Federico A. Bugni, Mengsi Gao, Filip Obradovic, Amilcar Velez
This paper studies a specific inference problem for a partially-identified parameter of interest with an interval identified set. We consider the favorable situation in which a researcher has two possible estimators to construct the confidence interval proposed in Imbens and Manski (2004) and Stoye (2009), and one is more efficient than the other. While the literature shows that both estimators deliver asymptotically exact confidence intervals for the parameter of interest, their inference in terms of statistical power is not compared. One would expect that using the more efficient estimator would result in more powerful inference. We formally prove this result.
{"title":"On the power properties of inference for parameters with interval identified sets","authors":"Federico A. Bugni, Mengsi Gao, Filip Obradovic, Amilcar Velez","doi":"arxiv-2407.20386","DOIUrl":"https://doi.org/arxiv-2407.20386","url":null,"abstract":"This paper studies a specific inference problem for a partially-identified\u0000parameter of interest with an interval identified set. We consider the\u0000favorable situation in which a researcher has two possible estimators to\u0000construct the confidence interval proposed in Imbens and Manski (2004) and\u0000Stoye (2009), and one is more efficient than the other. While the literature\u0000shows that both estimators deliver asymptotically exact confidence intervals\u0000for the parameter of interest, their inference in terms of statistical power is\u0000not compared. One would expect that using the more efficient estimator would\u0000result in more powerful inference. We formally prove this result.","PeriodicalId":501293,"journal":{"name":"arXiv - ECON - Econometrics","volume":"129 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-07-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141865020","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}
Reducing financial risk is of paramount importance to investors, financial institutions, and corporations. Since the pioneering contribution of Johnson (1960), the optimal hedge ratio based on futures is regularly utilized. The current paper suggests an explicit and efficient method for testing the null hypothesis of a symmetric optimal hedge ratio against an asymmetric alternative one within a multivariate setting. If the null is rejected, the position dependent optimal hedge ratios can be estimated via the suggested model. This approach is expected to enhance the accuracy of the implemented hedging strategies compared to the standard methods since it accounts for the fact that the source of risk depends on whether the investor is a buyer or a seller of the risky asset. An application is provided using spot and futures prices of Bitcoin. The results strongly support the view that the optimal hedge ratio for this cryptocurrency is position dependent. The investor that is long in Bitcoin has a much higher conditional optimal hedge ratio compared to the one that is short in the asset. The difference between the two conditional optimal hedge ratios is statistically significant, which has important repercussions for implementing risk management strategies.
{"title":"Testing for the Asymmetric Optimal Hedge Ratios: With an Application to Bitcoin","authors":"Abdulnasser Hatemi-J","doi":"arxiv-2407.19932","DOIUrl":"https://doi.org/arxiv-2407.19932","url":null,"abstract":"Reducing financial risk is of paramount importance to investors, financial\u0000institutions, and corporations. Since the pioneering contribution of Johnson\u0000(1960), the optimal hedge ratio based on futures is regularly utilized. The\u0000current paper suggests an explicit and efficient method for testing the null\u0000hypothesis of a symmetric optimal hedge ratio against an asymmetric alternative\u0000one within a multivariate setting. If the null is rejected, the position\u0000dependent optimal hedge ratios can be estimated via the suggested model. This\u0000approach is expected to enhance the accuracy of the implemented hedging\u0000strategies compared to the standard methods since it accounts for the fact that\u0000the source of risk depends on whether the investor is a buyer or a seller of\u0000the risky asset. An application is provided using spot and futures prices of\u0000Bitcoin. The results strongly support the view that the optimal hedge ratio for\u0000this cryptocurrency is position dependent. The investor that is long in Bitcoin\u0000has a much higher conditional optimal hedge ratio compared to the one that is\u0000short in the asset. The difference between the two conditional optimal hedge\u0000ratios is statistically significant, which has important repercussions for\u0000implementing risk management strategies.","PeriodicalId":501293,"journal":{"name":"arXiv - ECON - Econometrics","volume":"200 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-07-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141865025","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}
In this paper we examine the existence of heterogeneity within a group, in panels with latent grouping structure. The assumption of within group homogeneity is prevalent in this literature, implying that the formation of groups alleviates cross-sectional heterogeneity, regardless of the prior knowledge of groups. While the latter hypothesis makes inference powerful, it can be often restrictive. We allow for models with richer heterogeneity that can be found both in the cross-section and within a group, without imposing the simple assumption that all groups must be heterogeneous. We further contribute to the method proposed by cite{su2016identifying}, by showing that the model parameters can be consistently estimated and the groups, while unknown, can be identifiable in the presence of different types of heterogeneity. Within the same framework we consider the validity of assuming both cross-sectional and within group homogeneity, using testing procedures. Simulations demonstrate good finite-sample performance of the approach in both classification and estimation, while empirical applications across several datasets provide evidence of multiple clusters, as well as reject the hypothesis of within group homogeneity.
