Value relevance of accounting numbers and sustainability information in Europe: Empirical evidence from nonfinancial companies
歐洲會計數位和可持續性資訊的價值相關性——來自非金融公司的經驗證據

Alessandro Migliavacca 亞歷山德羅·米利亞瓦卡Department of Economics and Statistics "Cognetti de'Martiis", Università degli Studi di Torino, Italy
義大利都靈大學經濟與統計系 “Cognetti de'Martiis”

ARTICLE INFO 文章資訊

JEL codes: JEL 代碼：

M41  M41 系列
G32  G32 系列
Q56  問 56

Keywords: 關鍵字：

Value relevance  值相關性
Accounting numbers  會計編號
Book values  賬麵價值
Earnings  收益
ESG ratings  ESG 評級
Sustainability information
可持續發展資訊
ERC Panel:  ERC 面板：
SH1_5

Abstract 抽象

Information is considered value relevant when used by financial markets in equity valuation and is reflected in market values. The value relevance of different information, such as accounting numbers, changes according to shifts in investors’ needs. Given the rising importance of environmental and social agendas for policymakers and practitioners, this study examines whether accounting numbers have lost explanatory power and sustainability information has become value relevant. The analysis focuses on 3025 nonfinancial companies operating in the eurozone from 2005 to 2020. This study makes several contributions to the extant literature. First, analyzing the trend in the value relevance of financial accounting numbers, the results point to an overall decrease in the explanatory power of book values and earnings, particularly for environmental, social, and governance (ESG)rated companies. However, the results indicate that ESG ratings have not gained value relevance over time. These findings have important implications for policymakers and practitioners. In line with the concept of “double materiality,” implementing ESG agendas will be possible only by redirecting corporate investment decision and increasing awareness of sustainability issues. This paper documents that such a shift has not yet been accomplished, despite the increasing commitment of the European Union to the ESG agenda.
當金融市場在股票估值中使用資訊時，信息被視為與價值相關，並反映在市場價值中。不同資訊（例如會計數位）的價值相關性會根據投資者需求的變化而變化。鑒於環境和社會議程對政策制定者和從業者的重要性日益增加，本研究研究了會計數位是否已經失去了解釋力，可持續性資訊是否變得具有價值相關性。該分析側重於 2005 年至 2020 年在歐元區運營的 3025 家非金融公司。這項研究對現存文獻做出了多項貢獻。首先，分析財務會計數字的價值相關性趨勢，結果表明賬麵價值和收益的解釋力總體下降，尤其是對於環境、社會和治理（ESG）評級的公司。然而，結果表明，ESG 評級並未隨著時間的推移而獲得價值相關性。這些發現對政策制定者和從業者具有重要意義。根據「雙重重要性」的概念，只有改變企業投資決策的方向和提高對可持續發展問題的認識，才能實施 ESG 議程。本文記錄了儘管歐盟對 ESG 議程的承諾越來越大，但這種轉變尚未完成。

1. Introduction 1. 引言

Since the Lisbon Treaty came into force, and even more after the Paris Agreement and Sustainable Development Goals, the sustainability of nonfinancial companies has become key to the European Union (EU)'s agenda. EU institutions are increasingly incorporating environmental and social aspects into regulations., including the Nonfinancial Reporting Directive in 2014, the Sustainable Finance Disclosure Regulation in 2019, the Corporate Sustainability Reporting Directive in 2022, and the Corporate Sustainability Due Diligence Directive in 2023, and action plans, such as the 2018 Action Plan on Sustainable Finance and the 2019 Green Deal. In parallel, public awareness of environmental and social issues has grown alongside fund flows toward sustainable activities (Cornell, 2021; Matos, 2020), and professional investors are increasingly integrating sustainability information into their decision-making processes (e.g., Amel-Zadeh & Serafeim, 2018; Global Sustainable Investment Alliance, 2020). In a survey of finance professionals, AmelZadeh and Serafeim (2018) reported that

82 %

of the sample considered sustainability information in their investment decisions, and

