Financial reporting, public services and local executives’ re-electability in Indonesia

Abstract This study aims to analyze the effects of audit findings, follow-up of audit recommendations, quality of financial reporting and public services on the incumbent electability. Using subnational elections data in Indonesia in 2015 and 2017, the final sample amounted to 291 observations, which consisted of 11 provincial elections, 55 municipal elections, and 255 district elections. The results showed a negative effect of audit findings on the incumbent electability, as hypothesized. Further findings present the positive influences of the follow-up of audit recommendations, financial reporting quality on incumbent electability. The last result implies the originality of the study in which non-financial performance factor as proxied by the quality of public services in subnational entities in Indonesia are used as the basis for political decision-making by constituents to elect the incumbent or challenger. The practical implications are the need for subnational executives to pay attention to the importance of accounting and auditing practices as well as public service in increasing local community satisfaction and trust. The theoretical findings are insightful in providing mechanisms of government accounting management and the usefulness of public sector information in influencing electoral democracy in local governments in Indonesia. The novelty of this study is in the empirical examination to link auditing practices and its political consequences.


PUBLIC INTEREST STATEMENT
What is the basis for voters to re-elect the incumbent executives in the second term? This study found that in Indonesian democracy, local communities also use audit reports as a basis for decision-making in voting, apart from assessing the quality of public services. Information related to the quality of local government financial reports, the number of audit findings and follow-up audit recommendations contained in the audit report are used as a basis for the community to assess the extent of accountability for local government financial management and to "punish" or "appreciate" incumbents. High quality of financial reports, low audit findings and follow-up audit recommendations are some factors of the incumbent reelectability. Therefore, it is important for local executives to pay attention to accounting functions in public services, as well as to pay attention to audit results in realizing a quality democracy.

Introduction
From a theoretical perspective of positive public choice, supply and demand are carried out by power holders and constituencies. Here, supply is made by power holders/resource managers, either directly elected, such as the legislative and executive, or indirectly, such as organizational units under the executive, while demand are made by the constituents. Mueller (1976) states that the decision function in electoral context is a quantification of voter reaction (R) obtained from the multiplication of the potential benefits of a public good by selecting a particular candidate (B) and the probability of additional benefits being elected the best candidate (P), and reduced by public procurement costs, such as information gathering costs (C), plus personal and voter benefits (D). This is mathematically formulated by R = BP-C + D.
Based on this theory, in determining their choice, voters will rationally use the information received, for example, related to the quality of financial management and public services (James & John, 2007;Boyne et al., 2009). This will encourage them to select candidates they perceive to be of the best benefit to their constituents. Such conditions are only possible in democratic societies, including in Indonesia, which adheres to a direct democratic system for the people to determine their executive in the local government through direct elections. This is a manifestation of democracy to openly give equal opportunities for people to directly elect or be elected as government leaders. Voters' rationality will be realized by using the available channels to obtain information to determine their choice (James & John, 2007). One of the important information available is the auditing reports to weigh the credibility of the incumbent and the reputation of governance (Banerjee et al., 2011;Strachan, 2014). Audit findings and the quality of local government financial reporting are related to the accounting function in producing reliable and relevant financial information. The results of previous research indicate that there is still debate regarding the usefulness of government financial reporting. Although several previous findings (Groot & Budding, 2008;Mack & Ryan, 2006;Steccolini, 2004) indicate the normative function of government financial reporting to meet regulatory needs and achieve accountability alone, Ingram and Copeland (1981), Alijarde (1997), Lin and Raman (1998), and Copley and Douthett (2014) found that local government financial reporting and/or the information presented therein has been useful in making decisions by the director of finance, government management, and other stakeholders such as financial entities and communities both in property investment and in re-electing incumbents. Basically, the auditing report provide a detailed assessment of the quality of regional financial management. In addition, voters can directly assess the extent to which the quality of public services provided by a local government, through health services, education, environmental licensing, and infrastructure (Adi et al., 2016;Johari et al., 2018;Purwanto, 2019). The auditing reports, quality of reporting and public services is also indicative information for voters to choose the incumbent, or the challenger. Therefore, it is very interesting to analyze the extent to which information about the quality of financial management and the quality of public services in subnational governments in Indonesia is used as an indicator of the electability of the incumbent in the second-term election.
