id
int64
28.8k
36k
text
stringlengths
44
3.03k
translation
stringlengths
36
2.47k
29,872
Considering a continuous random variable Y together with a continuous random vector X, I propose a nonparametric estimator f^(.|x) for the conditional density of Y given X=x. This estimator takes the form of an exponential series whose coefficients T = (T1,...,TJ) are the solution of a system of nonlinear equations that depends on an estimator of the conditional expectation E[p(Y)|X=x], where p(.) is a J-dimensional vector of basis functions. A key feature is that E[p(Y)|X=x] is estimated by generalized random forest (Athey, Tibshirani, and Wager, 2019), targeting the heterogeneity of T across x. I show that f^(.|x) is uniformly consistent and asymptotically normal, while allowing J to grow to infinity. I also provide a standard error formula to construct asymptotically valid confidence intervals. Results from Monte Carlo experiments and an empirical illustration are provided.
با توجه به یک متغیر تصادفی مداوم y همراه با یک تصادفی مداوم وکتور X ، من یک برآوردگر غیر پارامتری f^(. | x) را برای شرطی پیشنهاد می کنم چگالی y داده شده x = x.این برآوردگر به شکل یک سری نمایی می شود ضرایب آن T = (T1 ، ... ، TJ) راه حل یک سیستم غیرخطی است معادله هایی که به یک برآوردگر انتظار مشروط بستگی دارد E [p (y) | x = x] ، جایی که p (.) یک بردار j بعدی از توابع پایه است.یک کلید ویژگی این است که E [P (y) | x = x] توسط جنگل تصادفی عمومی تخمین زده می شود (Athey ، Tibshirani ، and Wager ، 2019) ، هدف قرار دادن ناهمگونی T در سراسر x.من نشان می دهم که f^(. | x) به طور یکنواخت سازگار و بدون علامت طبیعی است ، در حالی که اجازه می دهد j برای رشد به بی نهایت.من همچنین یک فرمول خطای استاندارد برای ساخت ارائه می دهم فواصل اعتماد به نفس معتبر بدون علامت.نتایج حاصل از آزمایش مونت کارلو و یک تصویر تجربی ارائه شده است.
29,873
This paper develops unified asymptotic distribution theory for dynamic quantile predictive regressions which is useful when examining quantile predictability in stock returns under possible presence of nonstationarity.
این مقاله نظریه توزیع یکپارچه یکپارچه را برای پویا توسعه می دهد رگرسیون پیش بینی کننده کمی که هنگام بررسی کمی مفید است پیش بینی پذیری در بازده سهام تحت حضور احتمالی غیرقانونی بودن.
29,874
Stock prices often react sluggishly to news, producing gradual jumps and jump delays. Econometricians typically treat these sluggish reactions as microstructure effects and settle for a coarse sampling grid to guard against them. Synchronizing mistimed stock returns on a fine sampling grid allows us to automatically detect noisy jumps and better approximate the true common jumps in related stock prices.
قیمت سهام غالباً به طرز لاغر به اخبار واکنش نشان می دهد ، پرش های تدریجی و پرش را تولید می کند تأخیراقتصاد سنجی معمولاً با این واکنشهای لاغر مانند رفتار می کند اثرات ریزساختار و برای یک شبکه نمونه گیری درشت برای محافظت در برابر آن مستقر شوید آنهاهمگام سازی بازده سهام اشتباه در یک شبکه نمونه گیری خوب به ما امکان می دهد به طور خودکار پرش های پر سر و صدا را تشخیص دهید و پرش های مشترک واقعی را تقریبی کنید در قیمت سهام مرتبط
29,875
To tackle difficulties for theoretical studies in situations involving nonsmooth functions, we propose a sequence of infinitely differentiable functions to approximate the nonsmooth function under consideration. A rate of approximation is established and an illustration of its application is then provided.
برای مقابله با مشکلات برای مطالعات نظری در مواقع توابع غیر نرم ، ما دنباله ای از بی نهایت متفاوت را پیشنهاد می کنیم توابع برای تقریبی عملکرد غیر نرم در نظر گرفته شده است.یک نرخ تقریب ایجاد شده و نمونه ای از کاربرد آن در آن زمان است ارائه شده
29,876
We propose a bootstrap generalized Hausman test for the correct specification of unobserved heterogeneity in both linear and nonlinear fixed-effects panel data models. We consider as null hypotheses two scenarios in which the unobserved heterogeneity is either time-invariant or specified as additive individual and time effects. We contrast the standard fixed-effects estimators with the recently developed two-way grouped fixed-effects estimator, that is consistent in the presence of time-varying heterogeneity under minimal specification and distributional assumptions for the unobserved effects. The Hausman test exploits the general formulation for the variance of the vector of contrasts and critical values are computed via parametric percentile bootstrap, so as to account for the non-centrality of the asymptotic chi-square distribution arising from the incidental parameters and approximation biases. Monte Carlo evidence shows that the test has correct size and good power properties. We provide two empirical applications to illustrate the proposed test: the first one is based on a linear model for the determinants of the wage of working women and the second analyzes the trade extensive margin.
ما برای مشخصات صحیح یک تست Hausman تعمیم یافته Bootstrap را پیشنهاد می کنیم ناهمگونی بدون نظارت در پانل اثرات ثابت خطی و غیرخطی مدل های داده.ما فرضیه های تهی را دو سناریو در نظر می گیریم که در آن ناهمگونی بدون نظارت یا متغیر است یا به عنوان افزودنی مشخص می شود اثرات فردی و زمانی.ما برآوردگرهای عوارض ثابت استاندارد را تضاد می کنیم با برآوردگر اثرات ثابت دو طرفه که اخیراً توسعه یافته است ، یعنی سازگار در حضور ناهمگونی متغیر زمان تحت حداقل مشخصات و فرضیات توزیع برای اثرات غیرقابل کنترل.در آزمون هاسمن از فرمولاسیون کلی برای واریانس بردار سوء استفاده می کند تضادها و مقادیر بحرانی از طریق بوت استرپ صدک پارامتری محاسبه می شوند ، به گونه ای که عدم محوریت مجذور مجسمه مجانبی را به خود اختصاص دهد توزیع ناشی از پارامترهای حادثه ای و تعصب تقریبی. شواهد مونت کارلو نشان می دهد که این تست دارای اندازه و قدرت مناسب است خواصما دو برنامه تجربی برای نشان دادن پیشنهادی ارائه می دهیم آزمون: اولین مورد بر اساس یک مدل خطی برای عوامل تعیین کننده دستمزد است از زنان شاغل و دومین تحلیل حاشیه تجارت.
29,890
Dynamic factor models have been developed out of the need of analyzing and forecasting time series in increasingly high dimensions. While mathematical statisticians faced with inference problems in high-dimensional observation spaces were focusing on the so-called spiked-model-asymptotics, econometricians adopted an entirely and considerably more effective asymptotic approach, rooted in the factor models originally considered in psychometrics. The so-called dynamic factor model methods, in two decades, has grown into a wide and successful body of techniques that are widely used in central banks, financial institutions, economic and statistical institutes. The objective of this chapter is not an extensive survey of the topic but a sketch of its historical growth, with emphasis on the various assumptions and interpretations, and a family tree of its main variants.
مدلهای فاکتور پویا از نیاز به تجزیه و تحلیل و پیش بینی سری زمانی در ابعاد فزاینده ای بالا.در حالی که ریاضی است آماری با مشاهده در مشاهدات با ابعاد بالا با مشکلات استنباط روبرو بودند فضاها در حال تمرکز بر روی به اصطلاح asmptotics-asmptotics-model spiked ، اقتصاد سنجی بودند یک رویکرد کاملاً مؤثرتر و کاملاً مؤثرتر اتخاذ کرد ، ریشه دار در مدل های فاکتور که در ابتدا در روان سنجی در نظر گرفته شده است.به اصطلاح روشهای مدل فاکتور پویا ، در دو دهه ، به یک گسترده و گسترده تبدیل شده است بدنه موفق تکنیک هایی که به طور گسترده در بانک های مرکزی استفاده می شود ، مالی نهادها ، مؤسسات اقتصادی و آماری.هدف از این فصل بررسی گسترده ای از موضوع نیست بلکه طرح تاریخی آن است رشد ، با تأکید بر فرضیات و تفسیرهای مختلف ، و الف درخت خانوادگی از انواع اصلی آن.
29,877
In this paper we reconsider the notion of optimality in estimation of partially identified models. We illustrate the general problem in the context of a semiparametric binary choice model with discrete covariates as an example of a model which is partially identified as shown in, e.g. Bierens and Hartog (1988). A set estimator for the regression coefficients in the model can be constructed by implementing the Maximum Score procedure proposed by Manski (1975). For many designs this procedure converges to the identified set for these parameters, and so in one sense is optimal. But as shown in Komarova (2013) for other cases the Maximum Score objective function gives an outer region of the identified set. This motivates alternative methods that are optimal in one sense that they converge to the identified region in all designs, and we propose and compare such procedures. One is a Hodges type estimator combining the Maximum Score estimator with existing procedures. A second is a two step estimator using a Maximum Score type objective function in the second step. Lastly we propose a new random set quantile estimator, motivated by definitions introduced in Molchanov (2006). Extensions of these ideas for the cross sectional model to static and dynamic discrete panel data models are also provided.
در این مقاله مفهوم بهینه را در تخمین تجدید نظر می کنیم مدلهای تا حدی شناسایی شده.ما مشکل کلی را در متن نشان می دهیم از یک مدل انتخاب باینری نیمهرامتری با متغیرهای متغیر گسسته به عنوان نمونه از مدلی که تا حدی مطابق با آن مشخص شده است ، به عنوان مثالBierens و Hartog (1988).یک برآوردگر تعیین شده برای ضرایب رگرسیون در مدل می تواند باشد ساخته شده با اجرای حداکثر نمره نمره پیشنهادی توسط مانسکی (1975).برای بسیاری از طرح ها این روش به مجموعه مشخص شده برای این پارامترها ، و بنابراین به یک معنا مطلوب است.اما همانطور که در کوماروا نشان داده شده است (2013) برای موارد دیگر عملکرد هدف حداکثر نمره بیرونی را نشان می دهد منطقه مجموعه مشخص شده.این روش های جایگزین را ایجاد می کند به یک معنا که به طور کلی به منطقه شناسایی شده همگرا می شوند طرح ها ، و ما چنین رویه هایی را پیشنهاد و مقایسه می کنیم.یکی از نوع هاجز است برآوردگر ترکیب حداکثر برآوردگر نمره با روشهای موجود.آ دوم یک برآوردگر دو مرحله ای با استفاده از یک عملکرد هدف حداکثر نوع نمره در مرحله دومدر آخر ما یک برآوردگر Quantile مجموعه تصادفی جدید را پیشنهاد می کنیم ، با انگیزه تعاریف معرفی شده در مولچانوف (2006).پسوند این ها ایده هایی برای مدل مقطعی به داده های پانل گسسته استاتیک و پویا مدل ها نیز ارائه شده اند.
29,878
This paper proposes a Lasso-based estimator which uses information embedded in the Moran statistic to develop a selection procedure called Moran's I Lasso (Mi-Lasso) to solve the Eigenvector Spatial Filtering (ESF) eigenvector selection problem. ESF uses a subset of eigenvectors from a spatial weights matrix to efficiently account for any omitted cross-sectional correlation terms in a classical linear regression framework, thus does not require the researcher to explicitly specify the spatial part of the underlying structural model. We derive performance bounds and show the necessary conditions for consistent eigenvector selection. The key advantages of the proposed estimator are that it is intuitive, theoretically grounded, and substantially faster than Lasso based on cross-validation or any proposed forward stepwise procedure. Our main simulation results show the proposed selection procedure performs well in finite samples. Compared to existing selection procedures, we find Mi-Lasso has one of the smallest biases and mean squared errors across a range of sample sizes and levels of spatial correlation. An application on house prices further demonstrates Mi-Lasso performs well compared to existing procedures.
در این مقاله یک برآوردگر مبتنی بر لاسو ارائه شده است که از اطلاعات تعبیه شده استفاده می کند در آمار موران برای تدوین یک روش انتخاب به نام Moran's I Lasso (mi-lasso) برای حل فیلتر فضا (ESF) eigenveector مشکل انتخابESF از زیر مجموعه ای از eigenvectors از وزن مکانی استفاده می کند ماتریس برای به حساب کردن هرگونه شرایط همبستگی مقطعی حذف شده در یک چارچوب رگرسیون خطی کلاسیک ، بنابراین نیازی به محقق برای مشخص کردن قسمت مکانی ساختاری اساسی مدل.ما مرزهای عملکرد را استخراج می کنیم و شرایط لازم را برای آن نشان می دهیم انتخاب eigenveector مداوم.مزایای اصلی برآوردگر پیشنهادی آیا این است که بصری ، از نظر تئوریک و از نظر قابل ملاحظه ای سریعتر از Lasso بر اساس اعتبارسنجی متقابل یا هر روش پیشنهادی به جلو به مرحله.ما نتایج شبیه سازی اصلی نشان می دهد که روش انتخاب پیشنهادی عملکرد خوبی دارد نمونه های محدود.در مقایسه با روشهای انتخاب موجود ، می بینیم که Mi-Lasso دارای یکی از کوچکترین تعصبات و میانگین خطاهای مربع در طیف وسیعی از نمونه ها اندازه و سطح همبستگی مکانی.درخواست بیشتر در مورد قیمت خانه نشان می دهد که Mi-Lasso در مقایسه با رویه های موجود عملکرد خوبی دارد.
29,879
We assess the finite sample performance of the conduct parameter test in homogeneous goods markets. Statistical power rises with an increase in the number of markets, a larger conduct parameter, and a stronger demand rotation instrument. However, even with a moderate number of markets and five firms, regardless of instrument strength and the utilization of optimal instruments, rejecting the null hypothesis of perfect competition remains challenging. Our findings indicate that empirical results that fail to reject perfect competition are a consequence of the limited number of markets rather than methodological deficiencies.
ما عملکرد نمونه محدود آزمون پارامتر رفتار را در ارزیابی می کنیم بازارهای کالاهای همگن.قدرت آماری با افزایش در تعداد بازارها ، یک پارامتر رفتاری بزرگتر و چرخش تقاضای قوی تر ابزار.با این حال ، حتی با تعداد متوسط ​​بازار و پنج شرکت ، صرف نظر از قدرت ابزار و استفاده از ابزارهای بهینه ، رد فرضیه تهی از رقابت کامل همچنان چالش برانگیز است.ما یافته ها نشان می دهد که نتایج تجربی که نتوانند کامل را رد کنند رقابت نتیجه ای از تعداد محدود بازارها به جای کمبودهای روش شناختی.
29,880
This paper proposes minimum distance inference for a structural parameter of interest, which is robust to the lack of identification of other structural nuisance parameters. Some choices of the weighting matrix lead to asymptotic chi-squared distributions with degrees of freedom that can be consistently estimated from the data, even under partial identification. In any case, knowledge of the level of under-identification is not required. We study the power of our robust test. Several examples show the wide applicability of the procedure and a Monte Carlo investigates its finite sample performance. Our identification-robust inference method can be applied to make inferences on both calibrated (fixed) parameters and any other structural parameter of interest. We illustrate the method's usefulness by applying it to a structural model on the non-neutrality of monetary policy, as in \cite{nakamura2018high}, where we empirically evaluate the validity of the calibrated parameters and we carry out robust inference on the slope of the Phillips curve and the information effect.
در این مقاله حداقل استنتاج فاصله برای یک پارامتر ساختاری از علاقه ، که به عدم شناسایی سایر ساختاری ها قوی است پارامترهای مزاحم.برخی از گزینه های ماتریس وزنه برداری منجر به مجانبی می شود توزیع های مجذور کای با درجه آزادی که می توانند به طور مداوم باشند تخمین زده می شود از داده ها ، حتی تحت شناسایی جزئی.در هر صورت، آگاهی از سطح شناسایی زیر مورد نیاز نیست.ما مطالعه می کنیم قدرت تست قوی ما.چندین مثال کاربرد گسترده ای را نشان می دهد روش و یک مونت کارلو عملکرد نمونه محدود آن را بررسی می کند.ما برای ایجاد استنباط می توان از روش استنتاج-روبست استفاده کرد هر دو پارامتر کالیبره شده (ثابت) و هر پارامتر ساختاری دیگر از علاقه.ما با استفاده از آن در ساختاری ، سودمندی روش را نشان می دهیم مدل در مورد عدم خنثی بودن سیاست های پولی ، مانند \ cite {nakamura2018high} ، جایی که ما به صورت تجربی اعتبار پارامترهای کالیبره شده و ما را ارزیابی می کنیم استنتاج قوی را در شیب منحنی فیلیپس و اثر اطلاعات
29,881
This paper studies the identification and estimation of a semiparametric binary network model in which the unobserved social characteristic is endogenous, that is, the unobserved individual characteristic influences both the binary outcome of interest and how links are formed within the network. The exact functional form of the latent social characteristic is not known. The proposed estimators are obtained based on matching pairs of agents whose network formation distributions are the same. The consistency and the asymptotic distribution of the estimators are proposed. The finite sample properties of the proposed estimators in a Monte-Carlo simulation are assessed. We conclude this study with an empirical application.
در این مقاله به بررسی شناسایی و تخمین یک نیمه پارامتری می پردازیم مدل شبکه باینری که در آن ویژگی اجتماعی بدون نظارت است درون زا ، یعنی ویژگی فردی بدون نظارت بر هر دو تأثیر می گذارد نتیجه باینری مورد علاقه و چگونگی شکل گیری پیوندها در شبکه.در شکل دقیق عملکردی از ویژگی های اجتماعی نهفته مشخص نیست.در برآوردگرهای پیشنهادی بر اساس جفت های تطبیق یافته مأمورین به دست می آیند توزیع تشکیل شبکه یکسان است.قوام و توزیع بدون علامت برآوردگرها ارائه شده است.نمونه محدود خواص برآوردگرهای پیشنهادی در یک شبیه سازی مونت کارلو ارزیابی می شود. ما این مطالعه را با یک برنامه تجربی نتیجه می گیریم.
29,882
Nonlinearity and endogeneity are common in causal effect studies with observational data. In this paper, we propose new estimation and inference procedures for nonparametric treatment effect functions with endogeneity and potentially high-dimensional covariates. The main innovation of this paper is the double bias correction procedure for the nonparametric instrumental variable (NPIV) model under high dimensions. We provide a useful uniform confidence band of the marginal effect function, defined as the derivative of the nonparametric treatment function. The asymptotic honesty of the confidence band is verified in theory. Simulations and an empirical study of air pollution and migration demonstrate the validity of our procedures.
غیرخطی بودن و درون زایی در مطالعات اثر علیت با داده های مشاهده ای.در این مقاله ، ما تخمین و استنباط جدیدی را پیشنهاد می کنیم روشهای مربوط به توابع اثر درمانی غیر پارامتری با درون زا و متغیرهای بالقوه با ابعاد بالا.نوآوری اصلی این مقاله این است روش اصلاح تعصب مضاعف برای ابزارهای غیرپارامتری مدل متغیر (NPIV) در ابعاد بالا.ما یک لباس مفید ارائه می دهیم باند اعتماد به نفس عملکرد اثر حاشیه ای ، به عنوان مشتق از عملکرد درمان غیرپارامتری.صداقت بدون علامت اعتماد به نفس باند در تئوری تأیید می شود.شبیه سازی ها و یک مطالعه تجربی از آلودگی هوا و مهاجرت اعتبار رویه های ما را نشان می دهد.
30,272
To what extent can agents with misspecified subjective models predict false correlations? We study an "analyst" who utilizes models that take the form of a recursive system of linear regression equations. The analyst fits each equation to minimize the sum of squared errors against an arbitrarily large sample. We characterize the maximal pairwise correlation that the analyst can predict given a generic objective covariance matrix, subject to the constraint that the estimated model does not distort the mean and variance of individual variables. We show that as the number of variables in the model grows, the false pairwise correlation can become arbitrarily close to one, regardless of the true correlation.
تا چه حد می تواند عوامل دارای مدلهای ذهنی غلط پیش بینی شده کاذب را پیش بینی کنند همبستگی ها؟ما یک "تحلیلگر" را مطالعه می کنیم که از مدلهایی استفاده می کند که شکل A را می گیرند سیستم بازگشتی معادلات رگرسیون خطی.تحلیلگر متناسب با هر معادله است برای به حداقل رساندن مجموع خطاهای مربع در برابر یک نمونه بزرگ خودسرانه.ما همبستگی حداکثر زوجی را که تحلیلگر می تواند پیش بینی کند ، مشخص می کند با توجه به یک ماتریس کواریانس هدف عمومی ، با توجه به محدودیت مدل برآورد شده میانگین و واریانس متغیرهای فردی را تحریف نمی کند. ما نشان می دهیم که با افزایش تعداد متغیرهای مدل ، جفت کاذب همبستگی بدون در نظر گرفتن واقعیت می تواند به طور دلخواه نزدیک به یک باشد همبستگی.
29,883
Generalized method of moments estimators based on higher-order moment conditions derived from independent shocks can be used to identify and estimate the simultaneous interaction in structural vector autoregressions. This study highlights two problems that arise when using these estimators in small samples. First, imprecise estimates of the asymptotically efficient weighting matrix and the asymptotic variance lead to volatile estimates and inaccurate inference. Second, many moment conditions lead to a small sample scaling bias towards innovations with a variance smaller than the normalizing unit variance assumption. To address the first problem, I propose utilizing the assumption of independent structural shocks to estimate the efficient weighting matrix and the variance of the estimator. For the second issue, I propose incorporating a continuously updated scaling term into the weighting matrix, eliminating the scaling bias. To demonstrate the effectiveness of these measures, I conducted a Monte Carlo simulation which shows a significant improvement in the performance of the estimator.
روش کلی برآوردگرهای لحظه ای بر اساس لحظه مرتبه بالاتر از شرایط حاصل از شوک های مستقل می توان برای شناسایی و تخمین استفاده کرد تعامل همزمان در اتورژهای بردار ساختاری.این مطالعه دو مشکل را که هنگام استفاده از این برآوردگرها در کوچک بوجود می آید ، برجسته می کند نمونه ها.اول ، برآوردهای نادرست از وزن بدون علامت کارآمد ماتریس و واریانس بدون علامت منجر به برآورد فرار و نادرست می شود استنتاجدوم ، بسیاری از شرایط لحظه ای منجر به تعصب مقیاس کوچک نمونه می شود به سمت نوآوری هایی با واریانس کوچکتر از واریانس واحد عادی سازی فرض.برای پرداختن به اولین مشکل ، من با استفاده از فرضیه پیشنهاد می کنم شوکهای ساختاری مستقل برای برآورد ماتریس وزن کارآمد و واریانس برآوردگر.برای شماره دوم ، من پیشنهاد می کنم مدت مقیاس بندی مداوم به روز شده در ماتریس وزنه برداری ، حذف تعصب مقیاسبرای نشان دادن اثربخشی این اقدامات ، من شبیه سازی مونت کارلو که پیشرفت قابل توجهی در عملکرد نشان می دهد از برآوردگر
29,884
A series of standard and penalized logistic regression models is employed to model and forecast the Great Recession and the Covid-19 recession in the US. These two recessions are scrutinized by closely examining the movement of five chosen predictors, their regression coefficients, and the predicted probabilities of recession. The empirical analysis explores the predictive content of numerous macroeconomic and financial indicators with respect to NBER recession indicator. The predictive ability of the underlying models is evaluated using a set of statistical evaluation metrics. The results strongly support the application of penalized logistic regression models in the area of recession prediction. Specifically, the analysis indicates that a mixed usage of different penalized logistic regression models over different forecast horizons largely outperform standard logistic regression models in the prediction of Great recession in the US, as they achieve higher predictive accuracy across 5 different forecast horizons. The Great Recession is largely predictable, whereas the Covid-19 recession remains unpredictable, given that the Covid-19 pandemic is a real exogenous event. The results are validated by constructing via principal component analysis (PCA) on a set of selected variables a recession indicator that suffers less from publication lags and exhibits a very high correlation with the NBER recession indicator.
