The Effect of Corporate Governance Regulation on the Profitability of Insurance Companies in Indonesia

. This study is driven by the growing importance of insurance companies in Indonesia and corporate governance as determinants of their profitability. Corporate governance is crucial in insurance companies as they provide instruments for risk transfers and savings mechanisms and promote investments in an economy. Our research examines the impact of implementing good corporate governance regulation POJK 73/2016 on the profitability of insurance companies in Indonesia. Using samples of all listed Indonesian insurance companies across 200 observations by employing LSDV panel data and 2SLS models, we found that the regulatory reform is negatively related to the company's profitability and thus reduces performance. The outcomes of this study advocate for the continuation of fine-tuning of existing regulations so it could deliver the stated objectives.


Introduction
The growing importance of the industry and the unraveling of several high-profile scandals have resulted in a renewed interest in corporate governance (CG) implementation in the Indonesian insurance sector.One of the cases that captured the public's attention is the Jiwasraya case.There were allegations of fraud and corruption involving two top executives of an insurance company with potential losses of public funds amounting to US$1.1 billion (Rp.16 trillion).With this unraveling, the Indonesian public now starts questioning how this could have happened and whether there were no principles of good CG that could monitor and prevent the events.Simultaneously, the public also begins to ponder the effectiveness of supervision carried out by the Indonesian financial services authority (OJK -Otoritas Jasa Keuangan).Over the years, numerous initiatives have been launched and proposed by the government to enhance the CG practice in Indonesia.It started with laying out the core principles of good governance in enacting the Corporate Law Number As stated in these regulations, the objectives are to increase the value and competitiveness of Indonesian insurance companies.
Nevertheless, poor governance practices are rampant among Indonesian insurance companies.Besides PT Jiwasraya Insurance, several scandals involving other large insurance companies, mainly state-owned enterprises (SOEs), are currently being investigated by Indonesian law enforcement agencies (Mukaromah 2020).To improve the situation, the government must continuously strengthen laws, improve CG regulations, and increase the enforcement of such laws and regulations.Simultaneously, companies must also act to improve the situation.CG practice must provide the appropriate market incentives and positively affect companies' market value.Particularly for insurance companies that play critical roles in transfer risk and generating incentives for savings and investment in the economy.Because this condition persists, it will negatively impact the industry and the overall being of the economy.To the best of the authors' knowledge, this would be the first paper that attempts to investigate the impact of CG regulations on listed insurance companies' profitability in Indonesia.This paper contributes to the general literature on CG and the insurance sector.Both the regulators and the insurance sector can utilize the implication of this study to fine-tune existing CG regulations and practices, which will improve the performance and value of their companies.