Empirically, previous research has extensively analyzed the factors that influence voter choice, including the determinants of incumbent re-electability in the second term. As mandated by Law Number 14 of 2008 concerning Public Information Disclosure, Regional Government Financial Reports (LKPD/Laporan Keuangan Pemerintah Daerah) and audit results from the Supreme Audit Agency (Badan Pemeriksa Keuangan/BPK) are some of the public information that must be published to the public, both through printed and electronic media. Thus, it is possible for the public to access government financial reports, to assess the extent to which the level of accountability and performance of regional financial management has been implemented by the incumbent. Therefore, the audit report which represents the level of accountability and performance of the incumbent regional financial management can affect the incumbent re-electability. Specifically, some of the determinants disclosed by previous studies such as information on accounting reports, and auditing results are important indicators to measure the re-electability of incumbent executives (Brender, 2003;Din et al., 2017;Ferraz & Finan, 2008Ingram & Copeland, 1981;James & John, 2007;Pereira et al., 2009). However, until now these studies are still limited to analyzing the effect of auditors "audit findings only, and no research has been found that analyzes the extent of the effect of follow-up on audit recommendations and the quality of local government financial reporting on incumbents" re-electability. In addition, in the context of auditing and the public sector, it is still rare to find research that analyzes the effect of the quality of public services as a valuable information for measuring non-financial performance and a complement to information on the financial performance of local governments on the electability of the incumbent. This motivates to test whether the audit findings, follow-up audit recommendations, the quality of financial reporting and the quality of public services affect the electability of the incumbent in the second-term election. Therefore, the study specifically examines the effect of audit findings and financial reports, while developing an empirical analysis of the effect of public service quality on incumbents' electability.
Consequently, the current paper seeks to make the following contributions to the existing literature. First, this study contributes theoretically to provide additional explanations for previous findings that analyze the extent to which audit reports affect the electability of incumbents (including, Ferraz & Finan, 2008Olken, 2007;Pereira et al., 2009). Second, this research contributes to the explanation of the effect of financial reporting and electoral politics. Similar to Daniels and Daniels (1991), and Alijarde (1997), this research also uses the assumption that financial reporting is not only used to comply with regulations, for realizing accountability or being used as a basis for stakeholder decision-making for economic interests and investment, but also as a basis for constituent's decision-making in determining their political preferences. This study conceptually provides an empirical examination of the usefulness of non-financial information of the quality of public services as decision-making basis for constituents to give or remove legitimacy for the executive in power. Fourth, this study contributes to broadening the insight into the mechanisms of financial management performance and non-financial indicators of public services as important determinants of electoral dynamics in subnational governance in Indonesia.

Background
In the administration of public sector, there is an important relationship between the implementation of regional elections, especially in Indonesia after the 1998 reforms were used by incumbents to win elections by improving public services. In the context of government-sector reform in Indonesia, incumbents also use financial reporting on regional budgets and auditing report to influence the electoral vote. Particularly in Indonesia, there is a tendency for local executives to improve the quality of LKPD through the achievement of Unqualified Opinion (WTP), aimed at increasing the accountability and credibility of local governments to increase electoral votes in the second-term election.
After the 1998 reforms, significant political and economic reforms have been carried out in order to realize a democratic life, strengthen decentralization, increase accountability, and create transparency in the public sector (Harun et al., 2015). As part of these reforms, since 2004, both the elections for members of the legislature and the elections for the President and Vice President have been carried out in a democratic manner, with direct electoral elections by the public, based on the principles of freedom, openness, honesty and fairness, which are held every 5 years. Likewise, regional elections, since 2005, are directly elected by the local. Initially, the candidates were only pairs of candidates proposed by political parties or coalitions of political parties, but since 2008, candidates can come from pairs individual candidates or through independent channels. The regulations also allow incumbents to become candidates in the second term. Since 2014, the implementation of regional elections in Indonesia has been well-regulated. One of the fundamental things is the stipulation of simultaneous elections for all subnational entities which will start in 2024.
In order to improve the quality of public services, government reform in Indonesia is conducted by strengthening decentralization and increasing the public participation and the private sector, including SOEs, both in the provision of facilities and public services to the community. To increase the accountability and responsiveness of services to the community, through the implementation of regional autonomy, the subnational government has been given the authority to organize public services, both basic services consisting of educational affairs, health, public works and spatial planning, public housing and community protection, as well as social and non-basic services, such as manpower affairs, women's empowerment and child protection, food, culture, land, environment, civil administration, family planning, transportation, communication and small and medium enterprises. The autonomy to the local authorities is also intended to improve public services to be carried out in a quality manner, which is marked by the implementation of public services that are integrated, transparent, sustainable, fair and affordable to the whole society.
As a result of regional autonomy, since 2004, financial balance between the central and subnational authorities has been also regulated. Local governments are given the authority to manage finances independently and collect revenues in accordance with the limits of their authority, such as local taxes and levies. In addition, related to regional financial management, subnational executives are required to prepare financial reports as a form of accountability for regional financial management, which is prepared based on government accounting standards. (Standar Akuntansi Pemerintahan/SAP) and examined by the BPK, based on a full accrual accounting system.
