Family business groups and earnings manipulation: An emerging economy perspective

Abstract

This paper examines the influence of family business groups' affiliated and non-affiliated firms on earnings manipulation on a sample of Pakistani listed firms for the period of 2014 to 2019. The sample of this paper consists of 323 listed firms from the Pakistan Stock Exchange. Data were manually collected from the annual reports of the companies and State Bank of Pakistan sources. OLS and The Panel data models are used to validate the hypotheses of the study. Two proxies of earnings management have been used: one is discretionary accruals and the other is real activity manipulation. The study findings show a negative relationship of family business groups' affiliated firms with accrual-based earnings manipulation. Additionally, the magnitude of earnings manipulation is more in non-affiliated firms as compared to family business group affiliated firms. This study provides rare and initial evidence of family business groups’ relationship with earnings manipulation in Pakistan; moreover, this study extends the literature on scarce literature related to Family business groups and earnings manipulation.

Keywords:

• Earnings manipulation
• family business groups
• corporate governance
• Pakistan Stock Exchange

PUBLIC INTEREST STATEMENT

Family business groups are a common phenomenon in South Asia (Bangladesh, India and Pakistan) and it is considered one of the most well-known features of the Asian stock markets. These family business groups play an important role in the development of emerging economies such as Pakistan. Business Groups may be defined as the aggregate of different companies financial and non-financial, private and public in different sectors more often controlled by a specific family through different proportions of ownership. In Pakistan, it is famously known as “Twenty-Two Families”. In this research paper, we have utilized Pakistani family business groups dominated corporate landscape to investigate the earnings manipulation practices in family business groups' affiliated firms. The study findings show that family business groups' affiliated firms are not engaged in manipulating accrual-based Earnings Management. Additionally, the magnitude of earnings management is more in non-affiliated or stand-alone firms as compared to family business group-affiliated firms.

1. Introduction

The objective of this research study is to provide empirical evidence that whether family business groups' affiliated firms engage more in earnings management (hereafter, EM) than unaffiliated or stand-alone firms in Pakistan. The study’s context is famous for concentrated ownership, family ownership and (famous) business groups. It is believed in the related literature that the pyramidal ownership structure (business groups) provides more opportunities for the controlling shareholders to engage in earnings manipulation in order to cover up its expropriating activates in business groups¹ (Beuselinck & Deloof, 2014; Kim & Yi, 2006; Muttakin et al., 2017). Business groups are a common phenomenon in South Asia (Bangladesh, India and Pakistan) and it is considered one of the most well-known features of the Asian stock markets. These business groups play an important role in the development of emerging economies such as Pakistan. Moreover, the related literature shows that business group affiliates use different methods to support poor performing affiliates such as cash and accrual-based related party transaction (propping), using transferring resources (tunneling) within groups and earnings manipulations² (Bertrand et al., 2002; Hussain & Safdar, 2018b; Jian & Wong, 2010). Beuselinck and Deloof (2014) added that structures of related-party transaction amongst the group affiliates provides not only extra opportunities but also extra instruments to group affiliates to manage its earnings more as compared to stand-alone companies. Therefore, the real financial position of a firm is masked through these related-party transactions via group affiliation, so to conclude the above findings, family business group firms have more incentives and opportunities to involve in EM as compare to non-affiliated group firms. According to Wang and Yung (2011) in the family firm’s governance, structure is weak due to family firms' controlling rights, which offered them a stronger position to get private gains by using its voting rights and executive positions and expropriate the resources of minority shareholders.

In contrast to the aforementioned discussion, business groups could be seen as an efficient form of organization where resources are effectively utilized in the best interest of all the stakeholders. Many advantages are associated with Business group affiliation, these may be easy access to finance for the affiliated firms (Muttakin et al., 2017), knowledge and risk-sharing among the group affiliated firms (Claessens et al., 2006; Hamelin, 2011), no need for 100 percent controlling rights (Beuselinck & Deloof, 2014) and supporting the affiliates through related party transactions. Moreover, having discussed the pros and cons of business groups, it could be seen under two opposite hypothesis “the conflict of interest hypothesis” as suggested by (Berle & Means, 1932; Jensen & Meckling, 1976) and “the efficient transaction hypothesis” by (Gordon et al., 2004; H. Ullah, 2014). The conflict-of-interest hypothesis is related to the agency theory, which posits that there is the conflict between principals and agents (The Type I Agency Problem) and the majority (principal) and the minority (principal) shareholder (The Type II Agency Problem). This conflict of interest leads managers in the case of the Type I Agency Problem to engage in EM for compensation and debt convent hypothesis, whilst in the case of the Type II Agency Problem, controlling shareholders take benefits at the expense of minority shareholders (Marchini et al., 2018). Moreover, the efficient transaction hypothesis indicates that affiliated firms engage in related transaction in order to promptly fulfill their needs and consider a useful business transaction not only for the firm but also for its all stakeholders (Gordon et al., 2004; H. Ullah, 2014). So, keeping both points of view in mind, the objective of this research study will be to check that whether a group affiliated firm engages in EM to manage earnings more as compared to non-group firms and what is the extent of EM in the group's affiliated and non-affiliated firms.

Moreover, we have the following key motivations for performing this research study. First, the literature of corporate governance (hereafter, CG) and EM is very rich; however, the literature of EM with family business groups is not ripe and fully captured in the business group-related literature in general and Pakistan in particular. Second, although few studies examined the role of business groups in the earnings manipulation based on developed and undeveloped economies such as (Beuselinck & Deloof, 2014) in Belgium (Kim & Yi, 2006), in South Korea Muttakin et al. (2017), in Bangladesh and Sarkar et al. (2013) in India, this research study will focus on an emerging economy of Pakistan, which is an ideal setting for carrying out this research due to concentrated ownership, family ownership and (famous) family business groups. Third, the Pakistani context offers both sets of companies, which are group-affiliated and group-unaffiliated firms and they are easily identifiable, so it will offer an excellent setting to investigate this paper objectives. Fourth, Pakistani business groups have dominated the private sector in the Pakistan economy and have a esteem position in Pakistani economic and political landscapes (Saeed & Sameer, 2015). In addition to this, Pakistani family business groups have strong political links with different governments that these links may stimulate affiliated firms to manage earnings due to the fact that the politically connected firms have poor financial quality (Hashmi et al., 2018). Finally, in Pakistani family, business groups are researched in multiple angles such as Hussain (2019) & Ullah (2014) linked with business groups with tunneling, Ghani and Ashraf (2005), Ghani et al. (2011) and Ullah (2017) had linked business groups with firm performance, Bhutta and Suleman (2017) had linked business groups with the capital structure and Bhutta et al. (2016) had linked business groups with ownership concentration. To the best of our knowledge, there is no such study that has linked business groups with accrual-based and real activity earnings manipulation. Therefore, we are aiming to fill the aforementioned gap in the literature, by linking business group affiliation with earnings manipulation by taking not only accrual-based earnings manipulation but real activity manipulation as well in Pakistani context. Following the research gab and motivations of the study, the objectives of this study are twofold, first, to examine family business groups' affiliation impact on EM, second, to check the magnitude of EM in both family business group's affiliated firms and non-affiliated firms.

In order to examine the study’s hypothesis, achieve its objectives and address the established gap by studying group affiliation and earnings manipulation relationship in the unique Pakistani context, which is dominated by concentrated family ownership and group ownership, this research study has utilized the data of 323 listed firms over the period of 2014–2019. In these 323 companies, 187 are family business groups' affiliated firms', whilst 136 are non-affiliated or stand-alone firms. We have found out that there are 60 business groups working in Pakistan. The data for this study have been manually collected from various sources such as annual reports, concerned companies’ websites and the State Bank of Pakistan (hereafter, SBP) publications. The findings of the study show that group affiliation is negatively related to accrual-based earnings management, contrary to the results reported in the studies of Beuselinck & Deloof, 2014; Kim & Yi, 2006; Muttakin et al., 2017; however, this research study failed to find the evidence that group affiliation and real activity management is negatively related to each other. The possible explanation for the negative relationship between group affiliation and accrual-based EM is that controlling shareholders may not want to lose their family firms identity, reputation and value by closely monitoring the management. The negative relationship is robust with the random effect model that reports similar findings as the Ordinary least Square Regression. Moreover, the findings reveal that the magnitude of EM is more in stand-alone firms as compared to group-affiliated firms. In addition to this, the female directorship and board independence are lower in the group-affiliated firms as compared to stand-alone firms. Furthermore, we have checked the EM behavior in the financially constrained and unconstrained firms along with affiliated and non-affiliated firms and the findings of this research paper show that constrained firms do manage their earnings, whereas unconstrained firms did not manage their earnings. Similarly, the findings show that the family business group-affiliated firms are not financially constrained as compared to stand-alone firms and the results show that affiliated firms are negatively related to EM on constrained firms’ criteria, whilst stand-alone firms are positively related to EM.

This study contributes to the related literature in several ways. First, this research has extended the rare and limited literature related to family business groups and earnings manipulation. Second, this research study provides evidence from Pakistan, an emerging economy, which has an ideal setting for conducting such a type of study; therefore, this research study has utilized the ideal settings and filled the gap related to earnings manipulation in family business groups of Pakistan. Third, in this study, EM is linked with financially constrained and unconstrained along with affiliated and non-affiliated firms, which the previous literature is lacking considering family business groups. Last but not the least, the previous literature (Beuselinck & Deloof, 2014; Kim & Yi, 2006; Muttakin et al., 2017; Sarkar et al., 2013) has used only one proxy of accrual EM, whilst this research study examines the impact of group affiliation on both types of earnings manipulation proxies, which are accrual-based EM and real activity EM. Moreover, we find out that group affiliation is a negative relationship with EM, which is perceived to be positive in emerging economies. Overall, this paper supplied further evidence of business groups and EM literature; since there is less work on it, this study has shed more light on the EM issue and extended the understanding of EM in (family) business groups.

