Net zero target: implications for real earnings management

The adverse impact of climate change and extreme weather events has been well-documented in the literature (see, e.g. Do et al., 2021; Huynh et al., 2020; Shan and Gong, 2012; Truong et al., 2020) ^[1]. Climate change is a major global risk challenge due to its widespread and interconnected impacts on the environment, economy, health and security. In mitigating the adverse impact of climate change on human well-being and the environment, the Paris Agreement (COP21) proposed a 45% decrease in CO₂ emissions by 2030 and a net zero carbon emissions target, hereafter referred to as a net zero target by 2050 (United Nations, 2022). Given that the net zero target has become a global rallying cry and requires substantial financial support, it is crucial to investigate the economic outcomes and/or consequences for firms that commit to moving toward the net zero targets. As such, this study examines the real earnings management (REM) among firms subjected to regulatory changes associated with countries committing to a net zero target in which firms are headquartered.

From a theoretical perspective, there are three mechanisms through which national-level net zero commitments can influence managers’ REM decisions, namely, regulatory uncertainty and anticipated compliance costs, legitimacy pressures and stakeholder scrutiny and capital market expectations and transition risk pricing. First, national commitment to the net zero targets should signal to firms the likelihood of more stringent regulatory measures (e.g. carbon pricing, mandatory disclosure or industry-specific transition plans). This, in turn, creates uncertainty about future operating costs and performance benchmarks, and such uncertainty incentivizes managers to use REM to present a smoother earnings trajectory (Francis et al., 2004; Ilhan et al., 2023). Second, firms headquartered in countries with formal net zero targets should face heightened legitimacy threats if they are perceived as lagging in climate preparedness (Suchman, 1995; Patten, 2002). Managers may respond by engaging in REM to demonstrate short-term resilience, particularly in carbon-intensive sectors. Finally, there is strong evidence that investors in global capital markets increasingly factor national climate commitments into risk assessments and capital allocation (Bolton and Kacperczyk, 2021; Krueger et al., 2020). Thus, managers may engage in REM to meet heightened performance expectations under climate-constrained scenarios.

There is ample evidence that eliminating excessive carbon emissions requires substantial costs and significant funding, as operational transformations are necessary. It is crucial to consider how firms operating in countries that have committed to a net zero target would finance the operational changes needed to align with the target reduction in carbon emissions. From a managerial perspective, countries exert significant pressure on firms to contribute to moving toward the net zero targets ^[2]. Therefore, in transitioning to the net zero targets, managers of firms headquartered in countries that have committed to a net zero target could engage in REM.

REM activities provide ideal research setting for our empirical examination as they are relatively easier for managers to manage. Managers can use their operational and long-term investment discretion to engage in REM activities at any time during the financial year. Prior studies find that managers could use REM activities to provide cash savings, and simultaneously improve the firm’s current earnings outlook as a result of the reduction in capital expenditures or expenses related to those projects (Baber et al., 1991; Bushee, 1998; Graham et al., 2005). This cash availability could be used to finance investments in new technologies and infrastructure systems in the coming years that allow firms to reduce their carbon emissions ^[3]. In the context of net zero targets, managers could also engage in REM activities by reducing current capital expenditures and R&D that are less likely to have a direct effect on reducing carbon emissions and achieving net zero targets. These REM activities could provide cash savings, which may be used to finance new technologies and infrastructure systems in preparing firms to commit to a net zero target.

Based on the above notions, we investigate whether firms headquartered in countries that have committed to a net zero target engage more in REM. We predict that in achieving both a net zero target and a desired earnings outlook, firms headquartered in countries that have committed to a net zero target exhibit higher levels of REM. In addition, we also investigate whether corporate social responsibility (CSR) performance significantly affects the association between REM and the commitment to move toward a net zero target. The effect of CSR in relation to the net zero targets is particularly important given the growing concerns about firms’ sustainability (Mackey et al., 2007; Margolis and Walsh, 2003). The relation between CSR performance and REM is not intuitively straightforward. On the one hand, prior studies find that good CSR enhances a firm’s performance, thereby providing less incentive for managers to manage earnings via REM (Hong and Andersen, 2011; Kim et al., 2012; Scholtens and Kang, 2013). Conversely, other studies find that firms require substantial funds to enhance their CSR performance through various activities, thereby inducing managers to engage in more earnings management via REM (Boukattaya and Halaoua, 2022; Kim et al., 2019; Kuo et al., 2021). Due to the inconclusive findings on the association between CSR performance and REM, we conjecture that CSR performance has no significant effect on the association between firms’ REM and net zero target commitment.