The results of the audits conducted by the BPK are presented in an Audit Result Report (Laporan Hasil Pemeriksaan/LHP) and submitted to the president or respective subnational executives and representative institutions in accordance with their respective authorities. The LHP on the government financial report contains the opinion or professional statement of the examiner regarding the fairness of the financial information presented in the government financial report, which consists of four types of unqualified opinion, qualified opinion, adverse opinion, and disclaimer opinion. LHP also presents audit findings, responses of government officials who are responsible for findings, conclusions, recommendations and monitoring results on follow-up recommendations. Findings related to financial audits are presented into two groups related to the internal control system and to the compliance with legislation. Meanwhile, the results of monitoring on the follow-up of recommendations are presented into four groups, namely: (i) in accordance with the recommendation, (ii) not in accordance with the recommendation, (iii) not been followed up; and (iv) cannot be followed up. The implementation of follow-up recommendations on the results of the audit is the responsibility of the auditee. Meanwhile, the BPK only has the authority to monitor the implementation of follow-up actions, and is not allowed to be involved in completing the followup of the auditing report.
In addition to the external audit agency or BPK, in the institutional structure of Government in Indonesia, there is also a Government Internal Supervisory Apparatus (Aparat Pengawas Intern Pemerintah /APIP), which consists of the Supervisory Agency Finance and Development (BPKP), inspectorate general which functionally carry out internal supervision on the provincial, district, and municipal governments. Similarly, the main function of APIP is to carry out internal control through audits, reviews, evaluations, monitoring, and other supervisory activities in order to improve the implementation of internal control (Sistem Pengendalian Internal Pemerintahan/ SPIP). In order to ensure the creation of an effective control environment in government entities, the importance of making the role of APIP is emphasized, especially in the context of providing adequate efficiency and effectiveness of achieving the objectives of implementing the duties and functions of government agencies and providing early warning and increasing the effectiveness of risk management in carrying out the duties and functions of government agencies as well as maintaining and improving the quality of governance in the implementation of duties and functions. Within the scope of public sector governance in Indonesia, APIP has the main function to supervise all activities carried out by regional work units (Satuan Kerja Perangkat Daerah/SKPD) in the context of carrying out government tasks and functions, particularly activities funded by the regional budget. In addition, the inspectorate is also tasked with reviewing financial and performance reports in order to ensure the reliability of the information presented before being submitted by regional executives to the BPK, including supervising, monitoring developments as well as verify the suitability of follow-up with audit recommendations that have been provided by the BPK.

Theoretical literature review
The basic assumptions of public choice theory are developed from economics and political science which assume that individuals will show rational behavior by maximizing their utility (Mueller, 1976). In his article, Mueller (1976) describes the analysis of public choice theory consisting of normative public choice theory and positive public choice theory. Normative public choice theory is related to the analysis of the values and constitutional design of the implementation of the political process in order to create political institutions that are more efficient, fair and responsive and capable of creating prosperity. Meanwhile, the theory of positive public choice is related to the analysis of the observable behavior of related parties, such as voters, political parties, politicians, bureaucrats, other stakeholder groups in the political process, including in conducting political exchange transactions in non-market demand and supply.
In the perspective of positive public choice theory, voters are depicted as having relatively little information about the political and governance process. Although the main role of bureaucrats is to carry out the rules and regulations that have been established, bureaucrats are often involved in making policies and use their specific information to influence and limit the choices of legislators/ communities. Therefore, this then causes bureaucrats to have an incentive to make policies that are not fully known by the community or not in accordance with the wishes of the community, so that the impact is control-loss in the provision of public goods which is mainly a result of bureaucratic negligence, in the form of excess supply of public goods and inefficient production of public goods (Giroux & Shields, 1993). In addition, with an orientation that is only selfish, bureaucrats will use the authority they have in managing resources for the sole purpose of fulfilling their personal interests, including in determining public expenditure budgeting policies (Giroux & Shields, 1993;Schneider & Damanpour, 2002). Therefore, a mechanism is needed that can prevent distortion of information and simultaneously control the behavior of the subnational executives and/or representatives.
