Related party transactions and earnings management in family firms: the moderating role of board characteristics Open Access

Over recent decades, the literature has widely studied earnings manipulation practices due to the possible serious effects that these practices can generate for investors and, more in general, for a company's stakeholders (Chih et al., 2008). Research has also pointed out the possible harmful effect stakeholders may suffer from a firm's transactions with related parties, namely companies belonging to the group of the reporting entity, its managers, majority shareholders and their close family members (Ryngaert and Thomas, 2012).

Earnings management (EM) practices, together with opportunistic related party transactions (RPTs), played a pivotal role in serious financial scandals involving large companies such as Bank of Credit and Commerce, Enron, Adelphia, Worldcom, Parmalat and Cirio and Subprime Mortgage. This has led scholars to investigate the determinants of the phenomenon of EM. In particular, different governance aspects have been analyzed, such as board characteristics (Davidson et al., 2005; Klein, 2002; Kao and Chen, 2004), ownership concentration (Yeo et al., 2002), presence of institutional investors (Chung et al., 2002; Rajgopal et al., 2003) and family control (Prencipe et al., 2008; Jaggi et al., 2009; Stockmans et al., 2010; Achleitner et al., 2014; Chi et al., 2015; Martin et al., 2016). The latest stream of the literature has provided mixed results, suggesting a more grained analysis of family firms in order to take into account their heterogeneity. The effect of a family CEO (Yang, 2010; Ansari et al., 2021), family involvement in C-suite roles (Ferramosca and Allegrini, 2018), board independence (Prencipe and Bar-Yosef, 2011; Jaggi et al., 2009) and female directorship (Vieira and Madaleno, 2019; Abdullah and Ismail, 2016; Mnif and Cherif, 2020) on EM have been analyzed, pointing out that family firms' governance characteristics affect their propensity to manipulate earnings.

Conversely, a limited number of scholars has investigated the relation between RPTs and EM (Thomas et al., 2004; Munir et al., 2013; Haji-Abdullah and Wan-Hussin, 2015; Marchini et al., 2018; El-Helaly et al., 2018; Alhadab et al., 2020), and results are quite mixed, suggesting that this issue deserves further research. This stream of literature has also taken into consideration the effect of ownership concentration (Munir et al., 2013), auditor type (El-Helaly et al., 2018) and ownership structure in terms of ownership concentration, family ownership and institutional ownership (Alhadab et al., 2020) on the association between RPTs and EM. The last study did not find a significant effect of family ownership on the association. Nevertheless, the definition of family ownership it assumed leaves room for more grained studies as it bases on the presence of two or more board members sharing the same surname, without controlling for the actual family ownership of a firm. Moreover, to the best of our knowledge, no studies have analyzed the family effect on the association between RPTs and EM by taking into consideration family firms' heterogeneity. According to literature, board characteristics and the extent to which family members occupy board positions are a major source of heterogeneity in family businesses (Chua et al., 2012; Nordqvist et al., 2014; Diéguez-Soto and Martinez-Romero, 2019), so, given the literature gap mentioned above, our research question is: “Does family ownership control affect the association between RPTs and EM? If so, do board characteristics moderate the association?”

We address this issue by drawing on Agency theory integrated by the socioemotional wealth (SEW) construct. We study a sample of 170, ex 2019, Italian non-financial listed firms for the period 2014–2019. The Italian context is of interest as, although RPTs are strictly regulated, they have nonetheless been opportunistically used and this has led to big financial scandals (Marchini et al., 2018). Moreover, a large presence of family companies that have a long-lasting family control (Franks et al., 2012) makes Italy a good setting in which to focus the research on family businesses.

“Earnings management occurs when management uses judgment in financial reporting and in structuring transactions to alter financial reports to either mislead some stakeholders about the underlying economic performance of the company or to influence contractual outcomes that depend on reported accounting numbers” (Healy and Wahlen, 1999, p. 368). Managers might manipulate earnings by exploiting discretionary accruals (Kothari et al., 2005; Jones, 1991) or resorting to real EM (REM) by departing from “normal operational practices” (Roychowdhury, 2006, p. 337).

