首页 > 最新文献

arXiv - ECON - Econometrics最新文献

英文 中文
The Application of Green GDP and Its Impact on Global Economy and Environment: Analysis of GGDP based on SEEA model 绿色 GDP 的应用及其对全球经济和环境的影响:基于 SEEA 模型的 GGDP 分析
Pub Date : 2024-09-04 DOI: arxiv-2409.02642
Mingpu Ma
This paper presents an analysis of Green Gross Domestic Product (GGDP) usingthe System of Environmental-Economic Accounting (SEEA) model to evaluate itsimpact on global climate mitigation and economic health. GGDP is proposed as asuperior measure to tradi-tional GDP by incorporating natural resourceconsumption, environmental pollution control, and degradation factors. Thestudy develops a GGDP model and employs grey correlation analysis and greyprediction models to assess its relationship with these factors. Key findingsdemonstrate that replacing GDP with GGDP can positively influence climatechange, partic-ularly in reducing CO2 emissions and stabilizing globaltemperatures. The analysis further explores the implications of GGDP adoptionacross developed and developing countries, with specific predictions for Chinaand the United States. The results indicate a potential increase in economiclevels for developing countries, while developed nations may experi-ence adecrease. Additionally, the shift to GGDP is shown to significantly reducenatural re-source depletion and population growth rates in the United States,suggesting broader envi-ronmental and economic benefits. This paper highlightsthe universal applicability of the GGDP model and its potential to enhanceenvironmental and economic policies globally.
本文利用环境经济核算体系(SEEA)模型对绿色国内生产总值(GGDP)进行了分析,以评估其对全球气候减缓和经济健康的影响。通过纳入自然资源消耗、环境污染控制和退化因素,GGDP 被提出作为传统 GDP 的更优衡量标准。研究建立了 GGDP 模型,并采用灰色关联分析和灰色预测模型来评估其与这些因素的关系。主要研究结果表明,用 GGDP 取代 GDP 可以对气候变化产生积极影响,特别是在减少二氧化碳排放和稳定全球温度方面。分析进一步探讨了发达国家和发展中国家采用 GGDP 的影响,并对中国和美国进行了具体预测。结果表明,发展中国家的经济水平可能会提高,而发达国家的经济水平可能会降低。此外,在美国,向 GGDP 的转变可显著减少自然资源损耗和人口增长率,从而带来更广泛的环境和经济效益。本文强调了 GGDP 模型的普遍适用性及其在加强全球环境和经济政策方面的潜力。
{"title":"The Application of Green GDP and Its Impact on Global Economy and Environment: Analysis of GGDP based on SEEA model","authors":"Mingpu Ma","doi":"arxiv-2409.02642","DOIUrl":"https://doi.org/arxiv-2409.02642","url":null,"abstract":"This paper presents an analysis of Green Gross Domestic Product (GGDP) using\u0000the System of Environmental-Economic Accounting (SEEA) model to evaluate its\u0000impact on global climate mitigation and economic health. GGDP is proposed as a\u0000superior measure to tradi-tional GDP by incorporating natural resource\u0000consumption, environmental pollution control, and degradation factors. The\u0000study develops a GGDP model and employs grey correlation analysis and grey\u0000prediction models to assess its relationship with these factors. Key findings\u0000demonstrate that replacing GDP with GGDP can positively influence climate\u0000change, partic-ularly in reducing CO2 emissions and stabilizing global\u0000temperatures. The analysis further explores the implications of GGDP adoption\u0000across developed and developing countries, with specific predictions for China\u0000and the United States. The results indicate a potential increase in economic\u0000levels for developing countries, while developed nations may experi-ence a\u0000decrease. Additionally, the shift to GGDP is shown to significantly reduce\u0000natural re-source depletion and population growth rates in the United States,\u0000suggesting broader envi-ronmental and economic benefits. This paper highlights\u0000the universal applicability of the GGDP model and its potential to enhance\u0000environmental and economic policies globally.","PeriodicalId":501293,"journal":{"name":"arXiv - ECON - Econometrics","volume":"5 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142184139","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Distribution Regression Difference-In-Differences 分布 回归 差异
Pub Date : 2024-09-03 DOI: arxiv-2409.02311
Iván Fernández-Val, Jonas Meier, Aico van Vuuren, Francis Vella
We provide a simple distribution regression estimator for treatment effectsin the difference-in-differences (DiD) design. Our procedure is particularlyuseful when the treatment effect differs across the distribution of the outcomevariable. Our proposed estimator easily incorporates covariates and,importantly, can be extended to settings where the treatment potentiallyaffects the joint distribution of multiple outcomes. Our key identifyingrestriction is that the counterfactual distribution of the treated in theuntreated state has no interaction effect between treatment and time. Thisassumption results in a parallel trend assumption on a transformation of thedistribution. We highlight the relationship between our procedure andassumptions with the changes-in-changes approach of Athey and Imbens (2006). Wealso reexamine two existing empirical examples which highlight the utility ofour approach.