مجموعه ای از مدل های رگرسیون لجستیک استاندارد و مجازات شده به کار می رود مدل و پیش بینی رکود بزرگ و رکود اقتصادی Covid-19 در ایالات متحده. این دو رکود اقتصادی با بررسی دقیق حرکت پنج مورد بررسی می شوند پیش بینی کننده های منتخب ، ضرایب رگرسیون آنها و پیش بینی شده احتمالات رکود اقتصادی.تجزیه و تحلیل تجربی پیش بینی کننده است محتوای بسیاری از شاخص های کلان اقتصادی و مالی با توجه به NBER شاخص رکود اقتصادی.توانایی پیش بینی مدل های اساسی است ارزیابی با استفاده از مجموعه ای از معیارهای ارزیابی آماری.نتایج به شدت از استفاده از مدلهای رگرسیون لجستیک مجازات شده در منطقه پشتیبانی کنید پیش بینی رکود اقتصادی.به طور خاص ، تجزیه و تحلیل نشان می دهد که یک استفاده مختلط از مدل های رگرسیون لجستیک مجازات مختلف در پیش بینی های مختلف افق تا حد زیادی از مدلهای رگرسیون لجستیک استاندارد در پیش بینی رکود بزرگ در ایالات متحده ، زیرا آنها به پیش بینی بالاتر می رسند دقت در 5 افق مختلف پیش بینی.رکود بزرگ تا حد زیادی است قابل پیش بینی ، در حالی که رکود اقتصادی Covid-19 با توجه به این امر غیرقابل پیش بینی است همه گیر Covid-19 یک رویداد برونزا است.نتایج توسط تأیید می شود ساخت از طریق تجزیه و تحلیل مؤلفه اصلی (PCA) بر روی مجموعه ای از انتخاب شده متغیرها یک شاخص رکود اقتصادی که کمتر از تاخیر در انتشار رنج می برد و همبستگی بسیار بالایی با نشانگر رکود اقتصادی NBER نشان می دهد.
29,885
We propose a regression-based approach to estimate how individuals' expectations influence their responses to a counterfactual change. We provide conditions under which average partial effects based on regression estimates recover structural effects. We propose a practical three-step estimation method that relies on subjective beliefs data. We illustrate our approach in a model of consumption and saving, focusing on the impact of an income tax that not only changes current income but also affects beliefs about future income. Applying our approach to Italian survey data, we find that individuals' beliefs matter for evaluating the impact of tax policies on consumption decisions.
ما یک رویکرد مبتنی بر رگرسیون را برای برآورد چگونگی افراد پیشنهاد می کنیم انتظارات بر پاسخ آنها به یک تغییر خلاف واقع تأثیر می گذارد.ما فراهم می کنیم شرایطی که تحت آن به طور متوسط ​​اثرات جزئی بر اساس برآورد رگرسیون بازیابی اثرات ساختاری.ما یک روش تخمین سه مرحله ای عملی را پیشنهاد می کنیم این به داده های اعتقادات ذهنی متکی است.ما رویکرد خود را در یک مدل نشان می دهیم مصرف و پس انداز ، با تمرکز بر تأثیر مالیات بر درآمد که نه فقط درآمد فعلی را تغییر می دهد بلکه بر اعتقادات مربوط به درآمد آینده نیز تأثیر می گذارد. با استفاده از رویکرد ما به داده های نظرسنجی ایتالیایی ، ما می یابیم که اعتقادات افراد ماده برای ارزیابی تأثیر سیاست های مالیاتی در تصمیمات مصرف.
29,886
Under correlated heterogeneity, the commonly used two-way fixed effects estimator is biased and can lead to misleading inference. This paper proposes a new trimmed mean group (TMG) estimator which is consistent at the irregular rate of n^{1/3} even if the time dimension of the panel is as small as the number of its regressors. Extensions to panels with time effects are provided, and a Hausman-type test of correlated heterogeneity is proposed. Small sample properties of the TMG estimator (with and without time effects) are investigated by Monte Carlo experiments and shown to be satisfactory and perform better than other trimmed estimators proposed in the literature. The proposed test of correlated heterogeneity is also shown to have the correct size and satisfactory power. The utility of the TMG approach is illustrated with an empirical application.
تحت ناهمگونی همبسته ، اثرات ثابت دو طرفه معمولاً استفاده می شود برآوردگر مغرضانه است و می تواند منجر به استنباط گمراه کننده شود.این مقاله پیشنهاد می کند برآوردگر جدید گروه متوسط ​​(TMG) که در نامنظم سازگار است نرخ n^{1/3} حتی اگر بعد زمان پانل به اندازه کوچک باشد تعداد رگرسیون های آن.پسوند به پانل هایی با اثرات زمانی ارائه می شود ، و یک آزمایش از نوع هاوسمان از ناهمگونی همبسته ارائه شده است.نمونه کوچک خواص برآوردگر TMG (با و بدون اثر زمانی) هستند با آزمایش مونت کارلو مورد بررسی قرار گرفته و نشان داده شده است که رضایت بخش است و عملکرد بهتری نسبت به سایر برآوردگرهای اصلاح شده ارائه شده در ادبیات انجام دهید.در آزمون پیشنهادی ناهمگونی همبسته نیز نشان داده شده است که صحیح است اندازه و قدرت رضایت بخش.ابزار رویکرد TMG نشان داده شده است با یک برنامه تجربی
29,887
We combine two recently proposed nonparametric difference-in-differences methods, extending them to enable the examination of treatment effect heterogeneity in the staggered adoption setting using machine learning. The proposed method, machine learning difference-in-differences (MLDID), allows for estimation of time-varying conditional average treatment effects on the treated, which can be used to conduct detailed inference on drivers of treatment effect heterogeneity. We perform simulations to evaluate the performance of MLDID and find that it accurately identifies the true predictors of treatment effect heterogeneity. We then use MLDID to evaluate the heterogeneous impacts of Brazil's Family Health Program on infant mortality, and find those in poverty and urban locations experienced the impact of the policy more quickly than other subgroups.
ما دو تفاوت غیرپارامتری را که اخیراً ارائه شده است ترکیب می کنیم روشها ، گسترش آنها برای بررسی اثر درمانی ناهمگونی در محیط پذیرش مبهم با استفاده از یادگیری ماشین.در روش پیشنهادی ، تفاوت یادگیری ماشین در تمایز (MLDID) ، اجازه می دهد تخمین اثرات متوسط ​​درمانی مشروط با زمان متغیر بر روی تحت درمان ، که می تواند برای انجام استنتاج دقیق در رانندگان استفاده شود ناهمگونی اثر درمانی.ما شبیه سازی ها را برای ارزیابی عملکرد MLDID و دریابید که پیش بینی کننده های واقعی را با دقت مشخص می کند ناهمگونی اثر درمانی.سپس ما از MLDID برای ارزیابی استفاده می کنیم تأثیرات ناهمگن برنامه بهداشت خانواده برزیل بر مرگ و میر نوزادان ، و پیدا کردن آنها در فقر و مکان های شهری ، تأثیر آن را تجربه کرد سیاست سریعتر از سایر زیر گروه ها.
29,888
This paper studies the identification and estimation of a nonparametric nonseparable dyadic model where the structural function and the distribution of the unobservable random terms are assumed to be unknown. The identification and the estimation of the distribution of the unobservable random term are also proposed. I assume that the structural function is continuous and strictly increasing in the unobservable heterogeneity. I propose suitable normalization for the identification by allowing the structural function to have some desirable properties such as homogeneity of degree one in the unobservable random term and some of its observables. The consistency and the asymptotic distribution of the estimators are proposed. The finite sample properties of the proposed estimators in a Monte-Carlo simulation are assessed.
در این مقاله به بررسی شناسایی و تخمین غیر پارامتری می پردازیم مدل dyadic غیرقابل تفکیک که در آن عملکرد ساختاری و توزیع فرض بر این است که اصطلاحات تصادفی غیرقابل کنترل ناشناخته است.شناسایی و تخمین توزیع اصطلاح تصادفی غیرقابل کنترل نیز هست پیشنهاد شده.من فرض می کنم که عملکرد ساختاری مداوم و دقیق است افزایش در ناهمگونی غیرقابل کنترل.من عادی سازی مناسب را پیشنهاد می کنم برای شناسایی با اجازه دادن به عملکرد ساختاری مقداری خواص مطلوب مانند همگن درجه یک در غیرقابل کنترل اصطلاح تصادفی و برخی از مشاهدات آن.قوام و بدون علامت توزیع برآوردگرها پیشنهاد شده است.خواص نمونه محدود از برآوردگرهای پیشنهادی در یک شبیه سازی مونت کارلو ارزیابی می شوند.
29,889
We incorporate a version of a spike and slab prior, comprising a pointmass at zero ("spike") and a Normal distribution around zero ("slab") into a dynamic panel data framework to model coefficient heterogeneity. In addition to homogeneity and full heterogeneity, our specification can also capture sparse heterogeneity, that is, there is a core group of units that share common parameters and a set of deviators with idiosyncratic parameters. We fit a model with unobserved components to income data from the Panel Study of Income Dynamics. We find evidence for sparse heterogeneity for balanced panels composed of individuals with long employment histories.
ما نسخه ای از سنبله و دال را قبل از صفر ("سنبله") و توزیع عادی در حدود صفر ("دال") به یک پویا چارچوب داده پانل برای مدل سازی ناهمگونی ضریب.بعلاوه همگن بودن و ناهمگونی کامل ، مشخصات ما نیز می تواند پراکنده را ضبط کند ناهمگونی ، یعنی یک گروه اصلی از واحدها وجود دارد که مشترک مشترک هستند پارامترها و مجموعه ای از انحراف دهنده ها با پارامترهای ایدیوسنکراتیک.ما یک مدل متناسب هستیم با استفاده از مؤلفه های بدون نظارت به داده های درآمد از مطالعه پانل درآمد پویایی شناسی.ما شواهدی برای ناهمگونی پراکنده برای پانل های متعادل پیدا می کنیم متشکل از افراد دارای سابقه اشتغال طولانی.
29,891
Many empirical applications estimate causal effects of a continuous endogenous variable (treatment) using a binary instrument. Estimation is typically done through linear 2SLS. This approach requires a mean treatment change and causal interpretation requires the LATE-type monotonicity in the first stage. An alternative approach is to explore distributional changes in the treatment, where the first-stage restriction is treatment rank similarity. We propose causal estimands that are doubly robust in that they are valid under either of these two restrictions. We apply the doubly robust estimation to estimate the impacts of sleep on well-being. Our results corroborate the usual 2SLS estimates.
بسیاری از کاربردهای تجربی اثرات علّی مداوم را تخمین می زنند متغیر درون زا (درمان) با استفاده از یک ابزار باینری.تخمین است به طور معمول از طریق 2SL های خطی انجام می شود.این روش نیاز به یک درمان متوسط ​​دارد تغییر و تفسیر علی نیاز به یکنواختی از نوع اواخر در مرحله اول.یک رویکرد جایگزین برای کشف تغییرات توزیع در درمان ، که در آن محدودیت مرحله اول شباهت رتبه درمانی است. ما برآوردهای علی را پیشنهاد می کنیم که به طور مضاعف قوی هستند و از این نظر معتبر هستند هر یک از این دو محدودیت.ما تخمین دو برابر قوی را به تأثیرات خواب بر بهزیستی را تخمین بزنید.نتایج ما معمول را تأیید می کند تخمین 2SLS.
29,892
Livestreaming commerce, a hybrid of e-commerce and self-media, has expanded the broad spectrum of traditional sales performance determinants. To investigate the factors that contribute to the success of livestreaming commerce, we construct a longitudinal firm-level database with 19,175 observations, covering an entire livestreaming subsector. By comparing the forecasting accuracy of eight machine learning models, we identify a random forest model that provides the best prediction of gross merchandise volume (GMV). Furthermore, we utilize explainable artificial intelligence to open the black-box of machine learning model, discovering four new facts: 1) variables representing the popularity of livestreaming events are crucial features in predicting GMV. And voice attributes are more important than appearance; 2) popularity is a major determinant of sales for female hosts, while vocal aesthetics is more decisive for their male counterparts; 3) merits and drawbacks of the voice are not equally valued in the livestreaming market; 4) based on changes of comments, page views and likes, sales growth can be divided into three stages. Finally, we innovatively propose a 3D-SHAP diagram that demonstrates the relationship between predicting feature importance, target variable, and its predictors. This diagram identifies bottlenecks for both beginner and top livestreamers, providing insights into ways to optimize their sales performance.
بازرگانی Livestreaming ، ترکیبی از تجارت الکترونیکی و خود مددی ، گسترش یافته است طیف گسترده ای از عوامل تعیین کننده عملکرد فروش سنتی.به عواملی را که به موفقیت در زمینه زندگی کمک می کند بررسی کنید تجارت ، ما یک پایگاه داده در سطح شرکت طولی با 19،175 می سازیم مشاهدات ، پوشش کامل زیرزمینی را پوشش می دهد.با مقایسه دقت پیش بینی هشت مدل یادگیری ماشین ، ما یک تصادفی را شناسایی می کنیم مدل جنگلی که بهترین پیش بینی از حجم کالاهای ناخالص را ارائه می دهد (GMV).علاوه بر این ، ما از هوش مصنوعی قابل توضیح برای باز کردن استفاده می کنیم جعبه سیاه مدل یادگیری ماشین ، کشف چهار واقعیت جدید: 1) متغیرها نمایندگی از محبوبیت وقایع زنده از ویژگی های مهم در پیش بینی GMV.و ویژگی های صوتی از ظاهر مهمتر هستند.2) محبوبیت تعیین کننده اصلی فروش برای میزبان زن است ، در حالی که آواز زیبایی شناسی برای همتایان مرد خود تعیین کننده تر است.3) شایستگی و اشکالاتی از صدا در بازار زندگی به همان اندازه ارزش ندارد.4) بر اساس تغییرات نظرات ، بازدید و موارد مشابه ، رشد فروش را می توان تقسیم کرد به سه مرحلهسرانجام ، ما به طور نوآورانه یک نمودار سه بعدی را پیشنهاد می کنیم که رابطه بین پیش بینی اهمیت ویژگی ، هدف را نشان می دهد متغیر ، و پیش بینی کننده آن.این نمودار تنگناها را برای هر دو شناسایی می کند مبتدی و برتر Livestreamers ، ارائه بینش در مورد راه هایی برای بهینه سازی آنها عملکرد فروش.
29,893
This paper introduces new techniques for estimating, identifying and simulating mixed causal-noncausal invertible-noninvertible models. We propose a framework that integrates high-order cumulants, merging both the spectrum and bispectrum into a single estimation function. The model that most adequately represents the data under the assumption that the error term is i.i.d. is selected. Our Monte Carlo study reveals unbiased parameter estimates and a high frequency with which correct models are identified. We illustrate our strategy through an empirical analysis of returns from 24 Fama-French emerging market stock portfolios. The findings suggest that each portfolio displays noncausal dynamics, producing white noise residuals devoid of conditional heteroscedastic effects.
در این مقاله تکنیک های جدیدی برای برآورد ، شناسایی و شبیه سازی مدلهای غیر قابل برگشت غیر قابل برگشت علی و علی.ما پیشنهاد می کنیم چارچوبی که مقادیر مرتبه بالا را ادغام می کند ، هر دو طیف را ادغام می کند bispectrum در یک عملکرد تخمین واحد.مدلی که به اندازه کافی داده ها را تحت این فرض نشان می دهد که اصطلاح خطا i.i.d است.است انتخاب شد.مطالعه مونت کارلو ما برآورد پارامتر بی طرفانه و زیاد را نشان می دهد فرکانس که مدل های صحیح با آنها مشخص می شوند.ما استراتژی خود را نشان می دهیم از طریق تجزیه و تحلیل تجربی از بازده از 24 بازار در حال ظهور Fama-French اوراق بهادار سهام.این یافته ها حاکی از آن است که هر نمونه کارها به طور غیرحضوری نمایش می دهد دینامیک ، تولید باقیمانده های سر و صدای سفید عاری از هتروسکوپی مشروط اثرات
29,895
This paper discusses the partial identification of treatment effects in sample selection models when the exclusion restriction fails and the monotonicity assumption in the selection effect does not hold exactly, both of which are key challenges in applying the existing methodologies. Our approach builds on Lee's (2009) procedure, who considers partial identification under the monotonicity assumption, but we assume only a stochastic (and weaker) version of monotonicity, which depends on a prespecified parameter $\vartheta$ that represents researchers' belief in the plausibility of the monotonicity. Under this assumption, we show that we can still obtain useful bounds even when the monotonic behavioral model does not strictly hold. Our procedure is useful when empirical researchers anticipate that a small fraction of the population will not behave monotonically in selection; it can also be an effective tool for performing sensitivity analysis or examining the identification power of the monotonicity assumption. Our procedure is easily extendable to other related settings; we also provide the identification result of the marginal treatment effects setting as an important application. Moreover, we show that the bounds can still be obtained even in the absence of the knowledge of $\vartheta$ under the semiparametric models that nest the classical probit and logit selection models.
در این مقاله به شناسایی جزئی اثرات درمانی در مدل های انتخاب نمونه هنگامی که محدودیت محرومیت از بین می رود و فرض یکنواختی در اثر انتخاب دقیقاً وجود ندارد ، هر دو که در استفاده از روشهای موجود چالش های اساسی هستند.رویکرد ما بر اساس روش لی (2009) ، که شناسایی جزئی را در زیر می داند فرض یکنواختی ، اما ما فقط یک تصادفی (و ضعیف تر) فرض می کنیم نسخه یکنواختی ، که بستگی به یک پارامتر از پیش تعیین شده $ \ vartheta $ دارد این نشان دهنده اعتقاد محققان به محتمل بودن یکنواختی است. تحت این فرض ، ما نشان می دهیم که هنوز هم می توانیم مرزهای مفیدی را بدست آوریم مدل رفتاری یکنواخت به شدت نگهداری نمی شود.روش ما مفید است هنگامی که محققان تجربی پیش بینی می کنند که بخش کوچکی از جمعیت در انتخاب یکنواخت رفتار نخواهد کرد.همچنین می تواند یک ابزار مؤثر باشد برای انجام تجزیه و تحلیل حساسیت یا بررسی قدرت شناسایی فرض یکنواختی.روش ما به راحتی قابل گسترش است تنظیمات مرتبط ؛ما همچنین نتیجه شناسایی حاشیه را ارائه می دهیم تنظیم اثرات درمانی به عنوان یک کاربرد مهم.علاوه بر این ، ما این را نشان می دهیم مرزها حتی در صورت عدم دانش نیز می توانند به دست بیایند $ \ vartheta $ تحت مدلهای نیمهرامتری که پروب کلاسیک را لانه می کنند و مدل های انتخاب ورود به سیستم.
29,896
This paper studies quasi Bayesian estimation and uncertainty quantification for an unknown function that is identified by a nonparametric conditional moment restriction. We derive contraction rates for a class of Gaussian process priors. Furthermore, we provide conditions under which a Bernstein von Mises theorem holds for the quasi-posterior distribution. As a consequence, we show that optimally weighted quasi-Bayes credible sets have exact asymptotic frequentist coverage.
در این مقاله برآورد شبه بیزی و کمیت عدم اطمینان بررسی شده است برای یک عملکرد ناشناخته که توسط یک شرط غیرپارامتری مشخص می شود محدودیت لحظه ای.ما نرخ انقباض را برای یک کلاس از روند گاوسی استخراج می کنیم مقدماتعلاوه بر این ، ما شرایطی را فراهم می کنیم که تحت آن یک برنشتاین فون میزس قضیه برای توزیع شبه خلفی در اختیار دارد.در نتیجه ، ما نشان می دهیم این مجموعه های معتبر شبه با وزن بهینه وزن دقیقاً بدون علامت دارند پوشش مکرر.
29,897
Can temporary subsidies to bundles induce long-run changes in demand due to learning about the relative quality of one of its constituent goods? This paper provides theoretical and experimental evidence on the role of this mechanism. Theoretically, we introduce a model where an agent learns about the quality of an innovation on an essential good through consumption. Our results show that the contemporaneous effect of a one-off subsidy to a bundle that contains the innovation may be decomposed into a direct price effect, and an indirect learning motive, whereby an agent leverages the discount to increase the informational bequest left to her future selves. We then assess the predictions of our theory in a randomised experiment in a ridesharing platform. The experiment provided two-week discounts for car trips integrating with a train or metro station (a bundle). Given the heavy-tailed nature of our data, we follow \cite{Athey2023} and, motivated by our theory, propose a semiparametric model for treatment effects that enables the construction of more efficient estimators. We introduce a statistically efficient estimator for our model by relying on L-moments, a robust alternative to standard moments. Our estimator immediately yields a specification test for the semiparametric model; moreover, in our adopted parametrisation, it can be easily computed through generalized least squares. Our empirical results indicate that a two-week 50\% discount on car trips integrating with train/metro leads to a contemporaneous increase in the demand for integrated rides, and, consistent with our learning model, persistent changes in the mean and dispersion of nonintegrated rides. These effects persist for over four months after the discount. A simple calibration of our model shows that around 40\% to 50\% of the estimated contemporaneous increase in integrated rides may be attributed to a learning motive.
آیا یارانه های موقت به بسته ها می تواند تغییرات طولانی مدت در تقاضا را به دلیل یادگیری در مورد کیفیت نسبی یکی از کالاهای تشکیل دهنده آن؟این کاغذ شواهد نظری و تجربی در مورد نقش این مکانیسم ارائه می دهد. از لحاظ تئوریکی ، ما مدلی را معرفی می کنیم که یک عامل در مورد کیفیت آن می آموزد نوآوری در مورد یک کالای اساسی از طریق مصرف.نتایج ما نشان می دهد که اثر معاصر یارانه یک طرفه به یک بسته نرم افزاری که حاوی نوآوری ممکن است به یک اثر قیمت مستقیم و یک غیرمستقیم تجزیه شود انگیزه یادگیری ، به موجب آن یک نماینده تخفیف را افزایش می دهد تا افزایش یابد Bequest اطلاعاتی که به آینده خود باقی مانده است.سپس پیش بینی ها را ارزیابی می کنیم نظریه ما در یک آزمایش تصادفی در یک بستر تحریک کننده.در آزمایش تخفیف های دو هفته ای برای سفرهای اتومبیل با قطار ارائه داد یا ایستگاه مترو (یک بسته نرم افزاری).با توجه به ماهیت سنگین داده های ما ، ما \ cite {athey2023} را دنبال کنید و با انگیزه نظریه ما ، یک نیمه پارامتری پیشنهاد دهید مدل برای اثرات درمانی که امکان ساخت کارآمدتر را فراهم می کند برآوردگرهاما یک برآوردگر از نظر آماری کارآمد برای مدل خود معرفی می کنیم با تکیه بر لحظه های L ، یک جایگزین قوی برای لحظه های استاندارد.برآوردگر ما بلافاصله یک تست مشخصات برای مدل نیمه پارامتری ارائه می دهد.علاوه بر این، در پارامتر سازی اتخاذ شده ما ، می توان آن را به راحتی از طریق تعمیم محاسبه کرد کمترین مربعات.نتایج تجربی ما نشان می دهد که تخفیف دو هفته 50 \ ٪ سفرهای اتومبیل با قطار/مترو منجر به افزایش معاصر در تقاضا برای سواری یکپارچه ، و مطابق با مدل یادگیری ما ، تغییرات مداوم در میانگین و پراکندگی سواری های غیر یکپارچه.اینها اثرات بیش از چهار ماه پس از تخفیف ادامه دارد.یک کالیبراسیون ساده از مدل ما نشان می دهد که حدود 40 \ ٪ تا 50 \ ٪ تخمین زده شده معاصر افزایش در سواری های یکپارچه ممکن است به یک انگیزه یادگیری نسبت داده شود.
29,898
Using a transformation of the autoregressive distributed lag model due to Bewley, a novel pooled Bewley (PB) estimator of long-run coefficients for dynamic panels with heterogeneous short-run dynamics is proposed. The PB estimator is directly comparable to the widely used Pooled Mean Group (PMG) estimator, and is shown to be consistent and asymptotically normal. Monte Carlo simulations show good small sample performance of PB compared to the existing estimators in the literature, namely PMG, panel dynamic OLS (PDOLS), and panel fully-modified OLS (FMOLS). Application of two bias-correction methods and a bootstrapping of critical values to conduct inference robust to cross-sectional dependence of errors are also considered. The utility of the PB estimator is illustrated in an empirical application to the aggregate consumption function.
با استفاده از تغییر مدل تاخیر توزیع شده خودجوش به دلیل Bewley ، یک برآوردگر رمان تلفیقی Bewley (PB) از ضرایب طولانی مدت برای پانل های پویا با دینامیک کوتاه ناهمگن ارائه شده است.PB برآوردگر به طور مستقیم با گروه میانگین جمع شده استفاده شده قابل مقایسه است (PMG) برآوردگر ، و نشان داده می شود که سازگار و بدون علامت طبیعی است.مونت کارلو شبیه سازی ها عملکرد نمونه کوچک PB را در مقایسه با موجود نشان می دهد برآوردگرها در ادبیات ، یعنی PMG ، پانل پویا (PDOL) و پانل OLS کاملاً اصلاح شده (fmols).کاربرد دو روش اصلاح تعصب و الف بوت استرپینگ مقادیر بحرانی برای انجام استنتاج قوی به مقطعی وابستگی خطاها نیز در نظر گرفته شده است.ابزار برآوردگر PB است در یک برنامه تجربی به عملکرد مصرف کل نشان داده شده است.
