چکیده
مقدمه
چارچوب رگرسیون چارکی دو مرحله ای
مطالعه شبیه سازی
کاربردها
نتیجه گیری
منابع
Abstract
Introduction
The two‑step quantile regression framework
Simulation study
Applications
Conclusion
References
چکیده
این مقاله یک رویکرد تحلیلی را پیشنهاد میکند که مکمل رگرسیون خطی دو مرحلهای سنتی و رگرسیون خطی یک مرحلهای پیشنهاد شده توسط چن و همکاران است. (J Account Res 56:751–796، 2018). استفاده از باقیمانده رگرسیون به عنوان متغیر وابسته در رگرسیون دوم روشی است که معمولاً در مطالعه حسابداری اختیاری استفاده می شود. چن و همکاران (J Account Res 56:751-796, 2018) پیشنهاد میکند که رگرسیون یک مرحلهای را برای جلوگیری از تعصب برآورد و خطای استنتاج اتخاذ کنید. با این حال، اثر سطح متوسط برآورد شده توسط رگرسیون OLS یک مرحلهای برای به دست آوردن طیف کلی رفتارهای حسابداری اختیاری کافی نیست و در نتیجه ممکن است کاربر خود را در ترسیم پیامدها گمراه کند. ما از رگرسیون کمی دو مرحلهای برای بررسی عوامل تعیینکننده حسابداری اختیاری مانند اقلام تعهدی اختیاری، هزینه اختیاری، تفاوتهای مالیاتی اختیاری دفتری و سرمایهگذاری غیرعادی در کمیتهای مختلف استفاده میکنیم. ما تفاوت بین رگرسیون یک مرحله ای و رگرسیون چند مرحله ای خود را با استفاده از چهار مطالعه حسابداری اختیاری مشترک نشان می دهیم. نتایج و پیامدهای ما، تا حدی، یافتههای متناقض بین نتایج رگرسیون OLS یک مرحلهای و کارهای ایجاد شده قبلی بر اساس رگرسیون دو مرحلهای را تطبیق میدهند.
توجه! این متن ترجمه ماشینی بوده و توسط مترجمین ای ترجمه، ترجمه نشده است.
Abstract
This paper proposes an analytical approach that complements the traditional two-step linear regression and one-single step linear regression suggested by Chen et al. (J Account Res 56:751–796, 2018). Using the regression residual as the dependent variable in a second regression is a procedure commonly used in studying discretionary accounting. Chen et al. (J Account Res 56:751–796, 2018) propose to adopt one-step regression to avoid estimation bias and inference error. However, the mean level effect estimated by one-step OLS regression is not sufficient to capture the overall spectrum of discretionary accounting behaviors and thus may mislead its user in drawing implications. We use two-stage quantile regression to examine determinants of discretionary accounting such as discretionary accruals, discretionary expense, discretionary book-tax differences, and abnormal investment in different quantiles. We illustrate the differences between the one-step regression and our two-step quantile regression using four common discretionary accounting studies. Our results and implications reconcile, to some extent, the contradictory findings between results of the one-step OLS regression and the previous established works based on two-step regression.
Introduction
Using the linear regression residual as the dependent variable in a second regression is a procedure commonly used in empirical accounting and fnance research. In a review paper about earnings quality, Dechow et al. (2010) document that “almost one hundred papers in [their] database use abnormal accruals as a measure of earnings quality and test predicted determinants or consequences. These studies test the joint hypothesis that the residual from an accruals model refects earnings management and that the predicted determinant induces earnings management or that earnings management has a predicted consequence”.1 Specifcally, in the frst-step regression, the researcher decomposes a dependent variable into its predicted and residual components, and the residuals are considered as the discretionary component or abnormality of the variable in question. Further examinations of the determinants of the discretionary component are then conducted. Such a procedure has become one of the standard approaches in accounting studies and is widely applied in research on discretionary accruals, real activities management, unexplained audit fees, and discretionary book-tax diference, among others. A thorough review of the popularity of the residual approach in accounting can be found in Chen et al. (2018).
Conclusion
The study of discretionary behavior, such as discretionary accruals, is one of the most important topics in accounting. However, Chen et al. (2018) and Christodoulou et al. (2018) point out that the traditional two-step approach provides biased estimates and incorrect inferences for discretionary accounting study. The single-step regression proposed by Chen et al. (2018) is a way to eliminate biases and restore correct inference. However, it also shows limitation in revealing the full picture of accounting behavior, since the linear regression only shows the conditional mean relation.
In this paper, we propose a novel two-step approach in which we conduct linear regression in the frst step and quantile regression in the second step. The main assumption is that the coefcients of the non-discretionary part of the variable in question are invariant across diferent quantiles, whereas the coefcients on the discretionary part vary to refect discretionary behavior.