我们为差分(DiD)设计中的治疗效果提供了一种简单的分布回归估计方法。当治疗效果随结果变量的分布而变化时,我们的方法尤其有用。我们提出的估计方法很容易纳入协变量,而且重要的是,可以扩展到治疗可能影响多个结果联合分布的情况。我们的关键识别限制条件是,未治疗状态下治疗者的反事实分布在治疗和时间之间没有交互影响。这一假设导致了对分布变换的平行趋势假设。我们强调了我们的程序和假设与 Athey 和 Imbens(2006 年)的 "变化中的变化 "方法之间的关系。我们还重新审查了两个现有的实证例子,它们突出了我们方法的实用性。
{"title":"Distribution Regression Difference-In-Differences","authors":"Iván Fernández-Val, Jonas Meier, Aico van Vuuren, Francis Vella","doi":"arxiv-2409.02311","DOIUrl":"https://doi.org/arxiv-2409.02311","url":null,"abstract":"We provide a simple distribution regression estimator for treatment effects\u0000in the difference-in-differences (DiD) design. Our procedure is particularly\u0000useful when the treatment effect differs across the distribution of the outcome\u0000variable. Our proposed estimator easily incorporates covariates and,\u0000importantly, can be extended to settings where the treatment potentially\u0000affects the joint distribution of multiple outcomes. Our key identifying\u0000restriction is that the counterfactual distribution of the treated in the\u0000untreated state has no interaction effect between treatment and time. This\u0000assumption results in a parallel trend assumption on a transformation of the\u0000distribution. We highlight the relationship between our procedure and\u0000assumptions with the changes-in-changes approach of Athey and Imbens (2006). We\u0000also reexamine two existing empirical examples which highlight the utility of\u0000our approach.","PeriodicalId":501293,"journal":{"name":"arXiv - ECON - Econometrics","volume":"11 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142184140","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Variable selection in convex nonparametric least squares via structured Lasso: An application to the Swedish electricity market 通过结构化 Lasso 在凸非参数最小二乘法中选择变量:瑞典电力市场的应用
Pub Date : 2024-09-03 DOI: arxiv-2409.01911
Zhiqiang Liao
We study the problem of variable selection in convex nonparametric leastsquares (CNLS). Whereas the least absolute shrinkage and selection operator(Lasso) is a popular technique for least squares, its variable selectionperformance is unknown in CNLS problems. In this work, we investigate theperformance of the Lasso CNLS estimator and find out it is usually unable toselect variables efficiently. Exploiting the unique structure of thesubgradients in CNLS, we develop a structured Lasso by combining $ell_1$-normand $ell_{infty}$-norm. To improve its predictive performance, we propose arelaxed version of the structured Lasso where we can control the twoeffects--variable selection and model shrinkage--using an additional tuningparameter. A Monte Carlo study is implemented to verify the finite sampleperformances of the proposed approaches. In the application of Swedishelectricity distribution networks, when the regression model is assumed to besemi-nonparametric, our methods are extended to the doubly penalized CNLSestimators. The results from the simulation and application confirm that theproposed structured Lasso performs favorably, generally leading to sparser andmore accurate predictive models, relative to the other variable selectionmethods in the literature.
