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（10.7-10.13）共发布 26 篇新工作论文，本期将会推送前 9 篇论文的题目与摘要，供读者学习。
学术财经研究团队翻译。
通货膨胀不确定性对家庭信念与行为的因果影响
Dimitris Georgarakos, Yuriy Gorodnichenko, Olivier Coibion, and Geoff Kenny #33014
Abstract: We implement a survey-based randomized information treatment that generates independent variation in the inflation expectations and the uncertainty about future inflation of European households. This variation allows us to assess how both first and second moments of inflation expectations separately affect subsequent household decisions. We document several key findings. First, higher inflation uncertainty leads households to reduce their subsequent durable goods purchases for several months, while a higher expected level of inflation increases them. Second, an increase in uncertainty about inflation induces households to tilt their portfolios towards safe and away from riskier asset holdings. Third, higher inflation uncertainty encourages household job search, leading to higher subsequent employment among the unemployed and less under-employment among the employed. Finally, we document that the level of inflation expectations has a different effect from uncertainty in inflation expectations and thus it is crucial to take into account both to measure their separate effects on decisions.
摘要：我们实施了一项基于调查的随机信息处理实验，生成了欧洲家庭对通货膨胀预期及未来通货膨胀不确定性的独立变化。这种变化使我们能够分别评估通货膨胀预期的第一时刻（即平均值）和第二时刻（即不确定性）对家庭后续决策的影响。我们记录了几项关键发现。首先，较高的通货膨胀不确定性导致家庭在接下来的几个月内减少耐用品的购买，而较高的通货膨胀预期水平则增加了这些购买。其次，通货膨胀不确定性的增加促使家庭将其投资组合向安全资产倾斜，减少对风险资产的持有。第三，较高的通货膨胀不确定性鼓励家庭寻找工作机会，导致失业家庭的就业率上升，已就业者的就业不足现象减少。最后，我们发现通货膨胀预期的水平与通货膨胀预期的不确定性对家庭决策的影响不同，因此在测量其对决策的单独影响时，必须同时考虑这两者。
工人是否低估了 COVID-19 风险？基于工资和死亡证明数据的证据
Cong T. Gian, Sumedha Gupta, Kosali I. Simon, Ryan Sullivan, and Coady Wing #33031
Abstract: When mortality risks of a job increase, economic theory predicts that wages will rise to compensate workers. COVID-19 became a new source of mortality risk from close contact with other workers and customers. Real wages have risen during the COVID-19 era, but research to date has been sparse on how much of this increase reflects compensating wage differentials for COVID-19 risk on the job. We use 2020- 2021 death certificate data which for the first time includes the decedent’s occupation and industry, together with other occupational and industry mortality for previous years from the Census of Fatal Occupational Injuries (CFOI) and wage data from the Current Population Survey (CPS) to examine whether compensating wage differentials for COVID-19 occupational risks are in line with prior estimates of Values of Statistical Life (VSL). First, we find that there are substantial differences in the compensating differentials associated with COVID-19 vs other sources of job-related mortality risk. Full time workers’ pay is higher by $24 per week in jobs with a 1 in 1,000 higher risk of COVID-19 mortality, but their pay is $320 higher in jobs with 1 in 1,000 higher risk of non-COVID-19 workplace mortality. The non-COVID-19 mortality wage premiums imply that workers trade off money and mortality risk using a VSL of about $18 million, which is near the upper range of the most cited VSL estimates in the literature. In contrast, the COVID-19 wage premium implies that workers make decisions using a VSL of the range $1.24 - $1.54 million, much lower than standard VSL measures. The results are consistent with workers substantially underestimating or undervaluing the risk of COVID-19 mortality.
