Accrual vs. Real Earnings Management in Internationally Diversified Firms: The Role of Institutional Supervision

Abstract

This study investigates whether internationally diversified firms substitute between accrual-based and real earnings management and examines how institutional supervision moderates this relationship. Drawing on a sample of Taiwanese firms listed on the Taiwan Stock Exchange from 2003 to 2016, we conduct regression analyses to test our hypothesis. We find that internationally diversified firms actively shift between accrual and real earnings management strategies depending on the constraints they face. Specifically, firms tend to rely more on accrual-based manipulation when information asymmetry is high and switch to real earnings management when accruals are more easily detected. We also show that stronger institutional supervision—measured by information transparency and investor protection—significantly curbs accrual-based earnings management. These findings reflect the higher volatility and agency problems associated with international operations, such as exposure to foreign risks and the distance between parent and subsidiary firms. By highlighting the conditions under which firms manage earnings and the supervisory mechanisms that constrain such behavior, this study offers practical insights for managers seeking to smooth earnings, investors aiming to evaluate firm transparency, and policymakers designing regulations to deter opportunistic financial reporting.

1. Introduction

In this study, we examine whether internationally diversified firms substitute between accrual-based and real earnings management, and how institutional supervision moderates this relationship. These questions are important because internationally diversified firms are subject to heightened uncertainty (He et al., 2024), including transaction and translation risk exposure (Al-Mallahma, 2025; Hagelin, 2003), varying legal and regulatory environments (Coeurderoy & Murray, 2008), and economic volatility in foreign markets. As a result, their earnings are more volatile than those of domestically focused firms. Additionally, the geographic and operational distance between parent and subsidiary units exacerbates agency conflicts (Doukas & Pantzalis, 2003; Purkayastha et al., 2022), creating strong incentives for earnings manipulation. Managers facing these pressures may strategically choose between accrual-based and real earnings management, depending on the intensity of oversight and reporting constraints. We investigate these issues using data from Taiwanese firms listed on the Taiwan Stock Exchange between 2003 and 2016. Taiwan presents a relevant and informative setting for several reasons. Taiwanese firms are increasingly engaged in international operations and face diverse cross-border risks. At the same time, Taiwan’s institutional environment offers meaningful variation in both investor protection and information transparency, enabling us to examine how these supervisory mechanisms affect firms’ earnings management behavior. This context enables us to observe how external oversight may constrain managerial discretion in the financial reporting process.

In this era of global competition, firms should expand their business operations globally to achieve diversification and gain a competitive edge. Importantly, firms can gain a competitive edge through international diversification, i.e., accessing cheap resources, specialized technology, and new market opportunities (Rugman & Verbeke, 2005; Song, 2014). Interestingly, managers employ distinct methods for manipulating earnings due to their different incentives. In this context, the seminal work of Schipper (1989) classified earnings management as either accrual-based earnings management or real earnings management. Some researchers focus on accrual-based earnings management (Fields et al., 2001), while others examine real activity manipulation (Cohen & Zarowin, 2010; Roychowdhury, 2006; Zang, 2012). Managers consider these two primary earnings management methods as alternatives, making it difficult for auditors and other regulators to detect. Due to the different characteristics of the methods, managers can engage in both methods simultaneously to manage earnings (Cohen & Zarowin, 2010). That is, managers actively shift between accrual-based and real earnings management. Therefore, this study aims to determine whether firms with an incentive to manipulate earnings would resort to real earnings management when accrual-based earnings management is limited. That is, it examines whether there are substitutions between accrual and real earnings management in internationally diversified firms.

More importantly, this study differs from previous studies (for example, Chin et al., 2009; Khanchel El Mehdi & Seboui, 2011) in a manner that examines the moderating role of information asymmetry and investor protection on the relationship between international diversification and earnings management. More specifically, this study examines the supervision mechanisms of international diversification, which may also influence the earnings management behavior of firms that are internationally diversified. Stricter supervision mechanisms limit managers’ opportunities for earnings manipulation. This study classifies supervision mechanisms as (1) investor firms’ outside supervision, whose level is captured by information transparency; and (2) investee firms’ outside supervision, using the investor protection of countries where investee firms are located, to determine the supervision level. Firms with greater transparency limit accrual-based earnings management (Beuselinck et al., 2019; Khanchel El Mehdi & Seboui, 2011; Lobo & Zhou, 2001). If internationally diversified firms disclose more information, investors can effectively supervise foreign operations and prevent managers from manipulating their earnings. In addition, strong investor protection can restrain attempts to manipulate earnings (H. Ali et al., 2022; Leuz et al., 2003).

The main findings of this study suggest that underlying Taiwanese internationally diversified firms utilize both accrual-based and real earnings management as alternative tools for earnings manipulation. Additionally, it is also found that stronger institutional mechanisms help deter accrual-based earnings management, as accruals can be easily detected. Moreover, in case of better information transparency and investor protection, firms engage in real earnings management, as this method is challenging to detect in internationally diversified firms.

This study makes significant contributions to the literature in at least two ways. First, this study aims to capture the shift in internationally diversified firms from accrual earnings management to real earnings management. Second, this study explores the moderating effects of supervisory mechanisms (information asymmetry and investor protection) on the relationship between international diversification and earnings management. The measures used in this study account for both investor and investee firms’ outside supervision, thereby allowing an examination of earnings management behavior under supervision. The findings of this study are helpful for managers, investors, and policymakers. Specifically, investors can assess the likelihood of earnings management in internationally diversified firms. Finally, policymakers can proactively design policies to understand the behavior of internationally diversified firms under weak governance mechanisms related to information quality and investor protection.

The remainder of this paper is organized as follows. Section 2 reviews prior studies and develops the hypotheses. Section 3 describes the methodology and data used in this study. Section 4 reports and discusses the findings of this study. Section 5 concludes this study.

2. Literature Review and Hypotheses Development

2.1. The Earnings Management Methods

Schipper (1989) and K. Ali et al. (2024) proposed two categories of earnings management: accrual-based earnings management and real earnings management. These two earnings management methods have distinct characteristics, and managers may choose different methods to manage earnings in various situations.

2.1.1. Accrual-Based Earnings Management

Accrual-based earnings management increases earnings to meet or beat the earnings targets and expectations of analysts’ forecasts, and only temporarily masks the firm’s actual economic performance (Dechow & Skinner, 2000; Du et al., 2024; Healy & Wahlen, 1999). Usually, companies will employ more accrual earnings management when they have more choices during the natural course of accounting. The use of accrual management is due to the flexibility of accounting rules provided by accounting standards, such as the allowance for discretionary accruals. However, Barton and Simko (2002) suggest that the effects of accounting choices are cumulative and reflected in the balance sheet at the end of the financial year. Although accrual-based earnings management can temporarily influence performance, its reverting effects would limit the ability to manipulate earnings later; that is, companies would be affected by the previously accumulated reverting effects (Abarbanell & Lehavy, 2003; Hunt et al., 1996). It means that earnings can be managed through discretionary accruals up to a certain limit. Accrual-based earnings management is conducted by deferring expenses until a later period or by realizing sales before the actual delivery. Thus, accrual-based earnings management is not accomplished by changing the underlying operating activities but rather through the choices of accounting methods. It does not influence real cash flows while affecting accounting income.