我们研究了凸非参数最小二乘法(CNLS)中的变量选择问题。虽然最小绝对收缩和选择算子(Lasso)是一种常用的最小二乘法技术,但它在 CNLS 问题中的变量选择性能尚不清楚。在这项工作中,我们研究了 Lasso CNLS 估计器的性能,发现它通常无法有效地选择变量。利用 CNLS 中子梯度的独特结构,我们结合 $ell_1$-norm 和 $ell_{infty}$-norm 开发了一种结构化 Lasso。为了提高结构化拉索的预测性能,我们提出了结构化拉索的松弛版本,在这个版本中,我们可以通过额外的调整参数来控制变量选择和模型收缩这两种效应。通过蒙特卡罗研究验证了所提方法的有限样本性能。在瑞典配电网络的应用中,当假设回归模型为半非参数时,我们的方法扩展到了双重惩罚 CNLS 估计器。仿真和应用结果证实,与文献中的其他变量选择方法相比,所提出的结构化 Lasso 方法性能良好,通常能得到更稀疏、更准确的预测模型。
{"title":"Variable selection in convex nonparametric least squares via structured Lasso: An application to the Swedish electricity market","authors":"Zhiqiang Liao","doi":"arxiv-2409.01911","DOIUrl":"https://doi.org/arxiv-2409.01911","url":null,"abstract":"We study the problem of variable selection in convex nonparametric least\u0000squares (CNLS). Whereas the least absolute shrinkage and selection operator\u0000(Lasso) is a popular technique for least squares, its variable selection\u0000performance is unknown in CNLS problems. In this work, we investigate the\u0000performance of the Lasso CNLS estimator and find out it is usually unable to\u0000select variables efficiently. Exploiting the unique structure of the\u0000subgradients in CNLS, we develop a structured Lasso by combining $ell_1$-norm\u0000and $ell_{infty}$-norm. To improve its predictive performance, we propose a\u0000relaxed version of the structured Lasso where we can control the two\u0000effects--variable selection and model shrinkage--using an additional tuning\u0000parameter. A Monte Carlo study is implemented to verify the finite sample\u0000performances of the proposed approaches. In the application of Swedish\u0000electricity distribution networks, when the regression model is assumed to be\u0000semi-nonparametric, our methods are extended to the doubly penalized CNLS\u0000estimators. The results from the simulation and application confirm that the\u0000proposed structured Lasso performs favorably, generally leading to sparser and\u0000more accurate predictive models, relative to the other variable selection\u0000methods in the literature.","PeriodicalId":501293,"journal":{"name":"arXiv - ECON - Econometrics","volume":"141 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142184145","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Double Machine Learning at Scale to Predict Causal Impact of Customer Actions 规模化双重机器学习预测客户行为的因果影响
Pub Date : 2024-09-03 DOI: arxiv-2409.02332
Sushant More, Priya Kotwal, Sujith Chappidi, Dinesh Mandalapu, Chris Khawand
Causal Impact (CI) of customer actions are broadly used across the industryto inform both short- and long-term investment decisions of various types. Inthis paper, we apply the double machine learning (DML) methodology to estimatethe CI values across 100s of customer actions of business interest and 100s ofmillions of customers. We operationalize DML through a causal ML library basedon Spark with a flexible, JSON-driven model configuration approach to estimateCI at scale (i.e., across hundred of actions and millions of customers). Weoutline the DML methodology and implementation, and associated benefits overthe traditional potential outcomes based CI model. We show population-level aswell as customer-level CI values along with confidence intervals. Thevalidation metrics show a 2.2% gain over the baseline methods and a 2.5X gainin the computational time. Our contribution is to advance the scalableapplication of CI, while also providing an interface that allows fasterexperimentation, cross-platform support, ability to onboard new use cases, andimproves accessibility of underlying code for partner teams.
客户行为的因果影响(CI)被广泛应用于整个行业,为各种类型的短期和长期投资决策提供依据。在本文中,我们应用双重机器学习(DML)方法来估算企业感兴趣的数百种客户行为和数亿客户的 CI 值。我们通过基于 Spark 的因果 ML 库和灵活的 JSON 驱动型模型配置方法对 DML 进行操作,以大规模(即跨越数百个行为和数百万客户)估算 CI。我们概述了 DML 方法和实施,以及与传统的基于潜在结果的 CI 模型相比的相关优势。我们展示了人口级和客户级 CI 值以及置信区间。验证指标显示,与基线方法相比,DML 的收益为 2.2%,计算时间增加了 2.5 倍。我们的贡献在于推进了 CI 的可扩展应用,同时还提供了一个接口,允许快速实验、跨平台支持、加入新用例的能力,并提高了合作伙伴团队对底层代码的可访问性。
{"title":"Double Machine Learning at Scale to Predict Causal Impact of Customer Actions","authors":"Sushant More, Priya Kotwal, Sujith Chappidi, Dinesh Mandalapu, Chris Khawand","doi":"arxiv-2409.02332","DOIUrl":"https://doi.org/arxiv-2409.02332","url":null,"abstract":"Causal Impact (CI) of customer actions are broadly used across the industry\u0000to inform both short- and long-term investment decisions of various types. In\u0000this paper, we apply the double machine learning (DML) methodology to estimate\u0000the CI values across 100s of customer actions of business interest and 100s of\u0000millions of customers. We operationalize DML through a causal ML library based\u0000on Spark with a flexible, JSON-driven model configuration approach to estimate\u0000CI at scale (i.e., across hundred of actions and millions of customers). We\u0000outline the DML methodology and implementation, and associated benefits over\u0000the traditional potential outcomes based CI model. We show population-level as\u0000well as customer-level CI values along with confidence intervals. The\u0000validation metrics show a 2.2% gain over the baseline methods and a 2.5X gain\u0000in the computational time. Our contribution is to advance the scalable\u0000application of CI, while also providing an interface that allows faster\u0000experimentation, cross-platform support, ability to onboard new use cases, and\u0000improves accessibility of underlying code for partner teams.","PeriodicalId":501293,"journal":{"name":"arXiv - ECON - Econometrics","volume":"26 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142184142","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Double Machine Learning meets Panel Data -- Promises, Pitfalls, and Potential Solutions 双机器学习与面板数据 -- 前景、陷阱和潜在解决方案