摘要：当工作的死亡风险增加时，经济理论预测工资将上升，以补偿工人。COVID-19 成为因与其他工人和顾客密切接触而带来的新的死亡风险来源。在 COVID-19 期间，实际工资有所上升，但迄今为止关于这部分增长中有多少是因工作中 COVID-19 风险而产生的补偿性工资差异的研究较为稀少。我们利用2020-2021年死亡证明数据（首次包含死者的职业和行业信息），并结合《职业致命伤害普查》（CFOI）中往年职业和行业死亡率数据，以及《当前人口调查》（CPS）中的工资数据，来分析 COVID-19 职业风险的补偿性工资差异是否与之前的统计生命价值（VSL）估算一致。首先，我们发现 COVID-19 与其他工作相关死亡风险来源的补偿差异存在显著差异。对于全职工人而言，在 COVID-19 死亡风险提高1/1,000的工作中，工资每周仅增加24美元，而在非 COVID-19 的工作场所死亡风险提高1/1,000的工作中，工资则增加320美元。非 COVID-19 死亡的工资溢价表明，工人在金钱与死亡风险之间的权衡基于约1800万美元的 VSL，这接近文献中最常引用的 VSL 估计的上限。相比之下，COVID-19 的工资溢价表明，工人决策基于124万至154万美元的 VSL，这远低于标准的 VSL 测量值。结果表明，工人对 COVID-19 的死亡风险存在显著低估或低估了其价值。
热能脱碳：热泵及分时电价对能源需求的影响
Louise Bernard, Andy Hackett, Robert D. Metcalfe, and Andrew Schein #33036
Abstract: Heat pumps have been proposed as the leading technology in the electrification of domestic heat and therefore could play a crucial part in the transition to low-carbon energy systems. However, there is very little causal evidence of the impact of heat pumps on energy demand and the impact of marginal prices to help optimize energy demand with heat pumps. We leverage a staggered roll-out of heat pumps from Octopus Energy Group to show that: (1) heat pumps have a large impact on energy demand, on average causing a 90% reduction in home gas use and a 61% increase in home electricity use – overall, households reduced total energy demand by 40% and carbon dioxide emissions by 36% in 2024 (with an average of 68% emissions savings over the lifetime of the heat pump); (2) a time-of-use tariff designed for heat pumps can provide large demand flexibility benefits, halving electricity consumption during the evening peak to help balance the grid, and that load shifting is possible on the coldest days and from all building types in our sample; (3) the marginal value of public funds of the current UK heat pump subsidy is £1.24 (for every £1 spent by the Government). Overall, we find that heat pumps can meaningfully decarbonize heat and subsidies to encourage heat pumps can be welfare-enhancing.
摘要：热泵已被提出作为家庭供暖电气化的主要技术，因此在向低碳能源系统过渡中可能发挥关键作用。然而，目前关于热泵对能源需求的影响及边际价格如何帮助优化热泵能源需求的因果证据非常少。我们利用 Octopus Energy Group 推出的热泵逐步推广计划，展示了以下几点：（1）热泵对能源需求有重大影响，平均导致家庭天然气使用量减少90%，家庭电力使用量增加61%——总体上，家庭的总能源需求减少了40%，2024年二氧化碳排放减少了36%（热泵整个生命周期内的平均排放节省达到68%）；（2）针对热泵设计的分时电价可以提供很大的需求灵活性好处，使晚高峰时的电力消耗减少一半，帮助电网平衡，并且在我们样本中的所有建筑类型中，即使在最寒冷的日子也可以进行负载转移；（3）当前英国热泵补贴的公共资金边际价值为1.24英镑（即政府每花费1英镑可带来1.24英镑的价值）。总体而言，我们发现热泵可以显著实现供暖脱碳，鼓励热泵的补贴可以提高社会福利。1
理解跨社区代际流动性的异质性
Neil A. Cholli, Steven N. Durlauf, Rasmus Landersø, and Salvador Navarro #33035
Abstract: Recent research has uncovered large spatial heterogeneity in intergenerational mobility across neighborhoods in countries around the world. Yet there is little consensus on the reasons why mobility is high in some neighborhoods and low in others. This paper analyzes a generalized mobility model that examines the roles that families’ selection into neighborhoods and locational characteristics play in generating this spatial heterogeneity. We use administrative data from Denmark to decompose variation in mobility across nearly 300 larger and 2,000 smaller neighborhoods along these dimensions, accounting for sampling error. Families’ selection into neighborhoods and sampling error explain most observed heterogeneity across neighborhoods. Our generalized model explains most of the differences in mobility between neighborhoods, though a small but persistent difference remains between neighborhoods that our model cannot account for. An analysis of this “irreducible heterogeneity” suggests that neighborhoods exhibit multiple types in terms of their mobility effects.
摘要：最近的研究发现，全球各国社区间的代际流动性存在显著的空间异质性。然而，对于为何某些社区的流动性较高而其他社区较低，仍未达成共识。本文分析了一个广义的流动性模型，探讨了家庭选择进入某些社区以及社区位置特征在形成这种空间异质性中的作用。我们使用丹麦的行政数据，分解了近300个大社区和2000个小社区的流动性差异，考虑了抽样误差。家庭选择进入社区和抽样误差解释了大部分社区间观察到的异质性。我们的广义模型解释了大多数社区间流动性差异，尽管仍有一小部分持久的差异是模型无法解释的。对这种“不可简化的异质性”的分析表明，不同社区在流动性效果上表现出多种类型。
向参与者描述延迟接受机制与策略抗性：实验分析
Yannai A. Gonczarowski, Ori Heffetz, Guy Ishai, and Clayton Thomas #33020
Abstract: We conduct an incentivized lab experiment to test participants' ability to understand the DA matching mechanism and the strategyproofness property, conveyed in different ways. We find that while many participants can (using a novel GUI) learn DA's mechanics and calculate its outcomes, such understanding does not imply understanding of strategyproofness (as measured by specially designed tests). However, a novel menu description of strategyproofness conveys this property significantly better than other treatments. While behavioral effects are small on average, participants with levels of strategyproofness understanding above a certain threshold play the classical dominant strategy at very high rates.