The rest of this paper is organized in the following order: in the next section, the background of research settings is given, section three describes the related theory and literature and in the fourth section, hypotheses for the study are developed. In section five, the research design and methodology are given, whilst analysis and discussion on the results are given in section six. The study ends on conclusion and limitations.

2. Background of the research context/settings

Business groups mean a group of legally independent firms most often controlled by a business family in the industrial landscape of emerging economies (Pattnaik et al., 2018). Business groups are a common phenomenon in South Asia (Bangladesh, India and Pakistan) and they are considered one of the most well-known features of the Asian stock markets. These family business groups play a key role in the development of emerging economies such as Pakistan. The business groups’ origin in Pakistan can be traced back to Pakistan’s independence (partition, 1947). After the independence, the wealthiest Muslims family migrated to newly established Pakistan from the various parts of undivided India (Ahmad et al., 2018) and later due to the former President of Pakistan Ayub Khan’s (1958–1969) economic policies to promote private sector based on the free-market economy, these business groups have become in the 1960s and 1970s the leading business families (family business groups) in Pakistan (Papanek, 1972). Moreover, in 1968, Dr Mahbob-ul-Haq, a former Chief Economist of the Planning Commission of Pakistan, coined the word “Twenty-Two Families” for family business groups in Pakistan in a speech and added that these 22 families (family business groups) have domination in the economic and financial life in Pakistan by controlling two-third of industrial assets.

Furthermore, White (1974) added in his book that more than fifty percent of total assets of listed non-financial firms of the Karachi Stock Exchange are controlled by 43 business groups (families), whilst in the private sectors, these top 43 business families controlled 74 percent of total assets. In addition to the work of White (1974), in 1998, a book published by (Shahid-ur-Rahman, 1998) titled “Who Owns Pakistan” has extended the studies related to business groups operating in Pakistan and extended the number of family business groups by adding 22 more business families. Furthermore, Shahid-ur-Rahman, (1998) added that there are two sets of 22 families, the 22 families as mentioned by Dr Mahbob-ul-Haq the original 22 families and the present era (1997–98) 22 families. Shahid-r-Rahman (Shahid-ur-Rahman, 1998) added that out of total of 522 non-financial companies, the top 43 business groups in Pakistan own 212 companies that are controlling 43 percent of the total manufacturing assets; in addition to this, in financial companies, these 43 top business groups own 76 financial companies out of 175 listed financial companies. In addition to this, the ownership structure of Pakistan listed firms is mostly dominated by family business groups or family-owned business either directly by directors or promoters’ ownership or indirectly through associated companies’ ownership (ICMAP, 2011). Moreover, Haque and Husain (2021) investigated that 31 family business groups dominated the Karachi Stock Exchange (hereafter, KSE) 100 Index and there is extremely high connectivity between the board of directors of KSE 100 Index firms.

At present, we have found out that there are 60 family business groups operating in Pakistan and these business groups are collections of public and private companies. Some groups are as large as having 9–10 (such as Dewan Group and Saigol Group) affiliated listed firms on Pakistan Stock Exchange (hereafter, PSX) and as small as having zero or 1 (such as Fazal Group and Indus Group) listed firms on PSX. These business groups are simultaneously called “family business groups” or “Group of Companies”, often the group name starts with a leading/founder family member. Moreover, Pakistani family business groups work as a conglomerate and are diversified horizontally by having companies in multiple sectors and controlling these companies via common ownership, cross-holding and interlocking directors. By having discussed the background of family business groups in Pakistan, it is evident that family business groups have economic dominance in the corporate landscape in Pakistan. Moreover, due to its importance, many scholars/researchers have linked family business groups such as Hussain (2019) & Ullah (2014) with tunneling, Ghani and Ashraf (2005), Ghani et al. (2011) and W. Ullah (2017) had linked business groups with firm performance, Bhutta and Suleman (2017) had linked business groups with the capital structure and Bhutta et al. (2016) had linked business groups with ownership concentration. Therefore, there are limited theoretical discourse and empirical evidence on the family business group-affiliated firms and earnings manipulation, in general, and Pakistan, in particular, so Pakistan offers us an excellent setting to test the relationship of earnings manipulation in family business groups of Pakistan. Therefore, we are intending to utilize Pakistani family business group-dominated corporate landscape to investigate the earnings manipulation practices in family business group-affiliated firms.

3. Related theory and literature

The classical agency issue between shareholders and management (Type I) is valid in those contexts where the phenomenon of widespread ownership is prevailing and the gap of ownership and control is very wide such as UK and USA. However, in the case of Asia, Latin America and some parts of Europe, the gap of ownership and control is very low due to concentrated ownership and this concentration of ownership is in hands of a family, bank or the state (La Porta et al., 1999). Firms with concentrated or blockholders ownership have characterized with the absence of the classical agency Type I problem and have another agency problem, Type II, which is between majority shareholders and dispersed minority shareholders (Calabrò et al., 2017; Young et al., 2008). Hussain (2019) added that various concentrated ownership firms in emerging economies are affiliated with different business groups or Family Business Groups. Business groups are a common phenomenon in South Asia (Bangladesh, India and Pakistan) and they is considered one of the most well-known features of the Asian stock markets (Kouwenberg & Thontirawong, 2016). Business groups operate in the pyramidal ownership structure in these countries. According to Claessens et al. (2000), business groups operate under cross-shareholdings, boards interlocks, common management and pyramids ownership structure. A pyramidal business group is a collection of various independent firms usually controlled by a family through the pyramidal ownership structure at various levels (Almeida & Wolfenzon, 2006), in these groups, there is a vast difference between cash flow rights (hereafter, CFR) and control rights (hereafter, CR) which the family uses to facilitate its control. This difference between CFR and CR gives escalation to the severe agency problem between minority shareholders and controlling (family) shareholders, which may be used to expropriate minority shareholders. In addition to this, Basheer et al. (2021) reported that the Type II Agency Problem is more severe in family-owned business groups, which results in expropriation of minority shareholders rights. This expropriation may take place at the second or third level of the pyramid via tunneling, propping or related party transactions among affiliates; moreover, the management of these companies relates the propping, tunneling or related party transaction to EM.

The related literature shows that business group’s affiliates use different methods to support poor performing affiliates such as cash and accrual-based related party transaction (propping), using transferring resources (tunneling) within groups and earnings manipulations (Bertrand et al., 2002; Hussain & Safdar, 2018b; Jian & Wong, 2010). Beuselinck and Deloof (2014) added that structures of related-party transaction among the group affiliates provide not only extra opportunities but also extra instruments to group affiliates to manage their earnings more as compared to stand-alone companies. Therefore, the real financial position of a firm is masked through these related-party transactions via group affiliation and it is clear that family business groups' firms have more incentives and opportunities to involve in EM as compared to non-affiliated group firms. According to Wang and Yung (2011), in the family firm’s governance, the structure is weak due to family firms controlling rights, which offered them a stronger position to get private gains by using its voting rights and executive positions and expropriate the resources of minority shareholders. So, to conclude the above argument, the pyramidal ownership structure provides more opportunities for the controlling shareholders to engage in earnings manipulation in order to cover up its expropriating activities in family business groups and suffers from the Type II Agency Problem (Claessens et al., 2000; Young et al., 2008).

4. Hypothesis development of the study

The analysis of the Pakistani corporate sector shows that the majority of companies have concentrated ownership and have affiliation of family business groups (Hussain & Safdar, 2018b; Ikram & Naqvi, 2005; White, 1974). The ownership structure of family business groups is often structured in pyramids (pyramids ownership structure) and family business groups are known by cross-shareholding and common directorships in different affiliated family business firms. As compared to stand-alone firms, group-affiliated firms have more advantages and incentives than stand-alone firms. According to Muttakin et al. (2017), business group-affiliated firms have the position to easily obtain the valuable equipment, the expert and talented employees, financial resources and modern technology, whilst stand-alone firms may not do so. Moreover, through the transfer of resources with group affiliates (Khanna & Yafeh, 2005) and through sharing different types of assets, business groups can achieve economies of scale (Chang & Hong, 2000).

Furthermore, the related literature shows that business group affiliates use different methods to support poor performing affiliates such as cash and accrual-based related party transaction (propping), using transferring resources (tunneling) within groups (Bertrand et al., 2002; Hussain & Safdar, 2018b; Jian & Wong, 2010). Beuselinck and Deloof (2014) added that structures of related-party transaction among the affiliates provide not only extra opportunities but also extra instruments to the group affiliates to manage their earnings more as compared to stand-alone companies. Jian and Wong (2010) added that as compared to stand-alone companies, business group firms have more incentives and tools to engage in EM. Additionally, Business groups are known for the complicated ownership structure and this complicated structure leads business groups to questionable activities, which may harm widespread minority shareholders via EM (Ullah, 2014). Moreover, the studies of Beuselinck & Deloof, 2014; Kim & Yi, 2006; Muttakin et al., 2017 reported that business group affiliation has a positive relationship with accrual-based earnings manipulation; moreover, they added that business group affiliation tends affiliated firms to manage their earnings more than non-affiliated group firms. In addition to this, business group affiliation provides more incentives and chances to the majority of shareholders to manage earnings at the expense of minority shareholders because the family business group affiliated companies have more opportunities and incentives to divert resources from one affiliate to another affiliate as compared to non-group affiliated firms (Kim & Yi, 2006). So, to conclude the above findings, it is clear that family business group firms have more incentives and opportunities to involve in EM as compared to non-affiliated group firms. Based on the aforementioned discussion, we hypothesized the first hypothesis of this study in the following words:

H1a: Family Business Group-affiliated firms engage more in earnings management (accrual) than stand-alone firms.

H1b: Family Business Group-affiliated firms engage more in earnings management (Real Activity EM) than stand-alone firms.