29,899
In this paper, we consider estimation and inference for both the multi-index parameters and the link function involved in a class of semiparametric multi-index models via deep neural networks (DNNs). We contribute to the design of DNN by i) providing more transparency for practical implementation, ii) defining different types of sparsity, iii) showing the differentiability, iv) pointing out the set of effective parameters, and v) offering a new variant of rectified linear activation function (ReLU), etc. Asymptotic properties for the joint estimates of both the index parameters and the link functions are established, and a feasible procedure for the purpose of inference is also proposed. We conduct extensive numerical studies to examine the finite-sample performance of the estimation methods, and we also evaluate the empirical relevance and applicability of the proposed models and estimation methods to real data.
در این مقاله ، ما تخمین و استنباط را برای هر دو شاخص در نظر می گیریم پارامترها و عملکرد لینک درگیر در یک کلاس نیمه قطبی مدل های چند شاخص از طریق شبکه های عصبی عمیق (DNNS).ما به طراحی کمک می کنیم DNN توسط i) ارائه شفافیت بیشتر برای اجرای عملی ، ii) تعریف انواع مختلف پراکندگی ، iii) نشان دادن متفاوت بودن ، IV) با اشاره به مجموعه پارامترهای مؤثر ، و v) ارائه نوع جدیدی از عملکرد فعال سازی خطی اصلاح شده (RELU) ، و غیره. خصوصیات بدون علامت برای تخمین های مشترک از هر دو پارامتر شاخص و توابع پیوند تأسیس شده و یک روش عملی نیز به منظور استنباط نیز هست پیشنهاد شده.ما مطالعات عددی گسترده ای را برای بررسی نمونه محدود انجام می دهیم عملکرد روشهای تخمین ، و ما همچنین تجربی را ارزیابی می کنیم ارتباط و کاربرد مدلهای پیشنهادی و روشهای تخمین به داده های واقعی
29,900
This survey study discusses main aspects to optimal estimation methodologies for panel data regression models. In particular, we present current methodological developments for modeling stationary panel data as well as robust methods for estimation and inference in nonstationary panel data regression models. Some applications from the network econometrics and high dimensional statistics literature are also discussed within a stationary time series environment.
این مطالعه نظرسنجی جنبه های اصلی روشهای برآورد بهینه را مورد بحث قرار می دهد برای مدل های رگرسیون داده پانل.به طور خاص ، ما فعلی را ارائه می دهیم تحولات روش شناختی برای مدل سازی داده های پانل ثابت و همچنین روشهای قوی برای برآورد و استنباط در داده های پانل غیر ایستگاه مدل های رگرسیون.برخی از برنامه های کاربردی از اقتصاد سنجی شبکه و زیاد ادبیات آمار بعدی نیز در یک زمان ثابت مورد بحث قرار می گیرد محیط سری.
29,901
Naive maximum likelihood estimation of binary logit models with fixed effects leads to unreliable inference due to the incidental parameter problem. We study the case of three-dimensional panel data, where the model includes three sets of additive and overlapping unobserved effects. This encompasses models for network panel data, where senders and receivers maintain bilateral relationships over time, and fixed effects account for unobserved heterogeneity at the sender-time, receiver-time, and sender-receiver levels. In an asymptotic framework, where all three panel dimensions grow large at constant relative rates, we characterize the leading bias of the naive estimator. The inference problem we identify is particularly severe, as it is not possible to balance the order of the bias and the standard deviation. As a consequence, the naive estimator has a degenerating asymptotic distribution, which exacerbates the inference problem relative to other fixed effects estimators studied in the literature. To resolve the inference problem, we derive explicit expressions to debias the fixed effects estimator.
برآورد حداکثر احتمال زیاد مدل های ورود به سیستم باینری با اثرات ثابت به دلیل مشکل پارامتر حادثه منجر به استنباط غیرقابل اعتماد می شود.ما مطالعه میکنیم مورد داده های پانل سه بعدی ، که در آن مدل شامل سه مجموعه است اثرات افزودنی و همپوشانی بدون نظارت.این شامل مدل ها برای داده های پنل شبکه ، جایی که فرستنده ها و گیرنده ها دو طرفه را حفظ می کنند روابط با گذشت زمان ، و اثرات ثابت ناهمگونی بی نظیر را به خود اختصاص می دهد در سطح فرستنده ، زمان گیرنده و گیرنده فرستنده.در یک بدون علامت چارچوب ، جایی که هر سه بعد پانل در نسبی ثابت رشد می کنند نرخ ها ، ما تعصب پیشرو برآوردگر ساده لوح را توصیف می کنیم.استنباط مشکلی که ما شناسایی می کنیم به ویژه شدید است ، زیرا تعادل امکان پذیر نیست ترتیب تعصب و انحراف استاندارد.در نتیجه ، ساده لوح برآوردگر دارای توزیع بدون علامت انحطاط است ، که این را تشدید می کند مشکل استنباط نسبت به سایر برآوردگرهای اثرات ثابت که در آن مورد مطالعه قرار گرفته است ادبیات.برای حل مشکل استنتاج ، ما عبارات صریح را به دست می آوریم Debias برآوردگر جلوه های ثابت.
29,902
This paper considers the estimation of treatment effects in randomized experiments with complex experimental designs, including cases with interference between units. We develop a design-based estimation theory for arbitrary experimental designs. Our theory facilitates the analysis of many design-estimator pairs that researchers commonly employ in practice and provide procedures to consistently estimate asymptotic variance bounds. We propose new classes of estimators with favorable asymptotic properties from a design-based point of view. In addition, we propose a scalar measure of experimental complexity which can be linked to the design-based variance of the estimators. We demonstrate the performance of our estimators using simulated datasets based on an actual network experiment studying the effect of social networks on insurance adoptions.
در این مقاله تخمین اثرات درمانی به صورت تصادفی در نظر گرفته شده است آزمایش با طرح های آزمایشی پیچیده ، از جمله مواردی تداخل بین واحدها.ما یک تئوری تخمین مبتنی بر طراحی برای طرح های آزمایشی دلخواه.نظریه ما تجزیه و تحلیل بسیاری را تسهیل می کند جفت های تخمین طراحی که محققان معمولاً در عمل از آنها استفاده می کنند و ارائه می دهند روشها برای برآورد مداوم مرزهای واریانس بدون علامت.ما جدید پیشنهاد می کنیم کلاس های برآوردگرها با خصوصیات بدون علامت مطلوب از یک طراحی مبتنی بر نقطه نظر.علاوه بر این ، ما یک اندازه گیری مقیاس آزمایشی را پیشنهاد می کنیم پیچیدگی که می تواند با واریانس مبتنی بر طراحی برآوردگرها مرتبط باشد. ما عملکرد برآوردگرهای خود را با استفاده از مجموعه داده های شبیه سازی شده نشان می دهیم در یک آزمایش شبکه واقعی که تأثیر شبکه های اجتماعی را در آن مطالعه می کند تصویب بیمه.
29,915
We demonstrate that regression models can be estimated by working independently in a row-wise fashion. We document a simple procedure which allows for a wide class of econometric estimators to be implemented cumulatively, where, in the limit, estimators can be produced without ever storing more than a single line of data in a computer's memory. This result is useful in understanding the mechanics of many common regression models. These procedures can be used to speed up the computation of estimates computed via OLS, IV, Ridge regression, LASSO, Elastic Net, and Non-linear models including probit and logit, with all common modes of inference. This has implications for estimation and inference with `big data', where memory constraints may imply that working with all data at once is particularly costly. We additionally show that even with moderately sized datasets, this method can reduce computation time compared with traditional estimation routines.
ما نشان می دهیم که مدل های رگرسیون را می توان با کار تخمین زد به طور مستقل با روشی ردیف.ما یک روش ساده را مستند می کنیم که اجازه می دهد تا یک طبقه گسترده از برآوردگرهای اقتصاد سنجی اجرا شود تجمعی ، جایی که ، در حد مجاز ، برآوردگرها می توانند بدون هیچ وقت تولید شوند ذخیره بیش از یک خط از داده ها در حافظه رایانه.این نتیجه است در درک مکانیک بسیاری از مدلهای رگرسیون مشترک مفید است.اینها از روشها می توان برای سرعت بخشیدن به محاسبه تخمین های محاسبه شده از طریق استفاده کرد OLS ، IV ، رگرسیون ریج ، لاسو ، شبکه الاستیک و مدل های غیر خطی از جمله پروانه و ورود به سیستم ، با تمام روشهای مشترک استنباط.این پیامدهایی برای تخمین و استنباط با "داده های بزرگ" ، جایی که محدودیت های حافظه ممکن است دلالت کند این که کار با همه داده ها به طور همزمان بسیار پر هزینه است.ما علاوه بر این نشان می دهیم که حتی با مجموعه داده های متوسط ​​، این روش می تواند محاسبه را کاهش دهد زمان در مقایسه با روال تخمین سنتی.
29,903
This paper proposes a new method for estimating high-dimensional binary choice models. The model we consider is semiparametric, placing no distributional assumptions on the error term, allowing for heteroskedastic errors, and permitting endogenous regressors. Our proposed approaches extend the special regressor estimator originally proposed by Lewbel (2000). This estimator becomes impractical in high-dimensional settings due to the curse of dimensionality associated with high-dimensional conditional density estimation. To overcome this challenge, we introduce an innovative data-driven dimension reduction method for nonparametric kernel estimators, which constitutes the main innovation of this work. The method combines distance covariance-based screening with cross-validation (CV) procedures, rendering the special regressor estimation feasible in high dimensions. Using the new feasible conditional density estimator, we address the variable and moment (instrumental variable) selection problems for these models. We apply penalized least squares (LS) and Generalized Method of Moments (GMM) estimators with a smoothly clipped absolute deviation (SCAD) penalty. A comprehensive analysis of the oracle and asymptotic properties of these estimators is provided. Monte Carlo simulations are employed to demonstrate the effectiveness of our proposed procedures in finite sample scenarios.
در این مقاله یک روش جدید برای برآورد باینری با ابعاد بالا پیشنهاد شده است مدل های انتخاب.مدلی که ما در نظر می گیریم نیمه پارامتری است و NO را قرار می دهد فرضیات توزیع در اصطلاح خطا ، امکان هتروسیکاستیک را فراهم می کند خطاها ، و اجازه رگرسیون درون زا.رویکردهای پیشنهادی ما گسترش می یابد برآوردگر ویژه رگرسیون که در ابتدا توسط Lewbel (2000) پیشنهاد شده است.این برآوردگر در تنظیمات با ابعاد بالا به دلیل لعنت از ابعاد مرتبط با برآورد چگالی شرطی با ابعاد بالا. برای غلبه بر این چالش ، ما یک بعد داده مبتنی بر داده را معرفی می کنیم روش کاهش برای برآوردگرهای هسته غیر پارامتری ، که تشکیل دهنده آن است نوآوری اصلی این کار.این روش ترکیبی از مسافت کواریانس است غربالگری با روشهای معتبر متقاطع (CV) ، ارائه ویژه تخمین رگرسیون در ابعاد بالا امکان پذیر است.با استفاده از امکان پذیر جدید برآوردگر چگالی مشروط ، ما متغیر و لحظه را مورد بررسی قرار می دهیم (ابزاری متغیر) مشکلات انتخاب برای این مدل ها.ما حداقل مربعات مجازات شده را اعمال می کنیم (LS) و روش کلی برآوردگرهای لحظاتی (GMM) با یک هموار قطع شده مجازات انحراف مطلق (SCAD).تجزیه و تحلیل جامع از اوراکل و خصوصیات بدون علامت این برآوردگرها ارائه شده است.شبیه سازی مونت کارلو برای نشان دادن اثربخشی رویه های پیشنهادی ما در سناریوهای نمونه محدود.
29,904
This paper develops an asymptotic distribution theory for a two-stage instrumentation estimation approach in quantile predictive regressions when both generated covariates and persistent predictors are used. The generated covariates are obtained from an auxiliary quantile regression model and our main interest is the robust estimation and inference of the primary quantile predictive regression in which this generated covariate is added to the set of nonstationary regressors. We find that the proposed doubly IVX estimator is robust to the abstract degree of persistence regardless of the presence of generated regressor obtained from the first stage procedure. The asymptotic properties of the two-stage IVX estimator such as mixed Gaussianity are established while the asymptotic covariance matrix is adjusted to account for the first-step estimation error.
این مقاله یک تئوری توزیع بدون علامت برای یک مرحله ای ایجاد می کند رویکرد تخمین ابزار دقیق در رگرسیون پیش بینی کننده کمی هر دو متغیر متغیر و پیش بینی کننده مداوم استفاده می شود.تولید شده متغیرهای متغیر از یک مدل رگرسیون کمکی کمکی و ما به دست می آیند علاقه اصلی برآورد قوی و استنباط مقادیر اولیه است رگرسیون پیش بینی کننده ای که در آن این همبستگی تولید شده به مجموعه اضافه می شود رگرسیون های غیر ایستگاه.ما می دانیم که برآوردگر دو برابر پیشنهادی IVX است محکم به درجه انتزاعی پایداری صرف نظر از حضور رگرسیون تولید شده به دست آمده از روش مرحله اول.بدون علامت خواص برآوردگر دو مرحله ای IVX مانند Gaussianity مختلط هستند در حالی که ماتریس کواریانس بدون علامت تنظیم شده است به حساب می آید خطای تخمین مرحله اول.
29,905
Policymakers often desire a statistical treatment rule (STR) that determines a treatment assignment rule deployed in a future population from available data. With the true knowledge of the data generating process, the average treatment effect (ATE) is the key quantity characterizing the optimal treatment rule. Unfortunately, the ATE is often not point identified but partially identified. Presuming the partial identification of the ATE, this study conducts a local asymptotic analysis and develops the locally asymptotically minimax (LAM) STR. The analysis does not assume the full differentiability but the directional differentiability of the boundary functions of the identification region of the ATE. Accordingly, the study shows that the LAM STR differs from the plug-in STR. A simulation study also demonstrates that the LAM STR outperforms the plug-in STR.
سیاستگذاران غالباً خواستار یک قانون درمانی آماری (STR) هستند که تعیین می کند یک قانون واگذاری درمانی که در یک جمعیت آینده از موجود مستقر شده است داده ها.با دانش واقعی فرآیند تولید داده ها ، میانگین اثر درمانی (ATE) مقدار اصلی توصیف درمان بهینه است قانون.متأسفانه ، ATE اغلب مشخص نمی شود بلکه جزئی است شناخته شده است.با فرض شناسایی جزئی از ATE ، این مطالعه یک تجزیه و تحلیل بدون علامت محلی انجام می دهد و به صورت مجانبی محلی توسعه می یابد Minimax (LAM) STR.تجزیه و تحلیل تفاوت کامل را فرض نمی کند اما متفاوت بودن جهت عملکردهای مرزی منطقه شناسایی ATE.بر این اساس ، مطالعه نشان می دهد که Lam Str با خیابان افزونه متفاوت است.یک مطالعه شبیه سازی همچنین نشان می دهد که LAM Str از strong in str.
29,906
This study investigates the problem of individualizing treatment allocations using stated preferences for treatments. If individuals know in advance how the assignment will be individualized based on their stated preferences, they may state false preferences. We derive an individualized treatment rule (ITR) that maximizes welfare when individuals strategically state their preferences. We also show that the optimal ITR is strategy-proof, that is, individuals do not have a strong incentive to lie even if they know the optimal ITR a priori. Constructing the optimal ITR requires information on the distribution of true preferences and the average treatment effect conditioned on true preferences. In practice, the information must be identified and estimated from the data. As true preferences are hidden information, the identification is not straightforward. We discuss two experimental designs that allow the identification: strictly strategy-proof randomized controlled trials and doubly randomized preference trials. Under the presumption that data comes from one of these experiments, we develop data-dependent procedures for determining ITR, that is, statistical treatment rules (STRs). The maximum regret of the proposed STRs converges to zero at a rate of the square root of the sample size. An empirical application demonstrates our proposed STRs.
این مطالعه به بررسی مشکل شخصی سازی تخصیص های درمانی می پردازد با استفاده از ترجیحات بیان شده برای درمان.اگر افراد از قبل بدانند چگونه تکالیف بر اساس ترجیحات بیان شده آنها ، ممکن است شخصی شود ترجیحات دروغین را بیان کنید.ما یک قانون درمانی فردی (ITR) را به دست می آوریم که هنگامی که افراد از نظر استراتژیک ترجیحات خود را بیان می کنند ، رفاه را به حداکثر می رساند.ما همچنین نشان می دهد که ITR بهینه ضد استراتژی است ، یعنی افراد این کار را نمی کنند انگیزه محکمی برای دروغ گفتن داشته باشید حتی اگر آنها پیشینی بهینه را بدانند. ساخت ITR بهینه نیاز به اطلاعات در مورد توزیع TRUE دارد ترجیحات و میانگین اثر درمانی که مربوط به ترجیحات واقعی است. در عمل ، اطلاعات باید از داده ها شناسایی و برآورد شوند.مانند ترجیحات واقعی اطلاعات پنهان است ، شناسایی نیست مستقیمما در مورد دو طرح آزمایشی بحث می کنیم که اجازه می دهد شناسایی: کارآزمایی های کنترل شده تصادفی کاملاً ضد استراتژی و دو برابر کارآزمایی ترجیح تصادفی.با این فرض که داده ها از یکی از آنها تهیه شده است این آزمایشات ، ما روشهای وابسته به داده را برای تعیین ITR توسعه می دهیم ، یعنی قوانین درمانی آماری (STR).حداکثر پشیمانی پیشنهادی Strs با سرعت ریشه مربع اندازه نمونه به صفر همگرا می شود.در کاربرد تجربی STR های پیشنهادی ما را نشان می دهد.
29,907
Many causal parameters depend on a moment of the joint distribution of potential outcomes. Such parameters are especially relevant in policy evaluation settings, where noncompliance is common and accommodated through the model of Imbens & Angrist (1994). This paper shows that the sharp identified set for these parameters is an interval with endpoints characterized by the value of optimal transport problems. Sample analogue estimators are proposed based on the dual problem of optimal transport. These estimators are root-n consistent and converge in distribution under mild assumptions. Inference procedures based on the bootstrap are straightforward and computationally convenient. The ideas and estimators are demonstrated in an application revisiting the National Supported Work Demonstration job training program. I find suggestive evidence that workers who would see below average earnings without treatment tend to see above average benefits from treatment.
بسیاری از پارامترهای علی به لحظه توزیع مشترک بستگی دارد نتایج بالقوهچنین پارامترهایی به ویژه در خط مشی مرتبط هستند تنظیمات ارزیابی ، جایی که عدم انطباق از طریق مدل Imbens & Angrist (1994).این مقاله نشان می دهد که تیز مشخص شده است تنظیم برای این پارامترها یک بازه با نقاط پایانی است که توسط ارزش مشکلات بهینه حمل و نقل.برآوردگرهای آنالوگ نمونه پیشنهاد شده است بر اساس مشکل دوگانه حمل و نقل بهینه.این برآوردگرها root-n هستند سازگار و همگرا در توزیع تحت فرضیات خفیف.استنباط رویه های مبتنی بر بوت استرپ ساده و محاسباتی است راحت.ایده ها و برآوردگرها در یک برنامه نشان داده می شوند بازنگری در برنامه آموزش شغلی تظاهرات کار پشتیبانی ملی.من شواهد پیشنهادی را پیدا کنید که کارگرانی که زیر درآمد متوسط ​​را می بینند بدون درمان تمایل به افزایش مزایای متوسط ​​از درمان دارد.
29,908
This paper considers inference in first-price and second-price sealed-bid auctions with a large number of symmetric bidders having independent private values. Given the abundance of bidders in each auction, we propose an asymptotic framework in which the number of bidders diverges while the number of auctions remains fixed. This framework allows us to perform asymptotically exact inference on key model features using only transaction price data. Specifically, we examine inference on the expected utility of the auction winner, the expected revenue of the seller, and the tail properties of the valuation distribution. Simulations confirm the accuracy of our inference methods in finite samples. Finally, we also apply them to Hong Kong car license auction data.
در این مقاله استنباط در قیمت اول و قیمت دوم با قیمت بسته بندی شده در نظر گرفته شده است حراج هایی با تعداد زیادی از داوطلبان متقارن که دارای خصوصی مستقل هستند ارزش های.با توجه به فراوانی داوطلبان در هر حراج ، ما پیشنهاد می کنیم چارچوب بدون علامت که در آن تعداد داوطلبان در حالی که تعداد آنها متفاوت است حراج ها ثابت است.این چارچوب به ما امکان می دهد بدون علامت انجام دهیم استنتاج دقیق در ویژگی های مدل کلیدی با استفاده از فقط داده های قیمت معاملات. به طور خاص ، ما استنباط در مورد ابزار مورد انتظار حراج را بررسی می کنیم برنده ، درآمد مورد انتظار فروشنده و خواص دم توزیع ارزیابی.شبیه سازی صحت استنباط ما را تأیید می کند روشهای موجود در نمونه های محدود.سرانجام ، ما آنها را نیز در مجوز اتومبیل هنگ کنگ اعمال می کنیم داده های حراج
29,909
Time-Varying Parameters Vector Autoregressive (TVP-VAR) models are frequently used in economics to capture evolving relationships among the macroeconomic variables. However, TVP-VARs have the tendency of overfitting the data, resulting in inaccurate forecasts and imprecise estimates of typical objects of interests such as the impulse response functions. This paper introduces a Theory Coherent Time-Varying Parameters Vector Autoregressive Model (TC-TVP-VAR), which leverages on an arbitrary theoretical framework derived by an underlying economic theory to form a prior for the time varying parameters. This "theory coherent" shrinkage prior significantly improves inference precision and forecast accuracy over the standard TVP-VAR. Furthermore, the TC-TVP-VAR can be used to perform indirect posterior inference on the deep parameters of the underlying economic theory. The paper reveals that using the classical 3-equation New Keynesian block to form a prior for the TVP- VAR substantially enhances forecast accuracy of output growth and of the inflation rate in a standard model of monetary policy. Additionally, the paper shows that the TC-TVP-VAR can be used to address the inferential challenges during the Zero Lower Bound period.
پارامترهای متغیر زمان وکتور وکتور خودکار (TVP-VAR) اغلب در اقتصاد برای ضبط روابط در حال تحول بین اقتصاد کلان استفاده می شود متغیرهابا این حال ، VAR های TVP تمایل به بیش از حد داده ها دارند ، منجر به پیش بینی های نادرست و تخمین های نادرست از اشیاء معمولی علایق مانند توابع پاسخ Impulse.در این مقاله تئوری پارامترهای متغیر زمان منسجم مدل بردار اتورگرایی مدل (TC-TVP-VAR) ، که در یک چارچوب نظری دلخواه به دست می آید یک تئوری اقتصادی اساسی برای تشکیل یک پارامترهای مختلف در زمان. این انقباض "تئوری منسجم" قبل از آن به طور قابل توجهی استنباط را بهبود می بخشد دقت و دقت پیش بینی نسبت به TVP-VAR استاندارد.علاوه بر این ، TC-TVP-VAR می تواند برای انجام استنباط خلفی غیرمستقیم در اعماق استفاده شود پارامترهای تئوری اقتصادی اساسی.مقاله نشان می دهد که استفاده از 3 عادلانه بلوک کینزی جدید برای تشکیل قبلی برای TVP- var به طور قابل توجهی دقت پیش بینی رشد خروجی و تورم را افزایش می دهد نرخ در یک مدل استاندارد از سیاست پولی.علاوه بر این ، مقاله نشان می دهد که TC-TVP-VAR می تواند برای رفع چالش های استنباطی در طول استفاده شود دوره محدود پایین صفر.
29,910
We analyze the synthetic control (SC) method in panel data settings with many units. We assume the treatment assignment is based on unobserved heterogeneity and pre-treatment information, allowing for both strictly and sequentially exogenous assignment processes. We show that the critical property that determines the behavior of the SC method is the ability of input features to approximate the unobserved heterogeneity. Our results imply that the SC method delivers asymptotically normal estimators for a large class of linear panel data models as long as the number of pre-treatment periods is sufficiently large, making it a natural alternative to the Difference-in-Differences.
ما روش کنترل مصنوعی (SC) را در تنظیمات داده پانل با بسیاری از آنها تجزیه و تحلیل می کنیم واحدهاما فرض می کنیم که تکلیف درمانی مبتنی بر ناهمگونی بی نظیر است و اطلاعات قبل از درمان ، هر دو به طور دقیق و متوالی امکان پذیر است فرآیندهای واگذاری اگزوژن.ما نشان می دهیم که خاصیت بحرانی که تعیین رفتار روش SC توانایی ویژگی های ورودی به ناهمگونی بدون محافظت را تقریبی کنید.نتایج ما حاکی از آن است که روش SC برآوردگرهای بدون علامت طبیعی را برای یک کلاس بزرگ از پنل خطی ارائه می دهد مدل های داده تا زمانی که تعداد دوره های قبل از درمان به اندازه کافی باشد بزرگ ، آن را به یک جایگزین طبیعی برای تفاوت در اختلافات تبدیل می کند.
