents held more manual occupations, and increased in-group marriage and naming. These effects persist into the second generation and are not observed for non-ethnic neighbors. Consistent with the historical record, Poles organized communal life around neighborhood parishes, but Italians were less church-centered.
摘要：移民聚居区提供有价值的族裔便利设施，但可能会延缓同化进程。我们通过使用“族裔”天主教教堂的集中位置决策，研究了大规模移民时代的聚居区形成。在教堂开设之后，选定社区中的同族裔居民与基线特征相匹配的地区居民相比，经历了收入下降，但加强了社区联系。接受治疗的居民从事更多体力劳动职业，并增加了族内婚姻和命名。这些影响持续到第二代，并且在非族裔邻居中没有观察到。与历史记录一致，波兰人围绕社区教区组织社区生活，而意大利人则较少以教堂为中心。
往期精选：