Contrary to the above-mentioned view, the study of CG and family business group literature depicted two opposite views or hypotheses: “The Conflict of Interest Hypothesis” as suggested by Berle & Means, 1932; Jensen & Meckling, 1976 and “The Efficient Transaction Hypothesis” by Gordon et al., 2004; Ullah, 2014. According to first hypothesis, family business groups expropriate resources or underperformed, whilst on the second view, family business groups improve the financial performance of the affiliates or involve in propping activates (Hussain & Safdar, 2018b). According to the “The Efficient Transaction Hypothesis”, business groups timely fulfill their needs and take care of all their stakeholders whilst doing business transaction with different parties (including related parties) and therefore may curb management in manipulating earnings. Moreover, the said hypothesis could be linked with “Alignment Hypothesis”, which states that the interest of family insider management and outside dispersed shareholders is aligned, and both the parties want to increase the firm value and earnings manipulation on behalf management, which may harm the firm’s value and reputation. Furthermore, according to Kim and Yi (2006), business groups provide opportunities for affiliated group companies to create value by sharing different types of group resources. Moreover, Kim (2019) depicted that accrual-based earnings manipulation is very costly for insiders (family directors) when it is affiliated with business groups. Moreover, keeping in view the Agency Theory, it advocates that the general incentives for an entity to involve in EM may be the reward and different incentives, which inspire management to involve in earnings manipulation, but family business group-affiliated firms often have in the high level of management the family or founding family members, who are aligned with firm value and reputation and they may not be involved in EM. In addition to this, the studies of Kim, 2019; Mindzak & Zeng, 2018; Zhu, 2020 reported that business group affiliates are negatively related to accrual-based EM, therefore, the second hypothesis of this study is

H2a: Family Business Groups' affiliated firms is negatively related to earnings management (accrual) than stand-alone firms.

H2b: Family Business Groups' affiliated firms is negatively related to earnings management (Real Activity EM) than stand-alone firms

5. Methodology

5.1. Sample selection and data

The sample of this study comprises the non-financial listed firms of the PSX. The initial sample for this study was 543 listed firms. The financial firms (132) have been excluded due to the difference in the accruals procedure and accounting policies as mentioned by Chee & Tham, 2021; Peasnell et al., 2005; Saona et al., 2020 in the concerned literature. In addition to this, companies with missing annual reports, defaulter companies and those companies, which were delisted, were also excluded; in addition to this, state own enterprises and foreign companies are also excluded (96) and the remaining (323) companies’ data were used in this study for the observable period of 2014–2019 consisting of 1329 to 1927 firm years observations due to the unbalanced panel. In the 323 companies, 187 are family business groups's affiliated firms, whilst 136 are non-affiliated or stand-alone firms. We have found out that there are 60 business groups operating in Pakistan.

The details about sample selection procedures and sample distribution are given in Tables 1 and 2. The final sample consists of the data of 27 sectors of PSX after excluding the financial sectors.

Table 1. Procedures of sample selection

Total firms listed in the PSX in June 2021	551
Less: Financial firms	(132)
	419
Less: Annual Reports Missing, Defaulter and Delisted firms for the said period, State Own enterprises and Foreign Companies	(96)
Firms selected from the study	323

Table 2. Sample distribution by year and industry

	PSX Sector Name		Years						Total by Industry	%
S.No.		Code	2014	2015	2016	2017	2018	2019
1	Automobile Assembler	801	12	12	12	12	12	12	72	0.04
2	Automobile Parts & Accessories	802	7	7	7	7	7	7	42	0.02
3	Cable & Electrical Goods	803	5	5	5	5	5	5	30	0.02
4	Cement	804	20	20	20	20	20	20	120	0.06
5	Chemical	805	27	27	27	27	27	27	162	0.08
6	Engineering	808	11	11	11	11	11	11	66	0.03
7	Fertilizer	809	5	5	5	5	5	5	30	0.02
8	Food & Personal Products	810	17	17	17	17	17	17	102	0.05
9	Glass & Ceramics	811	8	8	8	8	8	8	48	0.02
10	Leather & Tanneries	816	1	1	1	1	1	1	6	0
11	Miscellaneous	818	10	10	10	10	10	10	60	0.03
12	Oil & Gas Exploration Co	820	4	4	4	4	4	4	24	0.01
13	Oil & Gas Marketing Co	821	8	8	8	8	8	8	48	0.02
14	Paper & Board	822	8	8	8	8	8	8	48	0.02
15	Pharmaceuticals	823	10	10	10	10	10	10	60	0.03
16	Power Generation & Distribution	824	14	14	14	14	14	14	84	0.04
17	Refinery	825	5	5	5	5	5	5	30	0.02
18	Sugar & Allied Industries	826	29	29	29	29	29	29	174	0.09
19	Synthetic & Rayon	827	8	8	8	8	8	8	48	0.02
20	Technology & Communication	828	13	13	13	13	13	13	78	0.04
21	Textile Composite	829	54	54	54	54	54	54	324	0.17
22	Textile Spinning	830	30	32	31	31	32	31	187	0.1
23	Textile Weaving	831	5	5	5	5	5	5	30	0.02
24	Tobacco	832	2	2	2	2	2	2	12	0.01
25	Transport	833	4	4	4	4	4	5	25	0.01
26	Vanaspati & Allied Industries	834	2	2	2	2	2	1	11	0.01
27	Woolen	835	1	1	1	1	1	1	6	0
	Total by Years		320	322	321	321	322	321	1927	100

The data for this study have been manually collected from the annual reports, concerned companies’ websites and SBP publications. For the Family Business groups, the studies of Hussain, 2019; Hussain & Safdar, 2018a; Shah, 2009 have been used as a preliminary sources; later, it was further searched and confirms form pattern of shareholdings, category of shareholders, ownership structure and concerned business groups' websites; later on, it was coded in excel. For the board of directors and financial data in addition to the above-mentioned sources, the SBP publications were used for data extraction.

5.2. Variables measurement

5.2.1. Dependent variable

The discretionary accrual estimated under Jones Model (Jones, 1991) is the proxy used in this research study for earnings manipulation, consistent with the studies of Beuselinck & Deloof, 2014; Kim & Yi, 2006; Muttakin et al., 2017. The Jones Model developed by Jones, 1991 is considered the standard model of discretionary accruals because every other model has been established by modifying the Jones Model. In addition to this, in the emerging economies, the said model has been used for estimating EM (Khan et al., 2020; Kontesa et al., 2021; Muttakin et al., 2017); it is estimated as follows

(1) $\frac{AC{C}_{IT}}{LAGT{A}_{IT}}={\beta }_o+{\beta }_1(1|\mathrm{L}\mathrm{A}\mathrm{G}\mathrm{T}{\mathrm{A}}_{\mathrm{I}\mathrm{T}})+\hspace{0.17em}{\beta }_2(\mathrm{\Delta }SALE{S}_{IT}|\mathrm{L}\mathrm{A}\mathrm{G}\mathrm{T}{\mathrm{A}}_{\mathrm{I}\mathrm{T}})+\hspace{0.17em}{\beta }_3(\mathrm{P}\mathrm{P}{\mathrm{E}}_{IT}|\mathrm{L}\mathrm{A}\mathrm{G}\mathrm{T}{\mathrm{A}}_{\mathrm{I}\mathrm{T}})+{ϵ}_{IT},$(1)

where ACC_IT is the accruals as calculated by subtracting cash flow from operation from net income after tax as reported in the annual report, $\mathrm{\Delta }SALE{S}_{IT}$ is the change in sales in t year, PPE_IT is the property plant and equipment for year t and $LAGT{A}_{IT-1}$ is the lagged value of total assets in year t.

Moreover, in this research study, Real Activity Earnings Manipulation (hereafter, REM) proxies have also been used in order to check its relationship with business group dummy. This study followed (Roychowdhury, 2006) for REM proxies, which are calculated in the following equations:

(2) $\frac{CF{O}_{IT}}{LAGT{A}_{IT}}={\beta }_o+{\beta }_1(1|\mathrm{L}\mathrm{A}\mathrm{G}\mathrm{T}{\mathrm{A}}_{\mathrm{I}\mathrm{T}})+\hspace{0.17em}{\beta }_2(SALE{S}_{IT}|\mathrm{L}\mathrm{A}\mathrm{G}\mathrm{T}{\mathrm{A}}_{\mathrm{I}\mathrm{T}})+\hspace{0.17em}{\beta }_3(\mathrm{\Delta }SALE{S}_{IT}|\mathrm{L}\mathrm{A}\mathrm{G}\mathrm{T}{\mathrm{A}}_{\mathrm{I}\mathrm{T}})+{ϵ}_{IT},$(2)

(3) $\frac{PRO{D}_{IT}}{LAGT{A}_{IT}}={\beta }_o+{\beta }_1(1|\mathrm{L}\mathrm{A}\mathrm{G}\mathrm{T}{\mathrm{A}}_{\mathrm{I}\mathrm{T}})+\hspace{0.17em}{\beta }_2(SALE{S}_{IT}|\mathrm{L}\mathrm{A}\mathrm{G}\mathrm{T}{\mathrm{A}}_{\mathrm{I}\mathrm{T}})+\hspace{0.17em}{\beta }_3(\mathrm{\Delta }SALE{S}_{IT}|\mathrm{L}\mathrm{A}\mathrm{G}\mathrm{T}{\mathrm{A}}_{\mathrm{I}\mathrm{T}})+{\beta }_2(\mathrm{\Delta }SALE{S}_{IT-1}|\mathrm{L}\mathrm{A}\mathrm{G}\mathrm{T}{\mathrm{A}}_{\mathrm{I}\mathrm{T}})+{ϵ}_{IT},$(3)

(4) $\frac{DISEX{P}_{IT}}{LAGT{A}_{IT}}={\beta }_o+{\beta }_1(1|\mathrm{L}\mathrm{A}\mathrm{G}\mathrm{T}{\mathrm{A}}_{\mathrm{I}\mathrm{T}})+\hspace{0.17em}{\beta }_2(SALE{S}_{IT}|\mathrm{L}\mathrm{A}\mathrm{G}\mathrm{T}{\mathrm{A}}_{\mathrm{I}\mathrm{T}})+\hspace{0.17em}{ϵ}_{IT},$(4)

where CFO_IT is the cash flow form operation for t year and $LAGT{A}_{IT-1}$ is the lagged value of total assets in year t. $SALE{S}_{IT}$ is the sales for the year t and $\mathrm{\Delta }SALE{S}_{IT}$ is the change in sales in year t. Moreover, PROD_IT is the production cost equal to the cost of goods sold minus delta inventory, whilst DISEXP_IT is the sum of Selling, Administrative, Marketing and Research and development expenditure.