Therefore, we study the effect of family control on the association between RPTs and different forms of EM: accrual-based earnings management (AEM), estimated by the Kothari model (Kothari et al., 2005), and REM estimated by abnormal levels of cashflow from operations, abnormal levels of production costs and abnormal levels of discretionary expenses (Roychowdhury, 2006). In doing so, we control for the effect of certain governance characteristics such as level of family ownership, CEO duality, weight of women and family members on the board, presence of a Big Four auditing firm and for a range of company attributes. The literature points out the growing interest on corporate governance in family businesses (Shekhar et al., 2021), and to the best of our knowledge, this is the first paper to provide a grained analyses of the moderating effect of family firm governance attributes on the association between RPTs and EM.

Our empirical evidence points out a different association between RPTs and EM in family and non-family firms. For family firms, we find a significant positive association between RPTs and downward AEM and REM perpetrated by abnormal discretionary expenses as well as a significant negative relationship between RPTs and abnormal production costs. For non-family firms, we provide evidence for a significant negative association between RPTs and absolute discretionary accruals.

Focusing on family businesses, we find that board characteristics moderate the relationship between RPTs and different types of earnings manipulations. In particular, we find that the presence of CEO duality, board gender diversity and the presence of the family on the board positively affect the association between RPTs and, respectively, REM implemented through sales manipulations, downward AEM and upward AEM.

This study adds to the stream of research on EM and RPTs as, according to Agency theory, our results suggest that abnormal levels of RPTs are used as a substitute, or together with, different types of EM. The contribution to family firm literature is twofold. On the one hand, we provide evidence that the concern for SEW preservation makes the above association different in family and non-family firms. On the other, our results point out family firms' heterogeneity in earnings manipulations showing that board characteristics differently moderate the association between RPTs and the various form of EM.

The remainder of the paper is organized as follows: section 2 provides the theoretical framework, the literary review and hypotheses development; section 3 describes the data and methods; section 4 provides results, robustness checks and the discussion; section 5 concludes with contributions, implications, limits of the study and suggestions for further research.

Most studies dealing with EM practices rely on agency theory to explain why companies engage in this unethical behavior (Paiva et al., 2016). Incentives for firms to manipulate earnings vary according to the different types of agency conflicts (Jensen and Meckling, 1976; Fama and Jensen, 1983). The first type of conflict arises between principal (shareholders) and agent (managers). Due to the separation between ownership and control, as occurs in publicly held corporations, shareholders find it difficult to constantly monitor managers' decisions. This leads to information asymmetries as well as increasing the risk that managers act in their own self-interest and harm shareholders. When managers' compensation is tied to company performance, managers may inflate or decrease earnings according to the need, respectively, to meet a performance target that will entitle them to a benefit or to transfer some of the profits to future periods if they expect the company to perform below the target to which their compensation is related.

Family firms are characterized by ownership concentration (Faccio and Lang, 2002), the owning family often risks most of its private wealth in the company (Bennedsen et al., 2010), and this is a motivation to monitor managers closely and thus prevent opportunistic behaviors. Furthermore, the fact that family members often hold top management positions fosters the alignment of interests between managers and the controlling family (Blanco-Mazagatos et al., 2016).

The second type of agency conflict occurs between majority and minority shareholders. When ownership is highly concentrated, as in family businesses, a controlling shareholder might exert pressure on managers' decision-making process (Demsetz and Lehn, 1985) to appropriate private benefits at the cost of minority shareholders (Villalonga and Amit, 2010; Faccio et al., 2001).

Agency conflicts in family firms are also shaped by non-financial goals. The family has a stock of affective values invested in the firm and its preservation becomes a priority, which makes family businesses different from non-family firms (Chua et al., 2015). Relative to non-family businesses, family firms attach great importance to other relevant values in addition to financial performance, such as family legacy, trust, commitment and reputation (Camilleri and Valeri, 2021).