{"title":"Heterogeneous Grouping Structures in Panel Data","authors":"Katerina Chrysikou, George Kapetanios","doi":"arxiv-2407.19509","DOIUrl":"https://doi.org/arxiv-2407.19509","url":null,"abstract":"In this paper we examine the existence of heterogeneity within a group, in\u0000panels with latent grouping structure. The assumption of within group\u0000homogeneity is prevalent in this literature, implying that the formation of\u0000groups alleviates cross-sectional heterogeneity, regardless of the prior\u0000knowledge of groups. While the latter hypothesis makes inference powerful, it\u0000can be often restrictive. We allow for models with richer heterogeneity that\u0000can be found both in the cross-section and within a group, without imposing the\u0000simple assumption that all groups must be heterogeneous. We further contribute\u0000to the method proposed by cite{su2016identifying}, by showing that the model\u0000parameters can be consistently estimated and the groups, while unknown, can be\u0000identifiable in the presence of different types of heterogeneity. Within the\u0000same framework we consider the validity of assuming both cross-sectional and\u0000within group homogeneity, using testing procedures. Simulations demonstrate\u0000good finite-sample performance of the approach in both classification and\u0000estimation, while empirical applications across several datasets provide\u0000evidence of multiple clusters, as well as reject the hypothesis of within group\u0000homogeneity.","PeriodicalId":501293,"journal":{"name":"arXiv - ECON - Econometrics","volume":"12 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141865022","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}
The potential impact of nonresponse on election polls is well known and frequently acknowledged. Yet measurement and reporting of polling error has focused solely on sampling error, represented by the margin of error of a poll. Survey statisticians have long recommended measurement of the total survey error of a sample estimate by its mean square error (MSE), which jointly measures sampling and non-sampling errors. Extending the conventional language of polling, we think it reasonable to use the square root of maximum MSE to measure the total margin of error. This paper demonstrates how to measure the potential impact of nonresponse using the concept of the total margin of error, which we argue should be a standard feature in the reporting of election poll results. We first show how to jointly measure statistical imprecision and response bias when a pollster lacks any knowledge of the candidate preferences of non-responders. We then extend the analysis to settings where the pollster has partial knowledge that bounds the preferences of non-responders.
{"title":"Accounting for Nonresponse in Election Polls: Total Margin of Error","authors":"Jeff Dominitz, Charles F. Manski","doi":"arxiv-2407.19339","DOIUrl":"https://doi.org/arxiv-2407.19339","url":null,"abstract":"The potential impact of nonresponse on election polls is well known and\u0000frequently acknowledged. Yet measurement and reporting of polling error has\u0000focused solely on sampling error, represented by the margin of error of a poll.\u0000Survey statisticians have long recommended measurement of the total survey\u0000error of a sample estimate by its mean square error (MSE), which jointly\u0000measures sampling and non-sampling errors. Extending the conventional language\u0000of polling, we think it reasonable to use the square root of maximum MSE to\u0000measure the total margin of error. This paper demonstrates how to measure the\u0000potential impact of nonresponse using the concept of the total margin of error,\u0000which we argue should be a standard feature in the reporting of election poll\u0000results. We first show how to jointly measure statistical imprecision and\u0000response bias when a pollster lacks any knowledge of the candidate preferences\u0000of non-responders. We then extend the analysis to settings where the pollster\u0000has partial knowledge that bounds the preferences of non-responders.","PeriodicalId":501293,"journal":{"name":"arXiv - ECON - Econometrics","volume":"169 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141865023","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}
Tests based on the $2$- and $infty$-norm have received considerable attention in high-dimensional testing problems, as they are powerful against dense and sparse alternatives, respectively. The power enhancement principle of Fan et al. (2015) combines these two norms to construct tests that are powerful against both types of alternatives. Nevertheless, the $2$- and $infty$-norm are just two out of the whole spectrum of $p$-norms that one can base a test on. In the context of testing whether a candidate parameter satisfies a large number of moment equalities, we construct a test that harnesses the strength of all $p$-norms with $pin[2, infty]$. As a result, this test consistent against strictly more alternatives than any test based on a single $p$-norm. In particular, our test is consistent against more alternatives than tests based on the $2$- and $infty$-norm, which is what most implementations of the power enhancement principle target. We illustrate the scope of our general results by using them to construct a test that simultaneously dominates the Anderson-Rubin test (based on $p=2$) and tests based on the $infty$-norm in terms of consistency in the linear instrumental variable model with many (weak) instruments.