5.2.2. Explanatory variables

This study defines Family Business Group Affiliation (GA) as a dummy variable taking the value of 1 if listed firms are affiliated with business groups and zero otherwise consistent with studies of (Beuselinck & Deloof, 2014; Kim & Yi, 2006; Mindzak & Zeng, 2018; Muttakin et al., 2017). In the board of directors’ variables, the board independence (BIND) is measured by dividing the number of independent directors by the board size; in the same way, the board expertise (BEXP) and female directors (BDIV) are also estimated. The Audit Committee variables such as AC independence (ACIND) are estimated by dividing the number of independent directors by the size of AC.

5.2.3. Control variables

Consistent with EM and Business group literature, this study has used a battery of control variables in order to tackle the miss-specification problem in the model of the study (Saona et al., 2020). These variables include Big 4 audit firms (BIG4), loss dummy LOSS(D), leverage(LEVE), absolute value return of assets (ABSROA), log of total assets (LNTA), delta sales (SALES), Inventory over assets, (INV), fixed assets over total assets (TANG), age of the firms (FAGE) and firm size(FSIZE). The details of the study’s variables are given in Appendix 1.

5.2.4. Hypothesis testing

In order to test the study’s hypotheses, the following equation will be used:

(5) $DR{M}_{it}=\theta deg+{\theta }_{1}G{A}_{it}+{\theta }_{2}BEX{P}_{it}+{\theta }_{3}BDI{V}_{it}+{\theta }_{4}ACIN{D}_{it}+{\theta }_{4}BIN{D}_{it}+\sum _{M}pmControl+{\mu }_{it},$(5)

(6) $RA{M}_{it}=\theta deg+{\theta }_{1}G{A}_{it}+{\theta }_{2}BEX{P}_{it}+{\theta }_{3}BDI{V}_{it}+{\theta }_{4}ACIN{D}_{it}+{\theta }_{4}BIN{D}_{it}+\sum _{M}pmControl+{\mu }_{it}.$(6)

In Equations 5(5) $DR{M}_{it}=\theta deg+{\theta }_{1}G{A}_{it}+{\theta }_{2}BEX{P}_{it}+{\theta }_{3}BDI{V}_{it}+{\theta }_{4}ACIN{D}_{it}+{\theta }_{4}BIN{D}_{it}+\sum _{M}pmControl+{\mu }_{it},$(5) and 6(6) $RA{M}_{it}=\theta deg+{\theta }_{1}G{A}_{it}+{\theta }_{2}BEX{P}_{it}+{\theta }_{3}BDI{V}_{it}+{\theta }_{4}ACIN{D}_{it}+{\theta }_{4}BIN{D}_{it}+\sum _{M}pmControl+{\mu }_{it}.$(6) , the dependent variableDRM is discretionary accrual calculated under the Jones, 1991 model, whilst RAM is proxies calculated under the Roychowdhury, 2006 model. The predictors are BIND for board independence, BEXP for board expertise, BDIV for the female directorship and ACIND for Audit Committee Independence; moreover, control is for a battery of control variables; the details are fully given in appendix 1.

6. Results of the study

6.1. Descriptive Statistics

In Table 3, the descriptive statistics of family business group-affiliated firms is given. The average value of discretionary accruals estimated under theJones Model is .10, whilst the min is .01 and the maximum is .27; on the other hand, the mean value of aggregate real activity manipulation is .09, whilst −2.39 is minimum and .96 is the maximum value. In addition to this, Independence in board is 17 percent on average, whilst female directorship is 08 percent and expert directors are 34 percent on average in family business group-affiliated firms

Table 3. Descriptive statistics of family business group-affiliated firms

	N	Mean	Median	SD	Min	Max
Panel A: Descriptive Statistics of Dependent Variables
JMDA	888	.101	.074	.083	.012	.27
CFO RM	888	−.002	.002	.125	−.218	.204
PROD RM	883	.012	.006	.158	−.256	.285
DIS ERM	1080	.16	.079	.63	−3.008	9.492
REALEM	880	.096	.058	.291	−2.396	.967
Panel B: Descriptive Statistics of Board of Directors Variables
BIND	1100	.173	.143	.078	.091	.333
BEXP	823	.344	.3	.153	.143	.571
BDIV	1102	.08	0	.118	0	.571
ACIND	1099	.292	.333	.134	0	.5
Panel C: Descriptive Statistics of Firm Characteristics Variables
TANG	1082	.444	.452	.214	.101	.797
FSIZE	1082	8.68	8.64	1.31	6.37	10.7
LEVE	1083	.538	.538	.205	.213	.924
FAGE	1122	37.0	35	15.0	15	63
ABSROA	1083	7.67	6.227	5.93	.009	19.8
INV	1087	371.794	3.52	3291.954	−16600	64165.68
LN TA	1097	15006.84	5370.9	26252.87	−70.38	417080
Panel D: Dummy Variables
Categories	1	0
BIG4	585(53%)	529(47%)
LOSS(D)	273(25%)	839(75%)
CEOD	44(4%)	1057(96%)

Moreover, audit committee independence is on average 29 percent, its minimum value is zero and the maximum value is 50 percent of the audit committee. The family business groups’ affiliated firms are 53 percent engage Big 4 audit firms for auditing, whilst 47 percent firms engage non-Big 4 audit firms. In addition to this, only 25 percent of affiliated firms have reported a loss for the period of 2014–2019, whilst 75 percent reported a profit for the said period.

Moreover, in the stand-alone firms, 51 percent used Big 4 audit firms, whilst 49 percent of firms used non-Big 4 audit firms. As compared to affiliated firms, stand-alone firms reported more loss; the figure of 32 % shows that 32 percent stand-alone firms reported loss in the study period ranging from 2014 to 2019.

In Table 4, the descriptive Statistics of stand-alone firms are presented. The average value of earnings management estimated under the Jones Model is .102 slightly high as compared to affiliated firms, whilst real earnings manipulation proxies values are high as compared to affiliated firms; this finding indicates that stand-alone firms manage earnings more as compared to affiliated firms. Additionally, board independence and female directorship are 18 and 14 on average, which is quite high as compared to affiliated firms, whilst on the other hand, the expert directors in the board has a mean value of .30, which is low as compared to affiliated firms.

Table 4. Descriptive statistics of stand-alone firms

	N	Mean	Median	SD	Min	Max
Panel A: Descriptive Statistics of Dependent Variables
JMDA	623	.102	.079	.082	.012	.27
CFO RM	623	.003	.007	.128	−.218	.204
PROD RM	611	−.005	−.007	.169	−.256	.285
DIS ERM	751	.152	.093	.496	.003	9.102
REALEM	608	.103	.07	.281	−.786	.964
Panel B: Descriptive Statistics of Board of Directors Variables
BIND	736	.184	.143	.083	.091	.333
BEXP	506	.307	.286	.144	.143	.571
BDIV	739	.143	.111	.173	0	1
ACIND	729	.294	.333	.144	0	.5
Panel C: Descriptive Statistics of Firm Characteristics Variables
TANG	754	.465	.469	.234	.101	.797
FSIZE	755	8.411	8.314	1.447	6.373	10.76
LEVE	755	.561	.564	.238	.213	.924
FAGE	798	37.353	33	15.758	15	63
ABSROA	755	8.24	7.40	6.24	.004	19.8
INV	762	1012.777	−2.383	9355.761	−29100	122000
LN TA	761	31315.15	4232.21	89305.81	0	766597
Panel D: Dummy Variables
Categories	1	0
BIG4	401(51%)	385(49%)
LOSS(D)	244(32%)	528(68%)
CEOD	17(2%)	724(98%)

In Table 5, the full sample descriptive statistics is given. The average mean value of the EM proxy is 10, whilst 0.1 is minimum and .27 is maximum. The real activity manipulation proxies range from mean .005 to .15 and the minimum ranges from 0.12 to .58, whilst the maximum ranges from .20 to 9.40. The average independence in Pakistani listed firms is .17, which means that out of ten board members, there are only 1.7 independent directors, which is quite low. Furthermore, the average number of expert members in the board of Pakistani firms is 33 percent. The female directorship is 10 percent on average in the sample size firms. At last, this table shows that 58 percent out of 323 firms are affiliated with family business groups, whilst 42 percent are stand-alone firms.