The concept of SEW (Gómez-Mejía et al., 2007, 2011), through its different dimensions, encompasses the various non-financial returns the family receives from exerting control over the firm (Berrone et al., 2012). Family control and influence over the firm gives the family visibility and status (Gavana et al., 2021). Organizations' social structure is characterized by the set of relationship among their members (Valeri, 2021, p. 4), and in family firms, relationships involve family members as well as non-family members (Miller et al., 2009). Family members are emotionally attached to the company, establish long-term relationships with employees and forge strong social ties with the community affected by the business (Berrone et al., 2010, 2012). It becomes crucial to pass on the firm to future generations, who will thus inherit and perpetuate the links created by the family, its values, reputation and social status (Berrone et al., 2012). Since the controlling family expects to derive both financial and non-financial returns from the firm, the incentive to monitor managers is greater in family businesses than in their non-family counterparts, lowering agency conflicts.

The desire to preserve SEW has different effects on agency conflicts of the second type (Martin et al., 2017), namely those between majority shareholders and minority shareholders. On the one hand, the determination to maintain control over the company may induce the family to take decisions that adversely affect the firm's performance, damaging minority shareholders (Gómez-Mejía et al., 2007). Furthermore, family-controlled firms may have an incentive to manage earnings downward to reduce dividends, retain resources in the firm and protect transgenerational control (Achleitner et al., 2014). On the other hand, family members experience a strong sense of identification with the firm, so that firm's reputation and family reputation are strictly linked (Deephouse and Jaskiewicz, 2013). The fact that the company image reflects on the family may reduce the incentive to undertake activities detrimental to minority shareholders, which, if uncovered, would compromise the reputation of the firm and the family (Gottardo and Moisello, 2018).

Research provides two alternative explanations on the role of related party transactions. The first is the efficient transaction hypothesis (Friedman et al., 2003; Pozzoli and Venuti, 2014), according to which RPTs are an effective way to reduce transaction costs through the formation of a group internal market (Claessens et al., 2006; Jian and Wong, 2010).

The second is the conflict of interests' transaction hypothesis (Pozzoli and Venuti, 2014), in which RPTs are a means through which managers and controlling shareholders (insiders) expropriate the firm's resources for private benefit, to the detriment of minority shareholders and creditors (outsiders) (Mohammed, 2020; Johnson et al., 2000).

Opportunistic RPTs are used to improve the firm's financial performance (Jian and Wong, 2010; Williams and Taylor, 2013), and there is a tendency for firms to use RPTs to boost financial performance during the pre-initial public offering (IPO) period (Aharony et al., 2010) causing a negative impact on post-IPO long-term performance (Chen et al., 2011).

Empirical research provides evidence of the propension for business group owners to transfer profits from companies where they have low cashflow rights, to companies where they have high cashflow rights (Bertrand et al., 2002). The controlling shareholder in a business group may exploit different types of shareholding structures (direct, indirect and pyramidal ownership) in order to gain private benefits via RPTs, making it harder to identify the related party (Ali et al., 2021). Furthermore, the literature, basing on quantitative analysis methods, suggests that transactions with related parties may be either a substitute (El-Helaly et al., 2018) or a complement to EM practices (Thomas et al., 2004; Marchini et al., 2018), as well as a standalone tool to manipulate earnings (Chen et al., 2011).

The literature has shown a different propensity to manipulate earnings in family and non-family businesses (Prencipe et al., 2008; Achleitner et al., 2014; Chi et al., 2015; Gavana et al., 2017; Gottardo and Moisello, 2018). In particular, there is evidence that family firms exhibit a lower propensity to resort to AEM that non-family firms (Ali et al., 2007; Tong, 2007; Jiraporn and DaDalt, 2009; Prencipe and Bar-Yosef, 2011; Martin et al., 2016).