Pub Date : 2024-09-02 DOI: arxiv-2409.01266
Jonathan Fuhr, Dominik Papies
Estimating causal effect using machine learning (ML) algorithms can help torelax functional form assumptions if used within appropriate frameworks.However, most of these frameworks assume settings with cross-sectional data,whereas researchers often have access to panel data, which in traditionalmethods helps to deal with unobserved heterogeneity between units. In thispaper, we explore how we can adapt double/debiased machine learning (DML)(Chernozhukov et al., 2018) for panel data in the presence of unobservedheterogeneity. This adaptation is challenging because DML's cross-fittingprocedure assumes independent data and the unobserved heterogeneity is notnecessarily additively separable in settings with nonlinear observedconfounding. We assess the performance of several intuitively appealingestimators in a variety of simulations. While we find violations of thecross-fitting assumptions to be largely inconsequential for the accuracy of theeffect estimates, many of the considered methods fail to adequately account forthe presence of unobserved heterogeneity. However, we find that usingpredictive models based on the correlated random effects approach (Mundlak,1978) within DML leads to accurate coefficient estimates across settings, givena sample size that is large relative to the number of observed confounders. Wealso show that the influence of the unobserved heterogeneity on the observedconfounders plays a significant role for the performance of most alternativemethods.
使用机器学习(ML)算法估计因果效应,如果在适当的框架内使用,可以帮助放松函数形式假设。然而,这些框架大多假设有横截面数据,而研究人员通常可以获得面板数据,在传统方法中,面板数据有助于处理单位间的未观察异质性。在本文中,我们将探讨如何在存在未观察异质性的情况下,针对面板数据调整双重/偏差机器学习(DML)(Chernozhukov 等人,2018 年)。这种调整具有挑战性,因为 DML 的交叉拟合过程假定数据是独立的,而在非线性观察混杂的情况下,未观察到的异质性并不一定是可加可分的。我们在各种模拟中评估了几种直观吸引人的估计方法的性能。虽然我们发现违反交叉拟合假设对效应估计的准确性基本没有影响,但许多考虑过的方法未能充分考虑到非观测异质性的存在。然而,我们发现,在 DML 中使用基于相关随机效应方法(Mundlak,1978 年)的预测模型,在样本量相对于观察到的混杂因素数量较大的情况下,可以在各种情况下获得准确的系数估计。我们还表明,未观察到的异质性对观察到的混杂因素的影响对大多数替代方法的性能起着重要作用。
{"title":"Double Machine Learning meets Panel Data -- Promises, Pitfalls, and Potential Solutions","authors":"Jonathan Fuhr, Dominik Papies","doi":"arxiv-2409.01266","DOIUrl":"https://doi.org/arxiv-2409.01266","url":null,"abstract":"Estimating causal effect using machine learning (ML) algorithms can help to\u0000relax functional form assumptions if used within appropriate frameworks.\u0000However, most of these frameworks assume settings with cross-sectional data,\u0000whereas researchers often have access to panel data, which in traditional\u0000methods helps to deal with unobserved heterogeneity between units. In this\u0000paper, we explore how we can adapt double/debiased machine learning (DML)\u0000(Chernozhukov et al., 2018) for panel data in the presence of unobserved\u0000heterogeneity. This adaptation is challenging because DML's cross-fitting\u0000procedure assumes independent data and the unobserved heterogeneity is not\u0000necessarily additively separable in settings with nonlinear observed\u0000confounding. We assess the performance of several intuitively appealing\u0000estimators in a variety of simulations. While we find violations of the\u0000cross-fitting assumptions to be largely inconsequential for the accuracy of the\u0000effect estimates, many of the considered methods fail to adequately account for\u0000the presence of unobserved heterogeneity. However, we find that using\u0000predictive models based on the correlated random effects approach (Mundlak,\u00001978) within DML leads to accurate coefficient estimates across settings, given\u0000a sample size that is large relative to the number of observed confounders. We\u0000also show that the influence of the unobserved heterogeneity on the observed\u0000confounders plays a significant role for the performance of most alternative\u0000methods.","PeriodicalId":501293,"journal":{"name":"arXiv - ECON - Econometrics","volume":"1583 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-09-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142184144","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Bandit Algorithms for Policy Learning: Methods, Implementation, and Welfare-performance 政策学习的 Bandit 算法:方法、实施和福利绩效
Pub Date : 2024-08-31 DOI: arxiv-2409.00379
Toru Kitagawa, Jeff Rowley