In the context of this mechanism, audits can function to provide actual information to the public regarding the administration of government finance. In this case, the local executive will do everything in its power to get unqualified auditing report to demonstrate the government's performance to the public. Apart from that, the audit also details mistakes in financial management, including assessing the potential for corruption. In this context, audits are very influential to become a means of legitimacy for incumbents to gain public support and be an authoritative source to gain votes during elections. On the other hand, negative auditing report will damage the reputation of local executives and be detrimental in elections due to poor public perceptions of financial management. Strachan (2014) specifically analyzes the relationship between information and performance of local government in Indonesia and voting behavior and finds that publishing audit findings can reduce the electability of corrupt incumbents. Likewise, Ferraz and Finan (2008) found that the publication of audit findings can provide an opportunity for challengers to be selected. Chong et al. (2013) also revealed that audit publication is one indicator that can provide information for voters to participate in local elections. Tran et al. (2020) state that the performance of public administration is influenced by information regarding the accountability of public organizations. Ferraz and Finan (2011) showed that with the magnitude of the benefits of the audit report in revealing corruption in local government, the role of the audit in improving the quality of government has been functioning effectively, especially in maintaining the incumbent behavior. Similarly, Olken (2007) on 600 village road projects in [2003][2004] in Indonesia found that monitoring by the community of projects in the village (bottom-up) can reduce the occurrence of corruption in labor costs, and external party audits (top-down) will be more effective in reducing corruption, not only corruption in labor costs, but also can identify corruption at other costs. The reduction in corruption is caused by audit comments (audit findings) produced by the external auditor that can be utilized by the community as supplementary information in providing formal punishment for village governments and as a basis for selecting village officials in subsequent elections, so as to maintain trust and avoid punishment from the community, the village officials as project managers will be more careful in implementing and reporting the results of project implementation in the village. In terms of the relevance of information in the audit report with evaluating government performance, Patanakul et al. (2016) found that the audit was based on the interpretation of an auditor and the external auditor was believed to have competence and free from various interests. The information in the audit report produced by the auditor was generally reliable, traced and could be a useful source of information in assessing corrupt practices and has political consequences for incumbent electability.
The political consequences of fraud/manipulation of financial statements in government will be greater than economic incentives, so that the understanding of institutions in the political context including the impact on elected officials will provide deeper insight into the problem of financial statement fraud in government (Stalebrink & Sacco, 2007). Regarding the political consequences of government financial reporting, several of which have been investigated by Ingram and Copeland (1981). Ingram and Copeland (1981) show that in general the accounting ratios contained in the financial statements of city governments affect public decision-making, particularly on the electability of regional administrators. Problems related to the quality of government financial reporting are the potential for financial reporting fraud/manipulation, costs that may arise due to fraudulent government financial reports (especially when misinformation is revealed) such as debt burdens, reduced trust and citizen welfare, and political trust (Stalebrink & Sacco, 2007). Daniels and Daniels (1991) and Alijarde (1997) found that the information contained in an accrual-based financial report was not even sufficient to evaluate the financial condition of a government organization as a whole and was not functioning optimally as basis for decision-making. Therefore, some additional information is needed that is suitable for the needs of each stakeholder group, including information related to non-financial and economic performance. Boyne and Law (1991) as cited in Pilcher and Radnor (2005) state that the effectiveness of accountability in practice is impossible to achieve without accurate information related to the local government performance, because in the absence of performance information, the concept of accountability and including processes democracy as a whole will become artificial.
Related to the quality of public services as non-financial performance indicators to the public regarding their leader behavior, public services can also influence the political decisions of stakeholders. James and John (2007) found that only when the quality of subnational government performance was poor, performance information negatively affected incumbent electability. However, the results of Nakamura and Kim (2010) study using 1,657 respondents in Japan and 800 respondents in South Korea found that information related to the quality of public services was significantly related to people's satisfaction and trust, which affected people's participation in democracy. Arvate (2013) shows that electoral competition (number of candidates) increases the supply of public goods (student, teacher, and immunization registrations) in the Brazilian subnational government environment, where the increase in the supply of public goods is greater at the time of ex-ante election, and the provision of public goods is higher in areas that have a greater number of voters. Balaguer-Coll et al. (2015) in Spain found that an increase in subnational government spending had a positive effect on the subnational executive electability, which when explored further, capital expenditure (such as the provision of public service infrastructure) had a greater positive effect than operational expenditure.

Empirical literature review and hypotheses development
Corruption will erode the legitimacy of democracy and lead to erosion of community support to the government (Seligson, 2006). Auditors have pivotal role to mark potential areas for corruption (Avis et al., 2018;Everett et al., 2007). Audit reports produced by auditors have been proven to have been used as a basis for evaluating and decision-making by stakeholders (Ferraz & Finan, 2008Olken, 2007;Patanakul et al., 2016;Pereira et al., 2009). Information in audit reports can be used to evaluate the government performance and serves as a source of information to assess the extent of corrupt practices that occur in government and become the basis for community political decisions (Ferraz & Finan, 2008Olken, 2007). Pereira et al. (2009) found that irregularities that occurred at the subnational government during the first term did not affect the likelihood of incumbent electability, and the number of investigations conducted by external audit institutions can reduce the likelihood of re-elected-executives. Ferraz and Finan (2008) found that audit reports that revealed the corrupt behavior of the local executives had a negative influence on the electability. Therefore, the audit findings disclosed in the audit report can form the basis for the community's assessment of the financial management carried out by the local government (Pamungkas et al., 2019). The negative relationship between audit findings and the incumbent re-electability is also mainly stated by Olken (2007), Ferraz and Finan (2008), and Pereira et al. (2009. This means that a high number of audit findings indicates that there are many violations and problems in the management of regional finances. This condition can be perceived as the incapacity to manage regional finances appropriately, effectively, efficiently and accountably. This perception is more likely to reduce public confidence in the incumbent, so that he does not re-elect the incumbent in the second-term election. Based on this description, the hypothesis is stated as follows:

H1. Audit findings negatively affect the incumbent electability in second term.