63 %

deemed it material to investment performance. Despite their critique of the usefulness of environmental, social, and governance (ESG)
自《里斯本條約》生效以來，以及《巴黎協定》和可持續發展目標之後，非金融公司的可持續性已成為歐盟（EU）議程的關鍵。歐盟機構越來越多地將環境和社會因素納入法規，包括 2014 年的《非財務報告指令》、2019 年的《可持續金融披露條例》、2022 年的《企業可持續發展報告指令》和 2023 年的《企業可持續發展盡職調查指令》，以及 2018 年《可持續金融行動計劃》和 2019 年《綠色協定》等行動計劃。與此同時，公眾對環境和社會問題的認識隨著資金流向可持續活動而提高（康奈爾大學，2021 年;Matos，2020），專業投資者越來越多地將可持續性資訊整合到他們的決策過程中（例如，Amel-Zadeh & Serafeim，2018;全球可持續投資聯盟，2020 年）。在一項針對金融專業人士的調查中，AmelZadeh 和 Serafeim （2018）報告稱，

82 %

樣本在他們的投資決策中考慮了可持續性資訊，並

63 %

認為它對投資業績很重要。儘管他們對環境、社會和治理（ESG）的有用性提出了批評
information, Cornell and Damodaran (2020) recognize that sustainability criteria are key for high-profile institutional investors.
資訊，康奈爾大學和 Damodaran （2020）認識到可持續性標準是知名機構投資者的關鍵。

From an economic perspective, markets are not informationally perfect and base their evaluations on a limited portion of all available information (Grossman & Stiglitz, 1980). A piece of information is value relevant when either it explains variation in share prices or it has a predicted association with equity market values (Amir et al., 1993; Barth et al., 2008; Onali et al., 2017) As the economy evolves, equity valuation processes consider different elements and information, and considers shifts in investors’ needs (Barth et al., 2023). In line with this theoretical perspective, the most recent literature posits that value relevance studies should not be based only on a limited number of items, such as earnings and book values. Research should consider a larger number of proxies for external influences, economic conditions, and additional nonaccounting data to capture overall firm performance (Barth et al., 2023; Dunham & Grandstaff, 2022). Hence, over time, equity valuations could have changed their association with accounting measures, such as earnings and book values, to make way for additional information utilizing alternative accounting measures (Barth et al., 2023) and sustainability performance (Jadoon et al., 2021).
從經濟角度來看，市場在資訊上並不完美，他們的評估基於所有可用資訊的有限部分（Grossman & Stiglitz， 1980）。當一條信息解釋了股價的變化或與股票市場價值有預測的關聯時，它就是與價值相關的（Amir et al.， 1993;Barth et al.， 2008;Onali et al.， 2017）隨著經濟的發展，股票估值過程會考慮不同的元素和資訊，並考慮投資者需求的變化（Barth et al.， 2023）。根據這一理論觀點，最近的文獻認為價值相關性研究不應僅基於有限數量的專案，例如收入和賬麵價值。研究應考慮外部影響、經濟狀況和其他非會計數據的大量代理，以捕捉公司的整體表現（Barth 等人，2023 年;Dunham & Grandstaff，2022 年）。因此，隨著時間的推移，股票估值可能會改變它們與會計指標（如收益和賬麵價值）的關聯，以利用替代會計指標（Barth 等人，2023 年）和可持續性績效（Jadoon 等人，2021 年）為其他資訊讓路。

Prior research has found mixed trends in the value relevance of
先前的研究發現，價值相關性的趨勢喜憂參半

accounting numbers (for a review, see Dunham & Grandstaff (2022). A large body of literature discusses the role of sustainability information on corporate performance (e.g., Giglio et al., 2021; Gillan et al., 2021). The few studies dedicated to the value relevance of sustainability information have been mainly conducted at the regional or local level (e. g., Cordazzo et al., 2020; Veltri & Silvestri, 2020). The literature indicates that to analyze the relevance of sustainability information, it is most appropriate to jointly consider economic performance and ESG information (Jadoon et al., 2021; Sidhoum & Serra, 2018).
會計數位（有關評論，請參閱 Dunham & Grandstaff （2022）。大量文獻討論了可持續性資訊對企業績效的作用（例如，Giglio 等人，2021 年;Gillan et al.， 2021）。少數專門研究可持續性資訊價值相關性的研究主要在區域或地方層面進行（例如，Cordazzo 等人，2020 年;Veltri & Silvestri，2020 年）。文獻表明，要分析可持續性資訊的相關性，最合適的做法是共同考慮經濟績效和 ESG 資訊（Jadoon 等人，2021 年;Sidhoum & Serra，2018 年）。