Static supervised learning-in which experimental data serves as a trainingsample for the estimation of an optimal treatment assignment policy-is acommonly assumed framework of policy learning. An arguably more realistic butchallenging scenario is a dynamic setting in which the planner performsexperimentation and exploitation simultaneously with subjects that arrivesequentially. This paper studies bandit algorithms for learning an optimalindividualised treatment assignment policy. Specifically, we studyapplicability of the EXP4.P (Exponential weighting for Exploration andExploitation with Experts) algorithm developed by Beygelzimer et al. (2011) topolicy learning. Assuming that the class of policies has a finiteVapnik-Chervonenkis dimension and that the number of subjects to be allocatedis known, we present a high probability welfare-regret bound of the algorithm.To implement the algorithm, we use an incremental enumeration algorithm forhyperplane arrangements. We perform extensive numerical analysis to assess thealgorithm's sensitivity to its tuning parameters and its welfare-regretperformance. Further simulation exercises are calibrated to the National JobTraining Partnership Act (JTPA) Study sample to determine how the algorithmperforms when applied to economic data. Our findings highlight variouscomputational challenges and suggest that the limited welfare gain from thealgorithm is due to substantial heterogeneity in causal effects in the JTPAdata.
静态监督学习--即实验数据作为估计最优治疗分配政策的训练样本--是一种常见的政策学习假设框架。一个可以说更现实但更具挑战性的场景是动态环境,在这种环境中,规划者同时对按顺序到达的受试者进行实验和开发。本文研究了学习最优个体化治疗分配政策的匪徒算法。具体来说,我们研究了 Beygelzimer 等人(2011 年)开发的 EXP4.P(Exponential weighting for Exploration andExploitation with Experts)算法在政策学习中的适用性。假定政策类别具有有限的 Vapnik-Chervonenkis 维度,且待分配的研究对象数量已知,我们提出了该算法的高概率福利-遗憾约束。为了实现该算法,我们使用了一种针对超平面安排的增量枚举算法。我们进行了大量的数值分析,以评估该算法对其调整参数的敏感性及其福利-遗憾表现。此外,我们还根据《国家就业培训合作法案》(JTPA)研究样本进行了进一步的模拟练习,以确定该算法在应用于经济数据时的表现。我们的研究结果凸显了各种计算挑战,并表明该算法的福利收益有限是由于 JTPA 数据中因果效应的巨大异质性造成的。
{"title":"Bandit Algorithms for Policy Learning: Methods, Implementation, and Welfare-performance","authors":"Toru Kitagawa, Jeff Rowley","doi":"arxiv-2409.00379","DOIUrl":"https://doi.org/arxiv-2409.00379","url":null,"abstract":"Static supervised learning-in which experimental data serves as a training\u0000sample for the estimation of an optimal treatment assignment policy-is a\u0000commonly assumed framework of policy learning. An arguably more realistic but\u0000challenging scenario is a dynamic setting in which the planner performs\u0000experimentation and exploitation simultaneously with subjects that arrive\u0000sequentially. This paper studies bandit algorithms for learning an optimal\u0000individualised treatment assignment policy. Specifically, we study\u0000applicability of the EXP4.P (Exponential weighting for Exploration and\u0000Exploitation with Experts) algorithm developed by Beygelzimer et al. (2011) to\u0000policy learning. Assuming that the class of policies has a finite\u0000Vapnik-Chervonenkis dimension and that the number of subjects to be allocated\u0000is known, we present a high probability welfare-regret bound of the algorithm.\u0000To implement the algorithm, we use an incremental enumeration algorithm for\u0000hyperplane arrangements. We perform extensive numerical analysis to assess the\u0000algorithm's sensitivity to its tuning parameters and its welfare-regret\u0000performance. Further simulation exercises are calibrated to the National Job\u0000Training Partnership Act (JTPA) Study sample to determine how the algorithm\u0000performs when applied to economic data. Our findings highlight various\u0000computational challenges and suggest that the limited welfare gain from the\u0000algorithm is due to substantial heterogeneity in causal effects in the JTPA\u0000data.","PeriodicalId":501293,"journal":{"name":"arXiv - ECON - Econometrics","volume":"22 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142184143","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Weighted Regression with Sybil Networks 利用 Sybil 网络进行加权回归
Pub Date : 2024-08-30 DOI: arxiv-2408.17426
Nihar Shah
In many online domains, Sybil networks -- or cases where a single userassumes multiple identities -- is a pervasive feature. This complicatesexperiments, as off-the-shelf regression estimators at least assume knownnetwork topologies (if not fully independent observations) when Sybil networktopologies in practice are often unknown. The literature has exclusivelyfocused on techniques to detect Sybil networks, leading many experimenters tosubsequently exclude suspected networks entirely before estimating treatmenteffects. I present a more efficient solution in the presence of these suspectedSybil networks: a weighted regression framework that applies weights based onthe probabilities that sets of observations are controlled by single actors. Ishow in the paper that the MSE-minimizing solution is to set the weight matrixequal to the inverse of the expected network topology. I demonstrate themethodology on simulated data, and then I apply the technique to a competitionwith suspected Sybil networks run on the Sui blockchain and show reductions inthe standard error of the estimate by 6 - 24%.