2.1.2. Real Earnings Management

On the other hand, real earnings management involves the manipulation of actual activities, which directly affects cash flow and may cause managers to forgo profits accrued to the firm. Real earnings management occurs through cash flow from operations, discretionary expenses, and production costs (A. M. Habib, 2024). Roychowdhury (2006) defines real earnings management as management activities (involving abnormal levels of cash flow from operations, production costs, and discretionary expenses) that deviate from standard business practices with the objective of meeting or beating earnings targets.

Real earnings management is costly because it reduces a firm’s real economic value (Ewert & Wagenhofer, 2005; A. Habib et al., 2022). For instance, an organization’s production capacity increases beyond optimal levels without planning, which unnecessarily increases the firm’s cost. Since real earnings management involves the manipulation of real activities, cash flow is directly affected, which may cause managers to forgo profits accruing to the firm due to self-beneficial motivations. It can help meet current earnings targets, but negatively affects firms’ future value (Roychowdhury, 2006). Additionally, Cohen and Zarowin (2010) partly support the view that the decline of post-SEO (seasoned equity offerings) operating performance is attributable to real earnings management. Likewise, Leggett et al. (2009) demonstrate that real earnings management is negatively related to subsequent period returns on assets and cash flows from operations.

In addition, Baber et al. (1991), Perry and Grinaker (1994), Roychowdhury (2006), and Cohen and Zarowin (2010) indicate that managers intend to manipulate real activities to meet last year’s earnings or analysts’ forecast benchmarks. Along the same lines, Gunny (2010) also states that managers manipulate real activities to meet earnings targets. For example, real earnings management expedites expenditures unnecessarily for income smoothing.

2.1.3. Choice of Earnings Management Methods

Earnings management methods have varying effects on firms, depending on their specific characteristics. For instance, real earnings management becomes the primary reason for operational inefficiency, value distortion, and negative cash flows. On the other hand, accrual earnings management reduces the quality of earnings and involves audit and regulatory risk because of the loss of trust of investors (Zang, 2012). Therefore, managers choose the optimal method to manage earnings by considering factors such as timing, environment, policies, and their influence on firms.

Roychowdhury (2006), Cohen et al. (2008), and Zang (2012) indicate that, although accrual-based earnings management does not affect real operations or cash flows, it is more likely to raise red flags that increase scrutiny, penalties, and lawsuits. Barton and Simko (2002) highlight that the effects of accrual management are likely to accumulate towards the balance sheet at the end of the year. Additionally, Kreutzfeldt and Wallace (1986) show that accrual items such as accounts receivable, inventory, accounts payable, and fixed assets are easier for auditors to detect malfeasance, which could result in both earnings restatements and lawsuits (Billings et al., 2021; Skinner, 1997). Relative to accrual-based earnings, real earnings management is more difficult to detect for shareholders, SEC regulators, and auditors (Bui, 2024). Similarly, the probability of lawsuits is less likely in real earnings management than in accrual-based earnings management (Elmawazini et al., 2024). Similarly, Chan et al. (2015) indicate that real earnings management is less risky than accrual management.

Thus, managers exhibit a greater propensity for real earnings management when the expected cost of accrual-based earnings management increases (Demski, 2004). Although the consequences of real activities can be economically significant to the firm, managers prefer real earnings management over accrual-based earnings management because the litigation risk associated with real earnings management is lower, given factors such as similarity to everyday business decisions and inherent uncertainty in the business environment. Another prominent reason for engaging in real earnings management is that they are less likely to be detected by auditors and regulators (Graham et al., 2006). On the same notion, Cohen et al. (2008) demonstrate that stricter supervision after Sarbanes-Oxley (SOX) Act of 2002 leads managers to engage in less accrual-based earnings management and resort to real earnings management, which is less likely to draw the attention of auditors or regulators. Likewise, Zang (2012) also suggests that firms will have a greater propensity for real earnings management when facing intense scrutiny and litigation risk.

Ewert and Wagenhofer (2005) demonstrate that real earnings management increases when tightening accounting standards make accrual-based earnings management more difficult. That is, when accounting flexibility is limited, firms tend to resort to real earnings management, and vice versa (Chan et al., 2015; Chi et al., 2011; Cohen et al., 2008). Accounting flexibility means a firm cannot make discretionary accruals more than a limit, because all effects are accumulated at the end of the financial year to the balance sheet of the company (Barton & Simko, 2002). Thus, if managers have an incentive to manage earnings, they would choose accrual-based earnings management rather than real earnings management, which could potentially lead to long-term value destruction (Zang, 2012). The study suggests that managers usually have a greater propensity towards accrual-based earnings management. However, when accounting flexibility is limited, they are more likely to shift to real earnings management. Real earnings management serves as a substitute for accrual-based earnings management and, as a last resort, is employed after accrual-based earnings management. Thus, we propose the following hypothesis:

H1: 

Real earnings management is a substitute for accrual-based earnings management.

2.2. Earnings Management of Internationally Diversified Firms

International diversification has recently become a crucial strategy for enhancing firms’ competitiveness. In this regard, Rugman and Verbeke (2005) suggest that internationally diversified firms can access cheaper resources and capitalize on national differences in production and sales compared with their local counterparts. These advantages could be in the form of less costly machinery, cheap labor, and advanced technological methods (Aharoni, 2024). Likewise, the studies of Fisch and Zschoche (2011) and Lee and Song (2012) show that internationally diversified firms can arbitrage markets by shifting production factors or transferring resources across countries. Similarly, Song (2014) indicates that multinationalism increases the value of firms with a high level of international diversification.

However, international diversification has some drawbacks. For instance, multinational firms face exposure to transaction risk. For example, internationally diversified firms deal with those transactions that are denominated in foreign currencies, and fluctuations of exchange rates can affect the cash flow of company. Moreover, parent companies consistently report their earnings in their home currency, while subsidiaries report their earnings in the foreign currency. The parent company will face translation risk when preparing the consolidated financial statements at the end of the year. Despite the exchange rate risk, diversified firms also face operational and political risks (Ellstrand et al., 2002), along with cultural differences. All these factors increase the information asymmetry between different stakeholders of internationally diversified firms (Yildiz, 2021). Although financial accounting standards require firms to disclose financial information on foreign operating divisions, the information investors obtain from financial reporting may be insufficient. Information asymmetry increases the possibility of earnings management because stakeholders do not have sufficient relevant information to monitor managers’ actions, giving rise to incentives and opportunities for earnings management (Richardson, 2000; Schipper, 1989). Additionally, the complication of international diversification prevents outside investors from obtaining overseas information, making it difficult for auditors to detect problems (Duru & Reeb, 2002). Similarly, Berrill et al. (2021) suggest that higher levels of international diversification are associated with increased opportunistic earnings management.

