The Unintended Consequence of Environmental Regulations on Earnings Management: Evidence from Emissions Trading Scheme in China

Abstract

This study extends the existing research on the impact of environmental regulations from an earnings management perspective. Using the difference-in-differences approach, the study contributes to the understanding of the relationship between corporate earnings management and the implementation of China’s carbon emissions trading program. In particular, the study finds a positive relationship between corporate earnings management and the implementation of China’s carbon emissions trading program. Furthermore, our analysis reveals that this positive correlation is much stronger for firms facing tighter financial constraints, higher information opacity, less intense competition, and higher pressure to reduce emissions. These findings illustrate the unintended consequences of market-based environmental regulations and provide new evidence for assessing the efficiency of much-promoted market-based environmental regulations.

1. Introduction

The escalating incidents of pollution and the urgent imperative to achieve carbon neutrality have heightened the pressure for environmental protection. Worldwide, policymakers have implemented environmental regulations to curb greenhouse gas (GHG) emissions. Since the 18th National People’s Congress, the Chinese government has promulgated 122 ecological regulations, aiming to catalyze corporate efforts in reducing carbon emissions (Ministry of Ecological Environment of the People’s Republic of China, https://www.mee.gov.cn/ywgz/zcghtjdd/sthjzc/202101/P020210128390855763868.pdf, accessed on 1 January 2021.). The implementation of these laws and regulations signifies China’s commitment to establishing a comprehensive system of both explicit and implicit environmental regulations, inclusive of public participation. Additionally, it underscores the continual elevation of environmental regulations [1,2].

The Paris Agreement set up goals for almost all of the countries for their emission mitigation targets. As a core tool to remedy the negative externalities of environmental pollution, environmental regulations can be divided into command and control regulations and market-based regulations. Emissions trading schemes (ETSs) are the most mentioned market-based environmental regulations and are shown to be cost effective [3]. They are one of the approaches to promote economic development and environmental protection while achieving the goal of sustainable development in a dialectical interaction.

ETSs, serving as a central tool to address the negative externalities of environmental pollution, stand out as being widely acknowledged and are proven to be cost effective [3]. They represent a dialectical and interactive approach to fostering economic development and environmental protection, aligning with sustainable development goals. The Statistical Review of World Energy (SRWE) reports that China’s carbon emissions reached 10,523 MMT in 2021, constituting the largest share (one quarter) of global carbon emissions. Notably, China’s ETS dwarfs the EU’s ETS, being twice its size. Consequently, the planning and operation of China’s ETS draw significant global interest. China’s experience with the ETS not only offers valuable insight for balancing economic development and environmental conservation but also serves as a guide for other developing countries facing similar challenges. These challenges encompass: (1) increased uncertainty regarding future economic growth rates, contributing to the complexity of establishing long-term emission caps; (2) limited pass-through of carbon costs due to tightly regulated electricity markets; (3) inadequate statistical data; and (4) a lack of policy coordination [4]. While research on ETSs in developing countries holds significance, the current available literature is insufficient [5].

Considering the performance of China’s stock market, which ranks at the bottom globally in 2023, with financial fraud significantly impacting shareholders’ confidence, the China Securities Regulatory Commission (CSRC) audit revealed that, in 2023, 156 A-share listed companies were investigated by the CSRC, with 70% of the incidents involving information disclosure (https://www.thepaper.cn/newsDetail_forward_25865439, accessed on 1 January 2024). Currently, Chinese enterprises continue to grapple with issues in their development process, including an emphasis on economic benefits over social benefits, prioritizing shareholders’ interests over stakeholders’ interests, and giving precedence to short-term interests over long-term development.

Earnings information is the most crucial data in financial disclosure, reflecting firms’ revenues and costs [6,7]. The degree of information asymmetry and the efficiency of resource allocation in capital markets are greatly impacted by the quality of earnings information [8]. Earnings information serves as a vital tool for maintaining the healthy ecology of the financial market and safeguarding investors’ legitimate interests [5]. However, corporate managers often engage in earnings management by manipulating accruals and structuring real transactions. Their goal is to influence the stock market’s view of the company’s value, boost their own compensation, minimize the risk of violating loan agreements, and avoid regulatory scrutiny [9,10]. Earnings management tendencies can result in a degradation of earnings information quality. The academic spotlight on how to diminish the likelihood of a decline in earnings information quality has intensified [11]. Francis et al. [12] assert that earnings information quality is shaped by elements like the external environment, firm attributes, and decision makers. As the major participants of the market, firms have to change their extensive growth strategies to adapt to domestic reforms and the complicated external environment [13].

Currently, there are two market-based environmental regulation approaches in China that have garnered significant attention: environmental, social, and governance (ESG) ratings and the emissions trading system (ETS). ESG ratings are considered a form of market soft regulation [14]. Researchers advocating for the constraining influence of ESG ratings on earnings management primarily explain the underlying mechanism through the lenses of information asymmetry and ethics. Nonfinancial disclosures can provide crucial information, aiding investors in evaluating firms’ financial performance. Therefore, it is reasonable to suggest that ESG ratings, as a form of nonfinancial information disclosure, can also limit earnings management by mitigating information asymmetry [15]. Similarly, Borralho et al. [16] note that companies emphasizing ESG ratings are likely to provide more extensive corporate disclosures, which can help external investors gain a more thorough understanding of the company. This suggests that ESG rating disclosures play a supervisory role in restricting earnings management [14].

However, the vast majority of traders participating in the carbon market in China are pilot enterprises identified by the government for mandatory inclusion in the list of environmentally compliant enterprises. The actual costs and benefits of these enterprises’ participation in carbon trading are difficult to obtain from annual reports or other public sources. However, it can be inferred that, since these companies do not participate voluntarily but are forced to join, the companies on the list face greater pressure to transform themselves and bear the costs of environmental compliance. At the same time, however, the carbon market provides a place for participants to buy and sell carbon allowances freely; therefore, one might contend that the trading of carbon emissions in China is more akin to a market mechanism with “strict regulation”. The government’s strict regulation is prone to problems such as “one-size-fits-all” and “rent-seeking”. Additionally, hard government regulation is characterized by inconsistent regulatory efforts, limited regulatory capacity, high regulatory costs, and low regulatory efficiency. Current research is mainly focus on the macro level [17] or in developed country settings [18], and some other research investigated the relationship between ETS implementation and firm value as well as its FDI activities [19,20,21]. Nevertheless, research on firms’ financial management response to the introduction of the ETS is still lacking. In this context, there is a need to explore novel pathways through which the CO₂ ETS affects the quality of earnings information and to identify the underlying mechanisms.