在许多在线领域,Sybil 网络(或单个用户假冒多个身份的情况)是一个普遍存在的特征。这使得实验变得复杂,因为现成的回归估计器至少假定网络拓扑结构是已知的(如果不是完全独立的观察结果),而实际上假冒网络拓扑结构往往是未知的。文献只关注检测假网络的技术,导致许多实验者在估计治疗效果之前不得不完全排除可疑网络。我提出了一种更有效的解决方案:加权回归框架,根据观测数据集受单个行为者控制的概率进行加权。我在文中指出,MSE 最小化的解决方案是将权重矩阵设置为预期网络拓扑结构的倒数。我在模拟数据上演示了这一方法,然后将该技术应用于在 Sui 区块链上运行的疑似 Sybil 网络竞赛,结果显示估计值的标准误差减少了 6 - 24%。
{"title":"Weighted Regression with Sybil Networks","authors":"Nihar Shah","doi":"arxiv-2408.17426","DOIUrl":"https://doi.org/arxiv-2408.17426","url":null,"abstract":"In many online domains, Sybil networks -- or cases where a single user\u0000assumes multiple identities -- is a pervasive feature. This complicates\u0000experiments, as off-the-shelf regression estimators at least assume known\u0000network topologies (if not fully independent observations) when Sybil network\u0000topologies in practice are often unknown. The literature has exclusively\u0000focused on techniques to detect Sybil networks, leading many experimenters to\u0000subsequently exclude suspected networks entirely before estimating treatment\u0000effects. I present a more efficient solution in the presence of these suspected\u0000Sybil networks: a weighted regression framework that applies weights based on\u0000the probabilities that sets of observations are controlled by single actors. I\u0000show in the paper that the MSE-minimizing solution is to set the weight matrix\u0000equal to the inverse of the expected network topology. I demonstrate the\u0000methodology on simulated data, and then I apply the technique to a competition\u0000with suspected Sybil networks run on the Sui blockchain and show reductions in\u0000the standard error of the estimate by 6 - 24%.","PeriodicalId":501293,"journal":{"name":"arXiv - ECON - Econometrics","volume":"24 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142184147","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
State Space Model of Realized Volatility under the Existence of Dependent Market Microstructure Noise 存在依赖性市场微观结构噪声时的已实现波动率状态空间模型
Pub Date : 2024-08-30 DOI: arxiv-2408.17187
Toru Yano
Volatility means the degree of variation of a stock price which is importantin finance. Realized Volatility (RV) is an estimator of the volatilitycalculated using high-frequency observed prices. RV has lately attractedconsiderable attention of econometrics and mathematical finance. However, it isknown that high-frequency data includes observation errors called marketmicrostructure noise (MN). Nagakura and Watanabe[2015] proposed a state spacemodel that resolves RV into true volatility and influence of MN. In this paper,we assume a dependent MN that autocorrelates and correlates with return asreported by Hansen and Lunde[2006] and extends the results of Nagakura andWatanabe[2015] and compare models by simulation and actual data.