To test these conjectures, we use a sample of firms listed on stock exchanges across 71 countries from 2011 to 2021, for which financial data are available in the Compustat Global and Worldscope databases. Using a sample of 249,711 firm-year observations, our results indicate that firms headquartered in countries that have committed to the net zero targets engage in more REM activities. Our results are consistent for all four measures of REM activities. In addition, our findings show that this relation is weaker in firms with better CSR, which is an indication of a firm’s readiness and willingness to align with the objectives associated with the net zero target commitment. The results are robust to alternative measures of REM and employment of the propensity score matching (PSM) methodology. We also use four levels of net zero target commitments to determine the intensity at the country level and find REM increases with an increase in the intensity of the net zero target commitment. Finally, we find that firms that voluntarily commit to the net zero targets do not exhibit higher REM, consistent with the notion that such firms are better prepared to respond to the regulatory changes imposed by the country where they are headquartered.

Our study makes several important contributions to the literature and practice in the following ways. First, our study contributes to the growing literature on how macrolevel climate policy can have microlevel reporting consequences on firms. While prior research has examined the effects of climate regulation on investment behavior (D’Arcangelo et al., 2023; Grewal et al., 2019), disclosure (Christensen et al., 2021), corporate innovation (Liu et al., 2024) or financing activities (Nguyen et al., 2025), to the best of our knowledge no study has directly linked national net zero target commitment, a cornerstone of global climate governance, to managers’ REM decisions. This unique linkage allows us to provide new insights into how global climate commitments shape firm-level financial reporting incentives.

REM is particularly salient in the context of net zero targets because, unlike accrual-based earnings management, it operates through real business decisions such as reducing R&D, delaying investments or cutting discretionary spending. Focusing on REM, we uncover a subtle but economically significant channel through which national climate commitments may unintentionally incentivize short-term profit preservation at the expense of long-term sustainability. This unique perspective allows us to highlight the tension between global net zero targets and firm-level financial reporting incentives, offering insights into microlevel externalities of climate macro policies.

Second, our study provides evidence of important differentiation between national-level and firm-level commitments in the context of net zero targets. Existing work often focuses on firm voluntary initiatives (e.g. voluntary carbon disclosure, internal carbon pricing or CSR commitments). Our paper highlights how country-level regulatory pressure cascades down to firm reporting practices, even in the absence of immediate enforcement. This is a novel angle that bridges the macrolevel policy literature with the microlevel accounting and the REM literature. In addition, we also contribute methodologically by constructing the intensity variable that captures not only whether a country has committed to a net zero target but also the strength of its commitment (law, policy, pledge, or proposal). This allows us to empirically differentiate between “symbolic” commitment and “substantive” commitment, providing richer evidence than a binary treatment indicator often used in the literature.

Third, our study also contributes to the CSR literature by showing that firms with higher CSR performance are in a better position to deal with the financial challenges as a result of the regulatory shocks imposed by governments in meeting the net zero targets, leading to a reduced likelihood of REM activities. Our findings demonstrate to corporate stakeholders and activists that higher levels of CSR performance play an important role in mitigating the impact of a country’s net zero target commitment on firms’ REM activities.

Finally, our findings are informative to firms’ insiders, investors and regulators in relation to the potential risks that arise due to managers’ discretionary choice in using REM as a strategy to achieve both the net zero target preparations and the desired earnings outlook. More importantly, the findings provide insights to investors and analysts in helping them to assess the true value of firms. In particular, many countries are now racing to achieve a net zero target, which may cause economic trade-offs between achieving the net zero target and the financial implications of the regulatory changes imposed on corporate entities.