{"title":"Enhanced power enhancements for testing many moment equalities: Beyond the $2$- and $infty$-norm","authors":"Anders Bredahl Kock, David Preinerstorfer","doi":"arxiv-2407.17888","DOIUrl":"https://doi.org/arxiv-2407.17888","url":null,"abstract":"Tests based on the $2$- and $infty$-norm have received considerable\u0000attention in high-dimensional testing problems, as they are powerful against\u0000dense and sparse alternatives, respectively. The power enhancement principle of\u0000Fan et al. (2015) combines these two norms to construct tests that are powerful\u0000against both types of alternatives. Nevertheless, the $2$- and $infty$-norm\u0000are just two out of the whole spectrum of $p$-norms that one can base a test\u0000on. In the context of testing whether a candidate parameter satisfies a large\u0000number of moment equalities, we construct a test that harnesses the strength of\u0000all $p$-norms with $pin[2, infty]$. As a result, this test consistent against\u0000strictly more alternatives than any test based on a single $p$-norm. In\u0000particular, our test is consistent against more alternatives than tests based\u0000on the $2$- and $infty$-norm, which is what most implementations of the power\u0000enhancement principle target. We illustrate the scope of our general results by using them to construct a\u0000test that simultaneously dominates the Anderson-Rubin test (based on $p=2$) and\u0000tests based on the $infty$-norm in terms of consistency in the linear\u0000instrumental variable model with many (weak) instruments.","PeriodicalId":501293,"journal":{"name":"arXiv - ECON - Econometrics","volume":"127 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141771968","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}
We use the exact finite sample likelihood and statistical decision theory to answer questions of ``why?'' and ``what should you have done?'' using data from randomized experiments and a utility function that prioritizes safety over efficacy. We propose a finite sample Bayesian decision rule and a finite sample maximum likelihood decision rule. We show that in finite samples from 2 to 50, it is possible for these rules to achieve better performance according to established maximin and maximum regret criteria than a rule based on the Boole-Frechet-Hoeffding bounds. We also propose a finite sample maximum likelihood criterion. We apply our rules and criterion to an actual clinical trial that yielded a promising estimate of efficacy, and our results point to safety as a reason for why results were mixed in subsequent trials.
{"title":"Starting Small: Prioritizing Safety over Efficacy in Randomized Experiments Using the Exact Finite Sample Likelihood","authors":"Neil Christy, A. E. Kowalski","doi":"arxiv-2407.18206","DOIUrl":"https://doi.org/arxiv-2407.18206","url":null,"abstract":"We use the exact finite sample likelihood and statistical decision theory to\u0000answer questions of ``why?'' and ``what should you have done?'' using data from\u0000randomized experiments and a utility function that prioritizes safety over\u0000efficacy. We propose a finite sample Bayesian decision rule and a finite sample\u0000maximum likelihood decision rule. We show that in finite samples from 2 to 50,\u0000it is possible for these rules to achieve better performance according to\u0000established maximin and maximum regret criteria than a rule based on the\u0000Boole-Frechet-Hoeffding bounds. We also propose a finite sample maximum\u0000likelihood criterion. We apply our rules and criterion to an actual clinical\u0000trial that yielded a promising estimate of efficacy, and our results point to\u0000safety as a reason for why results were mixed in subsequent trials.","PeriodicalId":501293,"journal":{"name":"arXiv - ECON - Econometrics","volume":"61 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141771966","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}
The most common approach to causal modelling is the potential outcomes framework due to Neyman and Rubin. In this framework, outcomes of counterfactual treatments are assumed to be well-defined. This metaphysical assumption is often thought to be problematic yet indispensable. The conventional approach relies not only on counterfactuals, but also on abstract notions of distributions and assumptions of independence that are not directly testable. In this paper, we construe causal inference as treatment-wise predictions for finite populations where all assumptions are testable; this means that one can not only test predictions themselves (without any fundamental problem), but also investigate sources of error when they fail. The new framework highlights the model-dependence of causal claims as well as the difference between statistical and scientific inference.