Literature Review
The majorities of past studies showed a positive association between CG mechanisms and companies' financial performance.These findings are congruent with the agency theory prediction that companies with good governance standards performed better due to lower agency costs and more effective monitoring mechanisms.Good CG also positively affects companies' profitability by building trust, transparency, and accountability vital for the company's stability and business integrity in the long term.An exemplified CG practice increases profitability as it has a positive impact on investors' valuation of cash holdings and hence on companies' value (Zhang 2011).A study conducted by Kanagaretnam et al. (2007) on Canadian and U.S. companies suggested that companies with better and effective governance related to the reduced information asymmetry.Other studies also showed similar findings that better-governed companies have higher accounting performance and market performance (Brown and Caylor 2006;Gompers et al. 2003).Outside the U.S., Sami et al. (2011) showed that higher quality of CG has a significant positive association with companies' performance and valuation in China.Meah and Chaudhory (2019) found a significant positive relationship between CG and the profitability of Bangladesh's manufacturing companies.In the context of insurance companies, Malik (2011), for example, showed that corporate factors practices are essential determinants of the profitability of insurance companies in Pakistan.Likewise, Tornyeva and Wereko (2012) found that soft governance factors such as recruitment policy, staff training, and development positively affect the performance of the companies in Ghana.Wanyama and Olweny (2013) showed that board members' composition and their skills and expertise are positively associated with companies' financial performance.In a study on eight Asian countries, Zainudin et al. (2018) concluded that firm-specific factors, including CG factors, positively affect the profitability of insurance companies.In Indonesian context, prior research found that CG measures related to board of directors and ownership structure significantly affect insurance companies' profitability (Trisilviani et al. 2020).Recent studies that examined the relationship between CG and performance of insurance companies (Azegele et  The financial services sector is a highly regulated industry.Well-designed and executed regulations prevent financial fraud and limit the risks financial institutions can take with consumer funds.However, Boot et al. (1999) argued that authorities need to finetune the regulations continuously and level the playing field in the financial sector.Besides, Klein (2012) stated that the government must utilize a cost-benefit calculation to formulate regulations to obtain an optimum net benefit resulting from the intervention.Durden and Pech (2006) found that the financial sector regulation might overwhelm and burdened companies with increased complexity and compliance requirement.
The increased complexity and high compliance costs would adversely impact the company's agility and competitive advantage.Lawack-Davids (2012) observed these regulatory burdens in South Africa's financial sector and attributed them to regulatory misalignment and overlap.Likewise, Gaganis and Pasiouras (2013) concluded that supervisory authorities' power and regulatory arrangement affect the soundness of financial companies, particularly insurance companies.Reducing the regulatory burden in the financial sector may not be an easy task due to the complexity and highly competitive nature of the industry.Therefore, to balance the needs for stability and consumer protection, scholars such as Helm (2006) argued that policymaker needs to be specific and focus on the timing, place, and regulation method in addressing regulatory burden.
In Indonesia, duties to supervising and regulating the financial sector reside with the OJK, which establishes rules (POJK) to guide how the incumbent companies should establish their prudential behavior and conduct in the industry.OJK enacted regulation number 73 in 2016 (POJK73/2016) for insurance companies that provides rules and guidelines on their CG.This regulation has 11 points of requirements to be adhered to by all listed insurance companies in Indonesia.One requirement directed in the POJK73/2016 explored in this paper is the director of compliance requirements for listed insurance companies.The regulation stated that the director for the compliance function must not have concurrent marketing, engineering, and financial responsibility and functions.In other words, it must be a particular additional post of director on the board of directors.The rule also stated that all insurance companies must comply with the regulation at the latest by three years after it was enacted or by May 2019.Nevertheless, as of December 2020, not all insurance companies complied with the requirement.In this research, we propose the following hypothesis: H1: There is a positive relationship between the enactment of the CG regulation of POJK73/2016 and the profitability of listed insurance companies in Indonesia.
Our approach of investigating the impact of CG regulation reform on companies' performance with panel data has been used previously by authors such as Black and Khanna (2007) and Atanasov et al. (2008).