The core issue explored in this paper is whether the CO₂ ETS, a market-based mechanism characterized by “strict regulation”, impacts corporate earnings management in a positive way similar to ESG ratings, or whether it has an “unintended” negative effect. The introduction of the CO₂ ETS provides us with a quasi-experimental setting to explore the impact of the CO₂ ETS on earnings management, given the unclear ex ante nature of this impact. Several studies have contended that participants in the CO₂ ETS tend to enhance firms’ disclosure [22] and reduce information asymmetry. However, in alignment with Hu et al. [23], we acknowledge the possibility that ETSs may elevate administrative costs. We also posit that compliance costs could be substantial due to the high expenses associated with cleaner production systems, and the market’s uncertainty regarding carbon prices may render managers more cautious about their returns. Consequently, there might be an inclination toward upward earnings management.

This study empirically investigates the influence of the CO₂ ETS on the quality of earnings management using a dataset of Shanghai and Shenzhen A-share listed companies that were required to participate in the CO₂ ETS from 2012 to 2021. The results indicate a positive relationship between the CO₂ ETS and earnings management practices. The mechanism study delves deeper, revealing that the reduction in earnings quality induced by the carbon trading mechanism primarily originates from the alleviation of financial constraints on firms, increased corporate opacity, and heightened market competition. We further discern varied impacts based on differences in emission reduction targets, noting that the effects are more pronounced in regions facing greater pressure to meet commitments and possessing higher emission reduction targets.

The main contributions and innovations of this paper are reflected in the following areas:

Firstly, our study contributes to the enrichment of research on firms’ strategic responses to “strict regulation” ETS by unveiling the mechanisms underlying these responses. While the impact of external governance environments, such as law, government intervention, and socio-culture, on firms’ decisions has been extensively explored, the external environment of environmental regulations has received comparatively less analysis [9,24].

Secondly, we introduce a novel perspective for evaluating the efficiency of the CO₂ ETS, reexamining the political cost theory through the lens of environmental regulations. Additionally, we propose a new avenue for capital market regulators to enhance the quality of earnings information within the existing formal institutional framework.

Thirdly, our paper offers the inaugural empirical investigation into the causal impact of the CO₂ ETS on corporate earnings management within a developing country context. Prior research has largely concentrated on exploring the connection between ESG ratings performance and earnings quality.

Furthermore, from a regulatory standpoint, our findings suggest that the mandatory inclusion of firms in CO₂ ETS programs imposes additional social and financial pressures on firms, intensifying their incentives to engage in earnings management. While acknowledging the potential for varying effects of mandating firms to join a CO₂ ETS in different institutional contexts, our results align with the perspective that external regulatory pressures can prompt firms to act in unintended ways [25].

2. Theoretical Analysis and Research Hypothesis

To align with domestic reforms and navigate the challenging external environment, businesses, as key players in the carbon market, need to shift their growth strategies [13]. Current research mainly focuses on the macro level [17] or developed country settings [18]. While some studies have investigated how the implementation of ETSs affects firm value and their FDI activities [19,20,21], there is still a lack of research on corporate financial management responses to ETSs.

Government regulation increases expected costs for firms [16], impacting accounting disclosure at the firm level. Firms have incentives to manipulate profits to circumvent government regulation and external oversight [26,27]. In China, there is no consensus on how corporate earnings management relates to external regulation. Some argue that earnings management is positively influenced by heightened government regulation [28], while other studies demonstrate a negative effect [29]. The effect of ETS on earnings management remains uncertain. This research aims to examine the extent and direction of earnings management at the firm level within the context of market-oriented environmental regulations in China. Can carbon emissions trading affect firms’ earnings management behavior? This paper analyzes five main aspects:

Firstly, the ETS mechanism imposes compliance costs and carbon risks on firms, negatively affecting their performance and capital-raising ability. Managers may engage in earnings management to conceal these adverse impacts. In our context, firms that are mandatorily listed are treated as the group under consideration. Once listed for treatment, firms must enhance transparency in their environmental information disclosure and adhere to stricter whole-process regulation under the monitoring, reporting, and verification (MRV) framework. Any revelation of environmental issues prompts reactions from the public, including governments, analysts, auditors, financial institutions, rating agencies, etc. Consequently, treated enterprises face greater pressure to meet compliance constraints of their quota subsidies, leading to increased associated costs. However, enterprises, especially listed companies, must maintain relatively stable profit levels to gain trust from external stakeholders (e.g., investors and financial institutions). In this context, how firms navigate the increasing transition costs is our primary concern, and we seek answers from the perspective of earnings management. Conventional wisdom suggests that managers choose to smooth the reporting of earnings around some predetermined level. Fudenberg and Tirole [30] propose the motivation behind this “smoothing” behavior as a trade-off between the present and the future. Their study intuitively demonstrates that, when present performance is lacking, managers are motivated to bring future earnings into the current period to decrease the chance of dismissal. On the other hand, if future performance is anticipated to decline, managers are more inclined to defer current earnings to future periods to avoid being fired later on. Building on this inference, we argue that managers tend to smooth profits in response to transition costs [31] and derive Hypothesis 1:

Hypothesis 1. 

The ETS has an impact on a company’s earnings management.

Secondly, the CO₂ ETS introduces a disturbance in the distribution of firms’ financial positions, prompting firms to redistribute their profits. We hypothesize that the impact of transition costs on financially constrained firms is more pronounced due to their limited ability to secure external financing. Enterprises operate within a specific institutional environment, inevitably influencing their economic behavior. Firms facing financial constraints are more likely to practice earnings management to attract external financial assistance than those without such constraints [32,33]. In China, earnings management is more prevalent among firms seeking to alleviate financing constraints. Wang and Zhang [34] demonstrate that managers aim to increase firm profits, and this intention is positively associated with firms’ access to bank loans. Zhou et al. [35] also identifies a positive correlation between earnings management and financial constraints, especially among growing firms. Carbon emissions trading may impact firms’ earnings management behavior in two ways. First, carbon credit trading may elevate firms’ operating costs, intensifying their financial pressure. To cope, enterprises may adopt earnings management behaviors, such as profit manipulation or delayed recognition of losses. Second, carbon emissions trading may alter firms’ investment decisions, influencing their earnings management behaviors. For instance, firms may opt for investments in clean technologies to lower carbon emissions and decrease the purchase cost of carbon credits. However, such investments may alter the firm’s financial position, thereby affecting earnings management behavior. Despite this, some scholars propose a different viewpoint. The study by Yu et al. [36] discovered that firms engaged in carbon emissions trading experienced a notable decrease in carbon emission intensity, which led to improved financial performance. This suggests that carbon emissions trading influences firms’ earnings management behavior by impacting their operating costs (i.e., carbon emission costs), but they did not explicitly test the hypothesis regarding earnings management. Building on these arguments, we anticipate the CO₂ ETS to exert a more substantial impact on firms with stronger financial constraints, exacerbating earnings management to secure increased access to finance. This leads to Hypothesis 2:

Hypothesis 2. 