Prior studies (Roychowdhury, 2006; Dechow and Sloan, 1991; Baber et al., 1991; Bushee, 1998) have demonstrated how opportunistic managers could use their discretion to engage in REM in maximizing their utility. In the context of our study, REM activities could help firms achieve the transformation required to support the net zero targets in the countries where they are headquartered, while satisfying a desired earnings outlook. However, from the perspective of investors, higher levels of firms’ REM activities could be value-destructive in the long run. Hence, our findings inform stakeholders, particularly investors, so they can exercise caution to closely monitor and evaluate how those firms that are committed to the net zero targets mitigate the negative impact of their REM activities.

The remainder of the paper is structured as follows. Section 2 discusses the literature related to the development of hypotheses. Section 3 presents the research design and data in this study, including the sample selection process, all variables and regression models. The primary empirical results are presented in Section 4, followed by robustness checks and additional analyses in Section 5. Section 6 concludes this study.

In 2015, 196 sovereign nations signed the Paris Agreement (COP21) with the objective of limiting the rise of global temperature to 1.5°C. This is a substantial challenge for countries to meet the Paris Agreement’s target because they need to reduce CO₂ emissions by 45% by 2030 and achieve the net zero target by 2050 (United Nations, 2022). The Energy & Climate Intelligence Unit (2022) reports that 126 global countries and regions, excluding the USA, have committed to achieving the net zero target with varying years and levels of commitment. The variations include enshrined in law, outlined in official policy documents, declared or pledged and under discussion or proposed. To achieve a net zero target, countries are required to transform their economy and industry, which involves collaboration from various sectors (PwC Global, 2022).

The heightened policies in achieving the net zero targets could result in firms with regulatory exposure facing downside risks, such as additional capital expenditures and expenses. However, firms with exposure opportunities may enjoy upside potential benefits from innovation activities (e.g. renewable energy and energy storage) (Sautner et al., 2023). Stern and Valero (2021) indicate that achieving the net zero transition requires significant investments in clean innovation, which is accompanied by investments in complementary assets, including sustainable infrastructure, and human, natural and social capital. According to an estimate from the World Economic Forum (2021), approximately $50tn in incremental investments are needed in breakthrough technologies by 2050, such as solar technology, decarbonization, energy efficiency and carbon capture solutions, among others.

From the perspective of regulatory shocks related to climate change, achieving the Paris goals requires a series of intensified policies. For example, carbon pricing in the form of an emissions trading system (ETS) or carbon tax has been widely regarded as the most cost-effective and flexible method to achieve carbon emission reduction in line with the objectives of the Paris Agreement (The World Bank, 2022; UNCC, 2022). Studies in the literature have shown that carbon trading increases firm costs. For example, Goulder et al. (2010) document that a cap-and-trade system increases the marginal cost of production. Cohen et al. (2020) found that firms with high carbon footprints and higher regulatory risks create more and better green technologies than others. Shell plc (2022, p. 17) stated that “developing our low-carbon products and services is subject to challenges which could have a material adverse effect on our earnings, cash flows and financial condition.” Another study by Long et al. (2023) reveals that the implementation of ETS in China increases compliance costs and carbon risks, thereby motivating firms to engage in earnings management.

While prior work on REM has largely examined financial reporting incentives in response to capital market pressures or regulatory oversight (Gunny, 2010; Roychowdhury, 2006; Zang, 2012) and sustainability research has focused on how CSR, ESG regulation or environmental policies influence disclosure quality, transparency and investment behavior (Christensen et al., 2021; Grewal et al., 2019), our study is the first to connect national net zero climate targets with firms’ REM decisions ^[4]. In transitioning to the net zero targets, we conjecture that firms headquartered in countries that have committed to a net zero target are more likely to invest in new technologies and infrastructure systems in the coming years that could help firms to reduce the carbon emissions.