29,911
This paper presents new econometric tools to unpack the treatment effect heterogeneity of punishing misdemeanor offenses on time-to-recidivism. We show how one can identify, estimate, and make inferences on the distributional, quantile, and average marginal treatment effects in setups where the treatment selection is endogenous and the outcome of interest, usually a duration variable, is potentially right-censored. We explore our proposed econometric methodology to evaluate the effect of fines and community service sentences as a form of punishment on time-to-recidivism in the State of S\~ao Paulo, Brazil, between 2010 and 2019, leveraging the as-if random assignment of judges to cases. Our results highlight substantial treatment effect heterogeneity that other tools are not meant to capture. For instance, we find that people whom most judges would punish take longer to recidivate as a consequence of the punishment, while people who would be punished only by strict judges recidivate at an earlier date than if they were not punished. This result suggests that designing sentencing guidelines that encourage strict judges to become more lenient could reduce recidivism.
در این مقاله ابزارهای جدید اقتصاد سنجی برای باز کردن اثر درمانی ارائه شده است ناهمگونی مجازات جرائم تخلف در زمان به رعایت.ما نشان می دهیم چگونه می توان در توزیع ، تخمین و استنتاج می کند ، اثرات کمی و متوسط ​​درمان حاشیه ای در تنظیماتی که در آن درمان است انتخاب درون زا است و نتیجه مورد علاقه ، معمولاً مدت زمان متغیر ، به طور بالقوه سانسور راست است.ما اقتصاد سنجی پیشنهادی خود را کشف می کنیم روش شناسی برای ارزیابی تأثیر جریمه ها و جملات خدمات جامعه به عنوان نوعی مجازات در زمان به رعایت در وضعیت S \ ~ ao paulo ، برزیل ، بین سالهای 2010 تا 2019 ، با استفاده از تکالیف تصادفی AS-IF داوران به مواردنتایج ما ناهمگونی اثر درمانی قابل توجهی را برجسته می کند ابزارهای دیگر برای ضبط نیست.به عنوان مثال ، ما می دانیم که افرادی که بیشتر قضات مجازات می کنند تا به عنوان یک نتیجه از مجازات ، در حالی که افرادی که فقط توسط قضات سختگیر مجازات می شوند مجازات می شوند در تاریخ زودتر از مجازات آنها.این نتیجه نشان می دهد که طراحی دستورالعمل های مجازات که قاضی سختگیرانه را ترغیب می کند تا بیشتر شوند ملایم می تواند عود مجدد را کاهش دهد.
29,912
Counterfactuals in equilibrium models are functions of the current state of the world, the exogenous change variables and the model parameters. Current practice treats the current state of the world, the observed data, as perfectly measured, but there is good reason to believe that they are measured with error. The main aim of this paper is to provide tools for quantifying uncertainty about counterfactuals, when the current state of the world is measured with error. I propose two methods, a Bayesian approach and an adversarial approach. Both methods are practical and theoretically justified. I apply the two methods to the application in Adao et al. (2017) and find non-trivial uncertainty about counterfactuals.
خلاف واقع در مدل های تعادل توابع وضعیت فعلی است جهان ، متغیرهای تغییر اگزوژن و پارامترهای مدل.جاری تمرین با وضعیت فعلی جهان ، داده های مشاهده شده ، به طور کامل رفتار می کند اندازه گیری شده است ، اما دلیل خوبی وجود دارد که باور کنیم با آنها اندازه گیری می شود خطاهدف اصلی این مقاله ارائه ابزاری برای تعیین کمیت است عدم اطمینان در مورد ضد موارد ، هنگامی که وضعیت فعلی جهان است با خطا اندازه گیری می شود.من دو روش را پیشنهاد می کنم ، یک رویکرد بیزی و یک رویکرد مخالف.هر دو روش عملی و از لحاظ نظری توجیه شده اند.من دو روش را در برنامه در Adao و همکاران اعمال کنید.(2017) و پیدا کردن عدم قطعیت غیر واقعی در مورد متقابل.
29,913
The matrix exponential spatial models exhibit similarities to the conventional spatial autoregressive model in spatial econometrics but offer analytical, computational, and interpretive advantages. This paper provides a comprehensive review of the literature on the estimation, inference, and model selection approaches for the cross-sectional matrix exponential spatial models. We discuss summary measures for the marginal effects of regressors and detail the matrix-vector product method for efficient estimation. Our aim is not only to summarize the main findings from the spatial econometric literature but also to make them more accessible to applied researchers. Additionally, we contribute to the literature by introducing some new results. We propose an M-estimation approach for models with heteroskedastic error terms and demonstrate that the resulting M-estimator is consistent and has an asymptotic normal distribution. We also consider some new results for model selection exercises. In a Monte Carlo study, we examine the finite sample properties of various estimators from the literature alongside the M-estimator.
مدل های مکانی ماتریس نمایی شباهت هایی با الگوی اتورگرایی مکانی متعارف در اقتصاد سنجی فضایی اما ارائه می دهد مزایای تحلیلی ، محاسباتی و تفسیری.در این مقاله یک بررسی جامع ادبیات در مورد تخمین ، استنباط و مدل رویکردهای انتخاب برای مدلهای مکانی ماتریس مقطعی. ما در مورد اقدامات خلاصه برای اثرات حاشیه ای رگرسیون و جزئیات بحث می کنیم روش محصول وکتور ماتریس برای برآورد کارآمد.هدف ما فقط نیست به طور خلاصه یافته های اصلی ادبیات اقتصادی اقتصادی بلکه همچنین برای دسترسی بیشتر آنها برای محققان کاربردی.علاوه بر این ، ما با معرفی برخی از نتایج جدید در ادبیات مشارکت کنید.ما پیشنهاد می کنیم رویکرد تخمین M برای مدلهایی با اصطلاحات خطای هتروسیکاستیکی و نشان می دهد که m-desitimator حاصل سازگار است و دارای یک علامت بدون علامت است توزیع نرمال.ما همچنین نتایج جدیدی را برای انتخاب مدل در نظر می گیریم تمریناتدر یک مطالعه مونت کارلو ، ما خواص نمونه محدود را بررسی می کنیم برآوردگرهای مختلفی از ادبیات در کنار M-desitimator.
29,914
This survey discusses the recent causal panel data literature. This recent literature has focused on credibly estimating causal effects of binary interventions in settings with longitudinal data, with an emphasis on practical advice for empirical researchers. It pays particular attention to heterogeneity in the causal effects, often in situations where few units are treated. The literature has extended earlier work on difference-in-differences or two-way-fixed-effect estimators and more generally incorporated factor models or interactive fixed effects. It has also developed novel methods using synthetic control approaches.
این نظرسنجی در مورد ادبیات داده های پانل علیت اخیر بحث می کند.این اخیر ادبیات بر برآورد معتبر تأثیرات باینری متمرکز شده است مداخلات در تنظیمات با داده های طولی ، با تأکید بر عملی مشاوره برای محققان تجربی.توجه ویژه ای به ناهمگونی می کند در اثرات علی ، غالباً در شرایطی که چند واحد تحت درمان قرار می گیرند.در ادبیات کار قبلی را در مورد تفاوت در اختلافات گسترش داده است یا برآوردگرهای دو طرفه ثابت و مدلهای عاملی به طور کلی گنجانیده شده یا اثرات ثابت تعاملی.همچنین روشهای جدیدی را با استفاده از آن ایجاد کرده است رویکردهای کنترل مصنوعی.
29,929
Determining whether Global Average Temperature (GAT) is an integrated process of order 1, I(1), or is a stationary process around a trend function is crucial for detection, attribution, impact and forecasting studies of climate change. In this paper, we investigate the nature of trends in GAT building on the analysis of individual temperature grids. Our 'micro-founded' evidence suggests that GAT is stationary around a non-linear deterministic trend in the form of a linear function with a one-period structural break. This break can be attributed to a combination of individual grid breaks and the standard aggregation method under acceleration in global warming. We illustrate our findings using simulations.
تعیین اینکه آیا دمای متوسط ​​جهانی (GAT) یک فرایند یکپارچه است از سفارش 1 ، i (1) ، یا یک فرآیند ثابت در اطراف یک عملکرد روند بسیار مهم است برای تشخیص ، انتساب ، تأثیر و مطالعات پیش بینی تغییرات آب و هوا. در این مقاله ، ما در مورد ماهیت روندها در ساختمان GAT در تجزیه و تحلیل شبکه های دمای فردی.شواهد "میکرو تأسیس" ما نشان می دهد اینکه GAT در یک روند قطعی غیر خطی به شکل a ثابت است عملکرد خطی با یک استراحت ساختاری یک دوره.این استراحت می تواند باشد نسبت به ترکیبی از شکستهای شبکه و استاندارد نسبت داده شده است روش تجمیع تحت شتاب در گرم شدن کره زمین.ما خود را نشان می دهیم یافته ها با استفاده از شبیه سازی ها.
29,916
This paper investigates how certain relationship between observed and counterfactual distributions serves as an identifying condition for treatment effects when the treatment is endogenous, and shows that this condition holds in a range of nonparametric models for treatment effects. To this end, we first provide a novel characterization of the prevalent assumption restricting treatment heterogeneity in the literature, namely rank similarity. Our characterization demonstrates the stringency of this assumption and allows us to relax it in an economically meaningful way, resulting in our identifying condition. It also justifies the quest of richer exogenous variations in the data (e.g., multi-valued or multiple instrumental variables) in exchange for weaker identifying conditions. The primary goal of this investigation is to provide empirical researchers with tools that are robust and easy to implement but still yield tight policy evaluations.
در این مقاله به بررسی چگونگی رابطه خاص بین مشاهده شده و توزیع های متقابل به عنوان یک وضعیت شناسایی برای درمان عمل می کند اثرات هنگام درمان درون زا است و نشان می دهد که این شرایط دارای آن است در طیف وسیعی از مدلهای غیر پارامتری برای اثرات درمانی.برای این منظور ، ابتدا خصوصیات جدیدی از فرض شایع را ارائه می دهد ناهمگونی درمانی در ادبیات ، یعنی شباهت رتبه بندی می شود.ما خصوصیات سختگیری این فرض را نشان می دهد و به ما امکان می دهد برای آرامش آن به روشی از نظر اقتصادی ، و در نتیجه شناسایی ما وضعیت.این همچنین تلاش تغییرات غنی تر اگزوژن را در آن توجیه می کند داده ها (به عنوان مثال ، متغیرهای ابزاری چند ارزشی یا چندگانه) در ازای آن شرایط شناسایی ضعیف تر.هدف اصلی این تحقیق این است که ابزارهایی را برای محققان تجربی فراهم کنید اما هنوز هم ارزیابی های سخت سیاست را انجام می دهد.
29,917
This paper proposes three novel test procedures that yield valid inference in an environment with many weak instrumental variables (MWIV). It is observed that the t statistic of the jackknife instrumental variable estimator (JIVE) has an asymptotic distribution that is identical to the two-stage-least squares (TSLS) t statistic in the just-identified environment. Consequently, test procedures that were valid for TSLS t are also valid for the JIVE t. Two such procedures, i.e., VtF and conditional Wald, are adapted directly. By exploiting a feature of MWIV environments, a third, more powerful, one-sided VtF-based test procedure can be obtained.
در این مقاله سه روش تست جدید ارائه شده است که استنتاج معتبر در محیطی با متغیرهای ابزاری ضعیف (MWIV).مشاهده می شود که آماری T از برآوردگر متغیر ابزار Jackknife (Jive) توزیع بدون علامت دارد که با مربع های دو مرحله ای یکسان است (TSLS) T آماری در محیط تازه شناسایی شده.در نتیجه ، آزمایش کنید رویه هایی که برای TSLS T معتبر بودند نیز برای Jive t معتبر هستند.دو مورد از این دست رویه ها ، یعنی VTF و والد شرطی ، مستقیماً اقتباس می شوند.با بهره برداری ویژگی محیط های MWIV ، یک سوم ، قدرتمندتر ، مبتنی بر VTF روش تست را می توان بدست آورد.
29,918
For the over-identified linear instrumental variables model, researchers commonly report the 2SLS estimate along with the robust standard error and seek to conduct inference with these quantities. If errors are homoskedastic, one can control the degree of inferential distortion using the first-stage F critical values from Stock and Yogo (2005), or use the robust-to-weak instruments Conditional Wald critical values of Moreira (2003). If errors are non-homoskedastic, these methods do not apply. We derive the generalization of Conditional Wald critical values that is robust to non-homoskedastic errors (e.g., heteroskedasticity or clustered variance structures), which can also be applied to nonlinear weakly-identified models (e.g. weakly-identified GMM).
برای مدل متغیرهای ابزاری خطی بیش از حد مشخص شده ، محققان معمولاً تخمین 2SLS را به همراه خطای استاندارد قوی گزارش دهید و به دنبال آن باشید برای استنباط با این مقادیر.اگر خطاها homoskedastic هستند ، یکی می تواند میزان اعوجاج استنباطی را با استفاده از مرحله اول F کنترل کند مقادیر بحرانی از سهام و یوگو (2005) ، یا استفاده از محکم و محکم ابزارها مقادیر بحرانی والد شرطی موریرا (2003).اگر خطا باشد این روشها غیر هومیوسستیک ، این روشها اعمال نمی شود.ما تعمیم می دهیم مقادیر بحرانی والد شرطی که برای خطاهای غیر هوموسیکاست قوی است (به عنوان مثال ، ناهمگونی یا ساختارهای واریانس خوشه ای) ، که می تواند باشد اعمال شده برای مدل های ضعیف غیر خطی (به عنوان مثال GMM ضعیف).
29,919
This paper addresses the challenge of identifying causal effects of nonbinary, ordered treatments with multiple binary instruments. Next to presenting novel insights into the widely-applied two-stage least squares estimand, I show that a weighted average of local average treatment effects for combined complier populations is identified under the limited monotonicity assumption. This novel causal parameter has an intuitive interpretation, offering an appealing alternative to two-stage least squares. I employ recent advances in causal machine learning for estimation. I further demonstrate how causal forests can be used to detect local violations of the underlying limited monotonicity assumption. The methodology is applied to study the impact of community nurseries on child health outcomes.
در این مقاله به چالش شناسایی اثرات علّی می پردازد درمان های غیربی ، سفارش داده شده با ابزارهای باینری متعدد.جنب ارائه بینش های رمان در مورد مربع های حداقل دو مرحله ای بسیار گسترده برآورد ، من نشان می دهم که میانگین وزنی از اثرات درمانی متوسط ​​محلی برای جمعیت ترکیبی ترکیبی تحت یکنواختی محدود مشخص می شود فرض.این پارامتر علی جدید دارای تفسیر شهودی است ، ارائه یک جایگزین جذاب برای حداقل مربعات دو مرحله ای.من اخیر استخدام می کنم پیشرفت در یادگیری ماشین علیت برای تخمین.من بیشتر نشان می دهم که چگونه از جنگل های علی می توان برای تشخیص تخلفات محلی محدوده محدود استفاده کرد فرض یکنواختی.این روش برای مطالعه تأثیر استفاده می شود مهد کودک های جامعه در مورد نتایج سلامت کودکان.
29,920
Canonical RD designs yield credible local estimates of the treatment effect at the cutoff under mild continuity assumptions, but they fail to identify treatment effects away from the cutoff without additional assumptions. The fundamental challenge of identifying treatment effects away from the cutoff is that the counterfactual outcome under the alternative treatment status is never observed. This paper aims to provide a methodological blueprint to identify treatment effects away from the cutoff in various empirical settings by offering a non-exhaustive list of assumptions on the counterfactual outcome. Instead of assuming the exact evolution of the counterfactual outcome, this paper bounds its variation using the data and sensitivity parameters. The proposed assumptions are weaker than those introduced previously in the literature, resulting in partially identified treatment effects that are less susceptible to assumption violations. This approach accommodates both single cutoff and multi-cutoff designs. The specific choice of the extrapolation assumption depends on the institutional background of each empirical application. Additionally, researchers are recommended to conduct sensitivity analysis on the chosen parameter and assess resulting shifts in conclusions. The paper compares the proposed identification results with results using previous methods via an empirical application and simulated data. It demonstrates that set identification yields a more credible conclusion about the sign of the treatment effect.
طرح های Canonical RD برآورد محلی معتبر از اثر درمانی در زمان قطع فرضیات استمرار خفیف ، اما آنها در شناسایی آنها ناکام هستند اثرات درمانی به دور از قطع بدون فرضیات اضافی.در چالش اساسی در شناسایی اثرات درمانی به دور از قطع است که نتیجه ضد عملی تحت وضعیت درمان جایگزین هرگز نیست مشاهده شده.این مقاله با هدف ارائه طرح روش شناختی برای شناسایی اثرات درمانی به دور از قطع در تنظیمات مختلف تجربی توسط ارائه یک لیست غیر اکسپرس از فرضیات در مورد نتیجه متقابل. به جای فرض تکامل دقیق نتیجه ضد خلاف واقع ، این کاغذ با استفاده از پارامترهای داده و حساسیت ، تغییرات آن را محدود می کند.در فرضیات پیشنهادی ضعیف تر از مواردی است که قبلاً در آن معرفی شده است ادبیات ، و در نتیجه اثرات درمانی جزئی که کمتر است مستعد تخلفات فرض است.این رویکرد هر دو واحد را در خود جای می دهد طرح های برش و چند برش.انتخاب خاص برون یابی فرض به پیشینه نهادی هر تجربی بستگی دارد کاربرد.علاوه بر این ، محققان برای انجام حساسیت توصیه می شود تجزیه و تحلیل در پارامتر انتخاب شده و ارزیابی تغییرات در نتیجه گیری. مقاله نتایج شناسایی پیشنهادی را با نتایج با استفاده از نتایج مقایسه می کند روشهای قبلی از طریق یک برنامه تجربی و داده های شبیه سازی شده.آی تی نشان می دهد که شناسایی مجموعه نتیجه گیری معتبرتر در مورد نشانه اثر درمانی.
29,921
The partially linear binary choice model can be used for estimating structural equations where nonlinearity may appear due to diminishing marginal returns, different life cycle regimes, or hectic physical phenomena. The inference procedure for this model based on the analytic asymptotic approximation could be unreliable in finite samples if the sample size is not sufficiently large. This paper proposes a bootstrap inference approach for the model. Monte Carlo simulations show that the proposed inference method performs well in finite samples compared to the procedure based on the asymptotic approximation.
از مدل انتخاب باینری جزئی خطی می توان برای تخمین استفاده کرد معادلات ساختاری که در آن غیرخطی ممکن است به دلیل کاهش حاشیه ظاهر شود بازده ، رژیم های مختلف چرخه زندگی یا پدیده های جسمی پرشور.در روش استنباط برای این مدل بر اساس بدون علامت تحلیلی اگر اندازه نمونه نباشد ، تقریب می تواند در نمونه های محدود غیرقابل اعتماد باشد به اندازه کافی بزرگدر این مقاله یک رویکرد استنتاج bootstrap برای مدل.شبیه سازی مونت کارلو نشان می دهد که روش استنباط پیشنهادی انجام می دهد خوب در نمونه های محدود در مقایسه با روش مبتنی بر بدون علامت تقریب
29,922
This paper expands traditional stochastic volatility models by allowing for time-varying skewness without imposing it. While dynamic asymmetry may capture the likely direction of future asset returns, it comes at the risk of leading to overparameterization. Our proposed approach mitigates this concern by leveraging sparsity-inducing priors to automatically selects the skewness parameter as being dynamic, static or zero in a data-driven framework. We consider two empirical applications. First, in a bond yield application, dynamic skewness captures interest rate cycles of monetary easing and tightening being partially explained by central banks' mandates. In an currency modeling framework, our model indicates no skewness in the carry factor after accounting for stochastic volatility which supports the idea of carry crashes being the result of volatility surges instead of dynamic skewness.
این مقاله با اجازه دادن به مدل های نوسانات تصادفی سنتی گسترش می یابد کمبود زمان بدون تحمیل آن.در حالی که عدم تقارن پویا ممکن است ضبط شود جهت احتمالی بازده دارایی آینده ، در معرض خطر رهبری است به پنجه بیش از حد.رویکرد پیشنهادی ما این نگرانی را توسط اهرم PRIOROR های القا کننده پراکنده برای انتخاب خودکار پوستی پارامتر به عنوان پویا ، استاتیک یا صفر در یک چارچوب داده محور.ما دو برنامه تجربی را در نظر بگیرید.اول ، در یک برنامه عملکرد اوراق قرضه ، پویا پویا چرخه نرخ بهره را از تسکین پولی و محکم کردن تا حدی توسط دستورات بانکهای مرکزی توضیح داده می شود.در یک ارز چارچوب مدل سازی ، مدل ما نشانگر عدم وجود ضریب حمل پس از آن است حسابداری برای نوسانات تصادفی که از ایده تصادفات پشتیبانی می کند به نظر می رسد که نتیجه نوسانات به جای کمبود پویا افزایش می یابد.
29,923
We introduce a new estimator, CRE-GMM, which exploits the correlated random effects (CRE) approach within the generalised method of moments (GMM), specifically applied to level equations, GMM-lev. It has the advantage of estimating the effect of measurable time-invariant covariates using all available information. This is not possible with GMM-dif, applied to the equations of each period transformed into first differences, while GMM-sys uses little information as it adds the equation in levels for only one period. The GMM-lev, by implying a two-component error term containing individual heterogeneity and shock, exposes the explanatory variables to possible double endogeneity. For example, the estimation of actual persistence could suffer from bias if instruments were correlated with the unit-specific error component. The CRE-GMM deals with double endogeneity, captures initial conditions and enhance inference. Monte Carlo simulations for different panel types and under different double endogeneity assumptions show the advantage of our approach. The empirical applications on production and R&D contribute to clarify the advantages of using CRE-GMM.
ما یک برآوردگر جدید ، CRE-GMM را معرفی می کنیم ، که از تصادفی همبسته سوء استفاده می کند رویکرد اثرات (CRE) در روش کلی لحظات (GMM) ، به طور خاص برای معادلات سطح ، GMM-LEV اعمال می شود.این مزیت را دارد تخمین تأثیر متغیرهای متغیر متغیر با استفاده از همه اطلاعات موجوداین با GMM-DIF امکان پذیر نیست ، معادلات هر دوره به اختلافات اول تبدیل شده است ، در حالی که GMM-SYS از آن استفاده می کند اطلاعات کمی به عنوان معادله در سطوح فقط برای یک دوره اضافه می کند.در GMM-LEV ، با دلالت بر یک اصطلاح خطای دو جزء حاوی فرد ناهمگونی و شوک ، متغیرهای توضیحی را در معرض دو برابر قرار می دهد درون زاییبه عنوان مثال ، تخمین پایداری واقعی می تواند رنج ببرد از تعصب اگر ابزارها با خطای خاص واحد ارتباط داشته باشند جزء.CRE-GMM با درون زایی مضاعف سروکار دارد ، اولیه را ضبط می کند شرایط و استنباط را تقویت می کند.شبیه سازی مونت کارلو برای پانل های مختلف انواع و تحت فرضیات مختلف درون زایی دوتایی مزیت را نشان می دهد رویکرد ما.برنامه های تجربی در تولید و تحقیق و توسعه به مزایای استفاده از CRE-GMM را روشن کنید.
29,924
A common approach to constructing a Synthetic Control unit is to fit on the outcome variable and covariates in pre-treatment time periods, but it has been shown by Ferman and Pinto (2021) that this approach does not provide asymptotic unbiasedness when the fit is imperfect and the number of controls is fixed. Many related panel methods have a similar limitation when the number of units is fixed. I introduce and evaluate a new method in which the Synthetic Control is constructed using a General Method of Moments approach where if the Synthetic Control satisfies the moment conditions it must have the same loadings on latent factors as the treated unit. I show that a Synthetic Control Estimator of this form will be asymptotically unbiased as the number of pre-treatment time periods goes to infinity, even when pre-treatment fit is imperfect and the set of controls is fixed. Furthermore, if both the number of pre-treatment and post-treatment time periods go to infinity, then averages of treatment effects can be consistently estimated and asymptotically valid inference can be conducted using a subsampling method. I conduct simulations and an empirical application to compare the performance of this method with existing approaches in the literature.