The CO₂ ETS has a greater impact on earnings management for more financially constrained firms.

Thirdly, carbon emissions trading provides additional information about a company’s business development, potentially ameliorating the information disadvantage faced by external investors and diminishing the likelihood of corporate earnings management. Studies argue that firms’ earnings management behavior will be curbed when the quality of information disclosed is improved. Firstly, carbon emissions trading may alter the market’s attitude toward firms’ information. Specifically, a company’s participation in carbon emissions trading may be construed by the market as an indication of the firm’s commitment to environmental protection, thereby enhancing the firm’s reputation and brand value. This information effect could reduce firms’ earnings management behavior. Secondly, carbon emission rights trading may influence corporate disclosure practices. Firms engaged in carbon emission rights trading may be compelled to disclose more environmental protection information, such as carbon emissions and carbon emission rights trading. Such information disclosure could further diminish firms’ earnings management behavior. In studies examining the link between information opacity and earnings management, the common approach is to follow the sequence: information disclosure—information asymmetry—earnings management [37,38,39]. Their model demonstrates that increased firm-level disclosure reduces information asymmetry. Research suggests that firms’ earnings management is restrained by information transparency, highlighting the interplay between information opacity and earnings manipulation [40,41,42]. There is consensus that the interaction between information opacity and earnings management is negatively correlated. Given that the CO₂ ETS mandates companies to reveal detailed information regarding their emissions and operations, potentially reducing their level of opacity, and assuming this reasoning applies to the CO₂ ETS context, we anticipate that the CO₂ ETS will have a more significant impact on companies with greater information opacity. Hence, we put forward Hypothesis 3:

Hypothesis 3. 

The CO₂ ETS has a greater impact on firms’ earnings management when firms’ information opacity is higher.

Fourthly, to a certain extent, the CO₂ ETS can be seen as similar to competition in product markets and functions as an alternative external regulatory mechanism. Carbon trading, on the one hand, exposes internal corporate data to the outside, increases transparency, and acts as a check on managerial behavior. On the other hand, by setting specific standards for corporate management, the CO₂ ETS motivates companies to improve their governance and risk management practices. Under the dual-action mechanism of carbon trading, the degree of corporate earnings management is reduced. Research has shown that competition in the product market can reduce information asymmetry, improve the oversight efficiency of external parties, tackle agency issues, and positively influence the management of corporate earnings. However, some researchers have also identified that product market competition can have a negative impact on the governance of earnings management. On the one hand, in a highly competitive setting, management may be more motivated to obscure financial details through earnings manipulation and limit information access for potential rivals. Heightened competition in the product market can squeeze a firm’s profit margins and impact its capacity to secure internal funding, prompting management to use earnings management strategies to satisfy financing needs. Product market competition is considered an external governance mechanism that can complement or serve as an alternative to certain internal governance mechanisms within the firm. In our sample, the firms are listed companies that were mandated to participate in the CO₂ ETS. Building on previous research, we hypothesize that the more competitive the market, the less latitude the listed companies have for engaging in earnings management. Therefore, we propose Hypothesis 4:

Hypothesis 4. 

The CO₂ ETS inhibits firms’ earnings management when they face more competition.

Fifthly, as per the implementation rules of the CO₂ ETS, firms exceeding the prescribed regional CO₂ standards are obligated to participate in the program. To circumvent environmental regulations, heavily polluting firms often resort to profit manipulation to alleviate regulatory pressures, meet external expectations, and gain public incentives [29]. On the one hand, heavily polluting firms naturally attract more attention from regulators and the public, leading to increased external pressures [43]. As the demand for reducing emission increases, companies are more inclined to manipulate their profits. On the other hand, Yuan et al. [44] argue that labor costs influence firms’ propensity for earnings management. Air pollution raises labor costs because workers demand higher compensation when the work environment suffers from severe air pollution [45,46]. As a result, the effect of carbon emission policies becomes more significant when pollution levels are elevated. The objective of the ETS is to reduce firms’ carbon emissions, and mandated firms must surrender a specific amount of carbon allowances within the compliance period. Firstly, in regions with more active carbon markets, firms tend to reduce earnings management as active carbon markets provide better conditions for firms to flexibly address environmental costs. Secondly, these firms are the actual implementers of achieving peak carbon, and we further suspect that whether local carbon has peaked or not will also impose different environmental pressures at the firm level. For example, in non-peak situations, the government may consider compressing allowances in carbon emissions trading, leading firms to compete for a limited number of allowances, thus engaging in activities that involve vacating production value or incurring real financial costs. Therefore, Hypothesis 5 is proposed here:

Hypothesis 5. 

The CO₂ ETS has a weaker impact on corporate earnings management in provinces with more active carbon markets and higher peak carbon completion.

3. Data and Methodology

3.1. Data and Sample Selection

The research goal of this study is to identify the unintended consequences of the CO₂ ETS. This study uses the implementation of the CO₂ ETS as a quasi-natural experiment and concentrates on companies listed on the Chinese A-share market. In a departure from current research, we uniquely use firm-level data to reveal these consequences. The firms in the sample are the listed firms that are regulated and actively involved in the CO₂ ETS. The list of firms involved and firm-level financial data are sourced from the Chinese Stock Market and Accounting Research Database (CSMAR), with the study covering the years 2012 through 2021.

3.2. Measurement for Earnings Management

As the nexus of a series of contracts, the contract design of enterprises is typically anchored in accounting figures, shaping optimal contract designs for managers. However, owing to the presence of bounded rationality and transaction costs, contracts are not omnipotent. Managers harbor private intentions when selecting accounting policies. In alignment with the approach of Dechow et al. [47] and Dechow and Skinner [48], we utilize discretionary accruals (DA) and real earnings management (TREM) as proxies for firms’ earnings management behavior. A higher value for both indicators indicates a higher probability that managers take advantage of the considerable flexibility in accounting policies to obscure the true economic activities and performance of the company, thus increasing the opacity of the company’s information.

Table 1. Definitions of variables.