Methods
Table 1 summarizes the definitions, concepts, and descriptions of variables used.This study uses Return on Assets (ROA) as companies' profitability measure, as it is commonly used in the literature, including in Indonesian insurance companies (Arifa and Ahmar 2017; Nuryanto et al. 2020; Siswanto and Hasanah 2019).This paper uses two variables of board size and board independence as control variables.Board size is an important CG factor in exercising internal control, leading to better financial performance (Dalton et al. 1998;Muttakin et al. 2012;Yasser et al. 2017).Jensen (1993) and Yermack (1996) showed that a small board is more effective.However, recent studies showed otherwise.A study by Muttakin et al. (2012) found that increases in board size improve the monitoring capabilities within a company.The board performs monitoring of the management as representative of shareholders and providing resources (Hillman and Dalziel 2003).Other studies (Dalton et al. 1998;Yasser et al. 2017;Kiel and Nicholson 2003).showed that increased board levels offered more experienced and skilled resources that advanced companies' profitability.Likewise, the proportion of board independence enhances companies' financial performance by improving the monitoring capacity and offering more experiences (Jensen 1993 Companies' financial performance also depends on their specific internal characteristics and environment (Love 2011).We use three independent variables of the company age, company size, and the volume of capital.Pástor and Pietro (2003) argued that investors' uncertainty lessens with the company's age.Literature also suggests that the variability of stock returns negatively relates to incorporation (Adams et al. 2005) and listing age (Cheng 2008).Lower profitability in aging companies may also be due to "inertia effects," where aging companies become inflexible and have difficulties fitting the rapidly changing environment (Barron et al. 1994).Similarly, Loderer and Waelchli (2010) showed that older age deteriorates companies' performance.Other studies also reported similar findings, such as Arora and Sharma (2016) and Meah and Chaudhory (2019) in India and Bangladesh, Öner Kaya (2015) in Turkish insurance companies, and Hidayati and Shofawati (2019) in Indonsian insurance companies.On the other hand, Coad et al. (2013) found a positive relationship between age and company profitability in the Spanish manufacturing sector.Meanwhile, Malik (2011) failed to find any significant relationship between profitability and companies' age in the insurance sector of Pakistan.
Studies on the determinants of insurance companies' profitability broadly suggest that companies' sizes positively relate to their profitability.We calculate company size as the natural logarithm of the total gross premium earned (Zainudin et al. 2018;Javaria et al. 2013).Browne and Hoyt (1995) argued that larger companies have greater capacities to dealing with uncertain market fluctuations.Larger companies also enjoy economies of scale and have greater capacities and resources, thus improving their financial performance (Browne and Hoyt 1995;Flamini et al. 2009).Malik (2011) showed that size is positively associated with the financial performance of insurance companies in Pakistan.Likewise, Zainudin et al. (2018) reported that larger size measured by premium earned translates into higher profits for insurance companies in eight Asian countries.Öner Kaya (2015) reported a positive relationship between company size and profitability for Turkey insurance companies.The size of a company's capital is an indicator of its financial strength and capital adequacy, as it represents funds contributed by shareholders.Scholars also perceive the size of capital as a solvency ratio, whereby a higher volume of capital means that the company is highly solvent (Zainudin et al. 2018).The volume of capital also indicates the ability of a company to cover financial losses and how it finances its assets (Hassan and Bashir 2012).Most studies that look at the insurance sector found that companies with more considerable capital have higher profitability (Malik 2011 where ROA is the return on total assets, BD_Size denotes the board of directors' size, BD_Ind indicates the board independence.The binary variable of CGR_OJK signifies the enactment of CG regulation reform of POJK73/2016, which takes a value of 1 in period 14 onwards (second quarter of 2019) or 0 for the periods before.We use three years to lag the variable as the OJK required all insurance companies to comply with the regulation only three years after its issuance or May 2019.The C_Age variable represents the company's age, C_Size denotes the company's size, and C_Capital indicates the size of the company's capital.
t and parameters to be term, while i denotes the selected insurance company (i -2016, --2020).The alternative to the LSDV panel model specified above is the random effects panel data model.The LSDV model allows for heterogeneity or individuality of the ten companies by enabling them to have their time-invariant intercept value.In contrast, the randomeffects model treats individual company-specific constant or intercept as a sample randomly drawn from populations absorbed in the error terms.We run Hausman's specification test to diagnose the better specification of the two-panel models, as suggested by Wooldridge (2012).Based on the test, the fixed effects or LSDV model appeared to be the more appropriate model for this study, as the Chi-squared value was 35.95 and the p-value was 0.0000.Thus, the Hausman test overwhelmingly suggests that the difference in coefficients between the LSDV and random panel data models is systematic, which leads us to choose the panel data fixed-effects model with the LSDV estimator as specified in Equation 1.We also verify our model specifications for a possible endogeneity problem.
Endogeneity may occur when investigating the relationship between company performance and corporate governance as causality can go either way: good governance causes higher performance or highperforming companies tend to have more effective management (Love 2011; Bahadur 2016).
To address possible endogeneity and check the robustness of our finding, we employ another approach called the two-stage least square (2SLS), which addresses a potential endogeneity problem by regressing the alleged endogenous variable with an instrumental variable.The predicted values of this regression then replaced the likely endogenous variable as an explanatory variable in the model of interest.The 2SLS model effectively addressed the endogeneity problem in the previous corporate governance studies (