This study investigates the change in the value relevance of economic and sustainability performance information (Jadoon et al., 2021), focusing on a sample of European nonfinancial firms from 2005 to 2020. Sustainability performance data are available for an increasing number of companies, however, data is still not available for most firms that disclose sustainability information (Eliwa et al., 2021). This study asks whether increasing environmental and social awareness has changed the value relevance of accounting numbers to make way for additional information. Further, focusing on companies with retrievable ESG performance data, this study investigates whether sustainability information has gained value relevance.
本研究調查了經濟和可持續性績效信息的價值相關性的變化（Jadoon et al.， 2021），重點關注 2005 年至 2020 年的歐洲非金融公司樣本。越來越多的公司可以使用可持續發展績效數據，但是，大多數披露可持續發展資訊的公司仍然無法獲得數據（Eliwa et al.， 2021）。本研究詢問了環境和社會意識的提高是否改變了會計數字的價值相關性，從而為更多資訊讓路。此外，本研究重點關注擁有可檢索 ESG 績效數據的公司，調查可持續發展資訊是否獲得了價值相關性。

This study contributes to the literature in two ways. First, the results provide evidence of an overall decrease in value relevance of accounting numbers from 2005 to 2020. Overall, adopting the Ohlson (1995) Price Regression Model, the findings suggest that accounting measures in 2020 have lower explanatory power for market values than those in 2005. This reduction affected both earnings and book value in companies with ESG ratings. By contrast, the reduction is only observed in book values in companies with no sustainability scores. This finding is consistent with the notion that investors’ information requirements change over time and that increasing concern about sustainability could have modified how investors value companies.
本研究以兩種方式為文獻做出貢獻。首先，結果提供了 2005 年至 2020 年會計數位價值相關性總體下降的證據。總體而言，採用 Ohlson （1995）價格回歸模型，研究結果表明，2020 年的會計指標對市場價值的解釋力低於 2005 年。這種減少影響了具有 ESG 評級的公司的收益和賬麵價值。相比之下，僅在沒有可持續發展評分的公司中觀察到賬麵價值的下降。這一發現與投資者的資訊需求隨時間變化的觀點一致，並且對可持續性的日益關注可能會改變投資者對公司的估值。

Second, the results show that including sustainability information proxied by ESG ratings in a Price Regression Model does not add explanatory power or modify the relationship between share prices and accounting measures. Thus, ESG ratings do not explain variations in market value.
其次，結果表明，在價格回歸模型中包括由 ESG 評級代理的可持續性資訊不會增加解釋力或修改股價與會計指標之間的關係。因此，ESG 評級並不能解釋市場價值的變化。

The findings of this study should be of interest to practitioners and policymakers. Transitioning to more sustainable economic paradigms and implementing the EU’s Taxonomies and Action Plans will be possible only by redirecting investment decision processes and increasing corporate awareness of sustainability issues, in line with the “double materiality” concept. This paper documents that such a shift has not yet been accomplished, despite the fact that the EU has recently been dedicated to the ESG agenda.
這項研究的結果應該引起從業者和政策制定者的興趣。只有根據「雙重重要性」概念，重新調整投資決策流程並提高企業對可持續發展問題的認識，才能過渡到更可持續的經濟範式並實施歐盟的分類法和行動計劃。本文記錄了這種轉變尚未完成，儘管歐盟最近致力於 ESG 議程。

The remainder of this paper is organized as follows. Section 2 reviews the literature on the value relevance of financial accounting and sustainability information, and outlines the research hypotheses. Section 3 presents the research design, sample selection process, and data. Section 4 reports the results of the analysis. Finally, Section 5 discusses the findings and concludes the paper.
本文的其餘部分組織如下。第 2 節回顧了有關財務會計和可持續性資訊的價值相關性的文獻，並概述了研究假設。第 3 部分介紹了研究設計、樣本選擇過程和數據。第 4 節報告了分析結果。最後，第 5 節討論了研究結果並總結了論文。

2. Literature review and research hypotheses
2. 文獻綜述和研究假設

2.1. Value relevance of accounting numbers
2.1. 會計數位的價值相關性

Financial accounting research has largely investigated the value relevance of book value, earnings, and other types of accounting information. A large body of literature has evolved since the first works published by Beaver (1968) and Ball and Brown (1968). This research has influenced policymakers to regulate financial statements and improve their completeness, comparability, and consistency with investors’ needs (e.g., Barth et al., 2023; Dunham & Grandstaff, 2022).
財務會計研究主要調查了賬麵價值、收益和其他類型的會計信息的價值相關性。自 Beaver （1968）和 Ball and Brown （1968）發表的第一篇作品以來，已經發展了大量文獻。這項研究影響了政策制定者監管財務報表並提高其完整性、可比性以及與投資者需求的一致性（例如，Barth 等人，2023 年;Dunham & Grandstaff，2022 年）。