یک رویکرد متداول برای ساخت یک واحد کنترل مصنوعی قرار گرفتن در آن است متغیر نتیجه و متغیر در دوره های قبل از درمان ، اما این بوده است توسط Ferman و Pinto (2021) نشان داده شده است که این رویکرد بدون علامت ارائه نمی دهد بی طرفی وقتی تناسب ناقص است و تعداد کنترل ها ثابت است. بسیاری از روشهای پانل مرتبط با تعداد واحدها محدودیت مشابهی دارند تعمیر شد.من روشی جدید را که در آن کنترل مصنوعی است معرفی و ارزیابی می کنم با استفاده از یک روش کلی از رویکرد لحظات ساخته شده است که در صورت کنترل مصنوعی لحظه ای را که باید یکسان داشته باشد را برآورده می کند بارگذاری بر روی فاکتورهای نهفته به عنوان واحد تحت درمان.من نشان می دهم که یک کنترل مصنوعی برآوردگر این فرم به عنوان تعداد بدون علامت خواهد بود به عنوان تعداد دوره های زمانی قبل از درمان به بی نهایت می رود ، حتی در صورت تناسب قبل از درمان ناقص و مجموعه کنترل ها ثابت است.علاوه بر این ، اگر هر دو تعداد دوره های قبل از درمان و پس از درمان به بی نهایت می روند ، سپس به طور متوسط ​​از اثرات درمانی را می توان به طور مداوم تخمین زده و بدون علامت معتبر است استنتاج را می توان با استفاده از یک روش نمونه برداری انجام داد.من شبیه سازی ها را انجام می دهم و یک برنامه تجربی برای مقایسه عملکرد این روش با رویکردهای موجود در ادبیات.
29,925
This paper proposes a general framework for inference on three types of almost dominances: Almost Lorenz dominance, almost inverse stochastic dominance, and almost stochastic dominance. We first generalize almost Lorenz dominance to almost upward and downward Lorenz dominances. We then provide a bootstrap inference procedure for the Lorenz dominance coefficients, which measure the degrees of almost Lorenz dominances. Furthermore, we propose almost upward and downward inverse stochastic dominances and provide inference on the inverse stochastic dominance coefficients. We also show that our results can easily be extended to almost stochastic dominance. Simulation studies demonstrate the finite sample properties of the proposed estimators and the bootstrap confidence intervals. We apply our methods to the inequality growth in the United Kingdom and find evidence for almost upward inverse stochastic dominance.
در این مقاله یک چارچوب کلی برای استنباط در سه نوع ارائه شده است تقریباً سلطه ها: تقریباً لورنز تسلط ، تقریبا معکوس تصادفی تسلط و تسلط تقریبا تصادفی.ما ابتدا تقریباً لورنز را تعمیم می دهیم تسلط بر سلطه های تقریباً به سمت بالا و رو به پایین لورنز.سپس ما یک روش استنتاج bootstrap برای ضرایب تسلط لورنز ، که درجه های سلطه تقریباً لورنز را اندازه گیری کنید.علاوه بر این ، ما تقریباً پیشنهاد می کنیم سلطه های تصادفی معکوس به سمت بالا و رو به پایین و استنباط در ضرایب تسلط تصادفی معکوس.ما همچنین نشان می دهیم که نتایج ما می تواند به راحتی به تسلط تقریبا تصادفی گسترش می یابد.مطالعات شبیه سازی خصوصیات نمونه محدود برآوردگرهای پیشنهادی و فواصل اعتماد به نفس بوت استرپ.ما روشهای خود را برای رشد نابرابری اعمال می کنیم در انگلستان و یافتن شواهدی برای تصادفی تقریباً رو به بالا تسلط
29,926
I develop the theory around using control functions to instrument hazard models, allowing the inclusion of endogenous (e.g., mismeasured) regressors. Simple discrete-data hazard models can be expressed as binary choice panel data models, and the widespread Prentice and Gloeckler (1978) discrete-data proportional hazards model can specifically be expressed as a complementary log-log model with time fixed effects. This allows me to recast it as GMM estimation and its instrumented version as sequential GMM estimation in a Z-estimation (non-classical GMM) framework; this framework can then be leveraged to establish asymptotic properties and sufficient conditions. Whilst this paper focuses on the Prentice and Gloeckler (1978) model, the methods and discussion developed here can be applied more generally to other hazard models and binary choice models. I also introduce my Stata command for estimating a complementary log-log model instrumented via control functions (available as ivcloglog on SSC), which allows practitioners to easily instrument the Prentice and Gloeckler (1978) model.
من تئوری پیرامون استفاده از توابع کنترل را برای خطر ابزار توسعه می دهم مدل ها ، اجازه می دهند تا رگرسیون های درون زا (به عنوان مثال ، سوء استفاده). مدل های خطرناک داده های گسسته ساده می توانند به عنوان داده های پانل انتخاب باینری بیان شوند مدل ها ، و گسترده Prentice و Gloeckler (1978) داده های گسسته مدل خطرات متناسب به طور خاص می تواند به عنوان یک مکمل بیان شود مدل ورود به سیستم با اثرات ثابت زمان.این به من اجازه می دهد تا آن را به عنوان GMM بازیابی کنم تخمین و نسخه ساز آن به عنوان تخمین پی در پی GMM در a چارچوب تخمین Z (GMM غیر کلاسیک) ؛این چارچوب می تواند باشد برای ایجاد خواص بدون علامت و شرایط کافی استفاده می شود.در حالی که این مقاله به مدل Prentice و Gloeckler (1978) ، روش ها و بحث و گفتگو در اینجا می تواند به طور کلی در سایر مدل های خطر اعمال شود و مدلهای انتخاب باینری.من همچنین دستور stata خود را برای برآورد a معرفی می کنم مدل log-log مکمل که از طریق توابع کنترل ابزار شده است (موجود است IvCloglog در SSC) ، که به پزشکان این امکان را می دهد تا به راحتی Prentice را سازگار کنند و مدل Gloeckler (1978).
29,927
When fitting a particular Economic model on a sample of data, the model may turn out to be heavily misspecified for some observations. This can happen because of unmodelled idiosyncratic events, such as an abrupt but short-lived change in policy. These outliers can significantly alter estimates and inferences. A robust estimation is desirable to limit their influence. For skewed data, this induces another bias which can also invalidate the estimation and inferences. This paper proposes a robust GMM estimator with a simple bias correction that does not degrade robustness significantly. The paper provides finite-sample robustness bounds, and asymptotic uniform equivalence with an oracle that discards all outliers. Consistency and asymptotic normality ensue from that result. An application to the "Price-Puzzle," which finds inflation increases when monetary policy tightens, illustrates the concerns and the method. The proposed estimator finds the intuitive result: tighter monetary policy leads to a decline in inflation.
هنگام قرار دادن یک مدل اقتصادی خاص بر روی نمونه ای از داده ها ، ممکن است مدل به نظر می رسد که برای برخی از مشاهدات به شدت غلط شناخته شده است.این می تواند اتفاق بیفتد به دلیل وقایع غیرمستقیم غیرمجاز ، مانند یک ناگهانی اما کوتاه مدت تغییر در سیاست.این مسافت ها می توانند به طور قابل توجهی تخمین ها را تغییر دهند و استنتاجتخمین قوی برای محدود کردن تأثیر آنها مطلوب است.برای داده های ناخوشایند ، این باعث تعصب دیگری می شود که می تواند تخمین را نیز باطل کند و استنتاجدر این مقاله یک برآوردگر قوی GMM با یک تعصب ساده پیشنهاد شده است تصحیح که استحکام را به میزان قابل توجهی تخریب نمی کند.مقاله ارائه می دهد مرزهای استحکام نمونه محدود ، و هم ارزی یکنواخت بدون علامت با یک اوراکل که همه دور از دسترس را دور می کند.قوام و عادی بودن بدون علامت از آن نتیجهبرنامه ای برای "قیمت گذاری قیمت" ، که تورم را پیدا می کند وقتی سیاست پولی محکم می شود ، نگرانی ها و موارد را نشان می دهد روش.برآوردگر پیشنهادی نتیجه شهودی را می یابد: پولی محکم تر سیاست منجر به کاهش تورم می شود.
29,928
We examine finite sample performance of the Generalized Covariance (GCov) residual-based specification test for semiparametric models with i.i.d. errors. The residual-based multivariate portmanteau test statistic follows asymptotically a $\chi^2$ distribution when the model is estimated by the GCov estimator. The test is shown to perform well in application to the univariate mixed causal-noncausal MAR, double autoregressive (DAR) and multivariate Vector Autoregressive (VAR) models. We also introduce a bootstrap procedure that provides the limiting distribution of the test statistic when the specification test is applied to a model estimated by the maximum likelihood, or the approximate or quasi-maximum likelihood under a parametric assumption on the error distribution.
ما عملکرد نمونه محدود از کواریانس عمومی (GCOV) را بررسی می کنیم تست مشخصات مبتنی بر باقیمانده برای مدلهای نیمهرامتری با I.I.D.خطا آمار آزمون چند متغیره Portmanteau مبتنی بر باقیمانده به شرح زیر است به صورت بدون علامت $ \ chi^2 $ توزیع هنگامی که مدل توسط GCOV تخمین زده می شود برآوردگرنشان داده شده است که این آزمایش در کاربرد تک متغیره عملکرد خوبی دارد وکتور علت و معلولی علت و معلولیت) ، دوتایی اتوراس (DAR) و چند متغیره مدل های خودکار (VAR).ما همچنین یک روش bootstrap را معرفی می کنیم که توزیع محدود کننده آمار آزمون را هنگام مشخصات فراهم می کند آزمون برای مدلی که با حداکثر احتمال تخمین زده می شود یا احتمال تقریبی یا شبه حداکثر تحت یک فرض پارامتری در توزیع خطا
29,930
This set of lecture notes discuss key concepts for the Structural Analysis of Vector Autoregressive models for the teaching of a course on Applied Macroeconometrics with Advanced Topics.
این مجموعه از یادداشت های سخنرانی در مورد مفاهیم کلیدی برای تجزیه و تحلیل ساختاری بحث می کنند مدل های اتورگرایی وکتور برای آموزش یک دوره در مورد کاربردی اقتصاد کلان با موضوعات پیشرفته.
29,934
This paper is concerned with the problem of variable selection in the presence of parameter instability when both the marginal effects of signals on the target variable and the correlations of the covariates in the active set could vary over time. We pose the issue of whether one should use weighted or unweighted observations at the variable selection stage in the presence of parameter instability, particularly when the number of potential covariates is large. We allow parameter instability to be continuous or discrete, subject to certain regularity conditions. We discuss the pros and cons of Lasso and the One Covariate at a time Multiple Testing (OCMT) method for variable selection and argue that OCMT has important advantages under parameter instability. We establish three main theorems on selection, estimation post selection, and in-sample fit. These theorems provide justification for using unweighted observations at the selection stage of OCMT and down-weighting of observations only at the forecasting stage. It is shown that OCMT delivers better forecasts, in mean squared error sense, as compared to Lasso, Adaptive Lasso and boosting both in Monte Carlo experiments as well as in 3 sets of empirical applications: forecasting monthly returns on 28 stocks from Dow Jones , forecasting quarterly output growths across 33 countries, and forecasting euro area output growth using surveys of professional forecasters.
این مقاله به مشکل انتخاب متغیر در وجود ناپایداری پارامتر هنگامی که هر دو اثر حاشیه سیگنال بر روی متغیر هدف و همبستگی متغیرهای متغیر در مجموعه فعال با گذشت زمان می تواند متفاوت باشد.ما این مسئله را مطرح می کنیم که آیا شخص باید از وزنه برداری استفاده کند یا مشاهدات بدون وزنی در مرحله انتخاب متغیر در حضور ناپایداری پارامتر ، به ویژه هنگامی که تعداد متغیرهای بالقوه باشد بزرگما اجازه می دهیم که ناپایداری پارامتر مداوم یا گسسته باشد ، منوط به شرایط منظم خاص.ما در مورد جوانب مثبت و منفی لاسو و یک متغیر متغیر در یک زمان تست چندگانه (OCMT) برای انتخاب متغیر و استدلال کنید که OCMT تحت بی ثباتی پارامتر مزایای مهمی دارد.ما سه قضیه اصلی در مورد انتخاب ، انتخاب پست تخمین ، و در نمونه مناسب است.این قضیه ها توجیهی برای استفاده از افراد بدون مجازات ارائه می دهند مشاهدات در مرحله انتخاب OCMT و وزن پایین مشاهدات فقط در مرحله پیش بینی.نشان داده شده است که OCMT پیش بینی های بهتری را ارائه می دهد ، در میانگین حس خطای مربع ، در مقایسه با لاسو ، لاسو تطبیقی ​​و تقویت هر دو در آزمایش های مونت کارلو و همچنین در 3 مجموعه برنامه های تجربی: پیش بینی بازده ماهانه در 28 سهام از داو جونز ، پیش بینی سه ماهه رشد خروجی در 33 کشور و پیش بینی رشد تولید منطقه یورو با استفاده از نظرسنجی از پیش بینی کنندگان حرفه ای.
29,935
We introduce a novel approach for comparing out-of-sample multi-step forecasts obtained from a pair of nested models that is based on the forecast encompassing principle. Our proposed approach relies on an alternative way of testing the population moment restriction implied by the forecast encompassing principle and that links the forecast errors from the two competing models in a particular way. Its key advantage is that it is able to bypass the variance degeneracy problem afflicting model based forecast comparisons across nested models. It results in a test statistic whose limiting distribution is standard normal and which is particularly simple to construct and can accommodate both single period and longer-horizon prediction comparisons. Inferences are also shown to be robust to different predictor types, including stationary, highly-persistent and purely deterministic processes. Finally, we illustrate the use of our proposed approach through an empirical application that explores the role of global inflation in enhancing individual country specific inflation forecasts.
ما یک رویکرد جدید برای مقایسه چند مرحله ای خارج از نمونه معرفی می کنیم پیش بینی های به دست آمده از یک جفت مدل تو در تو که مبتنی بر پیش بینی است اصل شاملرویکرد پیشنهادی ما به یک روش جایگزین متکی است آزمایش محدودیت لحظه جمعیت دلالت بر پیش بینی شامل اصل و این که خطاهای پیش بینی شده از دو مدل رقیب را در a پیوند می دهد روش خاصمزیت اصلی آن این است که قادر به دور زدن واریانس است مشکل انحطاط مدل نقصی مبتنی بر مقایسه پیش بینی بر روی تو در تو مدل ها.این منجر به یک آمار آزمون می شود که توزیع محدود کننده آن استاندارد است طبیعی است و به ویژه برای ساخت آن ساده است و می تواند هر دو را در خود جای دهد مقایسه پیش بینی دوره و افق طولانی تر.استنتاج ها نیز هستند نشان داده شده برای انواع مختلف پیش بینی کننده ، از جمله ثابت ، فرآیندهای بسیار مداوم و صرفاً قطعی.سرانجام ، ما تصویر می کنیم استفاده از رویکرد پیشنهادی ما از طریق یک برنامه تجربی که بررسی می کند نقش تورم جهانی در تقویت تورم خاص کشور فردی پیش بینی
29,936
Consumer choice modeling takes center stage as we delve into understanding how personal preferences of decision makers (customers) for products influence demand at the level of the individual. The contemporary choice theory is built upon the characteristics of the decision maker, alternatives available for the choice of the decision maker, the attributes of the available alternatives and decision rules that the decision maker uses to make a choice. The choice set in our research is represented by six major brands (products) of laundry detergents in the Japanese market. We use the panel data of the purchases of 98 households to which we apply the hierarchical probit model, facilitated by a Markov Chain Monte Carlo simulation (MCMC) in order to evaluate the brand values of six brands. The applied model also allows us to evaluate the tangible and intangible brand values. These evaluated metrics help us to assess the brands based on their tangible and intangible characteristics. Moreover, consumer choice modeling also provides a framework for assessing the environmental performance of laundry detergent brands as the model uses the information on components (physical attributes) of laundry detergents.
مدل سازی انتخاب مصرف کننده وقتی که ما به درک می پردازیم ، مرحله مرکزی را می گیرد چگونه ترجیحات شخصی تصمیم گیرندگان (مشتریان) برای تأثیر محصولات تقاضا در سطح فرد.نظریه انتخاب معاصر ساخته شده است با توجه به ویژگی های تصمیم گیرنده ، گزینه های موجود برای انتخاب تصمیم گیرنده ، ویژگی های گزینه های موجود و قوانین تصمیم گیری که تصمیم گیرنده برای انتخاب استفاده می کند.انتخاب در تحقیقات ما توسط شش مارک اصلی (محصولات) لباسشویی ارائه شده است مواد شوینده در بازار ژاپن.ما از داده های پانل خریدهای 98 استفاده می کنیم خانوارهایی که ما از مدل پروبیت سلسله مراتبی استفاده می کنیم ، که توسط a تسهیل می شود زنجیره مارکوف مونت کارلو شبیه سازی (MCMC) به منظور ارزیابی برند مقادیر شش مارک.مدل کاربردی همچنین به ما امکان ارزیابی ملموس را می دهد و ارزشهای نامحسوس برند.این معیارهای ارزیابی شده به ما کمک می کنند تا مارک های مبتنی بر ویژگی های ملموس و نامشهود آنها.علاوه بر این، مدل سازی انتخاب مصرف کننده همچنین چارچوبی برای ارزیابی عملکرد محیطی مارک های مواد شوینده لباسشویی به عنوان مدل از آن استفاده می کند اطلاعات مربوط به مؤلفه ها (ویژگی های فیزیکی) مواد شوینده لباسشویی.
29,937
We apply classical statistical decision theory to a large class of treatment choice problems with partial identification, revealing important theoretical and practical challenges but also interesting research opportunities. The challenges are: In a general class of problems with Gaussian likelihood, all decision rules are admissible; it is maximin-welfare optimal to ignore all data; and, for severe enough partial identification, there are infinitely many minimax-regret optimal decision rules, all of which sometimes randomize the policy recommendation. The opportunities are: We introduce a profiled regret criterion that can reveal important differences between rules and render some of them inadmissible; and we uniquely characterize the minimax-regret optimal rule that least frequently randomizes. We apply our results to aggregation of experimental estimates for policy adoption, to extrapolation of Local Average Treatment Effects, and to policy making in the presence of omitted variable bias.
ما تئوری تصمیم گیری آماری کلاسیک را برای طبقه بزرگی از درمان اعمال می کنیم مشکلات انتخاب با شناسایی جزئی ، آشکار کردن نظری مهم و چالش های عملی اما همچنین فرصت های جالب تحقیق.در چالش ها عبارتند از: در یک کلاس کلی از مشکلات با احتمال گاوسی ، همه قوانین تصمیم گیری قابل قبول است.نادیده گرفتن همه ماکسیمین بهینه است داده ها؛و برای شناسایی جزئی جزئی شدید ، بی نهایت تعداد زیادی وجود دارد قوانین تصمیم گیری بهینه Minimax-Regret ، که همه اینها گاهی اوقات تصادفی می شوند توصیه سیاست.فرصت ها این است: ما پشیمانی پروفایل را معرفی می کنیم معیاری که می تواند تفاوتهای مهمی بین قوانین و برخی از آنها را نشان دهد از آنها غیرقابل قبول ؛و ما منحصر به فرد بهینه Minimax-Regret را توصیف می کنیم قانون که حداقل اغلب تصادفی می شود.ما نتایج خود را برای تجمع اعمال می کنیم برآوردهای تجربی برای اتخاذ سیاست ، برای برون یابی میانگین محلی اثرات درمانی و سیاست گذاری در حضور متغیر حذف شده جانبداری.
29,938
Forecasting a key macroeconomic variable, consumer price index (CPI) inflation, for BRIC countries using economic policy uncertainty and geopolitical risk is a difficult proposition for policymakers at the central banks. This study proposes a novel filtered ensemble wavelet neural network (FEWNet) that can produce reliable long-term forecasts for CPI inflation. The proposal applies a maximum overlapping discrete wavelet transform to the CPI inflation series to obtain high-frequency and low-frequency signals. All the wavelet-transformed series and filtered exogenous variables are fed into downstream autoregressive neural networks to make the final ensemble forecast. Theoretically, we show that FEWNet reduces the empirical risk compared to single, fully connected neural networks. We also demonstrate that the rolling-window real-time forecasts obtained from the proposed algorithm are significantly more accurate than benchmark forecasting methods. Additionally, we use conformal prediction intervals to quantify the uncertainty associated with the forecasts generated by the proposed approach. The excellent performance of FEWNet can be attributed to its capacity to effectively capture non-linearities and long-range dependencies in the data through its adaptable architecture.
پیش بینی یک متغیر کلیدی اقتصادی کلان ، شاخص قیمت مصرف کننده (CPI) تورم ، برای کشورهای BRIC که از عدم اطمینان سیاست اقتصادی استفاده می کنند و خطر ژئوپلیتیکی یک پیشنهاد دشوار برای سیاست گذاران در مرکز است بانک ها.این مطالعه یک شبکه عصبی Wavelet Wavelet Ensemble رمان جدید را پیشنهاد می کند (Fewnet) که می تواند پیش بینی های بلند مدت قابل اعتماد برای تورم CPI ایجاد کند.در پیشنهاد حداکثر تحول موجک گسسته با هم همپوشانی را به CPI اعمال می کند سری تورم برای به دست آوردن سیگنال های با فرکانس بالا و با فرکانس پایین.همه سریال های تبدیل شده به موج و متغیرهای اگزوژن فیلتر شده در آن تغذیه می شوند شبکه های عصبی خودکار پایین دست برای پیش بینی گروه نهایی. از لحاظ تئوریکی ، ما نشان می دهیم که Fewnet خطر تجربی را در مقایسه با کاهش می دهد شبکه های عصبی تک و کاملاً متصل.ما همچنین نشان می دهیم که پیش بینی های زمان واقعی رول به دست آمده از الگوریتم پیشنهادی به طور قابل توجهی دقیق تر از روش های پیش بینی معیار.علاوه بر این ، ما از فواصل پیش بینی کنفورماسی برای تعیین کمیت عدم اطمینان مرتبط استفاده می کنیم با پیش بینی های ایجاد شده توسط رویکرد پیشنهادی.عالی عملکرد Leatnet را می توان به ظرفیت آن برای ضبط مؤثر نسبت داد غیر خطی ها و وابستگی های دوربرد در داده ها از طریق سازگار آن معماری.
30,370
Questionable research practices like HARKing or p-hacking have generated considerable recent interest throughout and beyond the scientific community. We subsume such practices involving secret data snooping that influences subsequent statistical inference under the term MESSing (manipulating evidence subject to snooping) and discuss, illustrate and quantify the possibly dramatic effects of several forms of MESSing using an empirical and a simple theoretical example. The empirical example uses numbers from the most popular German lottery, which seem to suggest that 13 is an unlucky number.
شیوه های تحقیق مشکوک مانند Harking یا P-Hacking ایجاد شده است علاقه قابل توجه اخیر در سراسر و فراتر از جامعه علمی.ما چنین شیوه هایی را شامل می شود که شامل داده های مخفی است که تأثیر می گذارد استنباط آماری متعاقب آن با اصطلاح Messing (دستکاری شواهد مشروط به snooping) و بحث ، تصویر سازی و کمیت احتمالاً چشمگیر تأثیرات چندین شکل از آشفتگی با استفاده از یک تجربی و یک نظری ساده مثال.مثال تجربی از شماره های محبوب ترین آلمانی استفاده می کند قرعه کشی ، که به نظر می رسد 13 عدد بدشانس است.
29,939
This paper studies identification for a wide range of nonlinear panel data models, including binary choice, ordered repsonse, and other types of limited dependent variable models. Our approach accommodates dynamic models with any number of lagged dependent variables as well as other types of (potentially contemporary) endogeneity. Our identification strategy relies on a partial stationarity condition, which not only allows for an unknown distribution of errors but also for temporal dependencies in errors. We derive partial identification results under flexible model specifications and provide additional support conditions for point identification. We demonstrate the robust finite-sample performance of our approach using Monte Carlo simulations, with static and dynamic ordered choice models as illustrative examples.
این مقاله به بررسی طیف گسترده ای از داده های پانل غیرخطی می پردازد مدل ها ، از جمله انتخاب باینری ، repsonse سفارش داده شده و سایر انواع محدود مدلهای متغیر وابسته.رویکرد ما مدل های پویا را با هر یک از آنها در خود جای می دهد تعداد متغیرهای وابسته به تاخیر و همچنین انواع دیگر (به طور بالقوه معاصر) درون زا.استراتژی شناسایی ما به یک جزئی متکی است شرایط ثابت بودن ، که نه تنها امکان توزیع ناشناخته را فراهم می کند خطاها بلکه برای وابستگی های زمانی در خطاها.ما جزئی می گیریم نتایج شناسایی تحت مشخصات مدل انعطاف پذیر و ارائه شرایط پشتیبانی اضافی برای شناسایی نقطه.ما نشان می دهیم عملکرد قوی نمونه محدود رویکرد ما با استفاده از شبیه سازی مونت کارلو ، با مدل های انتخابی مرتب و پویا به عنوان نمونه های مصور.