Data Collection
We collect data from the financial statements of the ten insurance companies listed on the Indonesia Stock Exchange (IDX) from 2016 to 2020.In total, we obtain 220 observations over the 20 quarters between the first quarter of 2016 and the fourth quarter of 2020.

Numerical Results
Table 2 presents the descriptive statistics of the variables.The average ROA for insurance companies in Indonesia is 0.0240, ranging from a minimum of -0.2160 to a maximum of 0.0984.Four out of 10 companies experienced negative profits during the sample period from close observation of the data set.One company consistently made negative profits between the third quarter of 2017 and the fourth quarter of 2019.The average number of directors and commissioners on board is 8 (exp 2.0662), ranging from a minimum of 5 (exp 1.6094) to a maximum of 12 (exp 2.4849).The average ratio of independent commissioners to the total commissioners is 0.5760, with a minimum of 0.3333 and a maximum of 0.7500.Therefore, the shows that every listed insurance company in Indonesia has at least one independent commissioner on board.The average company age is 48 years (exp 3.8626) and ranges from a minimum of 34 years (exp 3.5264) to a maximum of 67 years (exp 4.2047).The average company size is 23.3154 (Indonesian Rupiah (IDR) 729.3 billion), ranging from a minimum of 25.5661 (IDR 126,8 billion) to a maximum of 28.6375 (IDR 2,7 trillion).The average equity to total assets ratio for listed insurance companies in Indonesia is 0.3980 and ranges from a minimum of 0.1801 to a maximum of 0.6157.
We also explore the preliminary relationship between explanatory variables used in our study by calculating the Pearson correlation coefficients.The results in Table 3 show that the company's age and size correlates positively and significantly with the company performance proxy of ROA with coefficients of 0.2611 and 0.3188, respectively.However, the leading independent variable used in this paper to signify the CG regulation (CGR_OJK) correlates negatively and significantly with ROA.Moreover, there is a significant positive relationship between a company's age and company size, and ROA.The CGR_OJK correlates positively and significantly with the instrumental variable of the director of compliance variable (BD Comp).
Table 4 presents estimation results using the LSDV model.It is shown that the estimated coefficient of CGR_OJK is negative and statistically significant at the 5% level.This result implies that implementing the regulation decreases insurance companies' returns on assets by -0.0889.This finding rejects our hypothesis that there is a positive relationship between the enactment of POJK73/2016 and the profitability of listed insurance companies.As discussed in the literature review section, possible explanations of this unexpected result include that the new regulation overwhelmed and burdened insurance companies with increased complexity and compliance requirement, which in turn increased their cost and adversely affected their performance (Durden and Pech 2006; Lawack-Davids 2012).Moreover, it reduces flexibility and renders the decision-making process less efficient.Compliance requirements also significantly burden insurance companies' operating expenses, which adversely affected profitability.The compliance director requirement, for example, may cause a substantial extra budget for securing the services of a new director.Without the new regulation, insurance companies do not have to incur this extra budget as the function can be performed effectively and efficiently by the corporate secretary as the incumbent.In relation to agency theory, this finding thus suggests that this increase in monitoring costs (as one of the components of agency costs) is not compensated for with improved performance, indicating excessive monitoring mechanisms.
POJK 73/2016 turned out to be very similar to other previous CG regulations.This regulation appeared was enacted only to satisfy the requirement to have corporate governance regulation by the newly established FSA at that time.The change of authority from the Ministry of Finance (MoF) to OJK might have created a legal vacuum if there are no new CG regulations for insurance companies.At this juncture, OJK may have adopted the existing regulations into their issuance without any finetuning.As a result, there were only new mandatory actions for insurance companies instead of a new innovative set of corporate governance codes of conduct.It is not surprising if POJK73/2016 has adverse financial consequences for insurance companies in Indonesia.The regression results also indicate that board size (BD_Size) is significantly associated with insurance companies' profitability in Indonesia, while board independence (BD_Ind) does not significantly affect performance.A larger board size appeared to highly influence profitability as it enhanced the monitoring capacity of the management and helps to monitor management activities as suggested in the corporate governance literature.Similarly, company size (C_Size) and capital ratio (C_Capital) are positively and significantly related to companies' performance, while company age (C_Age) relates negatively to profitability.Thus, these findings confirmed previous similar studies on the impact of company characteristics on financial performance (Meah and