摘要：我们进行了一个有激励的实验室实验，以测试参与者对延迟接受（DA）匹配机制及其策略抗性（strategyproofness）特性的理解能力，传递方式有所不同。我们发现，尽管许多参与者能够通过一个新颖的图形用户界面（GUI）学习DA机制的运作并计算其结果，但这种理解并不意味着他们理解策略抗性（通过专门设计的测试衡量）。然而，一种新颖的菜单描述策略抗性的方式比其他方法显著更好地传达了这一特性。尽管行为效应平均来看较小，但当参与者对策略抗性的理解超过某一阈值时，他们以极高的频率采用经典的主导策略。
距离创新的远近是否成为人工智能采用的障碍？
Jennifer Hunt, Iain M. Cockburn, and James Bessen #33022
Abstract: Using our own data on Artificial Intelligence publications merged with Burning Glass vacancy data for 2007-2019, we investigate whether online vacancies for jobs requiring AI skills grow more slowly in U.S. locations farther from pre-2007 AI innovation hotspots. We find that a commuting zone which is an additional 200km (125 miles) from the closest AI hotspot has 17% lower growth in AI jobs' share of vacancies. This is driven by distance from AI papers rather than AI patents. Distance reduces growth in AI research jobs as well as in jobs adapting AI to new industries, as evidenced by strong effects for computer and mathematical researchers, developers of software applications, and the finance and insurance industry. 20% of the effect is explained by the presence of state borders between some commuting zones and their closest hotspot. This could reflect state borders impeding migration and thus flows of tacit knowledge. Distance does not capture difficulty of in-person or remote collaboration nor knowledge and personnel flows within multi-establishment firms hiring in computer occupations.
摘要：我们使用2007-2019年人工智能（AI）出版物数据与Burning Glass的招聘数据进行合并，研究在距离2007年前AI创新热点更远的美国地区，是否要求AI技能的在线职位空缺增长较慢。研究发现，离最近的AI热点每远200公里（约125英里），AI职位空缺的增长率会降低17%。这一现象主要由距离AI论文的影响推动，而非AI专利。距离不仅抑制了AI研究类工作的增长，还影响了将AI应用于新行业的工作，例如计算机和数学研究人员、软件应用程序开发人员，以及金融和保险行业。研究还发现，部分通勤区与其最近的AI热点之间的州界存在时，这种影响约20%可由此解释，可能反映了州界阻碍了移民和隐性知识的流动。距离并未反映亲自或远程合作的难度，也未反映计算机职业多机构公司内部的知识和人员流动问题。
COVID-19疫情期间劳动力市场的跨国分析
Robert Breunig, Wei Cheng, Laura Montenovo, Kyoung Hoon Lee, Bruce A. Weinberg, and Yinjunjie Zhang #33029
Abstract: The authors study employment outcomes during the early stages of the COVID-19 pandemic in eight countries with different case levels and policy responses: the United States, Australia, France, Denmark, Italy, South Korea, Spain, and Sweden. While the share of people not at work increased in all countries, safety net policies seem to influence whether people remained employed (but absent from work) versus unemployed or left the labor force. The authors find large employment decreases among middle-educated and young workers, increasing disparities in countries with the largest labor market declines. A variety of evidence suggests that labor demand was likely a larger driver of employment declines than labor supply and that stringent social distancing policies were sufficient to reduce employment even in the absence of high cases. Lastly, job characteristics - the importance of face-to-face interactions and the ability to work remotely - were closely related to labor market outcomes, with these relationships being stronger in countries with more cases.