In summary, the level of earnings management in firms with international diversification is higher than that of their counterparts for several reasons. First, internationally diversified firms must consolidate their annual reports at the end of the financial year and merge the performance of foreign operating divisions with that of the local business divisions. Thus, if a foreign unit has poor economic performance, the company is likely to engage in earnings management. Moreover, internationally diversified firms normally bear transaction and translation risk exposure and tend to manage their earnings. Stable policies of internationally diversified firms are warranted in situations where firms have stable earnings. However, as international diversification increases information asymmetry, firms tend to engage in accrual management rather than real earnings management. Following international diversification, managers reassess their earnings management methods. For firms with international diversification, it is challenging to detect accrual-based earnings management because of the complexity of their business operations (Berrill et al., 2021). Thus, firms tend to use the substitution method of real earnings management in the case of international diversification (Li et al., 2011; Roychowdhury, 2006). In addition, firms with international diversification emphasize competitiveness and profitability, so they would avoid real earnings management, which might limit the actual benefits to firms. This study proposes the following hypotheses:

H2a: 

International diversification is likely to increase accrual-based earnings management.

H2b: 

International diversification is likely to decrease the use of real earnings management.

2.3. Impact of Outside Supervision on the Earnings Management of Internationally Diversified Firms

Supervision mechanisms may affect the earnings management of firms with international diversification. When supervision mechanisms become stricter, managers’ behavior in manipulating earnings may change, as they prefer to choose a method that is harder for auditors, regulatory parties, and others to detect. For example, Chan et al. (2015) suggest that managers tend to shift from accrual-based earnings management to real earnings management as regulatory scrutiny or shareholder monitoring increases. These arguments are also in line with those of Ghaleb et al. (2022), who collected data from 1,056 observations of manufacturing companies in Bursa Malaysia, and found that the remuneration of non-executive directors and external audit fees help to mitigate earnings manipulation. In other words, they play a supervisory role in corporate governance. However, the supervision mechanisms in this study are classified as follows: (1) investor firms’ outside supervision, which is captured by information transparency; and (2) investee firms’ outside supervision, using the country’s legal traditions, such as the common law system, to ascertain the supervision level.

2.3.1. Information Transparency

Information transparency is a critical aspect of corporate governance. Yildiz (2021) highlights the importance of information among local and foreign firms, which have both benefits and costs because of information differences. In addition to financial reports, investors also seek information about business operations and other relevant news. Diamond and Verrecchia (1991) suggest that information disclosure can reduce information asymmetry, and the cost of capital for a firm can also be reduced through decreased investment risk. Firms’ expansion of disclosure reduces information asymmetry and hence increases stock liquidity in equity markets (Healy et al., 1999; Welker, 1995). Moreover, firms with policies promoting more informative disclosures have larger amounts and enhance the quality of analyst forecasts (M. H. Lang & Lundholm, 1996). If firms are undervalued, voluntary disclosures can help investors understand business strategies and reinforce their financial reporting credibility (Healy & Palepu, 1993).

Due to the complexity of multinational transactions, international diversification exacerbates the level of information asymmetry between firms and shareholders, thereby providing managers with more opportunities to engage in accrual-based earnings management. However, information transparency can hinder incentives for accrual-based earnings management. More information disclosure means that analysts are more likely to uncover accrual-based earnings management behaviors, and investors can make more informed decisions based on this information. Hence, when firms disclose more information, this leads to a reduction in information asymmetry, and managers engage in less accrual-based earnings management because of the limited extent to which they can manipulate accruals (Beuselinck et al., 2019; Lobo & Zhou, 2001).

If firms with international diversification reveal more information, investors can better understand the state of multinational investment, efficiently view foreign operations, and prevent managers from engaging in accrual manipulation. Therefore, managers could change their attempts to less visible earnings management methods (Beuselinck et al., 2019). To meet earnings targets, the reduction in accrual-based earnings management by internationally diversified firms causes managers to resort to real earnings management, as it is difficult for outsiders to scrutinize. However, this approach imposes greater long-term costs, which have negative consequences on future firm value. The following hypothesis was proposed:

H3a: 

Firms with international diversification and higher information transparency are more likely to decrease their use of accrual-based earnings management.

H3b: 

Firms with international diversification and higher information transparency are more likely to increase their use of real earnings management.

2.3.2. Investor Protection

Investor protection is defined as the extent to which laws protect investors’ rights and the strength of the legal institutions that facilitate law enforcement. The seminal work of Porta et al. (1998) indicates that the common law legal tradition provides greater investor protection due to flexibility in granting investor relief. For this reason of stronger rights protection, investors are willing to invest in the open and large capital markets of common-law countries (Ball et al., 2000; La Porta et al., 1997). Investor protection is a key factor that influences earnings management behavior. In this vein, Leuz et al. (2003) indicate that strong investor protection can restrain attempts to manipulate earnings because strong protection can limit insiders’ ability to acquire private control benefits, reducing their incentives to mask firm performance. That is, the level of earnings management in common-law countries is lower than that in code-law countries. Moreover, Burgstahler et al. (2006) found that strong legal systems are negatively associated with earnings management. Because countries with strong investor protection have stricter regulations, disclosure requirements, and high litigation risk environments, they have higher information quality and less earnings management (M. Lang et al., 2003, 2006).

In common-law countries, legal frameworks originate from court decisions through precedents and have a greater margin for providing investor protection (H. Ali et al., 2022). Moreover, common-law countries have stricter regulations and more rules of business operations and disclosure requirements than code-law countries. Firms investing in common-law countries must comply with local regulations and disclose financial reporting information according to local rules. Under the supervision of strong local laws, opportunities for accrual-based earnings management decrease. Internationally diversified firms investing in common-law countries face fewer attempts at accrual manipulation, considering intense supervision. It means that, due to strong investor protection, firm managers are likely to rely on real earnings management. Due to strict regulations in common-law countries, auditors are more likely to be sued than in code-law countries (Wingate, 1997), and auditing is more prudent. Therefore, managers tend to engage in real earnings management, which is difficult to achieve. Moreover, they undertake less accrual-based earnings management, which is easier to scrutinize and may lead to being sued. The following hypothesis was proposed:

H4a: 

Firms investing in countries with strong investor protection are likely to decrease their use of accrual-based earnings management.

H4b: 

Firms investing in countries with strong investor protection are likely to increase their use of real earnings management.