波动率是指股票价格的变化程度,在金融领域非常重要。实现波动率(RV)是利用高频观测价格计算出的波动率估计值。最近,RV 引起了计量经济学和数理金融学的极大关注。然而,众所周知,高频数据包含被称为市场微观结构噪声(MN)的观测误差。Nagakura 和 Watanabe[2015]提出了一种状态空间模型,将 RV 分解为真实波动率和 MN 的影响。本文假定 MN 与 Hansen 和 Lunde[2006]报告的收益率自相关和相关,并扩展了 Nagakura 和 Watanabe[2015]的结果,通过模拟和实际数据对模型进行比较。
{"title":"State Space Model of Realized Volatility under the Existence of Dependent Market Microstructure Noise","authors":"Toru Yano","doi":"arxiv-2408.17187","DOIUrl":"https://doi.org/arxiv-2408.17187","url":null,"abstract":"Volatility means the degree of variation of a stock price which is important\u0000in finance. Realized Volatility (RV) is an estimator of the volatility\u0000calculated using high-frequency observed prices. RV has lately attracted\u0000considerable attention of econometrics and mathematical finance. However, it is\u0000known that high-frequency data includes observation errors called market\u0000microstructure noise (MN). Nagakura and Watanabe[2015] proposed a state space\u0000model that resolves RV into true volatility and influence of MN. In this paper,\u0000we assume a dependent MN that autocorrelates and correlates with return as\u0000reported by Hansen and Lunde[2006] and extends the results of Nagakura and\u0000Watanabe[2015] and compare models by simulation and actual data.","PeriodicalId":501293,"journal":{"name":"arXiv - ECON - Econometrics","volume":"17 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142184146","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Sensitivity Analysis for Dynamic Discrete Choice Models 动态离散选择模型的敏感性分析
Pub Date : 2024-08-29 DOI: arxiv-2408.16330
Chun Pong Lau
In dynamic discrete choice models, some parameters, such as the discountfactor, are being fixed instead of being estimated. This paper proposes twosensitivity analysis procedures for dynamic discrete choice models with respectto the fixed parameters. First, I develop a local sensitivity measure thatestimates the change in the target parameter for a unit change in the fixedparameter. This measure is fast to compute as it does not require modelre-estimation. Second, I propose a global sensitivity analysis procedure thatuses model primitives to study the relationship between target parameters andfixed parameters. I show how to apply the sensitivity analysis procedures ofthis paper through two empirical applications.
在动态离散选择模型中,一些参数(如贴现因子)是固定的,而不是估计的。本文提出了动态离散选择模型对固定参数的两种灵敏度分析程序。首先,我开发了一种局部灵敏度度量,用于估计固定参数发生单位变化时目标参数的变化。由于不需要对模型进行重新估计,因此计算速度很快。其次,我提出了一种全局灵敏度分析程序,利用模型基元来研究目标参数和固定参数之间的关系。我通过两个经验应用来展示如何应用本文的灵敏度分析程序。
{"title":"Sensitivity Analysis for Dynamic Discrete Choice Models","authors":"Chun Pong Lau","doi":"arxiv-2408.16330","DOIUrl":"https://doi.org/arxiv-2408.16330","url":null,"abstract":"In dynamic discrete choice models, some parameters, such as the discount\u0000factor, are being fixed instead of being estimated. This paper proposes two\u0000sensitivity analysis procedures for dynamic discrete choice models with respect\u0000to the fixed parameters. First, I develop a local sensitivity measure that\u0000estimates the change in the target parameter for a unit change in the fixed\u0000parameter. This measure is fast to compute as it does not require model\u0000re-estimation. Second, I propose a global sensitivity analysis procedure that\u0000uses model primitives to study the relationship between target parameters and\u0000fixed parameters. I show how to apply the sensitivity analysis procedures of\u0000this paper through two empirical applications.","PeriodicalId":501293,"journal":{"name":"arXiv - ECON - Econometrics","volume":"34 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142184148","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Marginal homogeneity tests with panel data 面板数据的边际同质性检验
Pub Date : 2024-08-28 DOI: arxiv-2408.15862
Federico Bugni, Jackson Bunting, Muyang Ren
A panel dataset satisfies marginal homogeneity if the time-specific marginaldistributions are homogeneous or time-invariant. Marginal homogeneity isrelevant in economic settings such as dynamic discrete games. In this paper, wepropose several tests for the hypothesis of marginal homogeneity andinvestigate their properties. We consider an asymptotic framework in which thenumber of individuals n in the panel diverges, and the number of periods T isfixed. We implement our tests by comparing a studentized or non-studentizedT-sample version of the Cramer-von Mises statistic with a suitable criticalvalue. We propose three methods to construct the critical value: asymptoticapproximations, the bootstrap, and time permutations. We show that the firsttwo methods result in asymptotically exact hypothesis tests. The permutationtest based on a non-studentized statistic is asymptotically exact when T=2, butis asymptotically invalid when T>2. In contrast, the permutation test based ona studentized statistic is always asymptotically exact. Finally, under atime-exchangeability assumption, the permutation test is exact in finitesamples, both with and without studentization.