EM and firms' reputation are negatively associated (Rodriguez-Ariza et al., 2016; Salehi et al., 2020). The interests of families are closely related to the interests of the business, so that they would not be prone to managing earnings to avoid possible damage to the survival of the company and its reputation (Salvato and Moores, 2010; Borralho et al., 2020). Family firms who value SEW to a larger extent tend to avoid EM, which, if detected, would damage the reputation of the company and the family, putting at risk the survival of the firm for subsequent generations (Suprianto et al., 2019; Martin et al., 2016). Moreover, long-term focus limits pressure on a firm's management to achieve short-term performance goals, reducing the incentives for EM (Jiraporn and DaDalt, 2009). Nevertheless, there is evidence that family firms inflate earnings when economic performance is negative or decreasing in order to maintain lenders' trust, prevent their interference and preserve family control and influence (Stockmans et al., 2010).

Empirical findings on family firms' propensity to engage in REM are mixed and they depend mainly on the institutional settings considered by scholars (Ghaleb et al., 2020). Actually, REM benefits short-term cashflows to the detriment of long-term cashflows, resulting in a value decreasing activity (Roychowdhury, 2006). Since one of the main SEW priorities is to pass the business onto succeeding generations, family firms were found to resort less to REM than non-family businesses (Achleitner et al., 2014). Nevertheless, in countries characterized by weak investor protection the controlling family has a greater incentive to expropriate minority shareholders through REM, which is difficult to uncover and can be implemented during the year and not only at the reporting date (Eng et al., 2019; Razzaque et al., 2016).

As regards RPT research focusing on the nature of a firm's ownership has pointed out that RPTs have a negative effect on a firm's performance and that the effect is stronger for family firms (Mohammed, 2020), suggesting an opportunistic use of these transactions. There is also evidence (Kohlbeck et al., 2018) that family firms are more prone to enter into RPTs and that their financial performance declines when they report transactions with directors, officers and major shareholders. When opportunistic RPTs are put in place, managers may resort to EM practices to increase the benefits extracted or to disguise the expropriations perpetrated to the detriment of outsiders (Sarlak and Akbari, 2014; Marchini et al., 2018). Studies focusing on emerging markets have pointed out that RPTs exacerbate REM, but family ownership moderates this opportunistic association (Haji-Abdullah and Wan-Hussin, 2015). On the other hand, researchers focusing on Jordanian companies found a significant moderating effect of ownership concentration on the association between AEM and RPTs (Alhadab et al., 2020).

The empirical literature has provided evidence of an association between RPTs and EM, suggesting that transactions with related parties may be used as a substitute or a complement to other forms of earnings manipulations. Research has also pointed out an effect of family control on EM practices related to the peculiarities of agency conflicts and the relevance of non-financial goals in family firms. Basing ourselves on the theoretical framework and the above literature review, we posit the following:

EM practices differ according to family businesses' corporate governance attributes. The propensity for family firms to manage earnings is higher in the case of non-family CEOs, relative to family CEOs, as their compensation is more often linked to the financial performance of the firm to align the interests of family owners and CEOs (Yang, 2010). Moreover, because of a different level of family identification with the firm, companies created or inherited by a family present a higher quality of reported earnings relative to firms acquired by a family, especially when the CEO is a family member (Pazzaglia et al., 2013). Conversely, there is evidence that CEO duality may lower the board's effectiveness to provide oversight over managerial behavior, increasing EM (Chi et al., 2015). The literature also points out that founder CEOs tend to implement upward EM when they are close to reappointment to prevent minority shareholders from requesting a professional CEO to improve the firm's financial performance (Ansari et al., 2021).

On the one hand, the literature (Schulze et al., 2003) suggests that information asymmetry is lower when family members, rather than outsiders, perform monitoring as they communicate and cooperate more easily. From this perspective, family member participation would be an informal mechanism for enhancing internal corporate governance, especially in opaque markets (Fan and Yu, 2022). On the other hand, research has pointed out that the presence of family members on the board reduces the effectiveness of audit committees (Jaggi and Leung, 2007) and independent directors (Prencipe and Bar-Yosef, 2011; Jaggi et al., 2009) in limiting EM.