3. Research Design

3.1. Sample Selection

Financial data were obtained from the Taiwan Economic Journal (TEJ) finance database, while foreign investment data were collected from the Taiwan Market Observation Post System and annual financial reports. Information transparency data were obtained from the Information Disclosure and Evaluation System (IDES). More importantly, this study spans the period from 2003 to 2016 for two reasons. First, the IDES data became available in 2003. Second, technological breakthroughs emerged in Taiwan in 2016 (Chung, 2018), which may significantly alter the dynamics of the business landscape. Thus, this study was limited to the specific period from 2003 to 2016. Therefore, for consistency, this study limited the data to 2016. Since the Information Disclosure and Evaluation System began in 2003, our observations span the period from 2003 to 2016. To avoid bias, these observations are adjusted by the following: (1) Financial institutions or utility firms are excluded because their distinctive industry characteristics, financial structure, and rules to comply with are different from common industries; (2) firms are excluded if they are not on a calendar year basis; (3) foreign firms registered in tax heaven and TDR companies are excluded; (4) firms are excluded if the data of any variables are missing; (5) firms are excluded if R&D expenditure is less than 1 percent of total assets; and (6) firms with negative sales revenue are excluded. This screening process is consistent with that of a previous study by Prencipe (2012).

Table 1. Sample selection.

The Sample Period is from 2003 to 2016	Firm-Year Obs.
All firms listed in TSE/OTC	20,631
Financial institutions or utility firms	(730)
Firms are not on a calendar year basis	(48)
Registered in tax havens or TDR companies	(586)
Missing data for any variables	(4725)
R&D expenditure is less than 1 percent of total assets	(1562)
Negative sales revenue	(483)
Final observations	12,497

outlines the sample selection process, resulting in a final sample of 12,497 firm-year observations.

Table 2. Sample industries.

Code	Industry	Obs.	Percentage
11	Cement industry	149	1.192%
12	Food service industry	334	2.673%
13	Petrochemical industry	473	3.785%
14	Textile industry	772	6.177%
15	Mechatronic industry	1012	8.098%
16	Wire industry	142	1.136%
17	Chemistry industry	959	7.674%
18	Glass and ceramic industry	86	0.688%
19	Paper industry	95	0.760%
20	Iron and steel industry	542	4.337%
21	Wheel industry	167	1.336%
22	Car industry	85	0.680%
23	Electronic industry	6356	50.860%
25	Construction industry	339	2.713%
26	Transportation industry	305	2.441%
27	Tourism industry	95	0.760%
29	General merchandise industry	183	1.464%
99	Others	403	3.225%
Total		12,497	100%

presents the sample industries. The electronics industry has the most significant percentage in the sample, accounting for 50.86% of the total observations. The second is the mechatronics industry, and the third is the chemical industry.

3.2. Empirical Models

This study examines the choice of earnings management, whether international diversification affects this choice, and the impact of outside supervision on the earnings management practices of foreign investment firms. In the following two regression models, we use β₁ of accounting flexibility to test Hypothesis 1: the substitution between accrual and real earnings management, α₁ of international diversification (model 1) and β₂ of international diversification (model 2) to test the impacts of foreign investment on the uses of accrual and real earnings management separately (Hypothesis 2a and 2b). Moreover, α₄ and β₅ are used to test the moderating effect of information transparency on the relationship between international diversification and earnings management (both accrual and real). Simultaneously, α₅ and β₆ are used to test the moderating effect of investor protection on the nexus between international diversification and earnings management (both accruals and real).

$$\begin{array}{l}{AEM}_{it}={\alpha }_0+{\alpha }_1{ID}_{it}+{\alpha }_2{TRPA}_{it}+{\alpha }_3{COMLAW}_{it}+{\alpha }_4{ID}_{it}\times {TRPA}_{it}+{\alpha }_5{ID}_{it}\times {COMLAW}_{it}+{{\alpha }_{6}SIZE}_{it}\\ +{\alpha }_7\Delta {LEV}_{it}+{{\alpha }_{8}MTB}_{it}+{{\alpha }_{9}ROA}_{it}+{{\alpha }_{10}BIG4}_{it}+{{\alpha }_{11}COMP}_{it}+{\alpha }_{12}{YEAR}_{it}+{\epsilon }_{it}\end{array}$$

(1)

$$\begin{array}{l}{REM}_{it}={\beta }_0+{\beta }_1{\mathit{AF}}_{\mathit{it}}{+\beta }_2{ID}_{it}+{\beta }_3{TRPA}_{it}+{\beta }_4{COMLAW}_{it}+{\beta }_5{ID}_{it}\times {TRPA}_{it}+{\beta }_6{ID}_{it}\times {COMLAW}_{it}+{{\beta }_{7}SIZE}_{it}\\ +{\beta }_8{\Delta LEV}_{it}+{{\beta }_{9}MTB}_{it}+{{\beta }_{10}ROA}_{it}+{{\beta }_{11}BIG4}_{it}+{{\beta }_{12}COMP}_{it}+{\beta }_{13}{YEAR}_{it}+{\nu }_{it}\end{array}$$

(2)

where the variables are defined as follows.

AEMit	Discretionary accruals are measured by the Kothari et al. (2005) model.
REMit	Comprehensive real earnings management index, based on Cohen et al. (2008).
IDit	International diversification is calculated as the amount of foreign investment divided by total assets.
TRPAit	1 if the firm’s information is more transparent, zero otherwise.
COMLAWit	Calculated as the amount invested in common-law countries divided by the total amount of foreign investment.
AFit	Accounting flexibility is calculated as the ratio of the firm’s NOA to sales, divided by the industry median value.
SIZEit	Natural logarithm of total assets.
∆LEVit	Change in total liabilities divided by total assets.
MTBit	Market-to-book ratio. The market capitalization is divided by the book value of common equity.
ROAit	A firm’s return on assets is defined as the ratio of earnings before extraordinary items divided by total assets.
BIG4it	1 if the auditor is a Big 4 audit firm, zero otherwise.
COMPit	The sum of restricted stock grants in the current period and the aggregate number of shares held by executives at year-end (excluding stock options), scaled by the total outstanding shares of the firm.
YEARit	Year dummy variables, with the base year set to 1999.

3.3. Variable Definitions and Measures

(1)

Accrual-based earnings management (AEM)

This study follows Kothari et al.’s (2005) performance-adjustment model to estimate accrual-based earnings management by year and industry. The regression model is as follows:

$${\displaystyle \frac{{TA}_{i,t}}{A_{i,t-1}}}={\alpha }_0+{\alpha }_1\left({\displaystyle \frac{1}{A_{i,t-1}}}\right)+{\alpha }_2\left[{\displaystyle \frac{\left({∆REV}_{i,t}-{∆REC}_{i,t}\right)}{A_{i,t-1}}}\right]+{\alpha }_3\left({\displaystyle \frac{{PPE}_{i,t}}{A_{i,t-1}}}\right)+{\alpha }_4{ROA}_{i,t}+{\epsilon }_{it}$$

(3)

where the variables are defined as follows.