29,940
In this paper, we develop a generalized Difference-in-Differences model for discrete, ordered outcomes, building upon elements from a continuous Changes-in-Changes model. We focus on outcomes derived from self-reported survey data eliciting socially undesirable, illegal, or stigmatized behaviors like tax evasion, substance abuse, or domestic violence, where too many "false zeros", or more broadly, underreporting are likely. We provide characterizations for distributional parallel trends, a concept central to our approach, within a general threshold-crossing model framework. In cases where outcomes are assumed to be reported correctly, we propose a framework for identifying and estimating treatment effects across the entire distribution. This framework is then extended to modeling underreported outcomes, allowing the reporting decision to depend on treatment status. A simulation study documents the finite sample performance of the estimators. Applying our methodology, we investigate the impact of recreational marijuana legalization for adults in several U.S. states on the short-term consumption behavior of 8th-grade high-school students. The results indicate small, but significant increases in consumption probabilities at each level. These effects are further amplified upon accounting for misreporting.
در این مقاله ، ما یک مدل تفاوت در تمایز را ایجاد می کنیم نتایج گسسته ، سفارش داده شده ، بنا بر عناصر مداوم مدل تغییرات در تغییر.ما روی نتایج حاصل از گزارش خود تمرکز می کنیم داده های نظرسنجی را که از نظر اجتماعی رفتارهای نامطلوب ، غیرقانونی یا ننگ اجتماعی ایجاد می کنند ، بررسی کنید مانند فرار مالیاتی ، سوء مصرف مواد یا خشونت خانگی ، جایی که خیلی زیاد "نادرست است صفر "، یا به طور گسترده تر ، گزارش دادن به احتمال زیاد است. ما ارائه می دهیم خصوصیات برای روند موازی توزیع ، مفهومی برای ما رویکرد ، در یک چارچوب کلی مدل آستانه عبور.در مواردی که فرض بر این است که نتایج به درستی گزارش می شود ، ما یک چارچوب برای آن پیشنهاد می کنیم شناسایی و تخمین اثرات درمانی در کل توزیع. این چارچوب سپس به مدل سازی نتایج غیر گزارش شده گسترش می یابد ، اجازه می دهد تصمیم گزارش به وضعیت درمانی بستگی دارد.یک مطالعه شبیه سازی عملکرد نمونه محدود از برآوردگرها را مستند می کند.استفاده از ما روش شناسی ، ما تأثیر قانونی بودن ماری جوانا تفریحی را بررسی می کنیم برای بزرگسالان در چندین ایالت ایالات متحده در مورد رفتار مصرف کوتاه مدت دانش آموزان دبیرستانی کلاس هشتم.نتایج حاکی از کوچک ، اما قابل توجه است افزایش احتمال مصرف در هر سطح.این اثرات بیشتر است تقویت شده پس از حسابداری برای گزارش نادرست.
29,941
Inspired by the activity signature introduced by Todorov and Tauchen (2010), which was used to measure the activity of a semimartingale, this paper introduces the roughness signature function. The paper illustrates how it can be used to determine whether a discretely observed process is generated by a continuous process that is rougher than a Brownian motion, a pure-jump process, or a combination of the two. Further, if a continuous rough process is present, the function gives an estimate of the roughness index. This is done through an extensive simulation study, where we find that the roughness signature function works as expected on rough processes. We further derive some asymptotic properties of this new signature function. The function is applied empirically to three different volatility measures for the S&P500 index. The three measures are realized volatility, the VIX, and the option-extracted volatility estimator of Todorov (2019). The realized volatility and option-extracted volatility show signs of roughness, with the option-extracted volatility appearing smoother than the realized volatility, while the VIX appears to be driven by a continuous martingale with jumps.
با الهام از امضای فعالیت معرفی شده توسط تودوروف و Tauchen (2010) ، که برای اندازه گیری فعالیت یک semimartingale ، این مقاله استفاده شده است عملکرد امضای زبری را معرفی می کند.مقاله نشان می دهد که چگونه می تواند برای تعیین اینکه آیا یک فرآیند مشاهده شده توسط a ایجاد می شود استفاده می شود فرآیند مداوم که سخت تر از یک حرکت براون ، یک فرآیند پرش خالص است ، یا ترکیبی از این دو.علاوه بر این ، اگر یک فرآیند خشن مداوم وجود داشته باشد ، این عملکرد تخمین از شاخص زبری را ارائه می دهد.این از طریق یک انجام می شود مطالعه شبیه سازی گسترده ، جایی که می دانیم عملکرد زبری امضای همانطور که انتظار می رود در فرآیندهای خشن انجام شود.ما بیشتر مقداری بدون علامت به دست می آوریم خواص این عملکرد جدید امضا.این عملکرد به صورت تجربی اعمال می شود به سه اقدامات مختلف نوسانات برای شاخص S&P500.سه اقدام نوسانات تحقق یافته ، VIX و برآوردگر نوسانات گزینه ای است از تادوروف (2019).نوسانات تحقق یافته و نمایش نوسانات گزینه ای علائم زبری ، با نوسانات گزینه ای که نرم و صاف تر به نظر می رسد از نوسانات تحقق یافته ، در حالی که VIX به نظر می رسد توسط a مارتینگال مداوم با پرش.
29,942
This paper applies a regularization procedure called increasing rearrangement to monotonize Edgeworth and Cornish-Fisher expansions and any other related approximations of distribution and quantile functions of sample statistics. Besides satisfying the logical monotonicity, required of distribution and quantile functions, the procedure often delivers strikingly better approximations to the distribution and quantile functions of the sample mean than the original Edgeworth-Cornish-Fisher expansions.
در این مقاله یک روش منظم به نام افزایش مجدد تنظیم شده است برای یکنواختی گسترش Edgeworth و Cornish-Fisher و سایر موارد مرتبط تقریبی توزیع و عملکردهای کمی از آمار نمونه. علاوه بر رضایت یکنواختی منطقی ، مورد نیاز توزیع و عملکردهای کمی ، این روش اغلب به طرز چشمگیری بهتر می شود تقریبی به توزیع و عملکردهای کمی از نمونه از گسترش اصلی Edgeworth-Cornish-Fisher.
29,943
In this paper, we develop a new censored quantile instrumental variable (CQIV) estimator and describe its properties and computation. The CQIV estimator combines Powell (1986) censored quantile regression (CQR) to deal with censoring, with a control variable approach to incorporate endogenous regressors. The CQIV estimator is obtained in two stages that are non-additive in the unobservables. The first stage estimates a non-additive model with infinite dimensional parameters for the control variable, such as a quantile or distribution regression model. The second stage estimates a non-additive censored quantile regression model for the response variable of interest, including the estimated control variable to deal with endogeneity. For computation, we extend the algorithm for CQR developed by Chernozhukov and Hong (2002) to incorporate the estimation of the control variable. We give generic regularity conditions for asymptotic normality of the CQIV estimator and for the validity of resampling methods to approximate its asymptotic distribution. We verify these conditions for quantile and distribution regression estimation of the control variable. Our analysis covers two-stage (uncensored) quantile regression with non-additive first stage as an important special case. We illustrate the computation and applicability of the CQIV estimator with a Monte-Carlo numerical example and an empirical application on estimation of Engel curves for alcohol.
در این مقاله ، ما یک متغیر ابزاری کمی سانسور شده ایجاد می کنیم (CQIV) برآوردگر و خصوصیات و محاسبات آن را شرح داده و توصیف کنید.CQIV برآوردگر پاول (1986) رگرسیون کمی سانسور شده (CQR) را برای معامله ترکیب می کند با سانسور ، با یک روش متغیر کنترل برای ترکیب درون زا رگرسیونبرآوردگر CQIV در دو مرحله که غیر ادعایی هستند بدست می آید در Unobservables.مرحله اول یک مدل غیر افزودنی با پارامترهای بعدی نامحدود برای متغیر کنترل ، مانند کمی یا مدل رگرسیون توزیع.مرحله دوم غیر ادعایی را تخمین می زند مدل رگرسیون کمی سانسور شده برای متغیر پاسخ علاقه ، از جمله متغیر کنترل تخمین زده شده برای مقابله با درون زا.برای محاسبه ، ما الگوریتم CQR را که توسط Chernozhukov و Hong تهیه شده است گسترش می دهیم (2002) برای درج تخمین متغیر کنترل.ما عمومی می دهیم شرایط منظم برای عادی بودن بدون علامت برآوردگر CQIV و برای اعتبار روشهای تغییر شکل مجدد برای تقریب توزیع بدون علامت آن. ما این شرایط را برای برآورد رگرسیون کمی و توزیع تأیید می کنیم متغیر کنترلتجزیه و تحلیل ما دو مرحله (بدون سانسور) را پوشش می دهد رگرسیون با مرحله اول غیر افزودنی به عنوان یک مورد ویژه مهم.ما محاسبه و کاربرد برآوردگر CQIV را با a نشان دهید نمونه عددی مونت کارلو و یک کاربرد تجربی در تخمین منحنی های انگل برای الکل.
29,950
The frequentist method of simulated minimum distance (SMD) is widely used in economics to estimate complex models with an intractable likelihood. In other disciplines, a Bayesian approach known as Approximate Bayesian Computation (ABC) is far more popular. This paper connects these two seemingly related approaches to likelihood-free estimation by means of a Reverse Sampler that uses both optimization and importance weighting to target the posterior distribution. Its hybrid features enable us to analyze an ABC estimate from the perspective of SMD. We show that an ideal ABC estimate can be obtained as a weighted average of a sequence of SMD modes, each being the minimizer of the deviations between the data and the model. This contrasts with the SMD, which is the mode of the average deviations. Using stochastic expansions, we provide a general characterization of frequentist estimators and those based on Bayesian computations including Laplace-type estimators. Their differences are illustrated using analytical examples and a simulation study of the dynamic panel model.
روش مکرر حداقل فاصله شبیه سازی شده (SMD) به طور گسترده ای در آن استفاده می شود اقتصاد برای برآورد مدلهای پیچیده با احتمال غیرقابل تحمل.در دیگر رشته ها ، یک رویکرد بیزی که به عنوان محاسبات تقریبی بیزی شناخته می شود (ABC) بسیار محبوب تر است.این مقاله این دو به ظاهر مرتبط را به هم وصل می کند رویکردهای برآورد بدون احتمال با استفاده از یک نمونه معکوس که برای هدف قرار دادن خلفی هم از بهینه سازی و هم از وزن اهمیت استفاده می کند توزیعویژگی های ترکیبی آن ما را قادر می سازد تا یک تخمین ABC را از آن تجزیه و تحلیل کنیم چشم انداز SMD.ما نشان می دهیم که یک تخمین ایده آل ABC را می توان به عنوان یک به دست آورد میانگین وزنی دنباله ای از حالتهای SMD ، هر کدام مینیم کننده از انحراف بین داده ها و مدل.این در تضاد با SMD است ، که حالت انحراف متوسط ​​است.با استفاده از گسترش های تصادفی ، ما ارائه می دهیم خصوصیات کلی برآوردگرهای مکرر و موارد مبتنی بر محاسبات بیزی از جمله برآوردگرهای نوع لاپلاس.تفاوت های آنها است با استفاده از نمونه های تحلیلی و یک مطالعه شبیه سازی پویا نشان داده شده است مدل پانل.
29,944
In applications it is common that the exact form of a conditional expectation is unknown and having flexible functional forms can lead to improvements. Series method offers that by approximating the unknown function based on $k$ basis functions, where $k$ is allowed to grow with the sample size $n$. We consider series estimators for the conditional mean in light of: (i) sharp LLNs for matrices derived from the noncommutative Khinchin inequalities, (ii) bounds on the Lebesgue factor that controls the ratio between the $L^\infty$ and $L_2$-norms of approximation errors, (iii) maximal inequalities for processes whose entropy integrals diverge, and (iv) strong approximations to series-type processes. These technical tools allow us to contribute to the series literature, specifically the seminal work of Newey (1997), as follows. First, we weaken the condition on the number $k$ of approximating functions used in series estimation from the typical $k^2/n \to 0$ to $k/n \to 0$, up to log factors, which was available only for spline series before. Second, we derive $L_2$ rates and pointwise central limit theorems results when the approximation error vanishes. Under an incorrectly specified model, i.e. when the approximation error does not vanish, analogous results are also shown. Third, under stronger conditions we derive uniform rates and functional central limit theorems that hold if the approximation error vanishes or not. That is, we derive the strong approximation for the entire estimate of the nonparametric function. We derive uniform rates, Gaussian approximations, and uniform confidence bands for a wide collection of linear functionals of the conditional expectation function.
در برنامه ها معمول است که شکل دقیق یک انتظار مشروط ناشناخته است و داشتن اشکال عملکردی انعطاف پذیر می تواند منجر به پیشرفت شود. روش سری ارائه می دهد که با تقریب عملکرد ناشناخته بر اساس $ k $ توابع پایه ، جایی که $ k $ با اندازه نمونه $ n $ رشد می کند.ما با توجه به: (i) LLN های شارپ ، برآوردگرهای سری را برای میانگین مشروط در نظر بگیرید برای ماتریس های حاصل از نابرابری های غیر متغیر Khinchin ، (ب) مرزها در فاکتور Lebesgue که نسبت بین $ L^\ infty $ و $ l_2 $ خطاهای تقریب ، (iii) نابرابری های حداکثر برای فرآیندها انتگرال های آنتروپی واژگون می شوند ، و (IV) تقریب های قوی به نوع سری فرآیندها این ابزارهای فنی به ما امکان می دهد تا در ادبیات سریال مشارکت کنیم ، به طور خاص کار اصلی Newey (1997) ، به شرح زیر است.اول ، ما تضعیف می کنیم شرط در تعداد $ k $ عملکردهای تقریبی مورد استفاده در سری تخمین از $ k k^2/n \ به 0 $ k/n \ تا 0 $ ، تا فاکتورهای ورود که قبلاً فقط برای سری Spline در دسترس بود.دوم ، ما $ L_2 $ را استخراج می کنیم نرخ ها و قضیه های محدودیت مرکزی در هنگام خطای تقریب نتیجه می گیرند ناپدید می شودتحت یک مدل نادرست مشخص شده ، یعنی هنگامی که تقریب خطا از بین نمی رود ، نتایج مشابه نیز نشان داده شده است.سوم ، تحت قوی تر شرایطی که ما نرخ یکنواخت و قضایای محدودیت مرکزی عملکردی را به دست می آوریم که اگر خطای تقریب از بین می رود یا نه ، نگه دارید.یعنی ما قوی را استخراج می کنیم تقریب برای کل تخمین عملکرد غیر پارامتری. ما نرخ یکنواخت ، تقریب گاوسی و اعتماد به نفس یکنواخت را به دست می آوریم باندهای مجموعه گسترده ای از کارکردهای خطی مشروط عملکرد انتظار
29,945
In this article, we review quantile models with endogeneity. We focus on models that achieve identification through the use of instrumental variables and discuss conditions under which partial and point identification are obtained. We discuss key conditions, which include monotonicity and full-rank-type conditions, in detail. In providing this review, we update the identification results of Chernozhukov and Hansen (2005, Econometrica). We illustrate the modeling assumptions through economically motivated examples. We also briefly review the literature on estimation and inference. Key Words: identification, treatment effects, structural models, instrumental variables
در این مقاله ، ما مدلهای کمی را با درون زایی مرور می کنیم.ما روی تمرکز می کنیم مدلهایی که با استفاده از متغیرهای ابزاری به شناسایی می رسند و در مورد شرایطی که در آن شناسایی جزئی و نقطه ای قرار دارند بحث کنید به دست آمده.ما در مورد شرایط کلیدی بحث می کنیم ، که شامل یکنواختی و شرایط کامل از نوع ، با جزئیات.در ارائه این بررسی ، ما نتایج شناسایی Chernozhukov و Hansen (2005 ، Econometrica).ما فرضیات مدل سازی را از طریق نمونه های با انگیزه اقتصادی نشان دهید.ما همچنین به طور خلاصه ادبیات در مورد تخمین و استنباط را مرور کنید. کلمات کلیدی: شناسایی ، اثرات درمانی ، مدل های ساختاری ، ابزاری متغیرها
29,946
We derive fixed effects estimators of parameters and average partial effects in (possibly dynamic) nonlinear panel data models with individual and time effects. They cover logit, probit, ordered probit, Poisson and Tobit models that are important for many empirical applications in micro and macroeconomics. Our estimators use analytical and jackknife bias corrections to deal with the incidental parameter problem, and are asymptotically unbiased under asymptotic sequences where $N/T$ converges to a constant. We develop inference methods and show that they perform well in numerical examples.
ما برآوردگرهای اثرات ثابت پارامترها و میانگین اثرات جزئی را استخراج می کنیم در مدلهای داده پانل غیرخطی (احتمالاً پویا) با فردی و زمان اثراتآنها Logit ، Probit ، Probit ، Poisson و Tobit را پوشش می دهند این برای بسیاری از کاربردهای تجربی در اقتصاد خرد و کلان مهم است. برآوردگرهای ما از اصلاحات تعصب تحلیلی و jackknife برای مقابله با مشکل پارامتر حادثه ای ، و از نظر بدون علامت تحت مجانبی هستند توالی هایی که $ n/t $ به یک ثابت همگرا می شود.ما روشهای استنباط را توسعه می دهیم و نشان می دهد که آنها در نمونه های عددی عملکرد خوبی دارند.
29,947
This paper considers identification and estimation of ceteris paribus effects of continuous regressors in nonseparable panel models with time homogeneity. The effects of interest are derivatives of the average and quantile structural functions of the model. We find that these derivatives are identified with two time periods for "stayers", i.e. for individuals with the same regressor values in two time periods. We show that the identification results carry over to models that allow location and scale time effects. We propose nonparametric series methods and a weighted bootstrap scheme to estimate and make inference on the identified effects. The bootstrap proposed allows uniform inference for function-valued parameters such as quantile effects uniformly over a region of quantile indices and/or regressor values. An empirical application to Engel curve estimation with panel data illustrates the results.
در این مقاله ، شناسایی و تخمین اثرات paribus ceteris در نظر گرفته شده است رگرسیون مداوم در مدلهای پانل غیر قابل جدا شدن با همگن زمان. اثرات علاقه مشتقات ساختاری متوسط ​​و کمی است توابع مدل.ما می دانیم که این مشتقات با دو نفر مشخص می شوند دوره های زمانی برای "Stayers" ، یعنی برای افرادی که دارای همان ارزش های رگرسیون هستند در دو دوره زمانی.ما نشان می دهیم که نتایج شناسایی به مدلهایی که امکان مکان و اثرات زمانی را فراهم می کنند.ما غیرپارامتری را پیشنهاد می کنیم روش های سری و یک طرح بوت استرپ وزنه ای برای برآورد و استنباط در مورد اثرات مشخص شدهbootstrap پیشنهادی اجازه می دهد تا استنباط یکنواخت برای پارامترهای با ارزش عملکرد مانند اثرات کمی به طور یکنواخت در یک منطقه از شاخص های کمی و/یا مقادیر رگرسیون.یک برنامه تجربی برای انگل برآورد منحنی با داده های پانل نتایج را نشان می دهد.
29,948
We consider estimation and inference in panel data models with additive unobserved individual specific heterogeneity in a high dimensional setting. The setting allows the number of time varying regressors to be larger than the sample size. To make informative estimation and inference feasible, we require that the overall contribution of the time varying variables after eliminating the individual specific heterogeneity can be captured by a relatively small number of the available variables whose identities are unknown. This restriction allows the problem of estimation to proceed as a variable selection problem. Importantly, we treat the individual specific heterogeneity as fixed effects which allows this heterogeneity to be related to the observed time varying variables in an unspecified way and allows that this heterogeneity may be non-zero for all individuals. Within this framework, we provide procedures that give uniformly valid inference over a fixed subset of parameters in the canonical linear fixed effects model and over coefficients on a fixed vector of endogenous variables in panel data instrumental variables models with fixed effects and many instruments. An input to developing the properties of our proposed procedures is the use of a variant of the Lasso estimator that allows for a grouped data structure where data across groups are independent and dependence within groups is unrestricted. We provide formal conditions within this structure under which the proposed Lasso variant selects a sparse model with good approximation properties. We present simulation results in support of the theoretical developments and illustrate the use of the methods in an application aimed at estimating the effect of gun prevalence on crime rates.
ما تخمین و استنباط در مدل های داده پانل را با افزودنی در نظر می گیریم ناهمگونی خاص فردی در یک محیط ابعاد بالا.در تنظیم اجازه می دهد تا تعداد رگرسیونرهای مختلف از زمان بزرگتر باشد اندازهی نمونه.برای ایجاد تخمین آموزنده و استنباط عملی ، ما نیاز داریم که سهم کلی متغیرهای مختلف پس از از بین بردن ناهمگونی خاص فردی را می توان با یک نسبتاً کوچک اسیر کرد تعداد متغیرهای موجود که هویت آنها ناشناخته است.این محدودیت به مشکل تخمین اجازه می دهد تا به عنوان یک انتخاب متغیر ادامه یابد مسئله.نکته مهم ، ما با ناهمگونی خاص فردی به عنوان ثابت رفتار می کنیم اثراتی که اجازه می دهد این ناهمگونی مربوط به زمان مشاهده شده باشد متغیرهای مختلف به روشی نامشخص و اجازه می دهد تا این ناهمگونی ممکن است برای همه افراد غیر صفر باشید.در این چارچوب ، ما رویه هایی را ارائه می دهیم که استنتاج یکنواخت معتبر را بر روی یک زیر مجموعه ثابت از پارامترها در مدل جلوه های ثابت خطی متعارف و بیش از ضرایب در یک بردار ثابت از متغیرهای درون زا در داده های پانل متغیرهای ابزاری با ثابت اثرات و بسیاری از ابزارها.ورودی برای توسعه خواص ما روشهای پیشنهادی استفاده از یک نوع از برآوردگر Lasso است که اجازه می دهد برای یک ساختار داده گروهی که در آن داده ها در گروه ها مستقل هستند و وابستگی در گروه ها بدون محدودیت است.ما شرایط رسمی را در داخل ارائه می دهیم این ساختار که در آن نوع Lasso پیشنهادی یک مدل پراکنده را انتخاب می کند با خواص تقریبی خوب.ما نتایج شبیه سازی را در حمایت از ارائه می دهیم تحولات نظری و استفاده از روشها را در یک نشان می دهد کاربرد با هدف برآورد تأثیر شیوع اسلحه بر میزان جرم.
29,949
Factor structures or interactive effects are convenient devices to incorporate latent variables in panel data models. We consider fixed effect estimation of nonlinear panel single-index models with factor structures in the unobservables, which include logit, probit, ordered probit and Poisson specifications. We establish that fixed effect estimators of model parameters and average partial effects have normal distributions when the two dimensions of the panel grow large, but might suffer of incidental parameter bias. We show how models with factor structures can also be applied to capture important features of network data such as reciprocity, degree heterogeneity, homophily in latent variables and clustering. We illustrate this applicability with an empirical example to the estimation of a gravity equation of international trade between countries using a Poisson model with multiple factors.
ساختار فاکتور یا اثرات تعاملی دستگاه های مناسب برای متغیرهای نهفته را در مدل های داده پانل ترکیب کنید.ما اثر ثابت را در نظر می گیریم تخمین مدل های تک شاخص پانل غیرخطی با ساختار فاکتور در Unobservables ، که شامل ورود به سیستم ، Probit ، Probit و Poisson است مشخصات فنی.ما تعیین می کنیم که برآوردگرهای اثر ثابت از پارامترهای مدل و میانگین اثرات جزئی در هنگام دو بعد توزیع طبیعی دارد پانل بزرگ رشد می کند ، اما ممکن است از تعصب پارامتر حادثه رنج ببرد.ما نشان می دهیم چگونه می توان از مدل هایی با ساختار عاملی نیز برای گرفتن مهم استفاده کرد ویژگی های داده های شبکه مانند متقابل ، ناهمگونی درجه ، هموفیلی در متغیرهای نهفته و خوشه بندی.ما این کاربرد را با یک نشان می دهیم مثال تجربی برای برآورد معادله گرانش بین المللی تجارت بین کشورها با استفاده از یک مدل پواسون با چندین عامل.
29,951
In this note, we offer an approach to estimating causal/structural parameters in the presence of many instruments and controls based on methods for estimating sparse high-dimensional models. We use these high-dimensional methods to select both which instruments and which control variables to use. The approach we take extends BCCH2012, which covers selection of instruments for IV models with a small number of controls, and extends BCH2014, which covers selection of controls in models where the variable of interest is exogenous conditional on observables, to accommodate both a large number of controls and a large number of instruments. We illustrate the approach with a simulation and an empirical example. Technical supporting material is available in a supplementary online appendix.