The literature has investigated the influence of women directorship on EM in family firms. Findings point out that the presence of women on a board decreases EM (Vieira and Madaleno, 2019; Mnif and Cherif, 2020), although a previous study carried out in Malaysia showed that family ownership does not interact with women on boards to influence the propensity for EM (Abdullah and Ismail, 2016).

Regarding the effect of board characteristics on RPTs, there is evidence that family companies with multiple family members involved in the business are less prone to engage in RPTs than family firms in which only one family member owns and/or manages the business (Fan and Yu, 2022).

The above literature review points out that board characteristics influence EM practices in family firms as well as their propensity to engage in RPTs, therefore we expect board characteristics to affect how family businesses may associate RPTs and different forms of EM. We posit the following:

The data sample includes firms listed on the Italian stock exchange in the period 2014–2019. We exclude regulated industries such as banks, credit firms, brokers and insurance because of their unique accounting practices and regulations.

The final sample is given by 170 listed firms as in 2019 with available information on RPTs and 522 firm-year observations. We hand-collected the data on RPTs from the notes to the financial statements. The balance-sheet data are from Orbis, the global Bureau van Dijk database. Board characteristics and corporate governance are reconstructed based on information available on the Chamber of Commerce register and the Orbis database.

We examine the relation between accrual and REM activities and RPTs for family and non-family firms. We define a family firm as one where the ultimate owner is a family that owns at least 20% of common shares (Villalonga and Amit, 2010). We measure AEM with discretionary accruals (Dechow et al., 1995; Jones, 1991; Roychowdhury, 2006). We use three REM proxies: abnormal levels of cashflow from operations (RMCFO); abnormal levels of production costs (RMPROD); abnormal levels of discretionary expenses (RMDISX).

We measure discretionary accruals (DA) for each firm as the difference between total accruals and non-discretionary accruals. We estimate non-discretionary accruals by using the Kothari et al. model (2005):

where:

• A_i,t−1 = lagged total assets

• ΔREV_i,t = change in revenues in period t

• ΔREC_i,t = change in trade receivables in period t

• PPE_i,t = gross property, plant and equipment

• NI_i,t = net income

We estimate normal levels of cashflow from operations; normal levels of production costs and normal levels of discretionary expenses as in El-Helaly et al. (2018):

where:

The estimated residuals in the previous accruals and REM equations proxy for discretionary accruals (DA) and for the abnormal levels of cashflows from operations, production cost and discretionary expenses (RMCFO, RMPROD, RMDISX) ^[1]. In the models, we use the absolute value of discretionary accruals (ABSDA), or positive and negative discretionary accruals (PDA and NDA, respectively).

The explanatory and control variables are RPTs, board size (BSIZE), weight of woman on the board (WBD), CEO duality (CEOD), family ownership (FOWN), weight of family members on the board (FMBD), audit quality (BIG4), ROA, SIZE, leverage (LEV), firm age (AGE), market-to-book ratio (MKTBOOK), year and industry dummies.

Related party transactions are measured as the value of RPT scaled by total assets (Ryngaert and Thomas, 2012; Habib et al., 2017). RPT is the sum of sales, purchases and outstanding balances between the company and its related parties (El-Helaly et al., 2018; Alhadab et al., 2020).

Board size (BSIZE) is measured as the number of board members (Jensen, 1993). WBD is the weight of women as a percentage of board members (Nadeem et al., 2017; Cruz et al., 2019). CEO duality (CEOD) is a dummy with value of 1 if the CEO is also the chair of the board (Huybrechts et al., 2013). Family involvement is proxied by the family ownership (FOWN) and by the weight of family members sitting on the board of directors (FMBD).

Audit quality (BIG4) is a dummy variable that receives the value of 1 if the auditor is a member of the Big Four group and zero otherwise (Bennouri et al., 2015). We control for possible effects of net income on EM including ROA (Roychowdhury, 2006).