Nonetheless, research focusing on the evolution of the value relevance of accounting numbers reports mixed evidence (e.g., Barth et al., 2023; Dunham & Grandstaff, 2022). One stream of literature highlights the decreasing trends in the value relevance of earnings and book values
儘管如此，專注於會計數位價值相關性演變的研究報告了混合證據（例如，Barth 等人，2023 年;Dunham & Grandstaff，2022 年）。一連串文獻強調了收益和賬麵價值的價值相關性呈下降趨勢
(e.g., Brown et al., 1999; Lev & Gu, 2016). In contrast, other studies found no evidence of a reduction but rather a shift in value relevance between different types of information (e.g., Barth et al., 2023; Collins et al., 1997; Francis & Schipper, 1999). This divergence arises partly from different research designs.

^{1}

Other differences in value relevance trends have been linked to economic conditions and cycles, such as uncertainty (e.g., Loh & Stulz, 2018), economic crises (e.g., Adwan et al., 2020), financial bubbles (e.g., Morris & Alam, 2012), or volatility of market returns (e.g., Francis & Schipper, 1999; Song, 2015). Lastly, value relevance has been linked with country-level industry-specific factors (e.g., Demers & Lev, 2001). For example, these factors include the implementation of International Financial Reporting Standards (IFRS) (e.g., Barth et al., 2012; Callao et al., 2007), or other countryspecific economic environments and macroeconomic conditions (e.g., Barton et al., 2010).
（例如，Brown 等人，1999 年;Lev & Gu，2016 年）。相比之下，其他研究沒有發現不同類型資訊之間價值相關性發生變化的證據（例如，Barth 等人，2023 年;Collins et al.， 1997;Francis & Schipper， 1999）。這種分歧部分是由於不同的研究設計引起的。

^{1}

價值相關性趨勢的其他差異與經濟狀況和周期有關，如不確定性（例如（Loh & Stulz，2018）、經濟危機（例如，Adwan 等人，2020，金融泡沫（例如，Morris & Alam，2012，或市場回報的波動性（例如，Francis & Schipper，1999;Song，2015 年）。最後，價值相關性與國家層面的行業特定因素有關（例如，Demers & Lev，2001）。例如，這些因素包括國際財務報告準則（IFRS）的實施（例如，Barth 等人，2012 年;Callao et al.， 2007）或其他國家特定的經濟環境和巨集觀經濟條件（例如，Barton et al.， 2010）。

Several studies focused on European companies. For instance, Onali et al. (2017) investigated the impact of unobserved heterogeneity on value relevance regression models for a sample of firms between 2005 and 2013. Adwan et al. (2020) examined whether and how the adoption of fair value accounting moderates the changes in the value relevance of book values and earnings during crises, confirming prior literature that the value relevance of equity book values increases while that of net income decreases during financial crises. Harasheh et al. (2021) studied the Other Comprehensive Income (OCI) value relevance of EuroSTOXX50 companies over the 2010-2016 period. However, none of those studies examined whether or how the value relevance of accounting information changed over the periods considered.
幾項研究側重於歐洲公司。例如，Onali 等人（2017 年）調查了 2005 年至 2013 年間未觀察到的異質性對公司樣本的價值相關性回歸模型的影響。Adwan 等人（2020 年）研究了公允價值會計的採用是否以及如何調節危機期間賬麵價值和收益的價值相關性變化，證實了先前的文獻，即在金融危機期間，權益賬麵價值的價值相關性增加，而凈收入的價值相關性降低。Harasheh 等人（2021 年）研究了 2010-2016 年期間 EuroSTOXX50 公司的其他綜合收入（OCI）價值相關性。然而，這些研究都沒有檢查會計信息的價值相關性是否或如何在所考慮的期間內發生變化。

Therefore, extant literature does not identify a trend in the value relevance of accounting numbers, particularly when focusing on European nonfinancial companies. The first research question of this study aims to determine whether investors’ requirements changed during the last decade and whether accounting numbers lost part of their power to explain variations in share prices. Accordingly, the following null hypothesis is formulated:
因此，現存文獻並未確定會計數字的價值相關性趨勢，尤其是在關注歐洲非金融公司時。本研究的第一個研究問題旨在確定投資者的需求在過去十年中是否發生了變化，以及會計數位是否失去了部分解釋股價變化的能力。因此，提出了以下原假設：