در این یادداشت ، ما رویکردی برای برآورد پارامترهای علی و ساختاری ارائه می دهیم در حضور بسیاری از ابزارها و کنترل ها بر اساس روشهای تخمین مدل های پراکنده با ابعاد بالا.ما از این ابعاد بالا استفاده می کنیم روش هایی برای انتخاب هر دو ابزار و کنترل متغیرها برای استفاده. رویکردی که ما از آن استفاده می کنیم BCCH2012 را گسترش می دهد ، که شامل انتخاب ابزارها می شود برای مدل های IV با تعداد کمی کنترل ، و BCH2014 را گسترش می دهد ، که انتخاب کنترل در مدلهایی که متغیر علاقه است مشروط اگزوژن بر روی مشاهدات ، برای قرار دادن تعداد زیادی از آنها کنترل و تعداد زیادی ابزار.ما رویکرد را با a نشان می دهیم شبیه سازی و یک مثال تجربی.مواد پشتیبانی فنی در دسترس است در یک پیوست آنلاین تکمیلی.
29,952
We study Markov decision problems where the agent does not know the transition probability function mapping current states and actions to future states. The agent has a prior belief over a set of possible transition functions and updates beliefs using Bayes' rule. We allow her to be misspecified in the sense that the true transition probability function is not in the support of her prior. This problem is relevant in many economic settings but is usually not amenable to analysis by the researcher. We make the problem tractable by studying asymptotic behavior. We propose an equilibrium notion and provide conditions under which it characterizes steady state behavior. In the special case where the problem is static, equilibrium coincides with the single-agent version of Berk-Nash equilibrium (Esponda and Pouzo (2016)). We also discuss subtle issues that arise exclusively in dynamic settings due to the possibility of a negative value of experimentation.
ما مشکلات تصمیم مارکوف را مطالعه می کنیم که نماینده آن را نمی داند عملکرد احتمال انتقال نقشه برداری از حالتهای فعلی و اقدامات به آینده ایالت ها.نماینده نسبت به مجموعه ای از انتقال احتمالی اعتقاد قبلی دارد عملکرد و به روزرسانی باورها با استفاده از قانون Bayes.ما به او اجازه می دهیم که باشد اشتباه به این معنا که عملکرد احتمال انتقال واقعی نیست در حمایت از قبلی او.این مشکل در بسیاری از تنظیمات اقتصادی مرتبط است اما معمولاً توسط محقق قابل تجزیه و تحلیل نیست.ما مشکل را ایجاد می کنیم با مطالعه رفتار بدون علامت قابل کنترل است.ما یک مفهوم تعادل را پیشنهاد می کنیم و شرایطی را فراهم کنید که تحت آن رفتار حالت پایدار باشد.در مورد خاص که مشکل استاتیک است ، تعادل همزمان با نسخه تک عامل Berk-Nash تعادل (Esponda and Pouzo (2016)).ما همچنین در مورد موضوعات ظریف که منحصراً در تنظیمات پویا بوجود می آیند ، بحث کنید احتمال یک مقدار منفی آزمایش.
29,953
In this paper we study the problems of estimating heterogeneity in causal effects in experimental or observational studies and conducting inference about the magnitude of the differences in treatment effects across subsets of the population. In applications, our method provides a data-driven approach to determine which subpopulations have large or small treatment effects and to test hypotheses about the differences in these effects. For experiments, our method allows researchers to identify heterogeneity in treatment effects that was not specified in a pre-analysis plan, without concern about invalidating inference due to multiple testing. In most of the literature on supervised machine learning (e.g. regression trees, random forests, LASSO, etc.), the goal is to build a model of the relationship between a unit's attributes and an observed outcome. A prominent role in these methods is played by cross-validation which compares predictions to actual outcomes in test samples, in order to select the level of complexity of the model that provides the best predictive power. Our method is closely related, but it differs in that it is tailored for predicting causal effects of a treatment rather than a unit's outcome. The challenge is that the "ground truth" for a causal effect is not observed for any individual unit: we observe the unit with the treatment, or without the treatment, but not both at the same time. Thus, it is not obvious how to use cross-validation to determine whether a causal effect has been accurately predicted. We propose several novel cross-validation criteria for this problem and demonstrate through simulations the conditions under which they perform better than standard methods for the problem of causal effects. We then apply the method to a large-scale field experiment re-ranking results on a search engine.
در این مقاله مشکلات تخمین ناهمگونی در علت را بررسی می کنیم اثرات در مطالعات تجربی یا مشاهده ای و استنباط در مورد میزان تفاوت در اثرات درمانی در زیر مجموعه های جمعیتدر برنامه ها ، روش ما یک رویکرد داده محور را ارائه می دهد تعیین کنید که زیرمجموعات دارای اثرات درمانی بزرگ یا کوچک هستند و به فرضیه های آزمایش در مورد تفاوت در این اثرات.برای آزمایشات ، ما روش به محققان اجازه می دهد تا ناهمگونی در اثرات درمانی را شناسایی کنند در یک برنامه قبل از تجزیه و تحلیل ، بدون نگرانی در مورد باطل کردن مشخص نشده است استنتاج به دلیل آزمایش های متعدد.در بیشتر ادبیات مربوط به نظارت یادگیری ماشین (به عنوان مثال درختان رگرسیون ، جنگل های تصادفی ، لاسو و غیره) ، هدف ایجاد الگویی از رابطه بین ویژگی های یک واحد و یک نتیجه مشاهده شدهنقش برجسته ای در این روش ها توسط اعتبار سنجی متقابل که پیش بینی ها را با نتایج واقعی در نمونه های آزمایش مقایسه می کند ، به منظور انتخاب سطح پیچیدگی مدل که بهترین ها را ارائه می دهد قدرت پیش بینی کنندهروش ما از نزدیک مرتبط است ، اما از این نظر متفاوت است متناسب با پیش بینی اثرات علّی یک درمان به جای یک واحد نتیجهچالش این است که "حقیقت زمین" برای یک اثر علّی نیست برای هر واحد فردی مشاهده می شود: ما واحد را با درمان مشاهده می کنیم ، یا بدون درمان ، اما هر دو به طور همزمان نیست.بنابراین ، واضح نیست نحوه استفاده از اعتبارسنجی متقاطع برای تعیین اینکه آیا یک اثر علّی بوده است با دقت پیش بینی شده است.ما چندین معیار اعتبار سنجی متقابل جدید را برای این مشکل را نشان می دهد و از طریق شبیه سازی شرایطی را نشان می دهد که تحت آن آنها بهتر از روشهای استاندارد برای مشکل اثرات علی انجام می دهند.ما سپس روش را در یک آزمایش میدانی در مقیاس بزرگ قرار دهید موتور جستجو
29,954
This paper makes several important contributions to the literature about nonparametric instrumental variables (NPIV) estimation and inference on a structural function $h_0$ and its functionals. First, we derive sup-norm convergence rates for computationally simple sieve NPIV (series 2SLS) estimators of $h_0$ and its derivatives. Second, we derive a lower bound that describes the best possible (minimax) sup-norm rates of estimating $h_0$ and its derivatives, and show that the sieve NPIV estimator can attain the minimax rates when $h_0$ is approximated via a spline or wavelet sieve. Our optimal sup-norm rates surprisingly coincide with the optimal root-mean-squared rates for severely ill-posed problems, and are only a logarithmic factor slower than the optimal root-mean-squared rates for mildly ill-posed problems. Third, we use our sup-norm rates to establish the uniform Gaussian process strong approximations and the score bootstrap uniform confidence bands (UCBs) for collections of nonlinear functionals of $h_0$ under primitive conditions, allowing for mildly and severely ill-posed problems. Fourth, as applications, we obtain the first asymptotic pointwise and uniform inference results for plug-in sieve t-statistics of exact consumer surplus (CS) and deadweight loss (DL) welfare functionals under low-level conditions when demand is estimated via sieve NPIV. Empiricists could read our real data application of UCBs for exact CS and DL functionals of gasoline demand that reveals interesting patterns and is applicable to other markets.
در این مقاله چندین کمک مهم در ادبیات در مورد متغیرهای ابزاری غیر پارامتری (NPIV) و استنباط در a عملکرد ساختاری $ H_0 $ و عملکرد آن.اول ، ما sup-norm را استخراج می کنیم نرخ همگرایی برای SIEVE ساده محاسباتی NPIV (سری 2SLS) برآوردگرهای $ H_0 $ و مشتقات آن.دوم ، ما یک محدوده پایین را به دست می آوریم بهترین نرخ ممکن (Minimax) SUP-NORM تخمین H_0 $ را توصیف می کند و مشتقات آن ، و نشان می دهد که برآوردگر SIEVE NPIV می تواند به حداقل برسد نرخ هنگامی که $ h_0 $ از طریق یک غربال Spline یا Wavelet تقریب می یابد.بهینه ما نرخ SUP-NORM به طرز شگفت آور با نرخ های بهینه مربعات بهینه بهینه همزمان می شود برای مشکلات شدید در معرض خطر ، و فقط یک عامل لگاریتمی کندتر از است نرخ بهینه میانگین مربعات ریشه برای مشکلات خفیف در معرض.سوم ، ما از نرخ های SUP-norm ما برای ایجاد فرآیند یکنواخت گاوسی استفاده کنید تقریبی و باندهای اعتماد به نفس یکنواخت (UCBS) برای Bootstrap Score Bootstrap مجموعه عملکردهای غیرخطی $ H_0 $ در شرایط ابتدایی ، اجازه می دهد مشکلات خفیف و شدید بدبختی.چهارم ، به عنوان برنامه ها ، ما اولین نتایج بدون علامت و استنباط یکنواخت را برای پلاگین غربال T-Statistic از مازاد مصرف کننده دقیق (CS) و کاهش وزن مرده (DL) عملکردهای رفاهی در شرایط پایین هنگام تخمین تقاضا از طریق غربال NPIV.تجربی گرایان می توانند کاربرد داده های واقعی ما از UCB را بخوانند عملکرد دقیق CS و DL تقاضای بنزین که جالب است الگوهای و برای بازارهای دیگر کاربرد دارد.
29,955
Oversubscribed treatments are often allocated using randomized waiting lists. Applicants are ranked randomly, and treatment offers are made following that ranking until all seats are filled. To estimate causal effects, researchers often compare applicants getting and not getting an offer. We show that those two groups are not statistically comparable. Therefore, the estimator arising from that comparison is inconsistent. We propose a new estimator, and show that it is consistent. Finally, we revisit an application, and we show that using our estimator can lead to sizably different results from those obtained using the commonly used estimator.
درمان های بیش از حد با استفاده از لیست انتظار تصادفی اختصاص می یابد. متقاضیان به طور تصادفی رتبه بندی می شوند و پیشنهادات درمانی پس از آن انجام می شود رتبه بندی تا زمانی که همه صندلی ها پر شوند.برای برآورد اثرات علی ، محققان اغلب متقاضیان دریافت و دریافت پیشنهاد را مقایسه می کنند.ما نشان می دهیم که آنها دو گروه از نظر آماری قابل مقایسه نیستند.بنابراین ، برآوردگر بوجود می آید از آن مقایسه متناقض است.ما یک برآوردگر جدید پیشنهاد می کنیم و نشان می دهیم که سازگار استسرانجام ، ما یک برنامه را مجدداً مورد بررسی قرار می دهیم ، و استفاده از آن را نشان می دهیم برآوردگر ما می تواند منجر به نتایج قابل توجهی متفاوت از نتایج به دست آمده با استفاده شود برآوردگر متداول استفاده می شود.
29,967
The R package quantreg.nonpar implements nonparametric quantile regression methods to estimate and make inference on partially linear quantile models. quantreg.nonpar obtains point estimates of the conditional quantile function and its derivatives based on series approximations to the nonparametric part of the model. It also provides pointwise and uniform confidence intervals over a region of covariate values and/or quantile indices for the same functions using analytical and resampling methods. This paper serves as an introduction to the package and displays basic functionality of the functions contained within.
بسته R quantreg.nonpar رگرسیون کمی غیرپارامتری را پیاده سازی می کند روشهای تخمین و استنباط در مدلهای کمی خطی. quantreg.nonpar برآورد نقطه ای از عملکرد کمی مشروط را بدست می آورد و مشتقات آن بر اساس تقریب سری به قسمت غیرپارامتری از مدل.همچنین فواصل اعتماد به نفس و یکنواخت را از طریق a فراهم می کند منطقه مقادیر متغیر و/یا شاخص های کمی برای عملکردهای مشابه با استفاده از روشهای تحلیلی و تغییر شکل مجدد.این مقاله به عنوان مقدمه ای برای بسته بندی و عملکرد اصلی توابع موجود در آن را نشان می دهد.
29,956
The partial (ceteris paribus) effects of interest in nonlinear and interactive linear models are heterogeneous as they can vary dramatically with the underlying observed or unobserved covariates. Despite the apparent importance of heterogeneity, a common practice in modern empirical work is to largely ignore it by reporting average partial effects (or, at best, average effects for some groups). While average effects provide very convenient scalar summaries of typical effects, by definition they fail to reflect the entire variety of the heterogeneous effects. In order to discover these effects much more fully, we propose to estimate and report sorted effects -- a collection of estimated partial effects sorted in increasing order and indexed by percentiles. By construction the sorted effect curves completely represent and help visualize the range of the heterogeneous effects in one plot. They are as convenient and easy to report in practice as the conventional average partial effects. They also serve as a basis for classification analysis, where we divide the observational units into most or least affected groups and summarize their characteristics. We provide a quantification of uncertainty (standard errors and confidence bands) for the estimated sorted effects and related classification analysis, and provide confidence sets for the most and least affected groups. The derived statistical results rely on establishing key, new mathematical results on Hadamard differentiability of a multivariate sorting operator and a related classification operator, which are of independent interest. We apply the sorted effects method and classification analysis to demonstrate several striking patterns in the gender wage gap.
اثرات جزئی (ceteris paribus) علاقه به علاقه به غیرخطی و مدلهای خطی تعاملی ناهمگن هستند زیرا می توانند با چشمگیر متفاوت باشند متغیرهای زیربنایی مشاهده شده یا بدون نظارت.با وجود ظاهری اهمیت ناهمگونی ، یک روش معمول در کار تجربی مدرن است با گزارش متوسط ​​اثرات جزئی (یا در بهترین حالت ، متوسط ​​، آن را نادیده بگیرید اثرات برخی از گروه ها).در حالی که اثرات متوسط ​​مقیاس بسیار مناسب را ارائه می دهد خلاصه ای از جلوه های معمولی ، با تعریف آنها نتوانند کل را منعکس کنند تنوع اثرات ناهمگن.به منظور کشف این تأثیرات بسیار به طور کامل تر ، ما پیشنهاد می کنیم اثرات مرتب شده را تخمین و گزارش دهیم - مجموعه ای از اثرات جزئی تخمین زده شده در افزایش نظم و فهرست بندی شده توسط صدک هابا ساخت منحنی های اثر مرتب شده کاملاً نمایانگر و به تجسم دامنه اثرات ناهمگن در یک طرح کمک کنید.آنها مانند هستند گزارش مناسب و آسان در عمل به عنوان میانگین جزئی متعارف اثراتآنها همچنین به عنوان پایه ای برای تجزیه و تحلیل طبقه بندی ، جایی که ما هستیم واحدهای مشاهده ای را به اکثر یا کمترین گروه ها تقسیم کرده و خلاصه کنید خصوصیات آنهاما کمکی از عدم اطمینان (استاندارد) را ارائه می دهیم خطاها و باندهای اعتماد به نفس) برای اثرات مرتب شده و مربوط به آن تجزیه و تحلیل طبقه بندی ، و برای کمترین و کمترین مجموعه اعتماد به نفس ارائه می دهد گروه های آسیب دیدهنتایج آماری مشتق شده به ایجاد کلید ، جدید متکی است نتایج ریاضی در مورد متفاوت بودن هادامارد از یک مرتب سازی چند متغیره اپراتور و یک اپراتور طبقه بندی مرتبط ، که مستقل هستند علاقه.ما از روش اثرات مرتب شده و تجزیه و تحلیل طبقه بندی استفاده می کنیم چندین الگوی برجسته در شکاف دستمزد جنسیت نشان دهید.
29,957
In this paper, we propose a doubly robust method to present the heterogeneity of the average treatment effect with respect to observed covariates of interest. We consider a situation where a large number of covariates are needed for identifying the average treatment effect but the covariates of interest for analyzing heterogeneity are of much lower dimension. Our proposed estimator is doubly robust and avoids the curse of dimensionality. We propose a uniform confidence band that is easy to compute, and we illustrate its usefulness via Monte Carlo experiments and an application to the effects of smoking on birth weights.
در این مقاله ، ما یک روش مضاعف قوی برای ارائه ناهمگونی پیشنهاد می کنیم از میانگین اثر درمانی با توجه به متغیرهای مشاهده شده علاقه.ما وضعیتی را در نظر می گیریم که تعداد زیادی از متغیرهای متغیر مورد نیاز است برای شناسایی میانگین اثر درمانی اما متغیرهای مورد علاقه برای تجزیه و تحلیل ناهمگونی از ابعاد بسیار کمتری برخوردار است.برآوردگر پیشنهادی ما است دو برابر قوی و از لعنت ابعاد جلوگیری می کند.ما یکنواخت را پیشنهاد می کنیم باند اعتماد به نفس که محاسبه آن آسان است ، و ما سودمندی آن را از طریق نشان می دهیم آزمایش مونت کارلو و کاربردی در مورد اثرات سیگار کشیدن بر تولد وزن
29,958
In a unified framework, we provide estimators and confidence bands for a variety of treatment effects when the outcome of interest, typically a duration, is subjected to right censoring. Our methodology accommodates average, distributional, and quantile treatment effects under different identifying assumptions including unconfoundedness, local treatment effects, and nonlinear differences-in-differences. The proposed estimators are easy to implement, have close-form representation, are fully data-driven upon estimation of nuisance parameters, and do not rely on parametric distributional assumptions, shape restrictions, or on restricting the potential treatment effect heterogeneity across different subpopulations. These treatment effects results are obtained as a consequence of more general results on two-step Kaplan-Meier estimators that are of independent interest: we provide conditions for applying (i) uniform law of large numbers, (ii) functional central limit theorems, and (iii) we prove the validity of the ordinary nonparametric bootstrap in a two-step estimation procedure where the outcome of interest may be randomly censored.
در یک چارچوب یکپارچه ، ما برآوردگرها و باند های اعتماد به نفس را برای a ارائه می دهیم انواع مختلفی از اثرات درمانی هنگامی که نتیجه مورد علاقه ، به طور معمول a مدت زمان ، در معرض سانسور درست است.روش ما در جای خود قرار دارد اثرات درمانی متوسط ​​، توزیع و کمی تحت متفاوت شناسایی فرضیات از جمله عدم تمایل ، اثرات درمانی محلی ، و تفاوت های غیرخطی در اختلافات.برآوردگرهای پیشنهادی آسان هستند پیاده سازی ، نمایندگی از نزدیک ، کاملاً داده شده بر روی برآورد پارامترهای مزاحمت ، و به توزیع پارامتری متکی نیست فرضیات ، محدودیت های شکل یا محدود کردن درمان بالقوه ناهمگونی اثر در زیر گروه های مختلف.این اثرات درمانی نتایج به عنوان یک نتیجه از نتایج کلی تر در دو مرحله بدست می آید برآوردگرهای Kaplan-Meier که مورد علاقه مستقل هستند: ما شرایط را ارائه می دهیم برای اعمال (i) قانون یکنواخت تعداد زیاد ، (ب) حد اصلی عملکردی قضیه ها ، و (iii) ما اعتبار غیرپارامتری معمولی را اثبات می کنیم بوت استرپ در یک روش تخمین دو مرحله ای که نتیجه علاقه ممکن است به طور تصادفی سانسور شود.
29,959
In this review, we present econometric and statistical methods for analyzing randomized experiments. For basic experiments we stress randomization-based inference as opposed to sampling-based inference. In randomization-based inference, uncertainty in estimates arises naturally from the random assignment of the treatments, rather than from hypothesized sampling from a large population. We show how this perspective relates to regression analyses for randomized experiments. We discuss the analyses of stratified, paired, and clustered randomized experiments, and we stress the general efficiency gains from stratification. We also discuss complications in randomized experiments such as non-compliance. In the presence of non-compliance we contrast intention-to-treat analyses with instrumental variables analyses allowing for general treatment effect heterogeneity. We consider in detail estimation and inference for heterogeneous treatment effects in settings with (possibly many) covariates. These methods allow researchers to explore heterogeneity by identifying subpopulations with different treatment effects while maintaining the ability to construct valid confidence intervals. We also discuss optimal assignment to treatment based on covariates in such settings. Finally, we discuss estimation and inference in experiments in settings with interactions between units, both in general network settings and in settings where the population is partitioned into groups with all interactions contained within these groups.
در این بررسی ، ما روشهای اقتصاد سنجی و آماری را برای تجزیه و تحلیل ارائه می دهیم آزمایش های تصادفی.برای آزمایش های اساسی ما استرس مبتنی بر تصادفی است استنباط بر خلاف استنباط مبتنی بر نمونه گیری.مبتنی بر تصادفی استنباط ، عدم اطمینان در برآوردها به طور طبیعی از انتساب تصادفی ناشی می شود از درمانها ، به جای نمونه گیری فرضیه از یک بزرگ جمعیتما نشان می دهیم که چگونه این دیدگاه با تجزیه و تحلیل رگرسیون ارتباط دارد آزمایش های تصادفی.ما در مورد تحلیل های طبقه بندی شده ، زوج و آزمایش های تصادفی خوشه ای ، و ما بر سود کارآیی عمومی تأکید می کنیم از طبقه بندیما همچنین در مورد عوارض در آزمایشات تصادفی بحث می کنیم مانند عدم رعایتدر حضور عدم رعایت ما تضاد می کنیم تجزیه و تحلیل قصد برای درمان با تجزیه و تحلیل متغیرهای ابزاری امکان پذیر است ناهمگونی اثر درمانی عمومی.ما تخمین جزئیات را در نظر می گیریم و استنباط اثرات درمانی ناهمگن در تنظیمات با (احتمالاً بسیاری) متغیرهای متغیراین روشها به محققان اجازه می دهد تا ناهمگونی را توسط شناسایی زیرمجموعات با اثرات درمانی مختلف ضمن حفظ توانایی ساخت فواصل اطمینان معتبر.ما همچنین بهینه بحث می کنیم تکلیف به درمان بر اساس متغیرهای متغیر در چنین تنظیماتی.بالاخره ، ما در مورد تخمین و استنباط در آزمایشات در تنظیمات با تعامل بحث کنید بین واحدها ، هم در تنظیمات شبکه عمومی و هم در تنظیماتی که در آن جمعیت با تمام تعامل های موجود در گروه ها تقسیم می شود این گروه ها
29,960
In this paper we discuss recent developments in econometrics that we view as important for empirical researchers working on policy evaluation questions. We focus on three main areas, where in each case we highlight recommendations for applied work. First, we discuss new research on identification strategies in program evaluation, with particular focus on synthetic control methods, regression discontinuity, external validity, and the causal interpretation of regression methods. Second, we discuss various forms of supplementary analyses to make the identification strategies more credible. These include placebo analyses as well as sensitivity and robustness analyses. Third, we discuss recent advances in machine learning methods for causal effects. These advances include methods to adjust for differences between treated and control units in high-dimensional settings, and methods for identifying and estimating heterogeneous treatment effects.
در این مقاله ما در مورد تحولات اخیر در اقتصاد سنجی که به عنوان آنها مشاهده می کنیم بحث می کنیم برای محققان تجربی که در سوالات ارزیابی سیاست کار می کنند مهم است.ما روی سه حوزه اصلی تمرکز کنید ، جایی که در هر مورد توصیه هایی را برای آنها برجسته می کنیم کار کاربردیاول ، ما در مورد تحقیقات جدید در مورد استراتژی های شناسایی در ارزیابی برنامه ، با تمرکز ویژه بر روی روشهای کنترل مصنوعی ، ناپیوستگی رگرسیون ، اعتبار خارجی و تفسیر علی روشهای رگرسیون.دوم ، ما در مورد اشکال مختلف تجزیه و تحلیل تکمیلی بحث می کنیم برای اعتبار بیشتر استراتژی های شناسایی.این شامل دارونما است تجزیه و تحلیل و همچنین تجزیه و تحلیل حساسیت و استحکام.سوم ، ما بحث می کنیم پیشرفت های اخیر در روش های یادگیری ماشین برای اثرات علّی.این پیشرفت ها شامل روش هایی برای تنظیم تفاوت بین واحدهای تحت درمان و کنترل در تنظیمات با ابعاد بالا و روشهای شناسایی و تخمین اثرات درمانی ناهمگن.
29,974
In treatment allocation problems the individuals to be treated often arrive sequentially. We study a problem in which the policy maker is not only interested in the expected cumulative welfare but is also concerned about the uncertainty/risk of the treatment outcomes. At the outset, the total number of treatment assignments to be made may even be unknown. A sequential treatment policy which attains the minimax optimal regret is proposed. We also demonstrate that the expected number of suboptimal treatments only grows slowly in the number of treatments. Finally, we study a setting where outcomes are only observed with delay.