Table 5. Full sample descriptive statistics

	N	Mean	Median	SD	Min	Max
Panel A: Descriptive Statistics of Dependent Variables
JMDA	1515	.102	.076	.083	.012	.27
CFO RM	1515	0	.004	.126	−.218	.204
PROD RM	1498	.005	.004	.163	−.256	.285
DIS ERM	1841	.159	.085	.581	−3.008	9.492
REALEM	1492	.099	.063	.286	−2.396	.967
Panel B: Descriptive Statistics of Board of Directors Variables
BIND	1836	.177	.143	.08	.091	.333
BEXP	1329	.33	.286	.151	.143	.571
BDIV	1841	.106	0	.146	0	1
ACIND	1828	.293	.333	.138	0	.5
Panel C: Descriptive Statistics of Firm Characteristics Variables
TANG	1843	.453	.457	.223	.101	.797
FSIZE	1844	8.567	8.528	1.378	6.373	10.76
LEVE	1845	.547	.548	.219	.213	.924
FAGE	1933	37.227	34	15.359	15	63
ABSROA	1845	7.89	6.62	6.05	.004	19.8
INV	1857	632.833	−.22	6506.333	−29100	122000
LN TA	1869	21571.97	4940.66	60954.83	−70.38	766597
Panel D: Dummy Variables
Categories	1	0
BIG4	991(52%)	915(48%)
LOSS(D)	519(27%)	1370(73%)
CEOD	61(3%)	1781(97%)
GA	1122(58%)	816(42%)

6.2. Mean and maiden difference tests

In Table 6, the mean difference test of the Two-Sample t-test and median difference test of the Mann-Whitney Test are presented. In Panel A, dependent variables are given, among them, the real activity management proxies of abnormal discretionary expenses and abnormal production cost are significantly different in the family business groups' affiliated firms and non-affiliated firms. The Panel B board of directors’ variables are presented in which board independence, board expertise and board diversity are significantly different in the affiliated and non-affiliated firms.

Table 6. Mean and median test of differences

	Affiliated Firms		Non-Affiliated Firms		Test of difference
	Mean	Median	Mean	Median	t-stat.	z-stat.
Panel A: Dependent Variables
JMDA	.101	.074	.102	.079	.25	0.38
CFO RM	−.002	.002	.003	.007	.85	0.913
PROD RM	.012	.006	−.005	−.007	−1.95**	−1.97**
DIS ERM	.16	.079	.152	.093	−.3	2.35***
REALEM	.096	.058	.103	.07	.45	0.08
Panel B:Board of Directors Variables
BIND	.173	.143	.184	.143	2.85***	2.78***
BEXP	.344	.3	.307	.286	−4.35***	−4.10***
BDIV	.08	0	.143	.111	9.30***	8.19***
ACIND	.292	.333	.294	.333	.350	1.22
Panel C: Descriptive Statistics of Firm Characteristics Variables
TANG	.444	.452	.465	.469	2.00***	2.04**
FSIZE	8.687	8.643	8.411	8.314	−4.25***	−4.30***
LEVE	.538	.538	.561	.564	2.15***	1.96**
FAGE	37.094	35	37.353	33	.35	0.08
ABSROA	7.67	6.227	8.244	7.403	2.00***	1.79
BIG4	.474	0	.489	0	.65	0.64
LOSS(D)	.245	0	.316	0	3.4***	3.37***
CEOD	.039	0	.02	0	−2.05***	−2.00***
INV	371.794	3.52	1012.777	−2.383	2.1**	−1.06
LN TA	15006.84	5370.9	31315.15	4232.21	5.70***	−2.64***

Board independence and board female directors are high in the non-affiliated firms as compared to affiliated firms, whilst board expertise is higher in affiliated firms as compared to non-affiliated firms. Moreover, in the control variables, loss dummy, CEO duality and leverage are also significantly different in the affiliated and non-affiliated firms, CEO duality is more in the affiliated firms as compared to non-affiliated firms and if we look at the leverage and loss, Table 6 shows that leverage and loss are low in the family business group-affiliated firms as compared to non-affiliated firms. The findings of Table 6 show that stand-alone firms suffer more loss and have more leverage than family business groups' affiliated firms.

6.3. Pair-wise correlation matrix

In Table 7 and 8', the result of pair-wise correlation matrix is given; it shows the relationship between dependent and independent variables along with control variables. Its values indicate that there is no problem of multicollinearity because no value is close to .7 or more except for real activity manipulation proxies, which is theoretically correct and has no issue in the results. In addition to this, the dummy variable of GA (Group Affiliation) is negatively related to JMDA (discretionary accruals estimated under the jones Model) and this indicates that GA decreases earnings manipulation, which is also depicted in the regression results.

Table 7. The diagnostic tests before regression

Variance Inflation Factor
Variables	GA	TANG	ACIND	CEOD	FSIZE	LEVE	BIND	BEXP
VIF	1.331	1.625	1.765	1.177	2.16	1.524	1.911	1.431
1/VIF	0.751	0.616	0.567	0.85	0.463	0.656	0.523	0.699
Variables	FAGE	BIG4	LOSS(D)	BDIV	INV	LN_TA	ABSROA	Mean VIF
VIF	1.43	1.847	1.436	1.288	1.787	3.276	1.533	1.802
1/VIF	0.699	0.541	0.696	0.776	0.559	0.305	0.652	.

Breusch-Pagan/Cook-Weisberg Test for Heteroscedasticity

Ho: Constant variance

Variables: fitted values of abs-JM

χ2 (1) = 54.16

χ2 > chi2 = 0.0000

Wooldridge Test For Autocorrelation

H0: no first-order autocorrelation

F (1, 197) = 0.687

F-Statistics = 0.180

Table 8. Pair-wise correlation matrix

Variables	JMDA	CFO_RM	PROD_RM	DIS_ERM	REALEM	GA	TANG	ACIND	CEOD	FSIZE	LEVE	BIND	BEXP	FAGE	BIG4	LOSS(D)	BDIV	ACC	LN_TA	ABSROA
JMDA	1.000
CFO_RM	−0.10*	1.000
PROD_RM	0.087*	−0.784*	1.000
DIS_ERM	0.023	−0.028	0.007	1.000
REALEM	0.061*	−0.723*	0.829*	0.253*	1.000
GA	−0.07	−0.021	0.051*	0.008	−0.012	1.000
TANG	−0.17*	0.003	0.158*	−0.076*	0.106*	−0.04*	1.000
ACIND	−0.038	0.046	−0.088*	−0.006	−0.114*	−0.009	−0.013	1.000
CEOD	−0.016	0.003	0.009	−0.017	−0.012	0.047*	0.089*	−0.099*	1.000
FSIZE	0.007	0.213*	−0.172*	−0.165*	−0.347*	0.099*	−0.16*	0.125*	−0.05*	1.000
LEVE	0.092*	−0.192*	0.296*	0.004	0.197*	−0.05*	0.269*	−0.014	0.025	−0.043	1.000
BIND	0.015	0.057*	−0.120*	0.018	−0.096*	−0.06*	−0.039	0.569*	−0.07*	0.100*	0.002	1.000
BEXP	0.053	0.096*	−0.116*	−0.004	−0.116*	0.118*	−0.17*	0.079*	0.002	0.249*	−0.15*	0.197*	1.000
FAGE	−0.030	0.026	−0.076*	−0.014	−0.052*	−0.008	−0.10*	0.088*	−0.009	0.081*	−0.13*	0.067*	0.104*	1.000
BIG4	0.018	0.195*	−0.261*	−0.059*	−0.259*	−0.015	−0.24*	0.092*	−0.08*	0.445*	−0.21*	0.149*	0.282*	0.071*	1.000
LOSS(D)	−0.015	−0.338*	0.389*	0.067*	0.365*	−0.07*	0.326*	−0.045	0.070*	−0.33*	0.392*	−0.018	−0.15*	−0.04*	−0.29*	1.000
BDIV	−0.05*	−0.018	0.030	0.020	0.069*	−0.21*	0.150*	−0.059*	0.074*	−0.23*	0.068*	−0.04*	−0.17*	−0.08*	−0.165*	0.056*	1.000
INV	0.078*	−0.083*	−0.040	−0.011	−0.041	−0.04*	−0.159*	0.045	−0.027	0.143*	−0.109*	0.086*	0.023	0.065*	0.102*	−0.094*	−0.056*	1.000
LN_TA	−0.007	0.107*	−0.059*	−0.057*	−0.103*	−0.13*	−0.091*	0.080*	−0.036	0.389*	0.017	0.141*	0.043	0.117*	0.230*	−0.103*	−0.099*	0.301*	1.000
ABSROA	0.158*	0.374*	−0.452*	−0.019	−0.329*	−0.04*	−0.268*	0.122*	−0.084*	0.178*	−0.280*	0.109*	0.127*	0.047*	0.328*	−0.248*	−0.093*	0.163*	0.059*	1.000

shows significance at the 0.05 level

6.4. Multivariate regression results

As the objective of the study is to assess whether family business group affiliated firms are influential in changes of management EM practices as proxied by accruals and real activity manipulation. The data used for this study are the panel (unbalance), so various diagnostic tests are made prior to panel data estimation. Table 7 shows the results of multicollinearity, heteroscedasticity and autocorrelation. The results of the test show that there is no issue of multicollinearity and autocorrelation with the corresponding VIF values and F-Statistics values indicate no serious data issues; however, there is an issue of heteroscedasticity and the χ2 > chi2 value of 0.000 indicates that there is a heteroscedasticity problem and therefore, the robust standard error is used in regression analysis in order to tackle this issue. After the diagnostic test, this study has estimated Equation 5(5) $DR{M}_{it}=\theta deg+{\theta }_{1}G{A}_{it}+{\theta }_{2}BEX{P}_{it}+{\theta }_{3}BDI{V}_{it}+{\theta }_{4}ACIN{D}_{it}+{\theta }_{4}BIN{D}_{it}+\sum _{M}pmControl+{\mu }_{it},$(5) to test the study’s hypotheses.