如果特定时间的边际分布是同质或时间不变的,那么面板数据集就满足边际同质性。边际同质性与动态离散博弈等经济环境相关。本文提出了边际同质性假设的几种检验方法,并对其性质进行了研究。我们考虑了一个渐进框架,在这个框架中,面板中的个体数 n 是发散的,而周期数 T 是固定的。我们通过比较 Cramer-von Mises 统计量的学生化或非学生化 T 样本版本与合适的临界值来进行检验。我们提出了三种构建临界值的方法:渐近逼近法、自引导法和时间排列法。我们证明,前两种方法可以得出渐近精确的假设检验。当 T=2 时,基于非研究统计量的置换检验在渐近上是精确的,但当 T>2 时,置换检验在渐近上是无效的。最后,在时间可交换性假设下,无论有无学生化,置换检验在有限样本中都是精确的。
{"title":"Marginal homogeneity tests with panel data","authors":"Federico Bugni, Jackson Bunting, Muyang Ren","doi":"arxiv-2408.15862","DOIUrl":"https://doi.org/arxiv-2408.15862","url":null,"abstract":"A panel dataset satisfies marginal homogeneity if the time-specific marginal\u0000distributions are homogeneous or time-invariant. Marginal homogeneity is\u0000relevant in economic settings such as dynamic discrete games. In this paper, we\u0000propose several tests for the hypothesis of marginal homogeneity and\u0000investigate their properties. We consider an asymptotic framework in which the\u0000number of individuals n in the panel diverges, and the number of periods T is\u0000fixed. We implement our tests by comparing a studentized or non-studentized\u0000T-sample version of the Cramer-von Mises statistic with a suitable critical\u0000value. We propose three methods to construct the critical value: asymptotic\u0000approximations, the bootstrap, and time permutations. We show that the first\u0000two methods result in asymptotically exact hypothesis tests. The permutation\u0000test based on a non-studentized statistic is asymptotically exact when T=2, but\u0000is asymptotically invalid when T>2. In contrast, the permutation test based on\u0000a studentized statistic is always asymptotically exact. Finally, under a\u0000time-exchangeability assumption, the permutation test is exact in finite\u0000samples, both with and without studentization.","PeriodicalId":501293,"journal":{"name":"arXiv - ECON - Econometrics","volume":"11 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-08-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142184149","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
期刊
arXiv - ECON - Econometrics
全部 Acc. Chem. Res. ACS Applied Bio Materials ACS Appl. Electron. Mater. ACS Appl. Energy Mater. ACS Appl. Mater. Interfaces ACS Appl. Nano Mater. ACS Appl. Polym. Mater. ACS BIOMATER-SCI ENG ACS Catal. ACS Cent. Sci. ACS Chem. Biol. ACS Chemical Health & Safety ACS Chem. Neurosci. ACS Comb. Sci. ACS Earth Space Chem. ACS Energy Lett. ACS Infect. Dis. ACS Macro Lett. ACS Mater. Lett. ACS Med. Chem. Lett. ACS Nano ACS Omega ACS Photonics ACS Sens. ACS Sustainable Chem. Eng. ACS Synth. Biol. Anal. Chem. BIOCHEMISTRY-US Bioconjugate Chem. BIOMACROMOLECULES Chem. Res. Toxicol. Chem. Rev. Chem. Mater. CRYST GROWTH DES ENERG FUEL Environ. Sci. Technol. Environ. Sci. Technol. Lett. Eur. J. Inorg. Chem. IND ENG CHEM RES Inorg. Chem. J. Agric. Food. Chem. J. Chem. Eng. Data J. Chem. Educ. J. Chem. Inf. Model. J. Chem. Theory Comput. J. Med. Chem. J. Nat. Prod. J PROTEOME RES J. Am. Chem. Soc. LANGMUIR MACROMOLECULES Mol. Pharmaceutics Nano Lett. Org. Lett. ORG PROCESS RES DEV ORGANOMETALLICS J. Org. Chem. J. Phys. Chem. J. Phys. Chem. A J. Phys. Chem. B J. Phys. Chem. C J. Phys. Chem. Lett. Analyst Anal. Methods Biomater. Sci. Catal. Sci. Technol. Chem. Commun. Chem. Soc. Rev. CHEM EDUC RES PRACT CRYSTENGCOMM Dalton Trans. Energy Environ. Sci. ENVIRON SCI-NANO ENVIRON SCI-PROC IMP ENVIRON SCI-WAT RES Faraday Discuss. Food Funct. Green Chem. Inorg. Chem. Front. Integr. Biol. J. Anal. At. Spectrom. J. Mater. Chem. A J. Mater. Chem. B J. Mater. Chem. C Lab Chip Mater. Chem. Front. Mater. Horiz. MEDCHEMCOMM Metallomics Mol. Biosyst. Mol. Syst. Des. Eng. Nanoscale Nanoscale Horiz. Nat. Prod. Rep. New J. Chem. Org. Biomol. Chem. Org. Chem. Front. PHOTOCH PHOTOBIO SCI PCCP Polym. Chem.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1