Variables	Definitions
Discretional Accruals (DA)	$\begin{array}{l}{\displaystyle \frac{T{A}_{i,t}}{A_{i,t-1}}}={\beta }_0{\displaystyle \frac{1}{A_{i,t-1}}}+{\beta }_1{\displaystyle \frac{\mathsf{\Delta }RE{V}_{i,t}}{A_{i,t-1}}}+{\beta }_2\left({\displaystyle \frac{PP{E}_{i,t}}{A_{i,t-1}}}\right)+{\epsilon }_{i,t}\\ ND{A}_{i,t}={\widehat{\beta }}_0{\displaystyle \frac{1}{A_{i,t-1}}}+{\widehat{\beta }}_1{\displaystyle \frac{\mathsf{\Delta }RE{V}_{i,t}-\mathsf{\Delta }RE{C}_{i,t}}{A_{i,t-1}}}+{\widehat{\beta }}_2\left({\displaystyle \frac{PP{E}_{i,t}}{A_{i,t-1}}}\right)\\ D{A}_{i,t}={\displaystyle \frac{T{A}_{i,t}}{A_{i,t-1}}}-ND{A}_{i,t}\end{array}$ $\mathsf{\Delta }RE{V}_t$ refers to change in operating income, $\mathsf{\Delta }RE{C}_t$ refers to change in account receivable and PPE t refers to net fixed asset.
Real Earnings Management (TREM)	$\begin{array}{c}{\displaystyle \frac{CF{O}_{\mathrm{i}\mathrm{t}}}{A_{it-1}}}={\alpha }_0+{\alpha }_1{\displaystyle \frac{1}{A_{it-1}}}+{\alpha }_2{\displaystyle \frac{RE{V}_{\mathrm{i}\mathrm{t}}}{A_{it-1}}}+{\alpha }_3{\displaystyle \frac{\mathsf{\Delta }RE{V}_{\mathrm{i}\mathrm{t}}}{A_{it-1}}}+{\epsilon }_{it}\\ \begin{array}{c}{\displaystyle \frac{PRO{D}_{\mathrm{i}\mathrm{t}}}{A_{it-1}}}=b_0+b_1{\displaystyle \frac{1}{A_{it-1}}}+b_2{\displaystyle \frac{RE{V}_{\mathrm{i}\mathrm{t}}}{A_{it-1}}}+b_3{\displaystyle \frac{\mathsf{\Delta }RE{V}_{\mathrm{i}\mathrm{t}}}{A_{it-1}}}\\ +b_4{\displaystyle \frac{\mathsf{\Delta }RE{V}_{\mathrm{i}\mathrm{t}-1}}{A_{it-1}}}+{\epsilon }_{it}\end{array}\\ {\displaystyle \frac{DISEX{P}_{\mathrm{i}\mathrm{t}}}{A_{it-1}}}=c_0+c_1{\displaystyle \frac{1}{A_{it-1}}}+c_2{\displaystyle \frac{RE{V}_{\mathrm{i}\mathrm{t}-1}}{A_{it-1}}}+{\epsilon }_{it}\\ TRE{M}_{it}=(-1)A_{CFO}{O}_{\mathrm{i}\mathrm{t}}+A_{-}PRO{D}_{\mathrm{i}\mathrm{t}}+(-1)A_{-}DISEX{P}_{\mathrm{i}\mathrm{t}}\end{array}$
Size	Log(Total Assets) in year t−1
Leverage	Total Debt/Total Asset in year t−1
Free Cash flow	Operating Cash flow/Total Asset in year t−1
Tangibility	Tangible Assets/Total Asset in year t−1
Loss	A dummy variable equal to 1 when the firm has a negative net earnings in year t−1
Growth	Sales growth in year t−1
Duality	A dummy variable equal to 1 if the firm’s chairman and general manager are the same person
Independent Director	The number of independent directors/total directors in year t−1
Board	Number of board members in year t−1
Top1	Shareholding ratio of the largest shareholder in year t−1
Age	Year of listing in year t−1

provides definitions for the variables used in our analysis. The independent variables are discretional accruals and real earnings management, which reflect the earnings management practices of firms. Definitions for the control variables are also included, with all control variables lagged by one period.

3.3. Measurement for Control Variables

The choice of control variables follows the approach described in the research by Gao et al. [49]. The control variables encompass (1) Size; (2) Leverage; (3) Free Cash Flow; (4) Tangibility; (5) Loss; (6) Growth; (7) Duality: a dummy variable is set to 1 when the chairman and the general manager are the same individual; (8) Independent director; (9) Board; and (10) Age. Table 1 also includes comprehensive definitions for these variables.

3.4. Sample Statistics

In the sample selection process, we exclude all special treatment firms due to their unstable financial performance. Furthermore, we exclude observations from the initial public offering year due to the higher volatility associated with these companies. Observations with missing data are also removed from the sample. This process results in a total of 14,622 firm-year observations from 2791 firms. To tackle potential endogeneity issues, we apply Propensity Score Matching (PSM) with a caliper of 0.001 to identify a comparable sample for the treated groups. The details of the determinants model and matching efficiency are presented in

Table 2. Determinants model and matching efficiency.

Panel A				(1)
				Treat
Size				0.335 ***
				(0.035)
Lev				−0.469 **
				(0.220)
Free Cash Flow				0.786
				(0.583)
Tangibility				2.411 ***
				(0.203)
Loss				0.226 **
				(0.111)
Growth				−0.176 *
				(0.105)
Duality				0.294 ***
				(0.079)
Independent Director				1.710 **
				(0.697)
Board				0.051 **
				(0.023)
Top1				0.831 ***
				(0.230)
Age				0.003
				(0.006)
_cons				−12.069 ***
				(0.727)
N				14,622
R-Squared				0.062
Panel B	UnMatched	Mean			%reduct	t-test
Variable	Matched	Treated	Control	%bias	|bias|	t	p > |t|
Size	U	22.784	22.158	50.3		15.44	0.000
	M	22.63	22.653	−1.9	96.3	−0.4	0.687
Lev	U	0.44789	0.40711	20.8		6.2	0.000
	M	0.43962	0.44092	−0.7	96.8	−0.14	0.887
Free Cash Flow	U	0.06088	0.04849	19.6		5.71	0.000
	M	0.05888	0.06021	−2.1	89.3	−0.45	0.650
Tangibility	U	0.28483	0.20298	52.6		16.2	0.000
	M	0.26771	0.26029	4.8	90.9	1.02	0.309
Loss	U	0.1241	0.11219	3.7		1.13	0.259
	M	0.12346	0.12778	−1.3	63.7	−0.28	0.783
Growth	U	0.13736	0.17027	−9.8		−2.69	0.007
	M	0.14342	0.12626	5.1	47.9	1.18	0.237
Duality	U	0.28956	0.28722	0.5		0.16	0.877
	M	0.28507	0.31548	−6.7	−1202.4	−1.4	0.162
Independent Director	U	0.37972	0.37598	6.6		2.1	0.036
	M	0.37678	0.37851	−3.1	53.8	−0.65	0.515
Board	U	8.7911	8.453	19.3		6.32	0.000
	M	8.6981	8.6665	1.8	90.7	0.38	0.703
Top1	U	0.37013	0.33222	25.8		7.91	0.000
	M	0.36689	0.36095	4	84.3	0.84	0.399
Age	U	12.01	10.161	25.8		7.68	0.000
	M	11.557	11.691	−1.9	92.7	−0.39	0.696

Notes: (1) *, **, and *** indicate significance at the 10%, 5%, and 1% levels (two-tailed), respectively. (2) %reduct refers to the proportion of sample bias reduction.