High-quality government institutions, low corruption and the existence of law enforcement have a positive effect on people's satisfaction with democracy (Wagner et al., 2009). Likewise, when there is a change in the political environment, an increase in the role of local media, enforcement of rules related to financial audit and reporting and tighter measures to reduce obstacles to budget execution are found to have an effect on incumbent electability in the second term (Brender, 2003). This indicates that efforts to improve and enforce the law can increase public satisfaction and trust. Basically, audit findings can illustrate weaknesses in financial management, internal control structures, operational inefficiencies, non-compliance with accounting standards and financial reporting regulations, and corruption that occur in managing public revenues and expenditures (Cohen & Leventis, 2013;Dzikrullah et al., 2020;Johnson et al., 2012;Liu & Lin, 2012;Rahayu et al., 2020;Tumwebaze et al., 2018;Widyaningsih et al., 2019). For deficiencies/weaknesses/non-compliance/violations found in the audit process, the auditor will usually provide audit recommendations to the auditee in order to overcome these problems (Furqan et al., 2020). The purpose of providing audit recommendations can also be interpreted as an auditor's effort to demonstrate preventive measures for the auditee so that similar problems do not recur in the second term (Eckersley et al., 2014). The hope is that when the audit recommendations follow up is carried out properly by the auditee, the implementation of the internal control system and compliance with laws and regulations in the second term can be carried out more precisely, effectively and efficiently.
The legitimacy theory as stated by Suchman (1995) demonstrated that increasing organizational legitimacy will depend on the suitability of organizational behavior with the expectations of the parties around the organization, especially stakeholders, in carrying out its activities. Meanwhile, Power (2003) states that with the growing demands for transparency and accountability on organizations today, auditing will play an important role in producing legitimacy. Gabrini (2013) found that the existence of local government internal audits is more aimed at meeting the expectations of stakeholders, including preventing or blocking certain individual behaviors that are considered dangerous for the operational and strategic objectives of the organization. It can be said that apart from reflecting law enforcement efforts, the implementation of the follow-up of audit recommendations can also illustrate the commitment of local governments to streamline the internal audit function and achieve accountability for regional financial management. When the follow-up of audit recommendations is interpreted as a manifestation of the incumbent's commitment to upholding the law and realizing accountability, an increase in the follow-up of audit recommendations can have a positive impact on public satisfaction in government performance which leads to the re-election of the incumbent in the second-term election. The hypothesis proposed in this study is as follows: H2. Follow-up audit recommendations have a positive effect on the incumbent electability in second term.
Accountability is marked by the achievement of the objectives of the presentation of financial statements in providing relevant and reliable information and can be used as a basis for evaluation and decision-making. However, there are so many potential problems that might threaten the realization of accountability and reduce the value of post-reform accounting practices, including in the application of accrual, consolidated and fair value-based accounting systems (Stalebrink & Sacco, 2007). Harris et al. (2011) found that although most states had published their reports online, due to the low accessibility of the people, the quality of presentation and the procedure for publication of unstructured information, the published reports were not yet fully able contributing to managing relations between the government and citizens, and if this continues it can threaten the trust and participation of the community towards the government.
The problem related to the quality of government financial reporting is the potential for fraud/ manipulation of financial reporting, which can reduce investor/creditor confidence. Costs that may arise as a result of fraudulent government financial statements include an increase in the debt burden on future generations, reduced trust and citizen welfare, and decreased trust in political parties (Stalebrink & Sacco, 2007). Therefore, the political consequences caused by fraud/manipulation of financial statements on the government will be greater than the economic incentives. In the context of political institutions, the quality of financial reports will have an impact on the electability of officials and the integrity of the running government (Stalebrink & Sacco, 2007).
Therefore, when the quality of local government financial reporting is interpreted as the ability of the incumbent to manage and report local government finances, including ensuring fiscal sustainability. Here, improving the quality of local government financial reporting can increase public confidence in the incumbent, so that it will have an impact on the incumbent's vote acquisition and the incumbent's re-electability in the second-term election. Based on the description, the hypothesis proposed in this study is stated as follows: H3. The quality of government financial reporting has a positive effect on the incumbent electability.
Previous research found evidence indicating incumbent efforts to influence people's vote through the quality of public services or performance information. Saliterer and Korac (2013) show that despite having different motives, both regional officials and politicians use performance information for management/internal control purposes and the interests of their external accountability. Likewise, Arvate (2013) shows a relationship between proxied electoral competition and the number of candidates, increases the provision of public goods, such as student registration, teachers, and immunization and finds that the increase in the provision of public goods is greater at ex-ante elections, and higher in regions that have a larger number of voters. Balaguer-Coll et al. (2015) found that an increase in local government spending has a positive impact on the reelectability of regional officials.