H1: The value relevance of earnings and book values in the EU did not change from 2005 to 2020.
H1：從 2005 年到 2020 年，歐盟收益和賬麵價值的價值相關性沒有變化。

The first hypothesis does not directly focus on sustainability information because such data are produced by an increasing but still limited number of companies in the period under analysis. Therefore, this study first verifies whether increasing environmental and social awareness changes the value relevance of accounting numbers to make way for additional information. Moreover, this change can be highlighted by comparing the results for firms with and without sustainability performance data.
第一個假設並不直接關注可持續性資訊，因為在分析期間，此類數據是由不斷增加但仍然數量有限的公司產生的。因此，本研究首先驗證了環境和社會意識的提高是否會改變會計數字的價值相關性，從而為更多資訊讓路。此外，通過比較有和沒有可持續發展績效數據的公司的結果，可以突出這種變化。

2.2. Value relevance of sustainability information
2.2. 可持續性資訊的價值相關性

ESG, Corporate Social Responsibility (CSR), and other keywords for sustainability practices have been linked to several firm attributes, such as the cost of capital, risk, probability of default, and performance (e.g., Gillan et al., 2021; Li et al., 2022). Most studies focus on a single aspect of sustainability. These include environmental practices (e.g., Hassel et al., 2005; Palea et al., 2023), disclosure practices (e.g., Clarkson et al., 2013; Cormier & Magnan, 2007), and reporting standards (e.g.,
ESG、企業社會責任（CSR）和其他可持續發展實踐的關鍵詞與幾個公司屬性相關聯，例如資本成本、風險、違約概率和績效（例如，Gillan 等人，2021 年;Li et al.， 2022）。大多數研究都集中在可持續性的單一方面。這些包括環境實踐（例如，Hassel 等人，2005 年;Palea et al.， 2023）、披露做法（例如，Clarkson et al.， 2013;Cormier & Magnan， 2007）和報告標準（例如，

Baboukardos & Rimmel, 2016; Veltri & Silvestri, 2020). Studies focusing on a wider information set are usually limited in their geographical scope (e.g., Cardamone et al., 2012; Cordazzo et al., 2020; Ricci et al., 2020), sample size (e.g., Berthelot et al., 2012; de Villiers & Marques, 2016), or time frame (Zuraida et al., 2018).
Baboukardos & Rimmel， 2016;Veltri & Silvestri，2020 年）。專注於更廣泛資訊集的研究通常在其地理範圍上受到限制（例如，Cardamone 等人，2012 年;Cordazzo 等人，2020 年;Ricci et al.， 2020）、樣本量（例如，Berthelot et al.， 2012;de Villiers & Marques， 2016）或時間框架（Zuraida et al.， 2018）。

Among these topics, the research connecting stock financial performance and sustainability information dates back to the 1970s (e.g., Belkaoui, 1976; Moskowitz, 1972) and has hitherto evolved (Brooks & Oikonomou, 2018), albeit not to the extent of the “mainstream” value relevance literature. Some researchers have found that sustainability information and disclosure are either costly and detrimental to share prices (e.g., Cardamone et al., 2012; Moneva & Cuellar, 2009) or neutral and ineffective (e.g., Cordazzo et al., 2020). In contrast, other studies have found a positive relationship between sustainability performance and market value (e.g., Jadoon et al., 2021; Kaspereit & Lopatta, 2016; Miralles-Quirós et al., 2019).
在這些主題中，將股票財務表現和可持續性資訊聯繫起來的研究可以追溯到 1970 年代（例如，Belkaoui，1976 年;Moskowitz， 1972），並且已經發展至今（Brooks & Oikonomou， 2018），儘管沒有達到“主流”價值相關性文獻的程度。一些研究人員發現，可持續性資訊和披露要麼成本高昂，要麼對股價有害（例如，Cardamone 等人，2012 年;Moneva & Cuellar， 2009）或中立且無效（例如，Cordazzo et al.， 2020）。相比之下，其他研究發現可持續性績效與市場價值之間存在正相關關係（例如，Jadoon 等人，2021 年;Kaspereit & Lopatta， 2016;Miralles-Quirós 等人，2019 年）。