{"title":"Causal modelling without counterfactuals and individualised effects","authors":"Benedikt Höltgen, Robert C. Williamson","doi":"arxiv-2407.17385","DOIUrl":"https://doi.org/arxiv-2407.17385","url":null,"abstract":"The most common approach to causal modelling is the potential outcomes\u0000framework due to Neyman and Rubin. In this framework, outcomes of\u0000counterfactual treatments are assumed to be well-defined. This metaphysical\u0000assumption is often thought to be problematic yet indispensable. The\u0000conventional approach relies not only on counterfactuals, but also on abstract\u0000notions of distributions and assumptions of independence that are not directly\u0000testable. In this paper, we construe causal inference as treatment-wise\u0000predictions for finite populations where all assumptions are testable; this\u0000means that one can not only test predictions themselves (without any\u0000fundamental problem), but also investigate sources of error when they fail. The\u0000new framework highlights the model-dependence of causal claims as well as the\u0000difference between statistical and scientific inference.","PeriodicalId":501293,"journal":{"name":"arXiv - ECON - Econometrics","volume":"1 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141771969","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}
This paper proposes a new class of distributional causal quantities, referred to as the textit{outcome conditioned partial policy effects} (OCPPEs), to measure the textit{average} effect of a general counterfactual intervention of a target covariate on the individuals in different quantile ranges of the outcome distribution. The OCPPE approach is valuable in several aspects: (i) Unlike the unconditional quantile partial effect (UQPE) that is not $sqrt{n}$-estimable, an OCPPE is $sqrt{n}$-estimable. Analysts can use it to capture heterogeneity across the unconditional distribution of $Y$ as well as obtain accurate estimation of the aggregated effect at the upper and lower tails of $Y$. (ii) The semiparametric efficiency bound for an OCPPE is explicitly derived. (iii) We propose an efficient debiased estimator for OCPPE, and provide feasible uniform inference procedures for the OCPPE process. (iv) The efficient doubly robust score for an OCPPE can be used to optimize infinitesimal nudges to a continuous treatment by maximizing a quantile specific Empirical Welfare function. We illustrate the method by analyzing how anti-smoking policies impact low percentiles of live infants' birthweights.
{"title":"Identification and inference of outcome conditioned partial effects of general interventions","authors":"Zhengyu Zhang, Zequn Jin, Lihua Lin","doi":"arxiv-2407.16950","DOIUrl":"https://doi.org/arxiv-2407.16950","url":null,"abstract":"This paper proposes a new class of distributional causal quantities, referred\u0000to as the textit{outcome conditioned partial policy effects} (OCPPEs), to\u0000measure the textit{average} effect of a general counterfactual intervention of\u0000a target covariate on the individuals in different quantile ranges of the\u0000outcome distribution. The OCPPE approach is valuable in several aspects: (i) Unlike the\u0000unconditional quantile partial effect (UQPE) that is not $sqrt{n}$-estimable,\u0000an OCPPE is $sqrt{n}$-estimable. Analysts can use it to capture heterogeneity\u0000across the unconditional distribution of $Y$ as well as obtain accurate\u0000estimation of the aggregated effect at the upper and lower tails of $Y$. (ii)\u0000The semiparametric efficiency bound for an OCPPE is explicitly derived. (iii)\u0000We propose an efficient debiased estimator for OCPPE, and provide feasible\u0000uniform inference procedures for the OCPPE process. (iv) The efficient doubly\u0000robust score for an OCPPE can be used to optimize infinitesimal nudges to a\u0000continuous treatment by maximizing a quantile specific Empirical Welfare\u0000function. We illustrate the method by analyzing how anti-smoking policies\u0000impact low percentiles of live infants' birthweights.","PeriodicalId":501293,"journal":{"name":"arXiv - ECON - Econometrics","volume":"16 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141771971","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}
Florian Huber, Gary Koop, Massimiliano Marcellino, Tobias Scheckel
Commonly used priors for Vector Autoregressions (VARs) induce shrinkage on the autoregressive coefficients. Introducing shrinkage on the error covariance matrix is sometimes done but, in the vast majority of cases, without considering the network structure of the shocks and by placing the prior on the lower Cholesky factor of the precision matrix. In this paper, we propose a prior on the VAR error precision matrix directly. Our prior, which resembles a standard spike and slab prior, models variable inclusion probabilities through a stochastic block model that clusters shocks into groups. Within groups, the probability of having relations across group members is higher (inducing less sparsity) whereas relations across groups imply a lower probability that members of each group are conditionally related. We show in simulations that our approach recovers the true network structure well. Using a US macroeconomic data set, we illustrate how our approach can be used to cluster shocks together and that this feature leads to improved density forecasts.