. Following PSM, the dataset comprises 10,226 firm-year observations across 2662 firms. To reduce the effect of extreme values, all continuous variables are winsorized at the 1% level.

Our final sample comprises 2662 listed firms over the sample period from 2012 to 2021, with 10 firm-year observations included.

Table 3. Summary of descriptive statistics.

Panel A	N	Mean	SD	Min	Median	Max
Discretional Accruals	10,226	0.073	0.068	0.001	0.054	0.395
Real Earnings Management	10,226	−0.001	0.178	−0.594	0.008	0.482
Size	10,226	22.203	1.106	20.057	22.075	26.064
Leverage	10,226	0.405	0.195	0.054	0.395	0.852
Free Cash Flow	10,226	0.052	0.064	−0.139	0.05	0.238
Tangibility	10,226	0.204	0.133	0.002	0.187	0.68
Loss	10,226	0.117	0.321	0	0	1
Growth	10,226	0.153	0.343	−0.519	0.106	2.187
Duality	10,226	0.312	0.463	0	0	1
Independent Director	10,226	0.376	0.053	0.333	0.364	0.571
Board	10,226	8.462	1.537	5	9	14
Top1	10,226	0.336	0.141	0.088	0.317	0.729
SOE	10,226	0.318	0.466	0	0	1
Age	10,226	10.197	7.172	1	8	30
Panel B	Full sample
	Benchmark		Treat (CO2 ETS)		Total
Year	N	%	N	%	N	%
2012	51	5.72	364	3.9	415	4.06
2013	67	7.52	568	6.08	635	6.21
2014	71	7.97	646	6.92	717	7.01
2015	75	8.42	693	7.42	768	7.51
2016	70	7.86	721	7.72	791	7.74
2017	74	8.31	773	8.28	847	8.28
2018	75	8.42	854	9.15	929	9.08
2019	146	16.39	1565	16.76	1711	16.73
2020	152	17.06	1782	19.09	1934	18.91
2021	110	12.35	1369	14.67	1479	14.46
N	891	100	9335	100	10,226	100

(Panel A) offers an overview of the descriptive statistics for the firm-level observations across the entire sample, while Table 3 (Panel B) illustrates the yearly distribution of treatment and control firms. All continuous variables are winsorized at the 1% tails of their distributions. Our primary proxies are discretional accruals and real earnings management. Panel A shows that the averages of discretional accruals and real earnings management are 0.073 and −0.001, respectively. State-owned enterprises constitute 33.6% of the firms in our sample, with approximately 11.7% sales growth, 40.5% financial leverage, and 33.6% largest shareholding. Panel B shows the yearly distribution of the two groups, indicating a rising trend in the number of firms that received treatment. The treated group comprises 180 firms, accounting for 6.76% of the total sample. Due to the relatively small proportion, we use the PSM method to improve the comparability between the two groups and reduce endogenous problems to some degree. Although 2013 is generally assumed to be the initial year, the name list of firms required to participate in the CO₂ ETS is determined in 2012.

3.5. Difference-in-Difference

Our empirical strategy tests the hypothesis that the CO₂ ETS incentivizes firms to manage their earnings. It exploits variations in the treatment of regional carbon emissions schemes across both cross-section and time series, implementing difference-in-difference (DID) regression at the firm-year level. If the trends in earnings management for treated and untreated firms are parallel before the implementation of the CO₂ ETS, DID estimates can isolate the effects of the rule itself. The parallel trend test is shown in Figure 1, and the findings suggest that no significant difference exists between the treatment and control groups before the implementation of the CO₂ ETS. However, a significant difference occurs after the introduction of the CO₂ ETS, and the parallel trend test is passed.

3.6. Model Specification

We employ an ordinary least squares (OLS) regression model in our baseline analysis to examine the relationship between firms’ earnings management practices and market-driven environmental regulations. The correlation analysis aims to uncover any unintended consequences of the policy. The baseline regression equation is as follows:

$$Y_{it}={\alpha }_i+{\eta }_t+\beta {(Treat Post}_{it})+\gamma {(Controls}_{it-1})+{ϵ}_{it}$$

(1)

where $i$ refers to each individual firm, and t denotes the year. $Y$ represents DA and TREM as proxies for firms’ earnings management behavior. Firm controls ${(Controls}_{it-1})$ include the key inputs for commonly used firm-level financial measures, which have been clarified in previous section. ${\alpha }_i$ is used to control the firm fixed effect. $Treat$ equals 1 if a firm is in the predetermined regulated name list. $Post$ is also a dummy that equals 1 if the year is after the firm was included in the name list. In the baseline model, the key variable of interest is the coefficient $\beta$ on the interaction term. This coefficient reflects the difference in the change in earnings management between our treatment firms and benchmark firms following their mandatory participation in the CO₂ ETS.

4. Empirical Results

4.1. Main Result

Table 4. Baseline regression using PSM sample.

	(1)	(2)
	Discretional Accruals	Real Earnings Management
Treat × Post	0.024 ***	0.080 ***
	(0.006)	(0.015)
Size	−0.013 ***	0.020 ***
	(0.003)	(0.007)
Leverage	−0.001	−0.065 ***
	(0.011)	(0.025)
Free Cash Flow	−0.006	−0.191 ***
	(0.017)	(0.037)
Tangibility	0.000	−0.049 *
	(0.014)	(0.029)
Loss	0.002	−0.002
	(0.003)	(0.005)
Growth	0.005 **	−0.005
	(0.002)	(0.006)
Duality	0.003	0.003
	(0.003)	(0.006)
Independent Director	−0.001	−0.018
	(0.030)	(0.057)
Board	0.001	0.000
	(0.001)	(0.003)
Top1	0.000	0.074 *
	(0.016)	(0.040)
Age	0.002 ***	0.005 ***
	(0.001)	−(0.001)
_cons	0.335 ***	−0.420 **
	(0.078)	(0.171)
Firm	Yes	Yes
Year	Yes	Yes
Industry	Yes	Yes
N	10,226	10,226
R-Squared	0.030	0.044

Notes: Standard errors in parentheses * p < 0.1, ** p < 0.05, *** p < 0.01.

displays the empirical findings regarding the impact of the CO₂ ETS on firms’ earnings management. We have accounted for firm, year, and industry effects. The estimated coefficients of the interaction term Treat × Post in Columns (1) and (2) are significantly positive at the 1% level, suggesting that the implementation of the CO₂ ETS exerts a substantial positive influence on both discretional accruals and real earnings management.