Few studies have focused on European nonfinancial corporations. For instance, Moneva and Cuellar (2009) reported that while financial and environmental disclosures are value relevant, nonfinancial information is not, even though it increased slightly after the introduction of standards on environmental issues reporting. Cordazzo et al. (2020) analyzed the effects of the EU Nonfinancial Reporting Directive implementation in Italian listed companies. They found that sustainability information is not value relevant under voluntary or mandatory disclosure.
很少有研究關注歐洲非金融公司。例如，Moneva 和 Cuellar （2009）報告說，雖然財務和環境披露與價值相關，但非財務資訊則不相關，儘管在引入環境問題報告標準后它略有增加。Cordazzo 等人（2020 年）分析了歐盟非財務報告指令對義大利上市公司實施的影響。他們發現，在自願或強制披露下，可持續發展資訊與價值無關。

Permasatari and Narsa (2021) reported that sustainability reporting, separate from annual reports, is more value relevant than integrated reporting, although both could reinforce the relevance of accounting information. Jadoon et al. (2021) analyzed a sample of European firms and found that investors include corporate sustainability performance information in their valuations, along with financial and economic performance data.
Permasatari 和 Narsa （2021）報告稱，與年度報告分開的可持續發展報告比綜合報告更具價值相關性，儘管兩者都可以加強會計資訊的相關性。Jadoon 等人（2021 年）分析了歐洲公司的樣本，發現投資者在估值中包括企業可持續發展績效資訊，以及財務和經濟績效數據。

However, none of these studies specifically focused on the evolution of the explanatory power of sustainability information over time. Hence, the extant literature does not provide sufficient hints about the direction of trends in the value relevance of sustainability information. Therefore, this study tested the following null hypothesis:
然而，這些研究都沒有專門關注可持續性資訊的解釋力隨時間的演變。因此，現存文獻沒有提供關於可持續性資訊價值相關性趨勢方向的足夠提示。因此，本研究檢驗了以下原假設：

H2: The value relevance of ESG ratings in the EU did not change from 2005 to 2020.
H2：從 2005 年到 2020 年，歐盟 ESG 評級的價值相關性沒有變化。

3. Research design 3. 研究設計

As mentioned, sustainability information is being disclosed by an increasing but limited number of companies. Of these companies, even fewer have an ESG performance rating (Eliwa et al., 2021). Prior research found that finance professionals increasingly use sustainability information to make investment decisions (Amel-Zadeh & Serafeim, 2018; Christensen et al., 2022). To access such information, investors need dependable and accessible datasets (Dimson et al., 2020) to increase the reliability and comparability of the data, which are considered one of the main drivers of sustainable investment (Amel-Zadeh & Serafeim, 2018; Jonsdottir et al., 2022). Therefore, ESG ratings provide a simple yet informative element of a firm’s sustainability performance and are widely used by investors in their investment decision-making processes (Baker et al., 2022; Hübel & Scholz, 2020).
如前所述，披露可持續發展資訊的公司越來越多，但數量有限。在這些公司中，擁有 ESG 績效評級的公司更少（Eliwa 等人，2021 年）。先前的研究發現，金融專業人士越來越多地使用可持續性資訊來做出投資決策（Amel-Zadeh & Serafeim，2018;Christensen 等人，2022 年）。為了獲取這些資訊，投資者需要可靠且可訪問的數據集（Dimson 等人，2020 年）來提高數據的可靠性和可比性，這被認為是可持續投資的主要驅動力之一（Amel-Zadeh & Serafeim，2018 年;Jonsdottir et al.， 2022）。因此，ESG 評級為公司的可持續發展績效提供了一個簡單而翔實的元素，並被投資者在其投資決策過程中廣泛使用（Baker 等人，2022 年;Hübel & Scholz，2020 年）。

This study acknowledges that a large number of financial analytics companies issuing ESG ratings (Eccles et al., 2020), which usually adopt private and undisclosed methodologies that could lead to diverging results over a single firm’s scores (e.g., Atta-Darkua et al., 2020; Christensen et al., 2022; Dimson et al., 2020). Albeit limiting the study, this represents the most viable solution to keep the sample size as large as possible to improve inferences and reduce self-selection bias (Halbritter & Dorfleitner, 2015).
本研究承認，大量金融分析公司發佈 ESG 評級（Eccles 等人，2020 年），這些公司通常採用私人和未披露的方法，這可能導致單個公司評分的結果出現差異（例如，Atta-Darkua 等人，2020 年;Christensen 等人，2022 年;Dimson 等人，2020 年）。儘管限制了研究，但這代表了最可行的解決方案，以保持樣本量盡可能大，以改善推斷並減少自選擇偏差（Halbritter & Dorfleitner，2015）。