${\mathit{TA}}_{\mathit{it}}$	Total accruals are defined as earnings before extraordinary items and discontinued operations minus cash flows from operations.
${∆\mathit{REV}}_{\mathit{it}}$	Change in revenues.
${∆\mathrm{REC}}_{\mathit{it}}$	Change in accounts receivable.
${\mathit{PPE}}_{\mathit{it}}$	Gross value of property, plant, and equipment.
${\mathit{ROA}}_{it}$	Return on assets before tax, interest, and depreciation expenses.
$A_{it-1}$	Total assets in t − 1 year.

Following the previous study by Selimefendigil and Öner (2022), we computed the residuals of the above model using ordinary least squares (OLS) results. These residuals serve as a proxy for accrual-based earnings management. Finally, accrual-based earnings management (AEM) is the difference between total accruals and non-discretionary accruals.

(2)

Real earnings management (REM)

Following Roychowdhury (2006) and Cohen et al. (2008), abnormal levels of cash flow from operations, production costs, and discretionary expenses serve as proxies for real earnings management, capturing the state in which managers manipulate earnings through real activities. Based on Cohen et al.’s (2008) model, we calculate the residuals of the following models to estimate the abnormal levels of cash flows from operations (AbnCFO), abnormal production costs (AbnProd), and abnormal discretionary expenses (AbnDiscExp).

To compute the abnormal CFO, AbnProd, and AbnDisexp, we must compute regular cash flow from operations, normal production costs, and normal discretionary expenses (Al-Duais et al., 2022; Roychowdhury, 2006). AbnCFO is the difference between the actual and normal operating cash flow levels.

The residuals of AbnCFO are calculated using Model (4), which represents abnormal cash flow.

$${\displaystyle \frac{{CFO}_{it}}{{Assets}_{i,t-1}}}=a_{1t}\left({\displaystyle \frac{1}{{Assets}_{i,t-1}}}\right)+a_{2t}\left({\displaystyle \frac{Sales}{{Assets}_{i,t-1}}}\right)+a_{3t}\left({\displaystyle \frac{∆Sales}{{Assets}_{i,t-1}}}\right)+{\epsilon }_{it}$$

(4)

where the variables are defined as follows.

${\mathit{CFO}}_{\mathit{it}}$	Cash flows from operations.
${\mathit{Assets}}_{i,t-1}$	Total assets in t − 1year.
${\mathit{Sales}}_{\mathit{it}}$	Sales revenue.
${\Delta \mathit{Sales}}_{\mathit{it}}$	Change in sales revenue.

The production costs are the sum of the cost of goods sold and the change in inventory. We combined models (5) and (6) as model (7) and calculated AbnProd as the residuals of model (7).

$${\displaystyle \frac{{COGS}_{it}}{{Assets}_{i,t-1}}}=b_{1t}\left({\displaystyle \frac{1}{{Assets}_{i,t-1}}}\right)+b_{2t}\left({\displaystyle \frac{{Sales}_{it}}{{Assets}_{i,t-1}}}\right)+{\epsilon }_{it}$$

(5)

$${\displaystyle \frac{{∆INV}_{it}}{{Assets}_{i,t-1}}}=b_{1t}\left({\displaystyle \frac{1}{{Assets}_{i,t-1}}}\right)+b_{2t}\left({\displaystyle \frac{{∆Sales}_{it}}{{Assets}_{i,t-1}}}\right)+b_{3t}\left({\displaystyle \frac{{∆Sales}_{i,t-1}}{{Assets}_{i,t-1}}}\right)+{\epsilon }_{it}$$

(6)

$${\displaystyle \frac{{Prod}_{it}}{{Assets}_{i,t-1}}}=b_{1t}\left({\displaystyle \frac{1}{{Assets}_{i,t-1}}}\right)+b_{2t}\left({\displaystyle \frac{{Sales}_{i,t}}{{Assets}_{i,t-1}}}\right)+b_{3t}\left({\displaystyle \frac{{∆Sales}_{i,t}}{{Assets}_{i,t-1}}}\right)+b_{4t}\left({\displaystyle \frac{{∆Sales}_{i,t-1}}{{Assets}_{i,t-1}}}\right)+{\epsilon }_{i,t}$$

(7)

where the variables are defined as follows.

${\mathit{Prod}}_{\mathit{it}}$	The sum of the cost of goods sold and the change in inventory.
COGSit	Cost of goods sold.
∆INVit	Change in inventory.
${\mathit{Assets}}_{i,t-1}$	Total assets in t − 1 year.
${\mathit{Sales}}_{\mathit{it}}$	Sales revenue.
${\Delta \mathit{Sales}}_{\mathit{it}}$	Change in sales revenue.

AbnDiscexp is the difference between actual and normal levels of discretionary expenses. The residuals of AbnDiscexp were estimated using Model (8).

$${\displaystyle \frac{{Discexp}_{i,t}}{{Assets}_{i,t-1}}}=c_{1t}\left({\displaystyle \frac{1}{{Assets}_{i,t-1}}}\right)+c_{2t}\left({\displaystyle \frac{{Sales}_{i,t-1}}{{Assets}_{i,t-1}}}\right)+{\epsilon }_{it}$$

(8)

where the variables are defined as follows.

${\mathit{Discexp}}_{\mathit{it}}$	The sum of advertisement expenses, research and development expenses, and operating expenses.
${\mathit{Assets}}_{i,t-1}$	Total assets in t − 1 year.
Salesi,t−1	Sales revenue in t − 1 year.

To capture the comprehensive effect of real earnings management, we computed REM as the sum of the three standardized individual components (−StdAbnCFO + StdAbnProd − StdAbnDiscexp). Higher REM levels indicate higher levels of real earnings management.

(3)

Accounting flexibility (AF)

Barton and Simko (2002) suggest that the effects of accrual-based earnings management are cumulative and are reflected in the balance sheet, thereby overestimating net assets. Based on Barton and Simko (2002), the proxy for accounting flexibility is calculated as the ratio of the firm’s previous year’s net operating assets to its previous year’s sales. The higher the Beneish Balance Sheet (BBS) score, the lower the accounting flexibility.

$${BBS}_{i,t}={\displaystyle \frac{{NOA}_{i,t-1}}{{Sales}_{i,t-1}}}$$

(9)

where the variables are defined as follows.

${NOA}_{i,t-1}$	Ratio of net operating assets to sales at the beginning of the period. Net operating assets are calculated as shareholders’ equity less cash and marketable securities, plus total debt.
${Sale}_{i,t-1}$	Sales in t − 1 year.

Since accounting flexibility is likely to be highly dependent on the industry (DeFond, 2002), we use the firm’s industry as a benchmark. AF is calculated as the ratio of a firm’s NOA/sales to the industry median value. Higher AF levels indicate lower accounting flexibility than their industry counterparts.

$$\mathit{AF}={\displaystyle \frac{{\mathit{BBS}}_{i,t}}{{\mathit{Median}\mathit{Industry}\mathit{BBS}}_{i,t}}}\dots$$

(10)

where the variables are defined as follows.