The main regression results are presented in Tables 9 and 10 to test the hypotheses of the study. This research paper has utilized Equation 5(5) $DR{M}_{it}=\theta deg+{\theta }_{1}G{A}_{it}+{\theta }_{2}BEX{P}_{it}+{\theta }_{3}BDI{V}_{it}+{\theta }_{4}ACIN{D}_{it}+{\theta }_{4}BIN{D}_{it}+\sum _{M}pmControl+{\mu }_{it},$(5) as stated previously in order to analyze the relationship of the main dependent variables of JMDA (Jones Model Discretionary Accruals) estimated under the Jones, 1991 Model and the main predictor of GA (Family Business Group Affiliation Dummy) along with a buttery of control variables. In Table 9, Ordinary Least Square regression is carried out, the result in regressions (1) and (2) shows that there is a significant negative relationship of family business group-affiliated firms with discretionary accruals estimated under the Jones, 1991 model and this finding indicates that family business group-affiliated firms decrease earnings manipulation practices of management. Moreover, in Table 10, Random Effect regression is carried out after the Hausman Specification Test, which is shown in the last column of Table 10 and in regressions (3) and (4) as presented in Table 9, the GA is also negatively related to earnings manipulation proxy without robust standard error (3) and with robust standard error (4). Moreover, the negative relationship between GA and EM also supports the study’s hypothesis 2a, which says that family business groups' affiliated firms áre negatively related to EM (accrual). This negative relationship of family business group-affiliated firms and EM supports the findings of Hong et al., 2015; Mindzak & Zeng, 2018. According to Mindzak and Zeng (2018), group-affiliated firms are less engaged in accrual-based earnings manipulation as compared to non-group affiliated firms. However, the findings of this paper contradict the findings of Beuselinck & Deloof, 2014; Kim & Yi, 2006; Muttakin et al., 2017 and they reported that group-affiliated firms manage earnings more as compared to stand-alone firms.

Table 9. Regression results of ols

	(1)		(2)
Predicted Variable (JMDA)	The Jones Model (Jones, 1991)		The Jones Model (Jones, 1991)
	(Without Robust Stand-Error)		(With Robust Stand-Error)
Predictors	Coefficient	T-Statistics	Coefficient	T-Statistics
GA	−0.012***	(−2.18)	−0.012***	(−2.24)
BDIV	−0.045***	(−2.24)	−0.045***	(−2.27)
ACIND	−0.023	(−0.90)	−0.023	(−0.92)
BIND	0.034	(0.89)	0.034	(0.87)
BEXP	0.044***	(2.42)	0.044***	(2.28)
CEOD	−0.030**	(−1.75)	−0.030	(−1.61)
FSIZE	−0.004**	(−1.86)	−0.004**	(−1.89)
LEVE	0.024**	(1.79)	0.024**	(1.86)
TANG	−0.062***	(−4.62)	−0.062***	(−4.31)
FAGE	−0.000***	(−2.05)	−0.000***	(−2.18)
BIG4	0.003	(0.48)	0.003	(0.49)
LOSS(D)	−0.003	(−0.50)	−0.003	(−0.47)
INV	0.000***	(2.52)	0.000***	(2.22)
LN_TA	−0.000**	(−1.70)	−0.000	(−1.60)
ABSROA	0.002***	(4.33)	0.002***	(3.93)
Constant	0.208***	(7.78)	0.208***	(7.47)
Observations	1046		1046
Groups (Firms)	223		223
R-squared	0.248		0.248
F-Statistics	7.70		8.53
Year Dummy	Yes		Yes
Industry Dummy	Yes		Yes

Table 10. Regression results of the random effect

	(3)		(4)
Predicted Variable (JMDA)	The Jones Model (Jones, 1991)		The Jones Model (Jones, 1991)
	(Without Robust Stand-Error)		(With Robust Stand-Error)
Predictors	Coefficient	T-Statistics	Coefficient	T-Statistics
GA	−0.011**	(−1.83)	−0.011***	(−2.01)
BDIV	−0.041**	(−1.83)	−0.041***	(−2.04)
ACIND	−0.022	(−0.82)	−0.022	(−0.83)
CEOD	−0.025	(−1.34)	−0.025	(−1.56)
BIND	0.034	(0.83)	0.034	(0.84)
BEXP	0.043***	(2.11)	0.043***	(2.07)
TANG	−0.060***	(−4.04)	−0.060***	(−3.62)
FSIZE	−0.004	(−1.51)	−0.004	(−1.45)
LEVE	0.025**	(1.66)	0.025**	(1.69)
FAGE	−0.000**	(−1.68)	−0.000**	(−1.78)
BIG4	0.002	(0.32)	0.002	(0.35)
LOSS(D)	−0.005	(−0.79)	−0.005	(−0.79)
INV	0.000***	(2.14)	0.000***	(2.31)
LN_TA	−0.000	(−1.49)	−0.000	(−1.40)
ABSROA	0.002***	(4.16)	0.002***	(3.75)
Constant	0.203***	(6.73)	0.203***	(6.67)
Observations	1046		1046
Groups (Firms)	223		223
R-squared	0.24		0.24
Year Dummy	Yes		Yes
Industry Dummy	Yes		Yes
F-Statistics	250.45		1738.78
Hausman Test P-value	.128		.340

As we have mentioned earlier, business groups can be seen under two opposite hypotheses: “the conflict of interest hypothesis” as suggested by Berle & Means, 1932; Jensen & Meckling, 1976 and “the efficient transaction hypothesis”. Our study’s findings support “The Efficient Transaction Hypothesis”; the aforementioned hypothesis claims that business groups timely fulfill their needs and take care of all their stakeholders whilst doing a business transaction with different parties (including related parties) and therefore may curb management in manipulating in earnings. In addition to this, the negative relationship of a group affiliated firms and EM supports the “Alignment Hypothesis”, which states that the interest of family insider management and outside dispersed shareholders us aligned, and both the parties want to increase the firm value. As mentioned in the literature that earnings manipulation may harm firm value (Sarkar et al., 2013), the alignment hypothesis implies that family management with the majority stakes in business will curtail management from earnings manipulation due to their family reputation and future business value. Moreover, the study of Hussain & Safdar, 2018b reported that in Pakistan, there is a high degree of ownership concentration and most firms are family owned; furthermore, they added that 87 percent of listed firms in Pakistan Stock Exchange, which has 10 or more percent family ownership. So due to the ownership concentration and family shareholding, the monitoring of management is very high in the Pakistani firms and it discourages management to tunnel the resources (Waseemullah et al., 2021). Keeping in view the aforementioned discussion, our study’s findings are different from the studies of Byun et al., 2013; Muttakin et al., 2017 and similar to Mindzak & Zeng, 2018; Zhu, 2020 and they reported that business group-affiliated firms have a higher level of reporting quality and lower-level earnings management.

In Table 11, the Random Effect regression of independent variable GA with Real Activates manipulation proxies is shown. The results show that GA (group affiliated dummy) has a negative relationship with real activity manipulation proxies except for abnormal cash flow from operation but does not have significant relationship. Although the coefficients are negative, these findings are not similar to the findings of Tables 9 and 10 in which there is a negative relationship of GA (group affiliated dummy) with accrual-based earnings manipulation.

Table 11. Regression results's real activities em

	(5)		(6)		(7)		(8)		(9)		(10)		(11)		(12)
Predicted Variable	Abnormal CFO_RM		Abnormal CFO_RM		Abnormal PRO_RM		Abnormal PRO_RM		Abnormal DISEX_RM		Abnormal DISEX_RM		AGGR_ RM		AGGR_ RM
	(Without Robust Stand-Error)		(With Robust Stand-Error)		(Without Robust Stand-Error)		(With Robust Stand-Error)		(Without Robust Stand-Error)		(With Robust Stand-Error)		(Without Robust Stand-Error)		(With Robust Stand-Error)
Predictors	Coefficient	T-Stat	Coefficient	T-Stat	Coefficient	T-Stat	Coefficient	T-Stat	Coefficient	T-Stat	Coefficient	T-Stat	Coefficient	T-Stat	Coefficient	T-Stat
GA	0.005	(0.60)	0.005	(0.62)	−0.004	(−0.40)	−0.004	(−0.38)	−0.006	(−0.10)	−0.006	(−0.10)	−0.008	(−0.43)	−0.008	(−0.46)
BDIV	0.003	(0.08)	0.003	(0.08)	−0.036	(−0.94)	−0.036	(−0.89)	−0.115	(−0.62)	−0.115	(−0.58)	−0.023	(−0.34)	−0.023	(−0.37)
ACIND	−0.084***	(−2.31)	−0.084***	(−2.35)	0.073	(1.64)	0.073	(1.57)	0.091	(0.53)	0.091	(0.69)	0.051	(0.61)	0.051	(0.66)
CEOD	−0.003	(−0.13)	−0.003	(−0.14)	−0.021	(−0.69)	−0.021	(−0.70)	−0.048	(−0.41)	−0.048	(−1.11)	−0.065	(−1.14)	−0.065	(−1.09)
BIND	0.008	(0.15)	0.008	(0.16)	−0.088	(−1.29)	−0.088	(−1.26)	−0.085	(−0.29)	−0.085	(−0.23)	−0.074	(−0.58)	−0.074	(−0.62)
BEXP	0.028	(1.02)	0.028	(1.03)	−0.027	(−0.79)	−0.027	(−0.75)	0.017	(0.10)	0.017	(0.13)	0.029	(0.47)	0.029	(0.46)
TANG	0.127***	(6.29)	0.127***	(6.11)	−0.089***	(−3.53)	−0.089***	(−3.58)	−0.229**	(−1.85)	−0.229	(−1.43)	−0.138***	(−3.05)	−0.138***	(−3.32)
FSIZE	−0.001	(−0.38)	−0.001	(−0.38)	0.004	(0.79)	0.004	(0.76)	−0.120***	(−4.91)	−0.120***	(−2.28)	−0.049***	(−5.96)	−0.049***	(−6.34)
LEVE	−0.019	(−0.94)	−0.019	(−0.80)	0.023	(0.90)	0.023	(0.74)	−0.081	(−0.66)	−0.081	(−0.36)	0.030	(0.67)	0.030	(0.59)
FAGE	−0.001***	(−2.21)	−0.001***	(−2.43)	0.000	(0.62)	0.000	(0.59)	−0.002	(−0.94)	−0.002	(−1.17)	0.001	(1.37)	0.001	(1.40)
BIG4	−0.014	(−1.47)	−0.014	(−1.48)	0.022**	(1.89)	0.022***	(2.20)	0.066	(1.09)	0.066	(1.00)	0.018	(0.86)	0.018	(1.07)
LOSS(D)	−0.077***	(−8.24)	−0.077***	(−8.20)	0.095***	(8.34)	0.095***	(7.90)	0.113***	(2.51)	0.113	(1.24)	0.149***	(6.92)	0.149***	(6.64)
INV	−0.000***	(−9.58)	−0.000***	(−5.09)	0.000***	(3.05)	0.000***	(2.68)	−0.000	(−0.46)	−0.000	(−1.06)	0.000***	(2.37)	0.000***	(2.04)
LN_TA	0.000***	(5.14)	0.000***	(4.98)	−0.000	(−1.32)	−0.000	(−1.29)	−0.000**	(−1.71)	−0.000	(−1.35)	−0.000	(−0.25)	−0.000	(−0.24)
ABSROA	0.008***	(12.50)	0.008***	(10.26)	−0.010***	(−12.53)	−0.010***	(−10.59)	−0.002	(−0.69)	−0.002	(−0.64)	−0.013***	(−8.48)	−0.013***	(−7.69)
Constant	−0.058	(−1.42)	−0.058	(−1.43)	0.104***	(2.05)	0.104***	(2.10)	1.410***	(5.27)	1.410***	(2.61)	0.663***	(7.31)	0.663***	(7.82)
Observations	1046		1046		1029		1029		1235		1235		1029		1029
Groups (Firms)	223		223		223		223		228		228		223		223
Year Dummy	Y		Y		Y		Y		Y		Y		Y		Y
Industry Dummy	Y		Y		Y		Y		Y		Y		Y		Y
R-squared	0.38		0.38		0.45		0.45		.16		.16		.34		.34
F-Statistics	572.79		1014.00		669.57		1007.11		127.78		75.06		480.68		860.64