Furthermore, capital spending, such as provision of public service infrastructure, has a greater positive effect than operational spending. The role of non-financial performance information, such as public services, is more likely to have an impact on the economic decisions of stakeholders and on the political decisions of stakeholders. Therefore, the quality of public services can increase public satisfaction and motivation to participate in democracy. This can be an incentive for the incumbent to gain the trust of the public, so that it can increase the incumbent's electability in the second-term election. Based on this description, the hypothesis proposed in this study is stated as follows: H4. The quality of public services has a positive effect on the incumbent electability.

Research design
This research was conducted in Indonesia, especially in the 2015 and 2017 local elections. The year 2015 was the beginning of the implementation of local elections simultaneously in 269 regions, while in 2017 the elections were held simultaneously in 101 regions. There were 223 incumbents who participated in regional elections in 2015 and 82 in 2017. However, due to limited data (especially public service quality data not available for 14 regions), the final sample was 291 observations, consisting of 11 Governor Elections, 55 Mayor Elections, and 225 Regent Elections. From the participation, the incumbent who received the most votes was 174 incumbents or around 59.79% of the total Election, consisting of 4 or 36.36% in the gubernatorial election, 37 or 67.27% in the Mayor election, and 113 or 59.11% in the regent election. All data used in this study were sourced from the Indonesian Supreme Audit Agency in form of LKPD for audit findings data, follow-up on audit recommendations, quality of financial reporting and subnational government measurements, from the Ministry of Home Affairs to public service data, status and age of the subnational government, and from the Central Statistics Agency for other data (Table 1).
To answer the research problem and test the hypothesis, the empirical model in this study is as follows: INC tþ1 ¼β 0 þβ 1 FINDING tÀ 1 þβ 2 FOLUP tÀ 1 þβ 3 FINQ tÀ 1 þβ 4 PUBQ t þβ 5 SIZE t þβ 6 GRDP t þβ 7 AGES t þβ 8 MUN t þε t (1) Furthermore, Table 2 outlines step by step the total population to the selection of the final sample. For more detailed information, the study needs to first outline the operational definition of each variable, either dependent, independent and control ones. The variable of INC t+1 in this study refers to the incumbent electability measured by the percentage of votes compared to the total number of valid votes in the following year's election.
The variables of FINDING t , FOLUP t , and FINQ t are associated with the report of the external auditor's auditing report. FINDING t is measured by the number of audit findings disclosed by the auditor in the audit report of the LKPD. FOLUP t is measured by the percentage of the number of follow-up audit recommendations that have been carried out by the subnational government and in accordance with the audit recommendations divided by the total audit recommendations given by the BPK auditor. FINQ t refers to the LKPD quality and is measured by the audit opinion dummy variable, "1" for unqualified opinion (WTP), and "0" for the others. the results of evaluations of the implementation of the tasks and functions of the subnational government (specifically related to public services), with scores ranging from 0.00 to 4.00. The higher the SPM achievement score, the better the quality of public services at the subnational government.
The control variables are SIZE t , GRDP t , AGES t , and MUN t . The variable of SIZE t is measured by natural logarithm (Ln) of total assets of the subnational government. The size of the subnational government illustrates the extent of the availability of human resources or assets owned by the subnational government, both in the form of current assets, investments and fixed assets, including infrastructure and public service facilities, such as roads, irrigation, office buildings, school buildings and so forth. Therefore, it is suspected that government measures can provide satisfaction to the public on the incumbent performance in managing government assets, so as to increase public trust and have a positive impact on the incumbent electability. GRDP t is measured by the natural logarithm (Ln) of GRDP at a constant 2010 per capita price. GRDP per capita is the average income of a population in a district/municipality obtained from the number of GRDP divided by the number of residents in the area concerned, so as to describe the  prosperity/welfare of the population in an area in a certain period. Therefore, it is assumed that per-capita GRDP can have an impact on increasing community satisfaction with incumbent's performance in improving the economy and people's welfare, so that it will have an impact on the greater number of electability.
AGES t is measured by the number of years since the formation of subnational government. The age of the subnational government also shows the experience of the community in assessing the subnational government performance and the previous election. MUN t , is to measure subnational government status which is measured by category: "2" for province, "1" for municipality, and "0" for district. This variable represents the different types of subnational government status in Indonesia. Therefore, the incumbent electability in each region will be different based on the age and status of the subnational government.