${BBS}_{i,t}$	From model (9), the ratio of net operating assets to sales at the beginning of the period.
${\mathit{Median}\mathit{Industry}\mathit{BBS}}_{i,t}$	BBS is divided by the median of the same industry.

(4)

International diversification (ID)

International diversification is measured as the percentage of foreign assets (Ajay & Madhumathi, 2015). This study suggests that international diversification can help determine the extent to which firms invest in foreign markets and understand the conditions under which they make such investments. Therefore, ID is defined as the amount of foreign investment divided by total assets to measure the degree of international diversification (Doukas & Lang, 2003).

(5)

Supervision mechanism

(a)

Information transparency (TRPA)

International valuation institutions such as Standard & Poor’s and Credit Lyonnais Securities Asia announce the valuation results of firms’ information transparency, which expects companies to enhance voluntary information disclosure. In Taiwan, to reinforce corporate governance and ensure that information is fully disclosed to the public, the Securities and Futures Institute plans and implements an “Information disclosure and evaluation system,” which is authorized by the Taiwan Stock Exchange (TWSE) and the Over-the-counter (OTC) market. The system evaluates the transparency of information disclosure among firms listed on the TWSE and the OTC. Starting in 2003, the system reported only the names of firms with more transparent information. Since 2005, it has included all listed companies, and the results are classified into five grades: A+, A, B, C, and C-, ranked from most to least transparent. TRPA is a proxy for measuring the level of information transparency, where 1 indicates that the firm’s information is more transparent, and 0 otherwise. More transparent firms are defined as those (1) added to the name lists of 2003 and 2004 and (2) that received an A+ or A grade after 2005.

(b)

Investor protection (COMLAW)

Common law (COMLAW) is used to measure the level of legal protection for foreign investment. If the common-law ratio was greater than 0.5, it was set to 1; otherwise, it was set to 0. The common law ratio is calculated as the amount invested in common law countries divided by the total amount of foreign investment in those countries. The legal systems of common-law countries are based on precedents and court decisions, whereas civil-law countries are mainly based on codified statutes (Hsieh et al., 2020).

(6)

Control variables

(a)

Firm size (SIZE)

Myers and Majluf (1984) indicate that it is easier for firms with high information asymmetry to develop disadvantageous strategies such as abandoning beneficial investments, which could harm future performance. However, on the other hand, the amount of information available to investors from large firms is more extensive because larger firms have a richer information environment (Wiedman, 1996). Additionally, opportunities for earnings management through R&D cuts decrease (Bushee, 1998). Compared with small firms, earnings management is more complex to detect. Large firms tend to be held widely by institutional investors and are more closely monitored by analysts. In other words, the costs of earnings management for large firms may be greater. Therefore, we used the natural logarithm of total assets to control for the effect of firm size on earnings management.

(b)

Financial leverage (ΔLEV)

This study calculates the change in total liabilities divided by total assets, as a measure of financial leverage, to assess the level of capital and debt financing. Duke and Hunt (1990) documented that firms with higher debt levels are more likely to incur high costs for debt covenant violations. Because bondholders can restrict a firm’s investments and require it to meet specific financial targets, managers have incentives to manage earnings in a way that does not breach the contract. Bushee (1998) also shows that managers of firms with high financial leverage tend to cut R&D expenditure to increase earnings. As a result, the leverage ratio controls for the effect of a firm’s financial structure on earnings management.

(c)

Growth opportunities (MTB)

Skinner and Sloan (2002) indicate that firms with growth opportunities have a more negative evaluation by the stock market when they miss the earnings thresholds. This shows that a firm’s growth opportunities affect managers’ earnings management behavior (Roychowdhury, 2006). Firms with more growth opportunities are likely to manipulate earnings to embellish financial reports or raise funds. Furthermore, a high level of growth opportunities leads to high political costs and risks, which may encourage managers to manage earnings (AlNajjar & Riahi-Belkaoui, 2001). Consequently, the proxy for growth opportunities is the ratio of the market value of equity to the book value of equity based on Roychowdhury (Roychowdhury, 2006).

(d)

Firm performance (ROA)

Operating state affects earnings management incentives. Managers manipulate earnings to meet earnings benchmarks and maximize opportunistic firm value (Degeorge et al., 1999). Hence, we use return on assets to control for the effect of a firm’s performance on earnings management.

(e)

Audit quality (BIG4)

Audit quality affects earnings management. Higher audit quality constrains accrual-based earnings management, and thus, managers resort to real earnings management (Chi et al., 2011). Therefore, a Big 4 auditor is an indicator of control over the impact of audit quality on earnings management.

(f)

Compensation characteristics (COMP)

Cohen et al. (2008) add compensation variables to the model to capture the incentive effects of equity components of executive compensation. COMP is the sum of restricted stock grants in the current period and the aggregate number of shares held by executives at the year-end (excluding stock options), scaled by the total outstanding shares of the firm. Accordingly, this variable controls for the influence of executive compensation on earnings management.

(g)

Year dummy variables (YEAR)

YEAR is defined as a dummy variable with 2003 as the base year. That is, if the observations are from 2003, the dummies for the other control years are 0; if the observations are from 2016, TIME2016 is 1, and the others are 0.

4. Empirical Results

4.1. Descriptive Statistics

Table 3. Descriptive statistics.

	Mean	Q1	Median	Q3	Largest	Smallest	Std. Dev
AEM	0.019	−0.001	0.014	0.031	0.102	−0.210	0.021
REM	0.028	−0.081	0.025	0.121	0.212	−0.162	0.061
AF	1.340	0.713	1.016	1.512	3.116	0.227	1.012
ID	0.219	0.054	0.160	0.313	0.663	0.000	0.156
TRPA	0.245	0.000	0.000	0.000	1.000	0.000	0.418
COMLAW	0.836	1.000	1.000	1.000	1.000	0.000	0.368
SIZE	6.595	6.194	6.501	6.912	8.071	4.839	0.460
ΔLEV	0.006	−0.041	0.008	0.067	0.544	−1.323	0.121
MTB	1.574	0.860	1.386	1.920	8.682	0.090	1.112
ROA	0.039	0.008	0.047	0.091	0.521	−1.012	0.105
BIG4	0.848	1.000	1.000	1.000	1.000	0.000	0.359
COMP	0.246	0.144	0.215	0.318	0.813	0.013	0.131