Since we have 180 treated firms in the full sample, we use PSM to relieve the endogeneity problem. We follow the method of Kim et al. [50]. We begin by estimating a logit regression to identify the factors that determine whether a firm is treated, with the results shown in Table 2 (Panel A). Following this, we match the treatment firms with the control firms using the caliper matching method (with replacement). Covariates in the baseline regression are the matching variables used to get the corresponding control groups. The matching efficiency can be referred to in Table 2 (Panel B). After matching, the bias of the variables significantly drops. The findings indicate that there is no notable difference between the experimental group and the control group. Table 4 presents the results of the baseline regression using the PSM sample, showing that the CO₂ ETS exerts a positive and significant influence on firms’ earnings management, leading to a 2.4% rise in discretionary accruals and an 8% boost in real earnings management. We, therefore, find evidence to support Hypothesis 1.

Our results differ from the previous literature, which suggests that environmental regulations prohibit firms’ earnings management, arguing that external pressure mitigates firms’ earnings management due to monitoring effects [28]. Admittedly, the MRV process of the CO₂ ETS has a monitoring effect on individual firms, but we need to pay more attention to its potential signal effect on potentially unstable financial performance. Ball et al. [51] clarified three origins of firms’ earnings management: accountants value negative information more, lenders care more about economic losses, and economic loss relates to bad decisions that managers do not want to disclose. Based on this argument, we infer that managers may “internalize” transition costs by engaging in earnings management, whitewashing the profit. The main results of the study support our inference. Our results align with the findings of Long et al. [52], which indicated that firms subjected to emission regulations significantly engage in earnings management following the implementation of the carbon emissions trading system. They found that companies with lower cost pass-through capabilities and higher carbon emission intensities are more affected, driven by the compliance costs and carbon risks introduced by new environmental regulations. Similarly, Huang and Zhou’s [28] study on the impact of mandatory external regulation on corporate earnings management yielded comparable results in a regional sample, supporting the hypothesis of the political cost theory.

To ensure the reliability of our findings, we conduct multiple robustness checks, such as testing the parallel trend assumption (Figure 1) and performing placebo tests (Figure 2), all of which yield positive results. Figure 2 presents the placebo setting in the baseline model. By randomly sampling 500 times for the treated firms, we obtain a distribution plot of 500 coefficients. The means for these coefficients are zero. Black circles refer to the significance level. The majority of p-values are larger than 0.05. The placebo test is passed. The impact of the CO₂ ETS on earnings management is significant only within these firms.

4.2. Mechanism Analysis

4.2.1. Mechanism—Financial Constraints

If Hypothesis 2 is valid, the CO₂ ETS is likely to have a more pronounced effect on firms facing greater financial constraints, as they seek to improve their access to financing. We therefore include three proxies to measure financial constraints: KZ (the KZ index, which stands for Kaplan–Zingales Index [53], is the primary measure of financial constraints. It gauges a company’s dependence on external funding. Firms with elevated KZ scores are more likely to face financial challenges. We anticipate that the CO₂ ETS will have a more pronounced impact on companies with higher KZ scores); LFC (the second measurement of financial constraints proxy is the LFC index. The construction of the LFC index follows the research of Cleary [54] and Kuang [55]. We used Leverage, Net Working Capital (Working Capital-Currency-Short-term investment), ROE, Market to Book Value, and Dividend to formulate the LFC index. A higher LFC score implies a stricter financial constraints); and SA.

Table 5. Mechanism—financial constraints.

	(1)	(2)	(3)	(4)	(5)	(6)
	Discretional Accruals	Discretional Accruals	Discretional Accruals	Real Earnings Management	Real Earnings Management	Real Earnings Management
Treat × Post	0.021 ***	0.024 ***	0.029 ***	0.074 ***	0.080 ***	0.097 ***
	(0.006)	(0.006)	(0.007)	(0.015)	(0.015)	(0.016)
KZ	0.003 ***			−0.006 ***
	(0.001)			(0.002)
KZ × Treat × Post	0.007 *** (0.002)			0.019 *** (0.006)
LFC		−0.000			−0.000
		(0.000)			(0.000)
LFC × Treat × Post		0.110 * (0.065)			0.219 * (0.119)
SA			−0.002			0.080
			(0.026)			(0.055)
SA × Treat × Post			0.049 ** (0.023)			0.190 *** (0.051)
Controls	Yes	Yes	Yes	Yes	Yes	Yes
Firm	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes	Yes	Yes
N	10,226	10,226	10,226	10,226	10,226	10,226
R-Squared	0.033	0.030	0.030	0.045	0.045	0.045

Notes: Standard errors in parentheses * p < 0.1, ** p < 0.05, *** p < 0.01.

presents the results of this specification. The dependent variables are discretional accruals and real earnings management. The control variables remain unchanged. The results show that: (1) a higher KZ index refers to a higher financial constraint level, and the interaction term of KZ and Treat × Post captures the moderating effect of financial constraints; (2) a higher LFC score implies a stricter financial constraint, and LFC × Treat × Post captures the financial constraints mechanism; (3) the SA index tends to increase for samples with higher levels of financial constraints, and SA × Treat × Post is used to identify financial constraints impacting on earnings management. We find that all three financial constraints proxies strengthen the policy effect of the CO₂ ETS on firms’ earnings management, as expected. For all three proxies of financial constraints, the treatment effect is stronger in real earnings management than in discretional accruals.

The results confirm Hypothesis 2. Some scholars have pointed out that companies must increase environmental investments aimed at pollution reduction to correct for their non-compliant behaviors, such as purchasing pollution control equipment. This, in turn, puts pressure on their overall performance as they struggle to cope with rising environmental costs [56]. Consistent with our findings, Zhang et al. [57] also suggest that mandatory environmental regulations drive heavily polluting companies to increase environmental investments, making their managers more likely to engage in earnings management through real activities manipulation while simultaneously reducing discretionary accruals.