Firms are more likely to face increasing legitimacy and institutional threats and pressures if they are perceived as lagging in climate preparedness (see, e.g. Cho and Patten, 2007; DiMaggio and Powell, 1983; Suchman, 1995; Patten, 2002). These threats and pressures could lead firms to commit to the net zero targets. The transition to the net zero targets may induce managers to act opportunistically to manage earnings using REM by reducing current capital expenditures and R&D that are less likely to have a direct effect on reducing carbon emissions in achieving the net zero targets, with the expectation that future investments under a net zero framework will be costlier for firms ^[5]. As a result, this REM strategy could enable firms to increase their current earnings and cash availability. The cash availability due to cash savings as a result of REM could be used to finance investments in new technologies and infrastructure systems in the coming years that allow firms to reduce their carbon emissions in the future. In this case, REM could be used by managers to prepare firms for moving toward the net zero targets while helping managers to record higher current earnings ^[6].

Collectively, based on the above discussions, we predict that firms are more likely to manage earnings through REM in transitioning to net zero targets. Hence, this study develops the following testable hypothesis:

Carroll (1979) conceptualized corporate social performance as encompassing economic, ethical, legal and discretionary aspects of business performance, forming the four basic expectations of CSR. Connelly et al. (2011) argued that managers can adopt sustainable practices to signal corporate trustworthiness and prospects to stakeholders. Margolis and Walsh (2003) proposed that firms considering the interests of all stakeholders are likely to engage in nonprofit maximizing CSR activities (Mackey et al., 2007). Consistent with Carroll’s (1979) model, firms that move CSR tend to reduce or avoid earnings management through discretionary accruals and real activities manipulations (Hong and Andersen, 2011; Kim et al., 2012). Scholtens and Kang (2013) find a negative relationship between CSR and earnings management, measured by earnings smoothing and earnings aggressiveness, in an Asian context. These findings suggest that CSR-oriented firms meet ethical expectations and avoid opportunistic behavior by providing more transparent information and high-quality reporting.

However, some studies have documented a positive association between CSR and REM activities. For example, Kuo et al. (2021) suggested that CSR practices are positively associated with REM in firms where CEOs have higher shareholdings. A possible explanation for this finding is that higher shareholdings lead to stronger CEO alignment with the firm, incentivizing CEOs to engage in REM activities to enhance CSR performance and improve the firm’s reputation. Kim et al. (2019) found that socially responsible firms with state control and in regions with higher institutional development engage more in REM activities due to increased legitimacy risk.

Given the significant funds required for operational transformations for moving toward the net zero targets, firms are motivated to manage real activities that directly impact cash flows to increase earnings, such as accelerating sales revenue and delaying R&D and capital outlays to reduce expenses. In addition, managers may need significant funds to enhance the firm’s CSR performance, thereby inducing them to engage in more REM to report better earnings and cash flow (Boukattaya and Halaoua, 2022; Kim et al., 2019; Kuo et al., 2021). Consequently, it is plausible that firms with higher CSR performance are likely to engage in more REM for moving toward to the net zero transformation.

The discussion above suggests that higher CSR could either accelerate or mitigate the positive relationship between the commitment to the net zero targets and REM. Given the inconclusive findings on the association between CSR performance and REM, we conjecture that CSR performance has no significant effect on the association between firms’ REM and net zero targets. As such, we formulate the following nondirectional H2:

Panel A of Table 1 presents the sample selection process by country, which begins with an initial sample of 195 countries sourced from Nationsline. Subsequently, we remove observations related to 33 countries with missing data in the WorldScope and Compustat Global database and 35 countries without information on net zero target commitment. In addition, we exclude the USA from the main sample in our analysis to avoid potential dominance of results, given its significant market size and number of observations in the global study. As a result, the selection process yields a final sample of 126 countries. After excluding all missing observations related to the measures of dependent and control variables, the main sample comprises 71 unique countries.