Because the local elections observed in this study were held on 5 December 2015 and 15 February 2017, the auditor's report, as well as public service, per capita GRDP, and the size, age and status of the subnational government used were that issued 1 year before the local election (year t). Table 3 illustrates descriptive statistics for all variables. Around 59.79% the incumbent won the election with the mean of INCt variable being 47.54. This could have happened, because in the regional elections which had a minimum of 3 candidates, to win the regional elections, the candidate has to get the most votes with a minimum vote of around 35%. Meanwhile, the FOLUPt, FINQt, and PUBQt variables each showed a mean of 80.73, 0.42 and 2.32. This shows that the average sample has a high follow-up on audit recommendations, as well as good quality financial reporting and public services. The mean value of audit findings of 23.69 can be interpreted that the average sample has a low audit finding. The main variable statistical data indicate that in regions where the average number of audit findings is low, follow-up to audit recommendations is high, and the quality of financial reporting and public services is good, the average vote is around 50% in the second-term elections.

Empirical results and discussion
Related to the existence of control variables, in Table 3 shows that the mean values of SIZE t and GRDB t are 4,777.03 billion IDR and 9.46 million IDR respectively. The high mean government size is due to the inclusion of Jakarta Province in the sample (holding regional elections in 2017), with the highest asset value or around 442,977.22 billion IDR. However, because the size variable of the subnational government is not the main variable in this study, Jakarta Province is still included as a sample. Furthermore, the results of the correlation analysis between each variable are presented in Table 4. Table 4 shows that the INC t variable has a significant positive correlation with the FOLUP t , FINQ t and PUBQ t variables and a significant negative correlation with the FINDINGt variable. Table 4 also shows that there is a significant positive correlation between the INCt variable and the SIZE t and GRDP t variables. Hence, votes are not only correlated with audit findings, follow-up on audit recommendations, the quality of financial reporting and the quality of public services, but also correlates with the size and GRDP per capita of local governments. This is also consistent with the descriptive statistical data previously explained that in regions that have low audit findings and high follow-up audit recommendations as well as good quality financial reporting and public services, the average incumbent votes are also large in regional elections.
Hypothesis testing uses the ordinary least square method with the STATA-14 software program (Table 5).
In general, the results of multiple linear regression testing using robust standard errors indicate that the R-squared of 0.1095, the research model can explain the variation of 10.95% of the vote incumbent with a significance of (F-value) 1% and a mean of 1.38 VIF. Hence, the model does not have multicollinearity and reliability problems to be used in explaining variations in the incumbent electability.
In testing the first hypothesis, Table 5 shows that FINDING variable has a significant negative effect on the INC t variable, with a coefficient of −0.223, significant at the 5%. Thus, H1 was accepted. This means that an increase in the number of audit findings can have an impact on decreasing incumbent electability. The result is in line with Pereira et al. (2009) demonstrating auditing findings conducted by external parties is more likely to reduce the likelihood of re-elected-executives. Ferraz and Finan (2008) also found that audit reports that revealed the corrupt behavior of the local executives had a negative influence on the electability. Moreover, Strachan (2014) also found a link between publishing audit findings and its negative effect to the electability of corrupt incumbents.
Moreover, testing result of second hypothesis showed that FOLUP t variable has a significant positive effect on the INC t variable, with a coefficient of 0.153, p-value <0.05. Thus, H2 was accepted. This means that an increase in the suitability of the follow-up of the audit recommendations can have an impact on increasing incumbent electability. This is consistent with Stalebrink and Sacco (2007) and Brender (2003) stating that voters' trust will be influenced by the extent of accountability and the enforcement of rules in the management of regional finances. Local governments that have followed up in accordance with audit recommendations given by auditors, in the form of improvements to SPI (Sistem Pengendalian Intern) or enforcement of sanctions for irregularities/abuse in regional financial management in previous years, will get a positive response from the voters. Likewise, improving the quality of LKPD can increase the credibility of the subnational government, thereby affecting the voters' trust in the incumbents and increasing their electability.   Table 1. * **, **, * = significant p-value of 1%, 5%, 10%.
Statistical results also confirmed the third hypothesis. The result showed that FINQ t variable was found to have a significant positive effect on the INC t variable, with a coefficient of 3.238, p-value <0.1. Thus, H3 was accepted. This means that LKPD quality can have an impact on increasing the incumbent electability. This is in accordance with Harris et al. (2011) who found that the quality of financial reporting is of positive effect with the trust and participation of the community towards the government, including the possibility to re-elect the clean executives.
Finally, hypothesis testing of the effect of the quality of public service (PUBQ t ) showed that this variable was found to have a significant positive effect on the INC t variable, with a coefficient of 3.732, p-value <0.01. Thus, H4 was accepted. This means that when there is an increase in the quality of public services, the incumbent electability can also be increased. This is empirically proven that voters also base their decisions on the non-financial performance of the government, specifically related to the quality of public services. The results underline the importance of the information related to the quality of public services in increasing public satisfaction and trust as shown by Nakamura and Kim (2010). Furthermore, according to James and John (2007), poor service quality is likely to have a negative effect on electoral votes, and conversely, the qualified public service will positively influence the incumbent electability.