4.2.2. Mechanism—Firm Opacity

Within the sample, we expect that the impact of the CO₂ ETS should be stronger when firms’ opacity level is higher, as there is more manipulation space for managers. We further test this underlying mechanism (Hypothesis 3) by utilizing three proxies to measure firms’ opacity: Transparency Rating (the transparency rating is provided by CSMAR. The rating has four categories: excellence, good, pass, and fail. We assigned 4–1 to each category. The assessment included companies listed on both the Shenzhen Stock Exchange and the Shanghai Stock Exchange), Analysts’ Attention, and Big 4 Auditor (the “Big 4 Auditor” variable is set to 1 if the firm is audited by one of the four major international audit firms—Ernst & Young, KPMG, Deloitte, or PricewaterhouseCoopers).

Table 6. Mechanism—opacity.

	(1)	(2)	(3)	(4)	(5)	(6)
	Discretional Accruals	Discretional Accruals	Discretional Accruals	Real Earnings Management	Real Earnings Management	Real Earnings Management
Treat × Post	0.022 ***	0.025 ***	0.025 ***	0.077 ***	0.084 ***	0.082 ***
	(0.006)	(0.006)	(0.006)	(0.015)	(0.015)	(0.015)
Transparency Rating	−0.002			−0.004
	(0.002)			(0.003)
Transparency Rating × Treat × Post	−0.029 *** (0.005)			−0.044 *** (0.012)
Analysts Attention		0.003 ***			−0.007 ***
		(0.001)			(0.003)
Analysts Attention × Treat × Post		−0.008 ** (0.004)			−0.015 * (0.009)
Big4			−0.005			0.005
			(0.009)			(0.022)
Big4 × Treat × Post			−0.028 *			−0.074 *
			(0.016)			(0.040)
Contorls	Yes	Yes	Yes	Yes	Yes	Yes
Firm	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes	Yes	Yes
N	10,226	10,226	10,226	10,226	10,226	10,226
R-squared	0.035	0.031	0.030	0.047	0.046	0.044

Notes: Standard errors in parentheses * p < 0.1, ** p < 0.05, *** p < 0.01.

shows the influence of opacity using the three proxies above. Three proxies are used to depict firm opacity status. Columns (1) and (4) are the transparency rating proxy; Columns (2) and (5) are the results of specifications using the analysts’ attention proxy; and Columns (3) and (6) identify the opacity channel using the power of external audit. The coefficients of the interaction terms between proxies and Treat × Post are things captured in the opacity mechanism. The policy effect of the CO₂ ETS is more significant when the opacity level of the firm increases. The results are as expected and provide evidence to support Hypothesis 3, and the treatment effect is stronger in real earnings management as well.

According to the information asymmetry theory, on the one hand, a company’s participation in the ETS provides external stakeholders with more internal information about the company [58], aiding financial institutions and investors in assessing the firm’s risk and investment value. On the other hand, participation in the ETS draws greater public attention, particularly from analysts who possess informational resources and expertise [59]. As information intermediaries, analysts track and interpret corporate information, relaying it to external stakeholders through reports or earnings forecasts, thereby reducing the level of information asymmetry between company management and external stakeholders [60]. The lower the information asymmetry and the higher the transparency of a company, the greater the risk of earnings management being detected by management, which diminishes their motivation to engage in such practices [61]. Consequently, the quality of earnings improves.

4.2.3. Mechanism—Market Competition

In a perfectly competitive market, information is completely unimpeded, and investors and other stakeholders can easily obtain related information. There are few barriers for enterprises to enter or exit, and both buyers and sellers have the ability to fully obtain market information. The profit level among firms in the same industry tends to be consistent, and firms’ earnings can fully reflect their operating status. In such a competitive market, earnings management is prohibited due to decreased information asymmetry. We build on Peress’s findings [62], which demonstrate that companies operating in less concentrated industries (low competition) exhibit reduced volatility in cash flow and profit when considering external environmental regulations. Market competition acts as a powerful external regulatory tool that enhances the quality of information disclosure. As suggested by the structure-conduct-performance (SCP) theory, companies operating in less competitive markets tend to reap higher profits and have greater opportunities for earnings manipulation. Since profit is more transparent within firms in a more competitive setting, the space for earnings management is narrowed. If this argument holds, the possibility of listed companies engaging in earnings management will be reduced. We therefore expect a larger policy effect of the CO₂ ETS on earnings management in firms with lower market competition (Hypothesis 4). We use three proxies for market competition: HHI (the first proxy for market competition is the HHI index, which depicts the concentration of industries to reflect the dispersion and concentration of market size. A higher HHI index implies a more concentrated market. The calculation equation is: $HHI=\sum _{i=1}^n {\left({\displaystyle \frac{x_i}{x}}\right)}^2=\sum _{i=1}^n S_i^2$. Where $S_i^2$ refers to the square of the market share of individual firms. When HHI is equal to 1, the market is a monopoly); Number of Listed Firms in each industry (the second proxy for market competition is the number of listed companies within the industry. We follow the idea of [63,64,65]. We take the logarithm of the number of listed firms within the same industry (LNN), with a larger LNN implying a more competitive market); and Lerner index (the Lerner index is the third proxy for market competition. It assesses a company’s market power by comparing its price to its marginal cost. The Lerner index ranges from 0 to 1: a value near 0 indicates a market situation closer to perfect competition, while a value near 1 suggests a higher level of market power, approaching monopoly conditions).

Table 7. Mechanism—market competition.

	(1)	(2)	(3)	(4)	(5)	(6)
	Discretional Accruals	Discretional Accruals	Discretional Accruals	Real Earnings Management	Real Earnings Management	Real Earnings Management
Treat × Post	0.031 ***	0.030 ***	0.017 **	0.091 ***	0.094 ***	0.065 ***
	(0.007)	(0.007)	(0.008)	(0.015)	(0.016)	(0.018)
HHI	0.006			0.023
	(0.011)			(0.027)
HHI × Treat × Post	0.189 **			0.344 ***
	(0.076)			(0.130)
Listed Firms		0.003			0.006
		(0.004)			(0.008)
Listed Firms × Treat × Post		−0.014 * (0.008)			−0.033 ** (0.015)
Lerner Index			0.020 **			0.054 ***
			(0.008)			(0.019)
Lerner Index × Treat × Post			0.101 * (0.053)			0.193 * (0.101)
Controls	Yes	Yes	Yes	Yes	Yes	Yes
Firm	Yes	Yes	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes	Yes	Yes
N	10,226	10,226	10,226	10,226	10,226	10,226
R-squared	0.031	0.030	0.033	0.045	0.045	0.049

Notes: Standard errors in parentheses * p < 0.1, ** p < 0.05, *** p < 0.01.

displays our findings. This table shows how market competition affects the relationship between the CO₂ ETS and earnings management, using the PSM sample. We use the Herfindahl–Hirschman Index, the count of listed companies within the same industry, and the Lerner index to represent market competition levels. The coefficients of the interaction terms between these three measures and Treat × Post reflect the competitive dynamics. The higher the HHI index, the lower the number of listed firms, and higher Lerner index represents a more concentrated market. The direction of the coefficients of the interaction terms meets our original expectation and all are significant. Long et al. [52] suggest that companies with lower cost pass-through capabilities are more significantly impacted by ETSs. They hypothesize that firms operating in more competitive environments have poorer cost pass-through abilities, which is contrary to our research findings.