Panel B of Table 1 details the firm-level sample selection process for the REM model and the related control variables. The initial sample consists of all publicly listed firms with available financial data in the WorldScope and Compustat Global Database from 2011 to 2021. Starting the study in 2011 allows for the exclusion of the impact of the global financial crisis. The initial sample comprises 650,748 firm-year observations. Among these, 368,575 observations are excluded due to missing data for the REM model and control variables. In addition, we exclude 32,462 observations related to firms in the financial sector. Consequently, the selection process results in a final sample of 249,711 firm-year observations for the test of H1 and 28,685 firm-year observations for the test of H2, which examines the effect of CSR performance ^[7].

Panel C of Table 1 displays the sample distribution by industry, country and year. The industry distribution is categorized based on the Fama-French 48 industries, with the top 10 industries being explicitly listed, and all other industries combined under the label “All others.” Among the top 10 industries, 11.18% of the sample belongs to business services, 5.03% to machinery and 4.99% are chemical firms. In addition, 4.73% of the sample is composed of food product companies.

Regarding the country distribution, 16.49% of the sample firms are headquartered in China, while 14.04% and 13.26% of firm-year observations are from India and Japan, respectively. Firms from Australia account for 5.58% of the sample, and South Korea contributes 5.26% of the total sample. Panel C of Table 1 also presents the yearly distribution of the sample, indicating the percentage of observations for each year from 2011 (7.99%) to 2021 (9.71%). This demonstrates the gradual increase in observations over the period of the study.

The Energy & Climate Intelligence Unit (2022) provides data on the net zero target commitments for 126 countries. We manually collect the information on the net zero target commitment by each country from the websites of the Energy & Climate Intelligence Unit and Climate Home News. Climate Home News (2019b) offered the year of net zero target commitment for most countries in the sample, while for the remaining countries, the information is gathered through a manual search on their respective websites. The net zero target commitment varies in terms of the year of commitment and the level of commitment, which includes:

• in law;

• in official policy document;

• in declaration/pledge; and

• in discussion/proposed.

To test whether firms headquartered in countries committed to the net zero targets are more likely to engage in REM activities, we estimate the following regression model for H1:

where the subscripts i, j, c and t denote firm, industry, country and year, respectively. The regression model includes industry, country and year fixed effects, and uses appropriate indicator variables (INDUSTRY, COUNTRY and YEAR). We cluster the standard errors at the firm level. To test the impact of the net zero target commitment of a country on REM, this study uses a variable of interest, COUNTRY_NETZERO, a categorical variable coded 1 if the country where the firm is headquartered committed to a net zero target in that fiscal year, and 0 otherwise.

Next, we introduce an interaction variable of COUNTRY_NETZERO*CSR_i,t in the model to examine the incremental effect of a firm’s CSR performance on the relation between COUNTRY_NETZERO_i,t and REM. The following regression model is estimated to test H2:

CSR score is available in the Refinitiv database and comprises individual components related to environmental, social and corporate governance performance, and the overall CSR score. The variable of interest used is the interaction of country-level net zero target commitment and CSR performance (COUNTRY_NETZERO*CSR), to test for the incremental effect of CSR on the relation between country-level net zero target commitment and REM.

Because a country’s net zero target commitment is a regulatory shock for firms headquartered in that country, the setting of this study is less likely to be affected by endogeneity concerns. However, firms can voluntarily commit to the net zero targets. To address this concern, several additional tests are conducted, which include PSM and analysis based on firms’ commitment to the net zero targets.

We use four measures of REM following prior literature (e.g. Achleitner et al., 2014; Cohen and Zarowin, 2010; Garg, 2018; Ghaleb et al., 2020; Paredes and Wheatley, 2017; Roychowdhury, 2006). These four measures include abnormal cash flow from operations (REM_CFO), abnormal discretionary expenses (REM_DISEXP), abnormal production costs (REM_PRODCOST) and the sum of the three aforementioned measures (REM_SUM).

The first measure of REM is abnormal operating cash flow. To estimate this measure, the following cross-sectional regression (3) for each industry (based on Fama-French 48 industries) and year to capture the normal level of operating cash flows is used. The abnormal cash flow from operations for year t (REM_CFO_t) is the residual from regression (3):

where CFO_t is the cash flow from operations at the end of period t;

TA_t is the total assets at the end of period t; S_t is the sales during period t; and

The second measure of REM is abnormal discretionary expenses. We run cross-sectional regression (4) for each industry (based on Fama-French 48 industries) and year. The abnormal discretionary expenses for year t (REM_DISEXP_t) are the residuals from regression (4):

where DISEXP_t is the sum of R&D expenses, advertising expenses,and SG&A expenses at the end of period t; and. Other variables are as defined in Supplementary Material A.