Validation
Our primary analysis uses an LSDV panel data model for investigating the impact of POJK 73/2016 (CGR_OJK) on insurance companies' profitability.This model assumes that the relationship between the implementation of POJK 73/2016 and performance are not endogenously determined.However, several papers indicate that research related to CG is prone to endogeneity problems.To test our results and avoid possible endogeneity problems, we conducted an analysis using the 2SLS method.Specifically, we instrument the CGR_OJK variable with the BD_Comp variable, which signifies a compliance director's existence in an insurance company.We then regressed the dependent variable ROA on this instrument variable and other control variables and presented the results in Table 5.
The results of our robustness check confirm the results of our primary analysis.The enactment of POJK 73/2016 regulation instrumented by the existence of the compliance director variable (BD_Comp) is negatively and significantly affecting the profitability of insurance companies.The coefficient of the BD_Comp variable (-0.8442) is essentially the same as the estimated CGR_OJK coefficient in the LSDV model (-0.889).Nevertheless, the significant level of the BD_Comp is exceedingly higher at a 1% confidence level than that CGR_OJK of the LSDV model at a 5% level.From Table 5, The Durbin and Wu-Hausman diagnostics tests for the 2SLS model are significant, indicating that the 2SLS specification provides more consistent estimations than the LSDV specification.Therefore, we conclude that the robustness check confirmed our finding that the implementation of POJK 73/2016 negatively affects the profitability of listed insurance companies in Indonesia.

Conclusion
This study aims to examine the effect of the implementation of POJK 73/2016 on the performance of insurance companies in Indonesia.The primary method we used to test this relationship was the panel data LSDV regression method.We then checked for its robustness using the 2SLS method.Both the LSVD and 2SLS estimations show that the implementation of POJK 73/2016 has a significant adverse effect on the profitability of insurance companies in Indonesia, which lead us to the conclusion that there is a negative influence on the implementation of the rule to the level of insurance company profits.We recommend the regulator to pay attention to fine-tune existing regulations rather than issuing new ones, as excessive monitoring mechanism may hurt the company's profitability.A new regulation without adequate finetuning adds only to the companies' burdens instead of increasing their performance, undermining their level of profit.Additionally, the spirit of emphasising compliance in insurance companies is important to avoid future scandal or cases, nevertheless obliging the company to have a dedicated compliance director apparently may not be the best strategy as it deteriorates profitability.Thus, we also recommend keeping the compliance function, for example via compliance committee, rather than appointing a separate director.One shortcoming of our study is the limited sample of insurance companies and the relatively short span of the observation period.Besides, there is no breakdown of the new regulation, which may further explain the result.Therefore, we recommend a larger sample with more companies' characteristic data for further studies.

Table 1 .
Concept and Definition of Variables

Table 2 .
Summary of Descriptive Statistics

Table 4 .
Panel Data LSDV Regression Results