Note: 1. Variable definitions: AEM refers to accrual-based earnings management, as measured by the Kothari et al. (2005) model. REM is a comprehensive real earnings management index, based on the work of Cohen et al. (2008). AF is accounting flexibility, calculated as the ratio of the firm’s NOA/Sales to the industry median value. ID is international diversification, calculated as the amount of foreign investment divided by total assets. TRPA is 1 if the firm’s information is more transparent, 0 otherwise. COMLAW is 1 if the common law ratio is above 0.5, 0 otherwise, and the common law ratio is calculated as the amount invested in common-law countries to the total amount of foreign investment. SIZE is the natural logarithm of total assets. ΔLEV is the change in total liabilities divided by total assets. MTB is the market-to-book ratio, calculated as the market capitalization divided by the book value of common equity. ROA is return on assets, defined as the ratio of earnings before extraordinary items divided by total assets. BIG4 is 1 if the auditor is a Big 4 audit firm, 0 otherwise. COMP is the sum of restricted stock grants in the current period and the aggregate number of shares held by executives at the year-end (excluding stock options), scaled by the total outstanding shares of the firm. 2. Except for TRPA, COMLAW and BIG4, the variables are winsorized at the 2nd and 98th percentiles. 3. The sample spans from 2003 to 2016 and comprises 12,497 observations.

presents the descriptive statistics for the sample. The mean AEM was 0.019, and the median was 0.014. The mean REM score was 0.028, and the median was 0.025. These results indicate that right-skewed phenomena exist in the sample, with a few firms having relatively high AEM or REM.

4.2. Correlation Analysis

Table 4. Correlation coefficients of variables.

	AEM	REM	AF	ID	TRPA	COMLAW	SIZE	ΔLEV	MTB	ROA	BIG4	COMP
AEM	1
REM	0.67 *	1
AF	−0.14 *	0.29 **	1
ID	0.14 *	0.28 **	0.27 *	1
TRPA	−0.02	−0.02	−0.05 *	0.03	1
COMLAW	−0.06 *	−0.05 *	−0.05	−0.01	−0.04 *	1
SIZE	0.09 *	0.08 *	0.03 *	0.04 *	0.27 **	−0.06 **	1
ΔLEV	0.12 *	0.01	−0.07 *	0.06	0.09 *	−0.02	0.12 *	1
MTB	0.06 *	−0.07	−0.08 *	−0.04	0.06	−0.03 *	−0.06 *	0.13 *	1
ROA	−0.07	−0.12 *	−0.18 *	0.13 *	0.14 **	−0.09	0.16 **	0.25 *	0.37 *	1
BIG4	−0.13 *	0.14 *	0.03	0.05 *	0.07 **	−0.09	0.13 *	0.04 *	0.04 **	0.08 *	1
COMP	−0.17 *	−0.16 *	−0.16 *	−0.06 *	0.08	0.18 **	−0.17 *	0.08 *	0.06 *	0.07 *	0.06 *	1

Note: 1. The definitions of variables are the same as those in Table 3. 2. ** shows the significance at the 5% level and * shows the significance at the 10% level.

presents the Pearson’s correlation coefficients for the variables. By construction, most of the correlation coefficients were less than 0.3, indicating the degree of correlation within the tolerance level. AF is significantly negatively correlated with REM, which likely supports the hypothesis that accrual-based earnings management is an alternative method to real earnings management and vice versa.

4.3. Regression Analysis

Table 5. Regression analyses of accrual and real earnings management.

	Regression (1): AEM		Regression (2): REM
Const.	0.042 ***	(<0.001)	0.051 ***	(<0.001)
AF			−0.008 ***	(<0.001)
ID	0.021 **	(0.048)	−0.015 **	(0.012)
TRPA	−0.024 **	(0.015)	−0.033 *	(0.061)
COMLAW	−0.026 *	(0.071)	−0.017	(0.185)
ID×TRPA	−0.063 **	(0.015)	0.082 **	(0.013)
ID×COMLAW	−0.048 *	(0.063)	0.025 *	(0.127)
SIZE	0.023 **	(0.028)	0.016 *	(0.035)
ΔLEV	0.116 ***	(0.005)	−0.203 **	(0.015)
MTB	−0.041 **	(0.021)	−0.025 *	(0.081)
ROA	−0.027 ***	(0.009)	−0.064 **	(0.024)
BIG4	0.016 **	(0.012)	0.027 **	(0.013)
OWNER	−0.057 ***	(0.005)	−0.067 ***	(0.001)
YEARt	omitted	omitted	omitted	omitted
N	12,497		12,497
R-squared	0.097		0.095
Adj R-squared	0.093		0.092
F value	29.87		29.62

Note: 1. The definitions of variables are the same as those in Table 3. 2. *** significant at the 0.01 level; ** significant at the 0.05 level; * significant at the 0.1 level. 3. Numbers in parentheses are p-values.

shows the regression results for accrual-based earnings management and real earnings management. We use Regression (1) to test the influence of factors on accrual-based earnings management, and Regression (2) to test real earnings management.

4.3.1. The Substitution Between Accrual and Real Earnings Management

In the first analysis, we examine the substitution between accruals and real earnings management. As shown in Regression (2) in Table 5, the coefficient of accounting flexibility (AF) is −0.008. The p-value is less than 0.001, indicating that as the flexibility of accrual-based earnings management decreases, managers have a greater incentive to undertake real earnings management, which ultimately harms firms’ future value in the long run. The prominent reason for these findings is that when accounting flexibility is reduced, firms can consider limited accounting standards for evaluation. It presents the results on the relationship between accrual-based earnings management and real earnings management. The substitution between accrual-based earnings management and real earnings management supports Hypothesis 1. These results were consistent with those reported by Nguyen et al. (2023).

4.3.2. Earnings Management of Internationally Diversified Firms

In Table 5, the coefficients of ID present the impact of international diversification on accruals and real earnings management separately in Regressions (1) and (2). In Regression (1), the coefficient of ID is 0.021 (p-value = 0.048), which is statistically significant at the 5% level, suggesting that a higher level of international diversification is positively associated with accrual-based earnings management. Hypothesis 2a is supported: International diversification is likely to increase the use of accrual-based earnings management. The possible reason for these findings is that internationally diversified firms are characterized by complex business operations, which are scattered at different locations, and managers have a greater incentive to manipulate earnings through accrual-based earnings management (Khanchel El Mehdi & Seboui, 2011).

In Regression (2), the coefficient of ID is −0.015 (p-value = 0.012), which is positively significant at the 5% level, suggesting that higher international diversification is associated with a decrease in the use of real earnings management. Hypothesis 2b is supported: International diversification is likely to decrease the use of real earnings management. These results can be explained in such a way that internationally diversified firms face transaction exposure risk due to trade barriers, which hinder their real activities. Moreover, firms face translation risks when converting the currency of their subsidiary firms. Therefore, the cost of real earnings management may outweigh the benefits of earnings management in internationally diversified firms.

In sum, international diversification increases the use of accrual-based earnings management but decreases that of real earnings management. This result suggests that the complexity of international diversification makes it difficult for investors to read financial reports and understand a firm’s real operations, thereby increasing the information asymmetry (Masud et al., 2017). As international diversification increases, information asymmetry decreases, reducing the risk of auditors and regulators being deemed improper. Managers will decrease the use of real earnings management and shift to accrual-based management because real earnings management may ultimately harm a firm’s value.