در مشکلات تخصیص درمان ، افراد تحت درمان قرار می گیرند متوالیما مشکلی را مطالعه می کنیم که در آن سازنده سیاست نه تنها است علاقه مند به رفاه تجمعی مورد انتظار اما همچنین نگران این است عدم اطمینان/خطر نتایج درمان.در ابتدا ، تعداد کل تکالیف درمانی که باید انجام شود حتی ممکن است ناشناخته باشد.یک درمان متوالی سیاستی که حسرت بهینه Minimax را بدست می آورد ، پیشنهاد شده است.ما همچنین نشان می دهد که تعداد مورد انتظار درمان های زیر حد متوسط ​​فقط به آرامی رشد می کند در تعداد درمان هاسرانجام ، ما یک تنظیم را مطالعه می کنیم که نتایج آن است فقط با تأخیر مشاهده می شود.
29,961
Most modern supervised statistical/machine learning (ML) methods are explicitly designed to solve prediction problems very well. Achieving this goal does not imply that these methods automatically deliver good estimators of causal parameters. Examples of such parameters include individual regression coefficients, average treatment effects, average lifts, and demand or supply elasticities. In fact, estimates of such causal parameters obtained via naively plugging ML estimators into estimating equations for such parameters can behave very poorly due to the regularization bias. Fortunately, this regularization bias can be removed by solving auxiliary prediction problems via ML tools. Specifically, we can form an orthogonal score for the target low-dimensional parameter by combining auxiliary and main ML predictions. The score is then used to build a de-biased estimator of the target parameter which typically will converge at the fastest possible 1/root(n) rate and be approximately unbiased and normal, and from which valid confidence intervals for these parameters of interest may be constructed. The resulting method thus could be called a "double ML" method because it relies on estimating primary and auxiliary predictive models. In order to avoid overfitting, our construction also makes use of the K-fold sample splitting, which we call cross-fitting. This allows us to use a very broad set of ML predictive methods in solving the auxiliary and main prediction problems, such as random forest, lasso, ridge, deep neural nets, boosted trees, as well as various hybrids and aggregators of these methods.
بیشتر روشهای مدرن تحت نظارت آماری/ماشین (ML) هستند به صراحت برای حل مشکلات پیش بینی بسیار خوب طراحی شده است.دستیابی به این هدف دلالت بر این ندارد که این روش ها به طور خودکار برآوردگرهای خوبی از آن ارائه می دهند پارامترهای علی.نمونه هایی از چنین پارامترهایی شامل رگرسیون فردی است ضرایب ، میانگین اثرات درمانی ، بالابر متوسط ​​و تقاضا یا عرضه خاصیت ارتجاعیدر حقیقت ، برآورد چنین پارامترهای علّی به دست آمده از طریق ساده لوحانه وصل کردن برآوردگرهای ML به تخمین معادلات برای چنین پارامترهایی می تواند رفتار کند به دلیل تعصب منظم بسیار ضعیف است.خوشبختانه ، این تنظیم با حل مشکلات پیش بینی کمکی از طریق ابزارهای ML می توان تعصب را حذف کرد. به طور خاص ، ما می توانیم یک نمره متعامد را برای هدف پایین هدف تشکیل دهیم پارامتر با ترکیب پیش بینی های کمکی و اصلی ML.سپس نمره است برای ساختن یک برآوردگر مغرضانه از پارامتر هدف که به طور معمول استفاده می شود با سریعترین نرخ ممکن 1/ریشه (n) همگرا می شود و تقریباً خواهد بود بی طرفانه و عادی ، و از آن فواصل اطمینان معتبر برای اینها پارامترهای مورد علاقه ممکن است ساخته شود.روش نتیجه می تواند باشد یک روش "دو میلی لیتر" خوانده می شود زیرا به تخمین اولیه و مدل های پیش بینی کمکی.به منظور جلوگیری از بیش از حد ، ساخت و ساز ما همچنین از تقسیم نمونه K استفاده می کند ، که ما آن را متناسب با آن می نامیم. این به ما امکان می دهد از مجموعه ای بسیار گسترده از روشهای پیش بینی ML در حل استفاده کنیم مشکلات پیش بینی کمکی و اصلی مانند جنگل تصادفی ، لاسو ، ریج ، شبکه های عصبی عمیق ، درختان تقویت شده و همچنین هیبریدهای مختلف و جمع کننده های مختلف این روشها
29,962
This paper considers inference on fixed effects in a linear regression model estimated from network data. An important special case of our setup is the two-way regression model. This is a workhorse technique in the analysis of matched data sets, such as employer-employee or student-teacher panel data. We formalize how the structure of the network affects the accuracy with which the fixed effects can be estimated. This allows us to derive sufficient conditions on the network for consistent estimation and asymptotically-valid inference to be possible. Estimation of moments is also considered. We allow for general networks and our setup covers both the dense and sparse case. We provide numerical results for the estimation of teacher value-added models and regressions with occupational dummies.
در این مقاله استنتاج بر اثرات ثابت در یک مدل رگرسیون خطی در نظر گرفته شده است از داده های شبکه تخمین زده می شود.یک مورد مهم مهم از راه اندازی ما مواردی است مدل رگرسیون دو طرفه.این یک تکنیک اسب بخار در تجزیه و تحلیل است مجموعه داده های همسان ، مانند داده های کارفرما-کارمند یا داده های پانل دانشجویی.ما نحوه تأثیر ساختار شبکه بر روی صحت آن را رسمی کنید اثرات ثابت را می توان تخمین زد.این به ما امکان می دهد شرایط کافی را بدست آوریم در شبکه برای برآورد مداوم و استنتاج بدون علامت به ممکن بودن.تخمین لحظات نیز در نظر گرفته شده است.ما به ژنرال اجازه می دهیم شبکه ها و راه اندازی ما هر دو مورد متراکم و پراکنده را پوشش می دهد.ما فراهم می کنیم نتایج عددی برای برآورد مدلهای ارزش افزوده معلم و رگرسیون با آدمک های شغلی.
29,963
Quantile and quantile effect functions are important tools for descriptive and causal analyses due to their natural and intuitive interpretation. Existing inference methods for these functions do not apply to discrete random variables. This paper offers a simple, practical construction of simultaneous confidence bands for quantile and quantile effect functions of possibly discrete random variables. It is based on a natural transformation of simultaneous confidence bands for distribution functions, which are readily available for many problems. The construction is generic and does not depend on the nature of the underlying problem. It works in conjunction with parametric, semiparametric, and nonparametric modeling methods for observed and counterfactual distributions, and does not depend on the sampling scheme. We apply our method to characterize the distributional impact of insurance coverage on health care utilization and obtain the distributional decomposition of the racial test score gap. We find that universal insurance coverage increases the number of doctor visits across the entire distribution, and that the racial test score gap is small at early ages but grows with age due to socio economic factors affecting child development especially at the top of the distribution. These are new, interesting empirical findings that complement previous analyses that focused on mean effects only. In both applications, the outcomes of interest are discrete rendering existing inference methods invalid for obtaining uniform confidence bands for observed and counterfactual quantile functions and for their difference -- the quantile effects functions.
توابع جلوه ای کمی و کمی ابزارهای مهمی برای توصیفی هستند و تجزیه و تحلیل های علی به دلیل تفسیر طبیعی و شهودی آنها.موجود روشهای استنتاج برای این توابع برای تصادفی گسسته اعمال نمی شود متغیرهااین مقاله یک ساخت و ساز ساده و عملی همزمان را ارائه می دهد باند اعتماد به نفس برای عملکردهای کمی و کمی متغیرهای تصادفی گسسته.این مبتنی بر تحول طبیعی است باند اعتماد به نفس همزمان برای توابع توزیع ، که به راحتی هستند برای بسیاری از مشکلات موجود است.ساخت و ساز عمومی است و به آن بستگی ندارد ماهیت مشکل اساسی.این همراه با پارامتری کار می کند ، روشهای مدل سازی نیمه پارامتری و غیر پارامتری برای مشاهده شده و توزیع ضد خلاف ، و به طرح نمونه گیری بستگی ندارد.ما برای توصیف تأثیر توزیع بیمه از روش ما استفاده کنید پوشش در مورد استفاده از مراقبت های بهداشتی و به دست آوردن تجزیه توزیع از شکاف نمره آزمون نژادی.ما می دانیم که پوشش بیمه جهانی تعداد مراجعه به پزشک را در کل توزیع افزایش می دهد و آن شکاف نمره آزمون نژادی در سنین پایین کوچک است اما به دلیل افزایش سن به دلیل افزایش سن عوامل اقتصادی اجتماعی مؤثر بر رشد کودک به ویژه در صدر توزیعاینها یافته های تجربی جدید و جالب است که مکمل است تجزیه و تحلیل های قبلی که فقط روی اثرات میانگین متمرکز شده بودند.در هر دو برنامه ، نتایج علاقه گسسته است که روشهای استنباط موجود نامعتبر است برای به دست آوردن باند اعتماد به نفس یکنواخت برای کمی مشاهده شده و ضد خلاف توابع و برای تفاوت آنها - عملکرد جلوه های کمی.
29,964
We consider a variable selection problem for the prediction of binary outcomes. We study the best subset selection procedure by which the covariates are chosen by maximizing Manski (1975, 1985)'s maximum score objective function subject to a constraint on the maximal number of selected variables. We show that this procedure can be equivalently reformulated as solving a mixed integer optimization problem, which enables computation of the exact or an approximate solution with a definite approximation error bound. In terms of theoretical results, we obtain non-asymptotic upper and lower risk bounds when the dimension of potential covariates is possibly much larger than the sample size. Our upper and lower risk bounds are minimax rate-optimal when the maximal number of selected variables is fixed and does not increase with the sample size. We illustrate usefulness of the best subset binary prediction approach via Monte Carlo simulations and an empirical application of the work-trip transportation mode choice.
ما یک مشکل انتخاب متغیر را برای پیش بینی باینری در نظر می گیریم عواقب.ما بهترین روش انتخاب زیر مجموعه را که توسط آن متغیرهای متغیر است مطالعه می کنیم با حداکثر عملکرد Manski (1975 ، 1985) انتخاب شده است مشمول محدودیت در حداکثر تعداد متغیرهای انتخاب شده.ما نشان می دهیم اینکه این روش به عنوان حل یک عدد صحیح مخلوط می تواند به طور معادل اصلاح شود مشکل بهینه سازی ، که محاسبه دقیق یا تقریبی را امکان پذیر می کند راه حل با یک خطای تقریب قطعی محدود است.از نظر نظری نتایج ، ما مرزهای خطر فوقانی و پایین غیر متعارف را به دست می آوریم بعد متغیرهای بالقوه احتمالاً بسیار بزرگتر از اندازه نمونه است. مرزهای ریسک بالا و پایین ما در حد حداکثر نرخ بهینه است تعداد متغیرهای انتخاب شده ثابت است و با نمونه افزایش نمی یابد اندازه.ما سودمندی بهترین رویکرد پیش بینی باینری زیر مجموعه را نشان می دهیم از طریق شبیه سازی مونت کارلو و یک کاربرد تجربی از سفر انتخاب حالت حمل و نقل.
29,965
We present the Stata commands probitfe and logitfe, which estimate probit and logit panel data models with individual and/or time unobserved effects. Fixed effect panel data methods that estimate the unobserved effects can be severely biased because of the incidental parameter problem (Neyman and Scott, 1948). We tackle this problem by using the analytical and jackknife bias corrections derived in Fernandez-Val and Weidner (2016) for panels where the two dimensions ($N$ and $T$) are moderately large. We illustrate the commands with an empirical application to international trade and a Monte Carlo simulation calibrated to this application.
ما دستورات stata probitfe و logitfe را ارائه می دهیم ، که تخمین پروبیت و مدل های داده پانل را با اثرات غیرمجاز و/یا زمان انجام دهید.درست شد روشهای داده پانل اثر که تخمین می زند اثرات بدون نظارت می تواند به شدت باشد مغرضانه به دلیل مشکل پارامتر حادثه ای (Neyman and Scott ، 1948).ما با استفاده از اصلاحات تعصب تحلیلی و jackknife با این مشکل مقابله کنید مشتق شده در Fernandez-Val و Weidner (2016) برای پانل هایی که در آن دو بعد ($ n $ و $ t $) نسبتاً بزرگ هستند.ما دستورات را با یک نشان می دهیم کاربرد تجربی برای تجارت بین المللی و شبیه سازی مونت کارلو در این برنامه کالیبره شده است.
29,966
The Counterfactual package implements the estimation and inference methods of Chernozhukov, Fern\'andez-Val and Melly (2013) for counterfactual analysis. The counterfactual distributions considered are the result of changing either the marginal distribution of covariates related to the outcome variable of interest, or the conditional distribution of the outcome given the covariates. They can be applied to estimate quantile treatment effects and wage decompositions. This paper serves as an introduction to the package and displays basic functionality of the commands contained within.
بسته متقابل روشهای تخمین و استنباط را پیاده سازی می کند Chernozhukov ، Fern \ 'Andez-Val و Melly (2013) برای تجزیه و تحلیل متقابل.در توزیع های متقابل در نظر گرفته شده نتیجه تغییر یا تغییر است توزیع حاشیه ای متغیرهای مربوط به متغیر نتیجه علاقه یا توزیع مشروط نتیجه با توجه به متغیرهای متغیر. آنها را می توان برای برآورد اثرات درمانی و دستمزد کمی استفاده کرد تجزیهاین مقاله به عنوان مقدمه ای برای بسته و عملکرد اساسی دستورات موجود در آن را نشان می دهد.
29,968
This paper proposes new nonparametric diagnostic tools to assess the asymptotic validity of different treatment effects estimators that rely on the correct specification of the propensity score. We derive a particular restriction relating the propensity score distribution of treated and control groups, and develop specification tests based upon it. The resulting tests do not suffer from the "curse of dimensionality" when the vector of covariates is high-dimensional, are fully data-driven, do not require tuning parameters such as bandwidths, and are able to detect a broad class of local alternatives converging to the null at the parametric rate $n^{-1/2}$, with $n$ the sample size. We show that the use of an orthogonal projection on the tangent space of nuisance parameters facilitates the simulation of critical values by means of a multiplier bootstrap procedure, and can lead to power gains. The finite sample performance of the tests is examined by means of a Monte Carlo experiment and an empirical application. Open-source software is available for implementing the proposed tests.
در این مقاله ابزارهای تشخیصی غیر پارامتری جدید برای ارزیابی اعتبار بدون علامت از برآوردگرهای اثرات درمانی مختلف که به آنها متکی هستند مشخصات صحیح نمره گرایش.ما یک خاص را استخراج می کنیم محدودیت مربوط به توزیع نمره تمایل از درمان و کنترل گروه ها ، و تست های مشخصات را بر اساس آن تهیه کنید.آزمایشات حاصل انجام می شود هنگامی که بردار متغیرهای متغیر است ، از "نفرین ابعاد" رنج نمی برید با ابعاد بالا ، کاملاً داده محور هستند ، نیازی به پارامترهای تنظیم مانند ندارند به عنوان پهنای باند ، و قادر به تشخیص طبقه گسترده ای از گزینه های محلی هستند همگرا به تهی با نرخ پارامتری $ n^{-1/2} $ ، با $ n $ نمونه اندازه.ما نشان می دهیم که استفاده از یک پیش بینی متعامد در فضای مماس پارامترهای مزاحمت شبیه سازی مقادیر بحرانی را با استفاده از a تسهیل می کند روش bootstrap چند برابر ، و می تواند منجر به افزایش قدرت شود.نمونه محدود عملکرد تست ها با استفاده از یک آزمایش مونت کارلو و یک برنامه تجربینرم افزار منبع باز برای اجرای در دسترس است تست های پیشنهادی
29,969
We consider inference about coefficients on a small number of variables of interest in a linear panel data model with additive unobserved individual and time specific effects and a large number of additional time-varying confounding variables. We allow the number of these additional confounding variables to be larger than the sample size, and suppose that, in addition to unrestricted time and individual specific effects, these confounding variables are generated by a small number of common factors and high-dimensional weakly-dependent disturbances. We allow that both the factors and the disturbances are related to the outcome variable and other variables of interest. To make informative inference feasible, we impose that the contribution of the part of the confounding variables not captured by time specific effects, individual specific effects, or the common factors can be captured by a relatively small number of terms whose identities are unknown. Within this framework, we provide a convenient computational algorithm based on factor extraction followed by lasso regression for inference about parameters of interest and show that the resulting procedure has good asymptotic properties. We also provide a simple k-step bootstrap procedure that may be used to construct inferential statements about parameters of interest and prove its asymptotic validity. The proposed bootstrap may be of substantive independent interest outside of the present context as the proposed bootstrap may readily be adapted to other contexts involving inference after lasso variable selection and the proof of its validity requires some new technical arguments. We also provide simulation evidence about performance of our procedure and illustrate its use in two empirical applications.
ما استنتاج در مورد ضرایب را در تعداد کمی از متغیرها در نظر می گیریم علاقه به یک مدل داده پانل خطی با فرد بدون نظارت و افزودنی اثرات خاص زمان و تعداد زیادی از مخدوش کننده زمان اضافی متغیرهاما اجازه می دهیم تعداد این متغیرهای مخدوش اضافی باشد بزرگتر از اندازه نمونه ، و فرض کنید که علاوه بر زمان نامحدود و اثرات خاص فردی ، این متغیرهای مخدوش توسط a ایجاد می شوند تعداد کمی از عوامل مشترک و وابسته به ابعاد بالا اغتشاشات.ما اجازه می دهیم که هم عوامل و هم اختلالات مرتبط باشند به متغیر نتیجه و سایر متغیرهای مورد علاقه.برای ساختن آموزنده استنباط امکان پذیر است ، ما تحمیل می کنیم که سهم بخشی از متغیرهای گیج کننده که توسط اثرات خاص زمان ضبط نشده اند ، فردی اثرات خاص ، یا عوامل متداول را می توان با یک نسبتاً کوچک ضبط کرد تعداد اصطلاحاتی که هویت آنها ناشناخته است.در این چارچوب ، ما ارائه می دهیم یک الگوریتم محاسباتی مناسب مبتنی بر استخراج فاکتور و به دنبال آن رگرسیون Lasso برای استنباط در مورد پارامترهای مورد علاقه و نشان می دهد که روش حاصل از خواص بدون علامت خوبی برخوردار است.ما همچنین یک ساده ارائه می دهیم روش بوت استرپ k-step که ممکن است برای ساخت اظهارات استنباطی استفاده شود در مورد پارامترهای مورد علاقه و اثبات اعتبار بدون علامت آن.پیشنهاد bootstrap ممکن است در خارج از زمان مورد علاقه مستقل باشد زمینه به عنوان بوت استستراپ پیشنهادی به راحتی ممکن است با سایر زمینه ها سازگار شود شامل استنباط پس از انتخاب متغیر Lasso و اثبات آن اعتبار به برخی از استدلال های فنی جدید نیاز دارد.ما همچنین شبیه سازی را ارائه می دهیم شواهدی در مورد عملکرد روش ما و نشان دادن استفاده از آن به دو صورت برنامه های تجربی.
29,970
This article proposes different tests for treatment effect heterogeneity when the outcome of interest, typically a duration variable, may be right-censored. The proposed tests study whether a policy 1) has zero distributional (average) effect for all subpopulations defined by covariate values, and 2) has homogeneous average effect across different subpopulations. The proposed tests are based on two-step Kaplan-Meier integrals and do not rely on parametric distributional assumptions, shape restrictions, or on restricting the potential treatment effect heterogeneity across different subpopulations. Our framework is suitable not only to exogenous treatment allocation but can also account for treatment noncompliance - an important feature in many applications. The proposed tests are consistent against fixed alternatives, and can detect nonparametric alternatives converging to the null at the parametric $n^{-1/2}$-rate, $n$ being the sample size. Critical values are computed with the assistance of a multiplier bootstrap. The finite sample properties of the proposed tests are examined by means of a Monte Carlo study and an application about the effect of labor market programs on unemployment duration. Open-source software is available for implementing all proposed tests.
این مقاله آزمایش های مختلفی را برای ناهمگونی اثر درمانی در هنگام درمان ارائه می دهد نتیجه علاقه ، به طور معمول یک متغیر مدت زمان ، ممکن است سانسور شود. آزمون های پیشنهادی مطالعه می کنند که آیا یک خط مشی 1) دارای توزیع صفر (میانگین) است اثر برای همه زیرمجموعات تعریف شده توسط مقادیر متغیر ، و 2) اثر متوسط ​​همگن در زیر مجموعه های مختلف.تست های پیشنهادی مبتنی بر انتگرال های دو مرحله ای Kaplan-Meier است و به پارامتری تکیه نمی کنند فرضیات توزیع ، محدودیت های شکل یا محدود کردن پتانسیل ناهمگونی اثر درمانی در زیر گروه های مختلف.چارچوب ما نه تنها برای تخصیص درمان برون زا مناسب است بلکه می تواند به حساب بیاورد عدم رعایت درمان - یک ویژگی مهم در بسیاری از برنامه ها.در تست های پیشنهادی در برابر گزینه های ثابت سازگار هستند و می توانند تشخیص دهند گزینه های غیرپارامتری که در پارامتری به تهی همگرا هستند $ n^{-1/2} $-نرخ ، $ n $ اندازه نمونه است.مقادیر بحرانی با محاسبه می شوند کمک یک بوت استرپ چند برابر.خصوصیات نمونه محدود از آزمایش های پیشنهادی با استفاده از یک مطالعه مونت کارلو و یک کاربردی مورد بررسی قرار می گیرد در مورد تأثیر برنامه های بازار کار بر مدت بیکاری.متن باز نرم افزار برای اجرای تمام تست های پیشنهادی در دسترس است.
29,971
Extremal quantile regression, i.e. quantile regression applied to the tails of the conditional distribution, counts with an increasing number of economic and financial applications such as value-at-risk, production frontiers, determinants of low infant birth weights, and auction models. This chapter provides an overview of recent developments in the theory and empirics of extremal quantile regression. The advances in the theory have relied on the use of extreme value approximations to the law of the Koenker and Bassett (1978) quantile regression estimator. Extreme value laws not only have been shown to provide more accurate approximations than Gaussian laws at the tails, but also have served as the basis to develop bias corrected estimators and inference methods using simulation and suitable variations of bootstrap and subsampling. The applicability of these methods is illustrated with two empirical examples on conditional value-at-risk and financial contagion.
رگرسیون کمی افراطی ، یعنی رگرسیون کمی که به دم اعمال می شود از توزیع مشروط ، تعداد فزاینده ای از اقتصادی شمارش می شود و برنامه های مالی مانند ارزش در معرض خطر ، مرزهای تولید ، عوامل تعیین کننده وزن کم نوزادان و مدلهای حراج.این فصل مروری بر تحولات اخیر در تئوری و تجربی از رگرسیون کمی افراطی.پیشرفت های این تئوری به استفاده متکی بوده است تقریب ارزش شدید به قانون Koenker و Bassett (1978) برآوردگر رگرسیون کمی.قوانین ارزش شدید نه تنها نشان داده شده است تقریب دقیق تری نسبت به قوانین گاوسی در دم ، بلکه همچنین ارائه دهید به عنوان پایه ای برای توسعه برآوردگرهای اصلاح شده تعصب و استنباط خدمت کرده اند روشها با استفاده از شبیه سازی و تغییرات مناسب بوت استرپ و نمونه برداری. کاربرد این روشها با دو نمونه تجربی نشان داده شده است در مورد ارزش مشروط در معرض خطر و آلودگی مالی.
29,972
There is a large literature on semiparametric estimation of average treatment effects under unconfounded treatment assignment in settings with a fixed number of covariates. More recently attention has focused on settings with a large number of covariates. In this paper we extend lessons from the earlier literature to this new setting. We propose that in addition to reporting point estimates and standard errors, researchers report results from a number of supplementary analyses to assist in assessing the credibility of their estimates.
ادبیات بزرگی در مورد تخمین نیمه قطبی از درمان متوسط ​​وجود دارد اثرات تحت تکلیف درمانی غیرمجاز در تنظیمات با تعداد ثابت متغیرهای متغیراخیراً توجه به تنظیمات با یک بزرگ متمرکز شده است تعداد متغیرهای متغیر.در این مقاله درس های قبلی را گسترش می دهیم ادبیات به این تنظیم جدید.ما پیشنهاد می کنیم علاوه بر گزارش گزارش تخمین ها و خطاهای استاندارد ، محققان نتایج حاصل از تعدادی را گزارش می دهند تجزیه و تحلیل های تکمیلی برای کمک به ارزیابی اعتبار آنها تخمین ها