To obtain the third measure of REM abnormal production costs, we run regression (5) for each industry (based on Fama-French 48 industries) and year, to obtain the normal production costs. The abnormal production costs (REM_PRODCOST_t) are the residuals from regression (5):

where PROD_t is the sum of the cost of goods sold and the change in inventories at the end of period t. Other variables are as defined in Supplementary Material A.

Finally, we use the sum of the three proxies of REM above (REM_SUM) as the fourth measure. Considering the advantages and disadvantages of the REM measures, we test H1 and H2 using each of the four measures of REM as the variable of interest in separate regression models.

Our main independent variable of interest is COUNTRY_NETZERO, which represents the net zero target commitment at the country level. This variable is coded one if a country has committed to net zero targets either in law, or in an official policy document, or is in declaration/pledge, or proposed/in discussion, and zero otherwise.

In addition, we measure the intensity of the net zero target commitment based on the level of commitment. Our main underlying assumption is that the intensity of the regulatory change related to the net zero target commitment is higher for countries that have enacted the net zero targets in law compared to those in declaration/pledge. To operationalize this concept, we first code the four commitment levels as 4 (in law), 3 (policy document), 2 (declaration/pledge) and 1 (proposed/in discussion). Supplementary Material B provides a detailed explanation of the coding of the intensity variable. Next, we use the square of these scores (i.e. 16, 9, 4, 1 for in law, policy document, declaration/pledge and proposed/in discussion, respectively) of commitment as a proxy for the intensity of net zero target commitment on the firm (INTENSITY_NETZERO_SQUARE) ^[8].

The country-level financial control variables include GDP_GROWTH and GDP_PER_CAPITA to control the level of economic development. In addition, we include country-level corporate tax rate (CTR) to control for corporate tax rate legislated in each of the countries, as a higher (lower) statutory corporate tax rate is projected to have more (less) negative impact on the cash flows for firms headquartered in countries that commit to the net zero targets. Following Montenegro (2021), we include a number of national governance indicators such as corruption control (CORRUPT_CONTROL), government effectiveness (GOV_EFFECT), political stability and absence of violence/terrorism (POL_STABILITY), regulatory quality (REGULATE_QUAL), rule of law (RULE_LAW) and voice and accountability (VOICE_ACCOUNT). The models also control for climate risk (CLIMATE_RISK) and CO₂ emission per capita (CO2_PER_CAPITA). To control for the effect of COVID-19, we use a categorical variable COVID19, coded 1 if the financial year of observation is 2020 or 2021, and 0 otherwise. All variables are defined in Supplementary Material A.

We control the regression model with firm-level variables. Following Achleitner et al. (2014) and Ahmed et al. (2022), we include the log of total assets (SIZE), leverage (LEV) and return on assets (ROA) as control variables. We also include other control variables that may affect the association between the independent variables of interest and the dependent variables such as cash and cash equivalents (CASH), intangible assets (INTANG), whether a firm has foreign exchange transactions (FOREIGN), property, plant and equipment (PPE), research and development expense (R&D), selling, general and administrative expenses (SGA), and capital expenditures (CAPEX). We define all these variables in Supplementary Material A.

Table 2 presents the descriptive statistics for the REM model based on the full sample. The mean value for country-level target commitment (COUNTRY_NETZERO) is 0.15, suggesting that 15% of observations in the sample are headquartered in a country that committed to a net zero target. The mean (median) value for REM_CFO is 0.142 (0.077), for REM_DISEXP is 0.146 (0.080) and for REM_PRODCOST is 0.138 (0.089). Furthermore, the mean (median) value of REM_SUM, the sum of three individual measures of REM, is 0.427 (0.299) ^[9].