Firm size (SIZE) is the log of total assets (Muñoz-Bullon et al., 2018; Ginesti and Ossorio, 2021).

Financial leverage (LEV) is measured as total financial debt scaled by assets (Anagnostopoulou and Tsekrekos, 2017; Zamri et al., 2013).

We also control for firm age (AGE) and for the market-to-book ratio (MKTBOOK) (Marchini et al., 2018; Dechow et al., 2011).

Finally, we include dummies to control for year and industry effects.

For each proxy of accrual and REM, we run the model:

where:

• EM_i,t = DA, RMCFO, RMPROD, RMDISX

• RPT_i,t = related party transactions for firm i in period t scaled by assets

• X_k,i,t = control variables

To take into account the possible interaction effects between RPTs and CEOD, WBD and FMBD, for the family firm sub-sample, we implemented the augmented model:

We run generalized least squares (GLS) models separately for family and non-family firms, adjusting standard errors for clustering at the firm level in the presence of correlations across firms and controlling for time and industry fixed effects using dummies.

Table 1 shows the descriptive statistics for the full sample and the family and non-family firms separately. The mean values for the REM proxies RMCFO, RMPROD and RMDISX are close to zero but with significant differences between family and non-family firms. On the other hand, the value of discretionary accruals measured in absolute terms (ABSDA), or distinguishing negative (NDA) from positive discretionary accruals (PDA), are not statistically different for the two sub-samples. RPTs represent, on average, 7% of assets, but with marked differences between the two sub-samples as the average value for family firms is half the value of non-family firms. Board size and women on board are statistically different for family and non-family firms, and the same is true for CEO duality.

Table 2 shows that RPTs are correlated with abnormal cash flows from operations (RMCFO) and discretionary expenses (RMDISX). Abnormal cash flows from operations are negatively correlated with negative and positive discretionary accruals. The correlation matrix shows moderate to average associations between the variables, and we can rule out multicollinearity problems.

Tables 3 and 4 present the GLS panel results for equation (1) for each earning management proxy with firm's clustering to control for firm fixed effects and with year and industry fixed effects. We show a base model with the same explanatory variables for family and non-family firms, and a specific model of family firms where we add the variables measuring family involvement.

Table 5 presents the family firm results for the model in equation (2), introducing interaction effects of RPT with CEO duality, women on board and family members on board.

To verify that our results are not driven by the choice of measuring RPTs as a continuous variable, we re-run the above analysis with two different specifications for this variable. In the first specification, we measure RPTs using as an explanatory variable the estimated residual (ABNRPT) of the following equation:

where:

• A_i,t = total assets

• RPT_i,t = related party transactions for firm i in period t scaled by assets

• MKTBOOK = market-to-book ratio

• LEV = financial debts scaled by assets

• D_k = year and industry dummies

For the second specifications, we follow Ryngaert and Thomas (2012) who suggest using a dummy variable equal to 1 if the sum of RPT exceeds a certain threshold with respect to assets. In this case, we assumed as a cutoff point the sample median level of the RPTs to asset ratio, which is 8.2%. DRPT is a dummy that takes value of 1 if RPTs are equal or exceed 8.2% of assets and zero otherwise.

Tables 6–8 report the results for ABNRPT, while Tables 9–11 display the results using the dummy DRPT. The evidence based on these alternative measures of the RPTs variable by and large confirm the findings of the base model in Tables 3–5, pointing out the robustness of these results.

Our empirical findings suggest that family control affects the association between RPTs and EM in different ways, according to the type of earnings manipulation taken into account.