4.3.3. Impact of Outside Supervision on the Earnings Management of Internationally Diversified Firms

Supervision mechanisms of international diversification can affect earnings management in internationally diversified firms. More specifically, this study considers information transparency and investor protection as proxies for supervision mechanisms.

(1)

Information transparency

In Table 5, the coefficients of the interaction variable of international diversification and the information transparency dummy ($ID\times TRRA$) present the additional impact of international diversification on the accruals and real earnings management of more transparent firms, separately in Regressions (1) and (2). In Regression (1), the coefficient of $ID\times TRRA$ is −0.0063 (p-value = 0.015), indicating a negative and statistically significant effect at the 5% level, suggesting that the impact of international diversification on accrual-based earnings management is significantly less for firms with higher information transparency compared to those with less information transparency. The sum of the coefficients of $ID$ and $ID\times TRRA$ is −0.042 (the p-value of the Wald test is 0.031) and is negatively significant at the 5% level, suggesting that internationally diversified firms with high information transparency are likely to reduce the use of accrual-based methods for earnings manipulation. The primary reason for these findings is that, due to more transparent information in internationally diversified firms, accrual-based earnings management can be detected more easily by auditors. Thus, firms tend not to prefer this method for earnings management. Thus, Hypothesis 3a is supported. When firms disclose more transparent information, there is a lower chance of accrual management. These findings are in line with those of Masud et al. (2017), who indicate that diversified firms have fewer information asymmetry issues. These findings reject the asymmetric information hypothesis for the Taiwanese firms.

In Regression (2), the coefficient value of ID is significant and negative, −0.015 (p-value = 0.012), while the coefficient value of $ID\times TRRA$ is significant and positive, 0.082 (p-value = 0.013). These findings suggest that when information transparency is high, internationally diversified firms undertake real earnings management methods to manage their earnings. These arguments are also supported by the sum of the coefficients of ID and ID×TRRA is 0.097 (p-value of the Wald test is 0.012), which is positively significant at the 5% level, suggesting that high information transparency in internationally diversified firms is more likely to increase the use of real earnings management. Among other reasons, one possible explanation for these findings is that real earnings manipulation cannot be detected in the short run, and managers can meet or beat their earnings targets by adjusting the production scale, level of discretionary expenses, etc. The concept of Hypothesis 3b is supported. These findings are in line with the paper of Khanchel El Mehdi and Seboui (2011), who suggest that the geographical expansion of firm operations tends to increase earnings management due to the complexity of these operations. These results also support the agency theory argument.

To summarize, if internationally diversified firms are less transparent, they are likely to utilize accrual-based earnings manipulation methods. However, firms are likely to use real earnings management methods due to the increased transparency of information. The primary reason for this behavior is that users of financial statements can easily detect accrual earnings management. Consequently, managers tend to resort to real earnings management as their last option in a strict financial reporting environment, aiming to meet analyst forecasts and investors’ expectations of earnings.

(2)

Investor protection

This section examines the moderating effect of investor protection on the relationship between international diversification and earnings management techniques. In Table 5, the coefficients of the interaction variable between international diversification and the investor protection dummy (ID × COMLAW) present the additional impact of international diversification on accruals and real earnings management in firms with more investor protection, separately in Regression (1) and Regression (2). In Regression (1), the coefficient of ID × COMLAW was −0.048 (p-value = 0.063), indicating a negative and statistically significant effect at the 10% level. This finding suggests that the impact of international diversification on accrual-based earnings management for firms investing in countries with higher investor protection is significantly lower than that for firms investing in countries with less investor protection. The sum of the coefficients of ID and ID × COMLAW is −0.027 (the p-value of the Wald test is 0.068), indicating a negative and statistically significant effect at the 10% level. These results suggest that firms are less likely to employ accrual earnings manipulation techniques due to strong investor protection. These findings align with the argument that courts in common law countries are more flexible in protecting the rights of investors. Therefore, judges can impede the expropriation of minority investors. Thus, Hypothesis 4a is supported.

In Regression (2), the coefficient of ID × COMLAW is 0.025 (p-value = 0.127), which is insignificant, suggesting that the impact of international diversification on real earnings management for firms investing in countries with stronger investor protection is not higher than that for firms investing in countries with less investor protection. The sum of the coefficients of ID and ID × COMLAW is 0.01 (p-value of the Wilcoxon test is 0.232), which is also not significant. This could not prove that firms investing in countries with strong investor protection are likely to increase their use of real earnings management. Thus, Hypothesis 4b is not supported.

Overall, findings suggest that internationally diversified firms tend to reduce earnings manipulation through accrual or real earnings management due to strong investor protection through the courts. Common-law country rules are derived from precedents and court decisions, and they offer greater flexibility in terms of minority shareholders’ rights. Moreover, accrual earnings manipulation can be easily detected by shareholders, auditors, and Securities and Exchange Commission regulators. Thus, strong investor protection helps firms to protect the rights of minority shareholders and limit insiders from expropriating their wealth.

5. Conclusions

Using a sample of Taiwanese listed and OTC firms from 2003 to 2016, this study first examines the substitution between accruals and real earnings management. The primary objective is to determine whether firms with an incentive to manipulate earnings would resort to real earnings management when accrual-based earnings management is restricted. This study analyzes the impact of international diversification on substitution between accruals and real earnings management. Moreover, this study tests whether the supervision of international diversification, measured by information transparency and investor protection, has substitutive effects.

Empirical findings indicate a substitution between accrual-based earnings management and real earnings management. Managers usually have a greater propensity towards accrual-based earnings management; however, when accrual flexibility is limited, they are more likely to manage real activities. This study also indicates that international diversification increases information asymmetry, whereas the risk of auditors and regulators being deemed improper decreases. Firms with earnings management incentives tend to shift from real earnings management to accruals-based earnings management, which may ultimately damage their actual value.

Finally, the supervision mechanism of firms with international diversification, measured by information transparency and investor protection, is likely to decrease the use of accrual-based earnings management. Conversely, information transparency increases the use of real earnings management. As accrual-based earnings management behaviors are more likely to be discovered with increased information transparency, firms with earnings management incentives shift from accrual-based to real earnings management. Moreover, investor protection is likely to decrease accrual-based earnings management but does not increase real earnings management. Because auditors are more likely to be sued, and auditing is more prudent in the presence of strong investor protection, real earnings management is also deemed improper.

The findings have substantial implications for managers, investors, and policymakers. For example, firms can carefully extend their operations across boundaries by considering their regulatory frameworks and cultural values. Firms can proactively mitigate transaction and translation exposure by providing indirect guidance. Additionally, investors are better able to determine whether specified countries have civil or common-law systems. Regulatory bodies can develop frameworks for international investment based on the guidelines of this study.