摘要：作者研究了 COVID-19 疫情初期八个国家的就业结果，这些国家的疫情严重程度和政策应对有所不同，分别是美国、澳大利亚、法国、丹麦、意大利、韩国、西班牙和瑞典。尽管各国的非工作人口比例都有所增加，但安全网政策似乎影响了人们是继续受雇（但缺勤）还是失业或退出劳动力市场。研究发现，中等教育水平和年轻工人的就业减少幅度最大，且在劳动力市场下降最严重的国家，不平等现象加剧。多种证据表明，劳动力需求可能是就业下降的主要推动因素，而非劳动力供给，严格的社交距离政策即使在疫情病例不高的情况下，也足以导致就业减少。最后，工作特征——面对面互动的重要性以及远程工作的能力——与劳动力市场结果密切相关，这些关系在病例较多的国家中更为显著。1
在实验研究中利用生成式人工智能的12个最佳实践
Samuel Chang, Andrew Kennedy, Aaron Leonard, and John A. List #33025
Abstract: We provide twelve best practices and discuss how each practice can help researchers accurately, credibly, and ethically use Generative AI (GenAI) to enhance experimental research. We split the twelve practices into four areas. First, in the pre-treatment stage, we discuss how GenAI can aid in pre-registration procedures, data privacy concerns, and ethical considerations specific to GenAI usage. Second, in the design and implementation stage, we focus on GenAI’s role in identifying new channels of variation, piloting and documentation, and upholding the four exclusion restrictions. Third, in the analysis stage, we explore how prompting and training set bias can impact results as well as necessary steps to ensure replicability. Finally, we discuss forward-looking best practices that are likely to gain importance as GenAI evolves.
摘要：我们提供了十二条最佳实践，并讨论了每条实践如何帮助研究人员准确、可信且合乎伦理地使用生成式人工智能（GenAI）来提升实验研究。我们将这十二条实践分为四个领域。首先，在前处理阶段，我们讨论了GenAI如何在预注册程序、数据隐私问题以及与GenAI使用相关的伦理考量中提供帮助。其次，在设计和实施阶段，我们重点关注GenAI在识别新变化渠道、试点和文档记录，以及遵守四个排除限制方面的作用。第三，在分析阶段，我们探讨了提示词和训练集偏差如何影响结果，以及确保可重复性的必要步骤。最后，我们讨论了随着GenAI的发展，可能会变得更加重要的前瞻性最佳实践。
“购买美国货”的影响：电动汽车与《通胀削减法案》
Hunt Allcott, Reigner Kane, Maximilian S. Maydanchik, Joseph S. Shapiro, and Felix Tintelnot #33032
Abstract: We study electric vehicle (EV) tax credits in the US Inflation Reduction Act (IRA), the largest climate policy in US history, with three goals. First, we provide the first ex-post microeconomic welfare analysis of this central component of the IRA. Event studies around changes in eligibility for EV tax credits find that short-run economic incidence falls largely on consumers. Additionally, domestic content restrictions on tax credits for purchased vehicles have driven enormous shifts to leasing. Our equilibrium model shows that compared to pre-IRA policy, IRA EV credits generated $1.87 of US benefits per dollar spent in 2023, at taxpayer cost of $32,000 per additional EV sold. Compared to scenarios with no EV credits, however, the IRA EV credits created only $1.02 of benefits per dollar of government spending. Second, we characterize the gains from policies targeting heterogeneity in externalities across vehicles. We find that relative to uniform credits, differentiating credits across EVs according to their heterogeneous externalities would substantially increase policy benefits. Third, we quantify tradeoffs in the IRA EV credits between foreign and domestic welfare and between trade and the environment. We find that the IRA EV credits benefit the environment but undermine trade, since they decrease global carbon emissions but use profit shifting to decrease foreign producer surplus. A controversial IRA loophole that removes domestic content restrictions on tax credits for EV leases has negative domestic benefits.
摘要：我们研究了美国《通胀削减法案》（IRA）中电动汽车（EV）税收抵免的影响，这是美国历史上最大规模的气候政策，目标包括以下三方面。首先，我们提供了该法案核心部分的首次事后微观经济福利分析。通过对电动汽车税收抵免资格变化的事件研究发现，短期经济负担主要落在消费者身上。此外，针对购买车辆的税收抵免中的国内内容限制促使大量消费者转向租赁电动汽车。我们的均衡模型显示，与IRA出台前的政策相比，IRA的电动汽车税收抵免在2023年每花费1美元能为美国带来1.87美元的收益，而每售出一辆额外的电动汽车纳税人需承担32,000美元的成本。然而，与没有电动汽车税收抵免的情景相比，IRA的电动汽车税收抵免每花费1美元的政府支出仅创造了1.02美元的收益。其次，我们描述了针对不同车辆外部性差异的政策所带来的收益。我们发现，相较于统一的税收抵免，基于电动汽车的外部性差异来设定抵免额度能显著提升政策效益。最后，我们量化了IRA电动汽车税收抵免在国内外福利之间以及在贸易和环境之间的权衡。我们发现，IRA电动汽车税收抵免有利于环境，但对贸易不利，因为虽然它减少了全球碳排放，但通过利润转移降低了外国生产者的盈余。此外，IRA中有争议的漏洞，即取消了电动汽车租赁税收抵免的国内内容限制，对国内福利产生了负面影响。
往期精选：