4.2.4. Mechanism—Emission Mitigation Pressure

In this paper, we assess the pressure on companies to reduce emissions by measuring their firm-level carbon dioxide emissions. The research sample is categorized into two groups, those with high and low emission levels, determined by the median CO₂ emissions of the companies in the test.

Table 8. Mechanism—emission mitigation pressure.

Panel A	CO2 Emissions at Firm Level
	≥Median		<Median
	Discretional Accruals	Real Earnings Management	Discretional Accruals	Real Earnings Management
Treat × Post	0.025 ***	0.088 ***	0.006	0.032
	(0.009)	(0.023)	(0.006)	(0.022)
Controls	Yes	Yes	Yes	Yes
Firms	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes
N	4968	4968	5258	5258
R-Squared	0.032	0.046	0.041	0.060
Panel B	Level of Carbon Market Activity
	High		Low
	Discretional Accruals	Real Earnings Management	Discretional Accruals	Real Earnings Management
Treat × Post	0.018 *	0.096 ***	0.034 ***	0.027
	(0.010)	(0.021)	(0.012)	(0.043)
Controls	Yes	Yes	Yes	Yes
Firms	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes
N	2426	2426	1762	1762
R-Squared	0.044	0.077	0.065	0.106
Panel C	Status of Carbon Peaking at Province Level
	Peaked		On the way
	Discretional Accruals	Real Earnings Management	Discretional Accruals	Real Earnings Management
Treat × Post	0.013	0.045	0.023 ***	0.090 ***
	(0.016)	(0.045)	(0.006)	(0.016)
Controls	Yes	Yes	Yes	Yes
Firms	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes
N	875	875	9351	9351
R-Squared	0.075	0.172	0.032	0.040

Notes: Standard errors in parentheses * p < 0.1, *** p < 0.01.

shows the regression results, with Panel A illustrating a positive correlation between higher pollution levels and earnings management. To ensure the robustness of our findings, we also conduct alternative tests using carbon market activity levels and the carbon peaking mission completion status at the provincial level. We acquired the carbon peaking mission completion status from the Institute of Public and Environmental Affairs (IPE) and categorized our sample based on whether the areas have peaked or not. Low carbon market activity levels and non-peaking carbon peaking both indicate that the carbon market is not as active as expected.

In Panels B and C, it is apparent that, in areas experiencing high mitigation pressure (where carbon market activity is low and carbon peaking is ongoing), the inverse relationship between the cross-multiplier Treat × Post and the extent of corporate earnings management (Discretional Accruals and Real Earnings Management) is stronger compared to regions with lower reduction pressure. This indicates that, in regions facing greater emission reduction demands, the impact of the ETS on the escalation of corporate earnings management is more pronounced.

Our results successfully validate Hypothesis 5. With the implementation of the ETS, the carbon risk that companies face varies depending on their carbon intensity [66]. High carbon-intensive firms, which are heavily reliant on carbon-intensive fossil fuels such as coal and lignite, have a limited ability to switch to low-carbon alternatives. Consequently, they face greater carbon risk compared to low carbon-intensive firms due to carbon price volatility. This heightened pressure makes high carbon-intensive companies more inclined to manage earnings in order to avoid cash flow uncertainty and to maintain relationships with investors and creditors. Therefore, they are more likely to engage in earnings management than other firms [52].

5. Conclusions and Policy Implications

Using CO₂ emissions trading system (ETS) implementation as a quasi-experiment, this study analyzes its impact on earnings management practices among firms in China from 2012 to 2021. The findings indicate that: (1) Participation in the ETS encourages companies to engage in earnings management. Environmental regulations internalize the externalities of pollution, leading to increased costs and greater carbon risk for companies. To maintain investor enthusiasm and protect their own compensation incentives, managers may resort to earnings management. This aligns with the idea that increased environmental regulations amplify external public pressure on firms. (2) The CO₂ ETS will have a greater impact on companies facing stronger financial constraints, intensifying earnings management efforts to secure more financing opportunities. (3) The CO₂ ETS mandates that companies disclose detailed information about their emissions and operations, which may reduce their opacity. The CO₂ ETS is likely to have a more significant impact on companies with higher levels of information opacity. (4) Companies operating in less competitive markets tend to achieve higher profits and have more opportunities for earnings manipulation. The policy effects of the CO₂ ETS on earnings management are likely to be more pronounced for companies in less competitive markets. (5) In regions facing greater emission reduction pressure, the impact of the ETS on the escalation of corporate earnings management is more significant.

Our research provides valuable insights into both the fields of environmental policy and earnings management. It illuminates the unintended consequences of China’s highly anticipated CO₂ emissions trading program by examining firms’ earnings management. Our research shows that China’s CO₂ emissions trading system acts as a motivator for companies to practice earnings management. Based on the mechanism analysis, we propose two recommendations: (1) Capital market regulators should focus on firms’ actual business activities. While the carbon emissions trading mechanism plays a positive role in environmental improvement, it also carries the potential risk of exacerbating firms’ earnings management behavior. Therefore, regulators should consider formulating more integrated and comprehensive disclosure requirements that encompass corporate environmental behavior and pollution emission data. (2) Optimization of the external environment for enterprises. The carbon emissions trading system is closely linked with various external governance frameworks. This research indicates that the system has a stronger impact in regions with lower emission reduction pressures and is less significant in highly competitive product markets. Therefore, it is advisable for the government and relevant authorities to strengthen legal regulations, enhance the legal environment, develop a more comprehensive market-based factor allocation system, utilize the institutional environment to better supervise enterprises, address agency issues and information manipulation by companies, and improve investor protection and company information disclosure.

We anticipate that our findings will have a profound impact on the development of a unified national carbon trading market and will offer valuable insight for other countries exploring effective CO₂ emission reduction systems. Additionally, our findings provide empirical support for policymakers to prioritize the actual operational and financial behavior of firms when pursuing peak carbon and carbon-neutral commitments. Such emphasis aims to encourage sustainable corporate growth without compromising the external environment.