6.5. Additional analysis

In this section, we intend to test the role of financial constraints on earnings manipulation on the selected firms of the study; for this purpose, the Firm Size and Firm Age have been used to divide the sample into two sub-sample of constrained and unconstrained firms as used by Ellouze & Mnasri, 2020; Kurt, 2018.The criteria used for the division of constrained and unconstrained firms are the median of Firm Size and Firm age, below-median is considered constrained firms and above-median are unconstrained firms.

In Table 12, the regression result of constrained and unconstrained firms is given. In regressions 13 to 16, the regression of constrained and unconstrained firms based on the firm size is given, whilst from 17 to 20, the results are based on the firm age. The results show in Table 12 that constrained firms are positively related to EM proxied by Jones, 1991, the coefficient 0.035*** with T statistics (2.03) based on the firm size, whilst based on the firm age, it is positively related but not significant. Contrary to this, the unconstrained firms are negatively related with EM and the coefficient −0.035*** with T statistics (−2.03), which means that unconstrained firms do not manipulate earnings.

Table 12. Regression results of financially constrained and unconstrained firms with discretionary accruals (JMDA)

Predicted Variable	(13) JMDA		(14) JMDA		(15) JMDA		(16) JMDA		(17) JMDA		(18) JMDA		(19) JMDA		(20) JMDA
	FIRM SIZE								FIRM AGE
Predictors	Coef	T-Stat	Coef	T-Stat	Coef	T-Stat	Coef	T-Stat	Coef	T-Stat	Coef	T-Stat	Coef	T-Stat	Coef	T-Stat
CONSTRAINED FIRMS	—-	—-	0.035***	(2.03)	—-	—-	0.036***	(2.03)	—-	—-	0.008	(1.19)	—-	—-	0.009	(1.36)
UN CONSTRAINED FIRMS	−0.035***	(−2.03)	—-	—-	−0.036***	(−2.03)	—-	—-	−0.008	(−1.19)	—-	—-	−0.009	(−1.36)	—-	—-
STAND-ALONE FIRMS	—-	—-	—-	—-	—-	—-	0.011***	(2.03)	—-	—-	—-	—-	—-	—-	0.011***	(2.18)
AFFILIATED FIRMS WITH GROUPS	—-	—-	—-	—-	−0.011***	(−2.03)	—-	—-	—-	—-	—-	—-	−0.011***	(−2.18)	—-	—-
BDIV	−0.030	(−1.55)	−0.030	(−1.55)	−0.038**	(−1.90)	−0.038**	(−1.90)	−0.035**	(−1.77)	−0.035**	(−1.77)	−0.043***	(−2.16)	−0.043***	(−2.16)
BIND	0.043	(1.09)	0.043	(1.09)	0.040	(1.03)	0.040	(1.03)	0.036	(0.93)	0.036	(0.93)	0.034	(0.87)	0.034	(0.87)
BEXP	0.036**	(1.87)	0.036**	(1.87)	0.040***	(2.07)	0.040***	(2.07)	0.038**	(1.95)	0.038**	(1.95)	0.042***	(2.17)	0.042***	(2.17)
ACIND	−0.034	(−1.37)	−0.034	(−1.37)	−0.031	(−1.25)	−0.031	(−1.25)	−0.026	(−1.05)	−0.026	(−1.05)	−0.023	(−0.94)	−0.023	(−0.94)
CEOD	−0.030**	(−1.69)	−0.030**	(−1.69)	−0.028	(−1.56)	−0.028	(−1.56)	−0.031**	(−1.69)	−0.031**	(−1.69)	−0.029	(−1.58)	−0.029	(−1.58)
TANG	−0.061***	(−4.25)	−0.061***	(−4.25)	−0.062***	(−4.33)	−0.062***	(−4.33)	−0.059***	(−4.07)	−0.059***	(−4.07)	−0.060***	(−4.17)	−0.060***	(−4.17)
LEVE	0.029***	(2.26)	0.029***	(2.26)	0.028***	(2.17)	0.028***	(2.17)	0.026***	(1.99)	0.026***	(1.99)	0.025**	(1.89)	0.025**	(1.89)
BIG4	0.003	(0.43)	0.003	(0.43)	0.002	(0.36)	0.002	(0.36)	0.004	(0.61)	0.004	(0.61)	0.003	(0.52)	0.003	(0.52)
LOSS(D)	−0.003	(−0.37)	−0.003	(−0.37)	−0.003	(−0.47)	−0.003	(−0.47)	−0.002	(−0.33)	−0.002	(−0.33)	−0.003	(−0.43)	−0.003	(−0.43)
INV	0.000***	(2.30)	0.000***	(2.30)	0.000***	(2.32)	0.000***	(2.32)	0.000***	(2.22)	0.000***	(2.22)	0.000***	(2.23)	0.000***	(2.23)
LN_TA	−0.000	(−1.64)	−0.000	(−1.64)	−0.000**	(−1.81)	−0.000**	(−1.81)	−0.000	(−1.49)	−0.000	(−1.49)	−0.000	(−1.65)	−0.000	(−1.65)
ABSROA	0.002***	(4.08)	0.002***	(4.08)	0.002***	(4.02)	0.002***	(4.02)	0.002***	(4.08)	0.002***	(4.08)	0.002***	(4.01)	0.002***	(4.01)
Constant	0.119***	(2.85)	0.084	(1.49)	0.124***	(2.97)	0.077	(1.37)	0.190***	(6.84)	0.182***	(6.72)	0.196***	(7.10)	0.176***	(6.42)
Observations	1046		1046		1046		1046		1046		1046		1046		1046
R-squared	0.246		0.246		0.249		0.249		0.244		0.244		0.247		0.247
F-Statistics	7.966		7.966		8.146		8.146		8.039		8.039		8.239		8.239
Groups (Firms)	223		223		223		223		223		223		223		223
Year Dummy	Y		Y		Y		Y		Y		Y		Y		Y
Industry Dummy	Y		Y		Y		Y		Y		Y		Y		Y

Moreover, in regressions 15 and 16, the family business groups' affiliated and stand-alone firms are regressed with EM on the basis of the firm size, whilst in regressions 19 and 20, the family business groups' affiliated and stand-alone firms are regressed with EM on the firm age and the result shows that family business groups' affiliated firms are negatively related to EM based on the firm size as well as based on the firm age, whilst stand-alone firms are positively related to EM. The coefficient −0.011*** with T statistics (−2.03) shows the significantly negative relationship between EM in affiliated firms, whilst the coefficient 0.011*** with T statistics (2.03) shows the significantly positive relationship between EM and stand-alone firms. This result indicates that affiliated firms are not financially constrained, whilst stand-alone firms are financially constrained. Moreover, our result is in line with Saeed & Sameer, 2015; they reported in their study that affiliated firms are financially unconstrained, whilst non-affiliated firms or non-business groups firms are financially constrained. Moreover, Saeed and Sameer (2015) added that business group firms have enough internal sources to utilize internal capital and do not dependent on external debt as compared to non-business group firms who depend on external debt and are financially constrained. In addition to this, we have linked financially constrained and unconstrained firms with Real activity EM in Table 13 and the results reveal that the coefficients are negative for unconstrained firms and positive for unconstrained firms, but not significant at 5% in all regressions.