3.1. Empirical model 3.1. 經驗模型

The value relevance of accounting numbers is usually determined by adopting two models based on stock prices or returns (Onali et al., 2017). The Price Regression Model (Ohlson, 1995) regresses market value (i.e., stock price)

(P)

on equity book value per share

(B V)

and earnings per share (EPS), while the Return Regression Model (e.g., Francis & Schipper, 1999) regresses stock returns on earnings and changes in earnings.
會計數位的價值相關性通常是通過採用基於股票價格或回報的兩種模型來確定的（Onali et al.， 2017）。價格回歸模型（Ohlson，1995）根據每股

(B V)

權益賬麵價值和每股收益（EPS）

(P)

回歸市場價值（即股票價格），而回報回歸模型（例如，Francis & Schipper，1999）回歸股票收益回報和收益變化。

To explore the first research question and test H1, this study adapts Ohlson’s (1995) framework as the baseline empirical model, and the Return Regression Model serves as a robustness check. The baseline model is as follows:
為了探索第一個研究問題並測試 H1，本研究採用 Ohlson （1995）的框架作為基線實證模型，並採用回歸回歸模型作為穩健性檢查。基線模型如下所示：

P_{ith} = β_{0} + β_{1} B V_{ith} + β_{2} E P S_{ith} +

Industry

_{i} +

Country

_{i} +

Interim

_{h} + ε_{i t h}

P_{ith} = β_{0} + β_{1} B V_{ith} + β_{2} E P S_{ith} +

行業

_{i} +

國家

_{i} +

臨時

_{h} + ε_{i t h}

where

i

indicates firms,

t = 1, 2, \dots, 16

represents each year from 2005 to

2020, h

is an indicator of the half- and full-year data. Country (Country

_{i}

) and sector-effects (Industry

_{i}

) control for unobserved heterogeneity, and Interim

_{h}

is a dummy to control for fixed effects related to half-year data. The robustness checks include different baseline model specifications and alternative accounting numbers. In line with Barth and Clinch’s (2009) findings, all financial accounting variables are deflated by outstanding share numbers to avoid scale effects. The regressions include both half-year interim and annual report data.

^{2}

其中

i

表示公司，

t = 1, 2, \dots, 16

表示從 2005 年到

2020, h

的每一年，是半年和全年數據的指標。國家（Country

_{i}

）和行業效應（Industry

_{i}

）控制未觀察到的異質性，而 Interim

_{h}

是控制與半年數據相關的固定效應的虛擬物件。穩健性檢查包括不同的基線模型規範和替代會計編號。根據 Barth 和 Clinch （2009）的發現，所有財務會計變數都被已發行的股票數量所壓縮，以避免規模效應。回歸包括半年中期報告和年度報告數據。

^{2}

To test H2 and answer the second research question, the previous model was supplemented with ESG data as follows:
為了檢驗 H2 並回答第二個研究問題，前面的模型補充了 ESG 數據，如下所示：

\begin{aligned} P_{i t h} = & β_{0} + β_{1} B V_{i t h} + β_{2} E P S_{i t h} + β_{3} E S G_{i t} + {Industry}_{i} + {Country}_{i} + {Interim}_{h} \\ + ε_{ith} \end{aligned}

where

E S G

is sustainability performance information, proxied by ESG ratings.
其中

E S G

是可持續性績效資訊，由 ESG 評級代理。

In both models,

P_{i t h}

is determined as the mean of all share prices recorded between three and six months after the end of the fiscal period.

^{3}

All financial statement (deflated) variables are winsorized at the 1st and 99th percentiles to limit the impact of extreme outliers (e.g., Adwan et al., 2020; Barth et al., 2023). In the analyses, industry sectors are grouped according to Fama and French’s 48-industries classification (Fama & French, 1999, 2023b).
在這兩個模型中，

P_{i t h}

均被確定為財政期結束后三到六個月內記錄的所有股價的平均值。

^{3}

所有財務報表（縮小的）變數在第 1 個和第 99 個百分位數處縮水，以限制極端異常值的影響（例如，Adwan 等人，2020 年;Barth 等人，2023 年）。在分析中，根據 Fama 和 French 的 48 個行業分類（Fama 和 French，1999,2023b）對行業部門進行分組。

3.2. Methodology for testing trend hypotheses
3.2. 測試趨勢假設的方法

Both hypotheses require an analysis of variations in value relevance
這兩個假設都需要分析價值相關性的變化