Table 3 presents the Pearson correlation matrix based on the REM model. COUNTRY_NETZERO is significantly positively correlated with REM_DISEXP (coefficient = 0.06, p-value < 0.001), REM_PRODCOST (coefficient = 0.02, p-value < 0.001) and REM_SUM (coefficient = 0.02, p-value < 0.001).

Table 4 reports the regression results on the relation between net zero target commitment and REM. The results show that COUNTRY_NETZERO is positively associated with all four measures of REM and is statistically significant at the 1% level (REM_CFO: coefficient = 0.014, p-value < 0.01; REM_DISEXP: coefficient = 0.011, p-value < 0.01; REM_PRODCOST: coefficient = 0.007, p-value < 0.01; REM_SUM: coefficient = 0.032, p-value < 0.01) ^[10]. Collectively, the results provide support for H1 that firms headquartered in countries with the net zero target commitment have a higher likelihood of managing earnings through real activities ^[11]. The findings suggest that in transitioning toward the net zero targets, firms’ managers record higher earnings via REM by reducing current capital expenditures and/or R&D, which has less impact on reducing carbon emissions. This REM strategy has provided potential for higher cash availability in preparing firms to invest in new technologies and infrastructure systems in the coming years, which would help firms to reduce carbon emissions in the future in meeting the net zero targets.

In relation to the control variables, the results indicate that larger firms (SIZE) and firms with more property, plant and equipment (PPE) and intangible assets (INTANG) have a lower possibility of engaging in REM. In addition, firms with higher leverage (LEV), cash (CASH), research and development expenses (R&D) and capital expenditures (CAPEX) exhibit more REM^[12].

Table 5 presents the regression results of H2, which examines the effect of a firm’s CSR on the interaction between the net zero target commitment and REM activities. The coefficient estimates on the interaction between country-level net zero target commitment and CSR (COUNTRY_NETZERO*CSR) are negative and statistically significant for three of the four measures of REM (REM_CFO: coefficient = −0.034, p-value < 0.01; REM_DISEXP: coefficient = −0.032, p-value < 0.01; REM_SUM: coefficient = −0.071, p-value < 0.01).

The findings from this analysis suggest that the CSR performance of firms plays a significant role in mitigating the positive association between the net zero target commitment and REM activities. In other words, firms with higher CSR performance are better equipped to handle the regulatory shock of net zero targets in their home country, leading to a reduced likelihood of engaging in REM practices.

The comprehensive CSR score comprises three key components of performance: environmental, social and corporate governance. Each of these measures, both individually and collectively, contributes to the firm’s readiness and willingness to align with the objectives associated with the net zero target commitment. By demonstrating a commitment to sustainability and responsible business practices, these socially responsible firms are better prepared to face the challenges posed by the net zero target commitment and are less inclined to resort to REM as a coping mechanism ^[13].

Overall, our results are consistent with the theoretical framework developed by Carroll (1979), which posits that firms address their obligations to society through economic, legal, ethical and discretionary performance. In particular, our findings provide support to prior literature in showing that CSR-oriented firms appear to reduce or avoid earnings management through real activities manipulations.

To address endogeneity concerns, we use the PSM methodology. In the PSM approach, each observation in the subsample from countries with the net zero targets is matched with an observation from countries without the net zero targets, based on their closest propensity score within a caliper of 0.10, considering all control variables. The matched observations from both subsamples are then pooled to create the matched sample.

Table 6 presents the regression results for the matched sample, using the same dependent and independent variables as in the main model. The findings from the PSM sample, which includes 65,037 observations, are consistent with the main results. Specifically, COUNTRY_NETZERO is positively associated with all four measures of REM and is statistically significant when REM_CFO is used as the dependent variable (coefficient = 0.031). Similarly, for REM_DISEXP, REM_PRODCOST and REM_SUM, the coefficients are 0.022, 0.017 and 0.074, respectively and are statistically significant at the 1% level. Overall, the results from the PSM analysis provide further support to the main findings, reinforcing the positive association between the commitment to the net zero targets by countries and firms’ engagement in REM activities.