Table 13. Regression results of financially constrained and unconstrained firms with real activity manipulation (AG_RM)

Predicted Variable	(21) AG_RM		(22) AG_RM		(23) AG_RM		(24) AG_RM		(25) AG_RM		(26) AG_RM		(27) AG_RM		(28) AG_RM
	FIRM SIZE								FIRM AGE
Predictors	Coef	T-Stat	Coef	T-Stat	Coef	T-Stat	Coef	T-Stat	Coef	T-Stat	Coef	T-Stat	Coef	T-Stat	Coef	T-Stat
CONSTRAINED FIRMS	—-	—-	0.063	(1.34)	—-	—-	0.063	(1.35)	—-	—-	0.030	(1.61)	—-	—-	0.030**	(1.66)
UN CONSTRAINED FIRMS	−0.063	(−1.34)	—-	—-	−0.063	(−1.35)	—-	—-	−0.030	(−1.61)	—-	—-	−0.030**	(−1.66)	—-	—-
STAND-ALONE FIRMS	—-	—-	—-	—-	—-	—-	0.009	(0.71)	—-	—-	—-	—-	—-	—-	0.008	(0.64)
AFFILIATED FIRMS WITH GROUPS	—-	—-	—-	—-	−0.009	(−0.71)	—-	—-	—-	—-	—-	—-	−0.008	(−0.64)	—-	—-
BDIV	−0.042	(−0.92)	−0.042	(−0.92)	−0.049	(−1.04)	−0.049	(−1.04)	−0.052	(−1.11)	−0.052	(−1.11)	−0.058	(−1.23)	−0.058	(−1.23)
BIND	−0.047	(−0.53)	−0.047	(−0.53)	−0.048	(−0.55)	−0.048	(−0.55)	−0.038	(−0.43)	−0.038	(−0.43)	−0.039	(−0.45)	−0.039	(−0.45)
BEXP	0.095***	(2.16)	0.095***	(2.16)	0.098***	(2.21)	0.098***	(2.21)	0.091***	(2.07)	0.091***	(2.07)	0.094***	(2.11)	0.094***	(2.11)
ACIND	−0.051	(−0.81)	−0.051	(−0.81)	−0.049	(−0.78)	−0.049	(−0.78)	−0.075	(−1.22)	−0.075	(−1.22)	−0.073	(−1.19)	−0.073	(−1.19)
CEOD	−0.085***	(−1.97)	−0.085***	(−1.97)	−0.083**	(−1.92)	−0.083**	(−1.92)	−0.102***	(−2.24)	−0.102***	(−2.24)	−0.100***	(−2.20)	−0.100***	(−2.20)
LEVE	0.089***	(2.54)	0.089***	(2.54)	0.088***	(2.49)	0.088***	(2.49)	0.091***	(2.57)	0.091***	(2.57)	0.090***	(2.53)	0.090***	(2.53)
TANG	−0.099***	(−3.19)	−0.099***	(−3.19)	−0.100***	(−3.22)	−0.100***	(−3.22)	−0.104***	(−3.35)	−0.104***	(−3.35)	−0.105***	(−3.38)	−0.105***	(−3.38)
BIG4	−0.015	(−1.10)	−0.015	(−1.10)	−0.015	(−1.14)	−0.015	(−1.14)	−0.018	(−1.31)	−0.018	(−1.31)	−0.018	(−1.35)	−0.018	(−1.35)
LOSS(D)	0.049***	(2.48)	0.049***	(2.48)	0.048***	(2.46)	0.048***	(2.46)	0.049***	(2.45)	0.049***	(2.45)	0.048***	(2.43)	0.048***	(2.43)
INV	−0.000	(−0.10)	−0.000	(−0.10)	−0.000	(−0.10)	−0.000	(−0.10)	−0.000	(−0.39)	−0.000	(−0.39)	−0.000	(−0.39)	−0.000	(−0.39)
LN_TA	−0.000	(−0.14)	−0.000	(−0.14)	−0.000	(−0.22)	−0.000	(−0.22)	0.000	(0.61)	0.000	(0.61)	0.000	(0.55)	0.000	(0.55)
ABSROA	0.001	(1.17)	0.001	(1.17)	0.001	(1.15)	0.001	(1.15)	0.001	(1.19)	0.001	(1.19)	0.001	(1.17)	0.001	(1.17)
Constant	0.535***	(4.93)	0.472***	(3.20)	0.538***	(4.91)	0.466***	(3.20)	0.671***	(8.94)	0.641***	(9.13)	0.675***	(8.95)	0.636***	(9.03)
Observations	1029		1029		1029		1029		1029		1029		1029		1029
R-squared	0.226		0.226		0.226		0.226		0.220		0.220		0.220		0.220
F-Statistics	8.359		8.359		8.136		8.136		10.065		10.065		9.873		9.873
Groups (Firms)	223		223		223		223		223		223		223		223
Year Dummy	Y		Y		Y		Y		Y		Y		Y		Y
Industry Dummy	Y		Y		Y		Y		Y		Y		Y		Y

7. Conclusion

This research paper examines whether family business group affiliation affects the earnings manipulation in PSX listed firms for the period of 2014 to 2019. The final sample of 323 firms was used for analysis consisting of 187 group-affiliated firms and 136 non-affiliated firms. The study findings provide valuable insights into the business group influence on earnings manipulation. The results of the study show that family business group affiliation is negatively related to accrual -based earnings manipulation, which means that group affiliation reduces EM in Pakistani listed firms. However, the results of this study failed to provide evidence that family business group affiliation is negatively related to real activity manipulation. Moreover, the results of this study reveal that board independence and female directors are higher in stand-alone firms as compared to business group-affiliated firms. In addition to this, the board expertise is higher in affiliated firms as compared to non-affiliated firms and CEO duality is more in the affiliated firms as compared to non-affiliated firms. Furthermore, this study investigated that the leverage and loss are low in the family business group-affiliated firms as compared to non-affiliated firms and this finding indicates that stand-alone firms suffer more loss and have more leverage than family business groups' affiliated firms. In addition to this, our findings revealed that constrained firms are positively associated with EM as compared to financially unconstrained firms. Moreover, the study’s result indicates that affiliated firms are not financially constrained, whilst stand-alone firms are financially constrained.

Similar to other studies, the findings of this study should be prudently interpreted due to its limitations and the future research scholars could address the following limitations before extending the related research. This research study has included a battery of control variables related to board of directors and firm characteristics variables but omitted the ownership structure-related variables, which may/could affect the results. Furthermore, this research study has used only one discretionary accrual proxy, the Jones Model 1991 (Jones, 1991), and future scholars may consider other proxies as well. Additionally, this study considered only public listed companies affiliated with family business groups, although there are various group-affiliated private, companies which are out of the scope of this study. Despite aforementioned limitations, this research paper has added its part in the scarce literature by focusing on business groups and earnings manipulation, especially on the emerging economy perspective of Pakistan. Finally, it extends the line of research on Business Groups, Earnings Management and Corporate Governance.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The authors received no direct funding for this research.

Notes on contributors

Sattar Khan

Sattar Khan holds MS in Business and Administration and he is currently enrolled in Ph.D. Management in the Institute of Management Sciences, Peshawar, Pakistan. His area of research interest includes Financial Accounting, Audit Quality, Family Businesses, Earnings Manipulation and Corporate Governance. This Research Paper is a part of his Ph.D. Thesis.

Yasir Kamal received his Ph.D. in Finance from Shaheed Zulfikar Ali Bhutto Institute of Science and Technology (SZABIST), Islamabad. He has 18 years of teaching/research experience in various institutes/universities. His research interest includes Behavioral Finance, International Trade, Corporate Governance and Financial Econometrics. He is currently working as an Associate Professor at the Institute of Management Sciences, Peshawar, Pakistan. He has published various research papers in reputed Journals such as Global Business Review, Business & Economic Review, Asian Journal of Management Sciences and South African Journal of Business and Management.

Notes

1. Business Groups and Family Business Groups are used interchangeably.

2. Earnings Manipulation is Interchangeably used with Earnings Management.

Appendix 1

Variables Definitions

Table

Variables		Description
Dependent Variables
MCDA	Discretionary accruals calculated under the McNichols (2002) model.	Discretionary accruals calculated under the McNichols (2002) model.
REALEM	Aggregate real activity manipulation by adding the abnormal production cost with abnormal cash flow form operation and abnormal discretionary expenses	Aggregate real activity manipulation by the adding abnormal production cost with abnormal cash flow form operation and abnormal discretionary expenses, by taking into the account the studies by Mindzak & Zeng, 2018
DIS ERM	Abnormal discretionary expenses	Abnormal discretionary expenses estimated using the Roychowdhury, 2006 model.
RCFO	Abnormal cash flow form operation	Abnormal cash flow form operation estimated using the Roychowdhury, 2006 model.
PRO RM	Abnormal production cost	Abnormal production cost estimated using the Roychowdhury, 2006 model.
Independent Variables
GA	Family Business Group-Affiliated firms	A Dummy variable taking the value of one if a listed firm is associated with the family business group and if not, 0 will be taken.
Board of Directors Variables
BIND	Board Independence	Independent directors/ize of the board
BEXP	Board Expertise	Those directors who have accounting and finance-related education/experience by the size of the board
BDIV	Board Female directorship	Female directors/size of the board
ACIND	Audit Committee Independence	Number of directors in the audit committee/ize of the Audit Committee
CEOD	CEO Dual Position	A categorical variable taking the value of one if the Chairman and CEO are having the same position otherwise zero.
Control Variables
BIG4	BIG 4 Audit Firms	A firm audited by Big 4 audit firm will take 1 and others will take 0
ABSROA	Return on Assets	Earnings before tax/Total Assets, after that the absolute value is taken
LEVE	Leverage	Short-term and long-term Debt/Total Assets
LOSS(D)	Loss dummy	A Dummy variable, which is taking the value of 1 if the firm is reporting loss in the annual report other wise 0
FAGE	Age of the firm	Number of years of a firm since its incorporation.
LN TA	Log of Total Assets	Taking the log of the total assets value.
TANG	Tangibility	Fixed Assets/Total Assets
FSIZE	Size of the firm	Taking the log of Total Assets
INV	Inventory/ Total Assets	Inventory at the end of the period/Total Assets