E-mail address: alessandro.migliavacca@unito.it.
電子郵件地址：alessandro.migliavacca@unito.it。
https://doi.org/10.1016/j.intaccaudtax.2024.100620
$^{1}$ For instance, Collins et al. (1997) and Brown et al. (1999) analyzed the same sample over 40 years with different approaches. They respectively highlighted a slight increase and a decrease in the combined value relevance of book values and earnings over 40 years, with a shift in value relevance from earnings to book values. Similarly, Lev and Zarowin (1999) also found that the value relevance of earnings and book values was reduced over time, in contrast to earlier studies (e.g., Collins et al., 1997; Francis & Schipper, 1999), deeming the inconsistency due to the examination of different periods.
$^{1}$ 例如，Collins 等人（1997）和 Brown 等人（1999）用不同的方法分析了同一樣本 40 多年。它們分別強調了 40 年來賬麵價值和收益的綜合價值相關性略有增加和減少，價值相關性從收益轉變為賬麵價值。同樣，Lev 和 Zarowin （1999）也發現，與早期的研究相比，收入和賬麵價值的價值相關性隨著時間的推移而降低（例如，Collins 等人，1997 年;Francis & Schipper， 1999），認為由於對不同時期的審查存在不一致。
$^{2}$ Dunham and Grandstaff (2022) suggested that quarterly data capture more fluctuations than the usually adopted annual view of economic conditions, albeit annual data is provided with a more detailed audit process. Only halfyear data is used since quarterly financial reports in the EU was abolished by Directive 2013/50/EU to “reduce the administrative burden on smaller issuers.”.
$^{2}$ Dunham 和 Grandstaff （2022）建議，季度數據比通常採用的經濟狀況年度視圖捕捉到更多的波動，儘管年度數據提供了更詳細的審計過程。自歐盟的季度財務報告被第 2013/50/EU 號指令廢除以“減輕小型發行人的行政負擔”以來，僅使用半年數據。
$^{3}$ The justification for this choice is twofold. First, accounting information is usually released in the form of general financial reports at least three months and up to six months after the end of the fiscal period. For this reason, prior studies usually adopt the stock price at three months (e.g., Collins et al., 1997) or six months after the fiscal year-end (e.g., Barth et al., 2008; Onali et al., 2017). However, after its release, accounting information is also used as a historical base for price determination. Therefore, an average price should capture not only the value relevance of different information at its release but also over a short-term period of influence. $P_{i t 1}$ is the share price of the firm $i$ at the year $t$ recorded after the release of first semester interim data and is determined as the mean of the weekly prices recorded between September 30 and December 31 of the same year. $P_{i t 2}$ is the share price of the firm $i$ at the year $t$ recorded after the release of annual report data and is determined as the mean of weekly prices recorded between April 1 and June 30 of the following year $t + 1$ . In the regressions, $P_{i t 1}$ is correlated with interim data released after the first semester and $P_{i t 2}$ with annual report data. Robustness checks also include a test for point prices at three and six months.
$^{3}$ 這種選擇的理由有兩個。首先，會計資訊通常在財政期結束后至少三個月至最多六個月內以一般財務報告的形式發佈。出於這個原因，先前的研究通常採用財政年度結束后三個月（例如，Collins 等人，1997 年）或財政年度結束后六個月（例如，Barth 等人，2008 年;Onali et al.， 2017）。但是，在發佈后，會計資訊也被用作價格確定的歷史依據。因此，平均價格不僅應該捕捉不同資訊在其發佈時的價值相關性，還應該捕捉到短期影響的價值相關性。 $P_{i t 1}$ 是公司 $i$ 在第一半年中期數據發佈後記錄的年度 $t$ 的股價，確定為同年 9 月 30 日至 12 月 31 日期間記錄的每周價格的平均值。 $P_{i t 2}$ 是公司 $i$ 在年度報告數據發佈後記錄的年度 $t$ 的股價，並確定為每年 $t + 1$ 4 月 1 日至 6 月 30 日期間記錄的每周價格的平均值。在回歸中， $P_{i t 1}$ 與第一學期 $P_{i t 2}$ 后發佈的中期數據和年度報告數據相關。穩健性檢查還包括 3 個月和 6 個月時的點價測試。

Value relevance of accounting numbers and sustainability information in Europe: Empirical evidence from nonfinancial companies 歐洲會計數位和可持續性資訊的價值相關性——來自非金融公司的經驗證據