Our results show that family firms are less prone to engage in AEM than non-family businesses, although the difference is not significant. Nevertheless, they present significantly higher levels of REM perpetrated by the means of anticipated sales and overproduction. These results are consistent with the evidence that REM is difficult to discover and, given the reputational concerns related to the preservation of SEW, family firms might prefer this form of earnings manipulation (Lin and Shen, 2015; Calabrò et al., 2020). On the other hand, our findings point out that family firms are significantly less prone to perpetrate REM by the means of a reduction of discretionary expenses as it increases short-run performance but jeopardizes a firm's durability and the perpetuation of the family dynasty through the business (Berrone et al., 2012). The value of RPTs is significantly lower for family than for non-family firms, but when they engage in high levels of RPTs, they also tend to manipulate earnings by means of abnormal discretionary expenses, as indicated by the significant positive association between RPTs and RMDIX. When managers use RPTs in order to expropriate a firm's resources, they have incentives to manipulate earnings to maximize any benefits extracted or to mask such expropriation (Marchini et al., 2018; Jian, 2003).

Our findings show a negative significant relation between RPTs and RMPROD for family firms, suggesting that family companies strongly engaged in RPTs tend to use these transactions as a substitute for upward EM by the means of overproduction as the latter might be costly for shareholders. This type of REM jeopardizes SEW, as it implies a firm's shift from its normal business operations in order to reduce the fixed cost per unit and the average unit cost of production. This enhances present performance but reduces future operating results, by the means of bigger inventory costs and lowers liquidity (Roychowdhury, 2006; Eng et al., 2019).

We do not find a significant association between RPTs and RMCFO in family or in non-family businesses. This suggests that RPTs are not significantly related to earnings manipulations implemented through sales anticipations and that they might be used as a stand-alone type of EM in both types of company.

We find a significant relationship between RPTs and REM perpetrated by overproduction and discretionary expenses only for family firms. This supports the view that family control affects the above association (H1a) and suggests that non-family firms' higher level of RPTs might not be a complement to, or a substitute of REM, but an autonomous form of earnings manipulation.

As for the relation between RPTs and AEM, we find a negative relation between RPTs and absolute discretionary accruals only for non-family companies, suggesting that this type of firm tends to use the transactions with related parties as a substitute for AEM. Splitting the analyses of AEM into downward and upward EM, we find a significant positive relation between RPTs and downward EM only for family firms, partially supporting H1b. This association suggests that when family firms use RPTs in order to transfer financial benefits to the controlling shareholders – to the detriment of non-family shareholders – they also exploit the judgment inherent in some accounting choices to pursue, at the same time, their non-financial goals. In fact, by lowering earnings they reduce the outflow of resources in the form of dividends to the benefit of self-financing (Achleitner et al., 2014). In so doing, they avoid the use of external equity financing and ownership dilution as the preservation of family control on the business is a main aspect of SEW (Gomez-Mejia et al., 2007).

Controlling for the effect of family firms' board characteristics on the association between RPTs and EM, we find a significant positive association between the interaction, ‘CEO duality-RPTs and RMCFO’, supporting H2a. This indicates that, in the presence of CEO duality, high levels of RPTs are significantly used in association with earnings manipulations by means of sales anticipations. In accordance with previous research, we find that CEO duality is more common in family-controlled firms than in non-family companies (Cheung et al., 2005) and that the position is normally held by a family member (95.7% for our sample). CEO duality indicates a strong will of the owning family to maintain its control and influence on the business in order to preserve SEW. Under this view, the associated use of RPTs and REM is consistent with previous research pointing out that family CEOs might manipulate earnings to enhance performance and protect their appointment (Ansari et al., 2021), in this case in the dual role of CEO and chair of the board. On the other hand, CEO duality might indicate an opaque governance and a higher ability to extract private benefits from the business (Sharma, 2004; Claessens et al., 2002). Therefore, this result might suggest that, in cases of CEO duality when RPTs are used in order to extract benefits for the controlling family to the detriment of minority shareholders, REM might be used in order to mask or enhance this benefit (Marchini et al., 2018; Jian, 2003).

Our empirical evidence shows a significant positive relation between the interaction “family members on board-RPTs and upward AEM’. This means that, when family members sit on the board, RPTs are used in association with accrual-based earnings manipulations that increase performance. The literature reports that the presence of family members on the board weakens the effectiveness of audit committees in monitoring managerial exploitation of discretionary accruals (Jaggi and Leung, 2007). Consistently, our results suggest that when RPTs are opportunistically used, the presence of family members on the board facilitates the resort to AEM to raise performance in order to increase or disguise the benefit.