矢量自回归(VAR)常用的先验值会引起自回归系数的收缩。有时也会在误差协方差矩阵上引入收缩,但在绝大多数情况下,都没有考虑冲击的网络结构,而是将先验值置于精度矩阵的较低 Cholesky 因子上。在本文中,我们直接提出了 VAR 误差精度矩阵的先验值。我们的先验类似于标准的尖峰先验和板块先验,通过随机块模型对变量包含概率进行建模,将冲击聚类成组。在组内,组内成员之间存在关系的概率较高(导致较低的稀疏性),而组间关系则意味着每个组的成员之间存在条件关系的概率较低。我们的模拟结果表明,我们的方法很好地还原了真实的网络结构。通过使用美国宏观经济数据集,我们说明了如何使用我们的方法将冲击聚集在一起,并说明这一特征可以改善密度预测。
{"title":"Bayesian modelling of VAR precision matrices using stochastic block networks","authors":"Florian Huber, Gary Koop, Massimiliano Marcellino, Tobias Scheckel","doi":"arxiv-2407.16349","DOIUrl":"https://doi.org/arxiv-2407.16349","url":null,"abstract":"Commonly used priors for Vector Autoregressions (VARs) induce shrinkage on\u0000the autoregressive coefficients. Introducing shrinkage on the error covariance\u0000matrix is sometimes done but, in the vast majority of cases, without\u0000considering the network structure of the shocks and by placing the prior on the\u0000lower Cholesky factor of the precision matrix. In this paper, we propose a\u0000prior on the VAR error precision matrix directly. Our prior, which resembles a\u0000standard spike and slab prior, models variable inclusion probabilities through\u0000a stochastic block model that clusters shocks into groups. Within groups, the\u0000probability of having relations across group members is higher (inducing less\u0000sparsity) whereas relations across groups imply a lower probability that\u0000members of each group are conditionally related. We show in simulations that\u0000our approach recovers the true network structure well. Using a US macroeconomic\u0000data set, we illustrate how our approach can be used to cluster shocks together\u0000and that this feature leads to improved density forecasts.","PeriodicalId":501293,"journal":{"name":"arXiv - ECON - Econometrics","volume":"306 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141771829","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}
We propose a novel regression adjustment method designed for estimating distributional treatment effect parameters in randomized experiments. Randomized experiments have been extensively used to estimate treatment effects in various scientific fields. However, to gain deeper insights, it is essential to estimate distributional treatment effects rather than relying solely on average effects. Our approach incorporates pre-treatment covariates into a distributional regression framework, utilizing machine learning techniques to improve the precision of distributional treatment effect estimators. The proposed approach can be readily implemented with off-the-shelf machine learning methods and remains valid as long as the nuisance components are reasonably well estimated. Also, we establish the asymptotic properties of the proposed estimator and present a uniformly valid inference method. Through simulation results and real data analysis, we demonstrate the effectiveness of integrating machine learning techniques in reducing the variance of distributional treatment effect estimators in finite samples.
{"title":"Estimating Distributional Treatment Effects in Randomized Experiments: Machine Learning for Variance Reduction","authors":"Undral Byambadalai, Tatsushi Oka, Shota Yasui","doi":"arxiv-2407.16037","DOIUrl":"https://doi.org/arxiv-2407.16037","url":null,"abstract":"We propose a novel regression adjustment method designed for estimating\u0000distributional treatment effect parameters in randomized experiments.\u0000Randomized experiments have been extensively used to estimate treatment effects\u0000in various scientific fields. However, to gain deeper insights, it is essential\u0000to estimate distributional treatment effects rather than relying solely on\u0000average effects. Our approach incorporates pre-treatment covariates into a\u0000distributional regression framework, utilizing machine learning techniques to\u0000improve the precision of distributional treatment effect estimators. The\u0000proposed approach can be readily implemented with off-the-shelf machine\u0000learning methods and remains valid as long as the nuisance components are\u0000reasonably well estimated. Also, we establish the asymptotic properties of the\u0000proposed estimator and present a uniformly valid inference method. Through\u0000simulation results and real data analysis, we demonstrate the effectiveness of\u0000integrating machine learning techniques in reducing the variance of\u0000distributional treatment effect estimators in finite samples.","PeriodicalId":501293,"journal":{"name":"arXiv - ECON - Econometrics","volume":"20 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-07-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141771970","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}