According to H2b, we find a weakly significant association between the interaction “women on the board-RPTs and downward AEM”. Like previous literature, we find that the presence of women on the board, per se, exerts a lowering effect on the propensity to manipulate earnings using discretionary accrual (Vieira and Madaleno, 2019). On the other hand, our results suggest that when family firms' management engages in high levels of RPTs, extracting resources from the company, gender diversity tends to produce an opposite effect and might enhance the level of downward EM in order to reduce the exit of resources from the company via dividends or taxation. Given empirical evidence pointing out that female directors moderate the use of leverage in family firms (Huang and Kisgen, 2013; Poletti-Hughes and Martinez Garcia, 2022), this effect might be explained as the will to increase self-financing and limit the recourse to external financing.

This study is aimed at analyzing the relation between RPTs and different types of EM in family and non-family firms, as well as within family firms. It investigates the association between RPTs and different types of real and AEM for a sample of Italian family and non-family firms, for the period 2014–2019. This is one of the few papers studying this relationship in developed economies (Marchini et al., 2018; El-Helaly et al., 2018), and in doing so, it is the first research that deepens the analysis considering the effect of board characteristics in family businesses.

We add to the stream of literature that studies RPTs from an agency conflicts perspective as we provide evidence that companies use this type of transactions in association with, or as a substitute of, AEM or REM, controlling for certain governance and firm attributes. We contribute to SEW studies as, consistent with this view, we find that family control shapes agency conflicts and, as a result, the relationship between RPTs and EM in family businesses. We also extend family firm accounting research by highlighting the moderating role of board characteristics, in terms of CEO power, family member involvement and gender diversity, on the association between RPTs and the various forms of EM. In doing so, we also contribute to the growing stream of research on family firm governance (Shekhar et al., 2021) by providing evidence of the different effects of its attributes on earnings quality. We highlight the need to consider family firms as a heterogeneous group (Chi et al., 2015; Nordqvist et al., 2014; Diéguez-Soto and Martinez-Romero, 2019). In fact, taking into account board characteristics, we find that CEO duality, the weight of family members on board size and the presence of women on the board positively moderate, respectively, the association between RPTs and RMCFO, upward AEM and downward AEM. The last result also add to the stream of studies that analyses the effect of gender diversity in family firms' accounting choices (Abdullah and Ismail, 2016; Vieira and Madaleno, 2019; Mnif and Cherif, 2020).

These findings have practical implications for investors and auditors. They indicate that, although the level of RPTs, AEM and REM via abnormal discretionary expenses tends to be lower in family businesses, they might use opportunistic relationships between RPTs and EM, and board composition significantly influences the association. Moreover, the moderating effect of CEO duality and of family members weight on the board also suggest to regulators involved in issues related to corporate governance and RPTs a reflection on CEO power and board independence.

Our study has some limitations. It analyses the effect of family control on the relationship between RPTs and EM. Other SEW dimensions, such as the sense of identification with the company, concern for the firm's reputation and sense of dynasty might affect the association. We based our analyses on data contained in the financial statements, so we cannot measure these dimensions. Further research could extend the analysis to other SEW dimensions, collecting the above information by the means of surveys.

This explorative study analyses the effect of certain board characteristics (CEO duality, proportion of family members and women on board size) on the association between RPTs and EM. Further studies could deepen the analyses by adding more elements in terms of board diversity. A grained analysis of family members on the board in terms of generation, family branches they belong to and ownership stake held could better explain the effect of family businesses' heterogeneity on earnings manipulations.

Our research studies a single country sample in order to focus on the effect of family control and the moderating role of board characteristics on the association between RPTs and EM, preventing the confounding effects stemming from the use of different institutional settings. This might affect the generalizability of our findings, so future studies could analyze the possible effect of specific institutional aspects by extending the analysis to an international sample.