Variable Sign Expectations Coefficient /p-value
Cons making to elect an incumbent in the second term. This is consistent with Olken (2007) and Finan (2008, 2011), stating that the high number of audit findings expressed by auditors is perceived negatively by voters, so voters' trust in incumbents will decrease and cause voters not to re-elect incumbents in the second term Additionally, in testing the effect of control variables, the results showed that the SIZEt has a positive effect on the INCt with a coefficient of 1.274, p-value <0.05. This finding shows that the existence of subnational government assets, both current assets, investments and fixed assets, including public service infrastructure can also contribute to increasing public satisfaction with the incumbent's performance and increasing the electability. These findings confirm Arvate (2013) presenting a link the provision of public goods and vote. It was found that when the subnational government has assets, including the availability of infrastructure/goods of large public services, it can have an impact on the incumbent electability in the second-term election. In addition, the MUNt variable was also found to have a negative effect on the INCt variable with a coefficient of −3.382, p-value <0.1. This finding shows that there are significant differences in the incumbent electability in the elections in all subnational entities. The incumbent electability in district elections is greater than that at provincial and municipal elections. The district community satisfaction and trust on the incumbent performance compared to urban and provincial executives, which is likely to be caused by the plurality of the district community and/or the ease of accessing information related to the government performance.
In analyzing the sensitivity analysis of provinces, Table 6 showed that Jakarta Province has the highest asset value and almost one-third of the total assets of all samples. In addition, Table 5 also shows that there are differences in the incumbent electability based on the status of the subnational government. Therefore, because the average assets of the Provincial Government are greater than the Regencies and Cities, to test the sensitivity of the results of the hypothesis testing, sensitivity testing is carried out by removing from the sample of the Jakarta Province (model 1A) and the whole Province (model 1B (Table 6). Table 6 presents the results of sensitivity testing on the existence of provincial samples. The test results show that there are no differences in the results of the hypothesis testing of model 1A and model 1B with the main hypothesis testing. Although there are differences in the significance of some variables with previous testing, both audit findings, follow-up audit recommendations, the quality of LKPD, the quality of public services and the size of local governments have a significant effect on the incumbent electability. Only the subnational government status variable in Model 1B no longer significantly influences the incumbent electability. The research model is robust on the existence of provincial samples. The difference in the electability only occurred at the province with the district and municipal elections, but no difference was revealed between district and municipal elections. The result showed that there is no difference between the main test results with models 1A and 1B, and because the R-Square main test is greater than the 1A and 1B models, the subsequent analysis still uses elections in Jakarta Province and other Provinces as research samples.
Furthermore, the testing was conducted to analyze the effect of audit findings on incumbent electability. Table 7 presents the results of additional testing of the effect of each type of audit findings on the incumbent electability, using measurements of audit findings in the form of: (i) the number of audit findings related to weaknesses in the Internal Control System (FSPIt) in 1 C model; (ii) total audit findings related to total non-compliance with laws and regulations (FLAWt) in the 1D model; (iii) number of audit findings related to non-compliance with laws and regulations and resulting regional loss (FLOSSt) in Model 1E; (iv) the natural logarithm of the total IDR value of the audit findings related to non-compliance with laws and regulations (LnLAWt) in the 1F model; and (v) the natural logarithm of the IDR value of the audit findings related to non-compliance with laws and regulations and resulting in regional losses (LnLOSSt) in the 1 G model. The test results show that only the number of audit findings related to SPI weakness (FSPIt) has a significant negative effect on the incumbent electability, while audit findings related to non-compliance with legislation both in terms of the amount and value of the IDR, including those that cause corruption does not significantly influence the electability. Therefore, these findings confirm the results of Pereira et al. (2009) who previously found that there was no significant effect of irregularities that occurred in the subnational government on the possibility of re-elected incumbents. Audit findings related to non-compliance with laws and regulations, such as misuse and irregularities in the management and reporting of regional finances, including those related to potential losses to the region do not significantly influence the incumbent electability in the local elections. Voters only consider the number of audit findings, particularly those related to the weaknesses of the implementation of the SPI as a basis for the second-term electability.
Overall, the results indicate that voters will use information about the financial and nonfinancial performance of the subnational government in determining the extent of the incumbent performance. The performance evaluation will be the basis for voters to know the benefits that will be received by them when electing the subnational executives. When the subnational government can carry out its functions, especially in the management of regional finances and quality public services, it can be a basis for an incumbent to gain the votes in the second-term election. Under these conditions, as the assumption of public choice theory, as the party who has the authority to determine policies or the party who has more information, the subnational executives can utilize the authority/information for the purpose of influencing the political choices of the community, electability and personal benefit only, so there is potential for agency problems, in the form of moral hazard or information asymmetry. Therefore, the results emphasizes the importance of auditing in producing legitimacy, particularly the existence of internal audit of the subnational government and external audit in meeting the expectations of stakeholders (Setyaningrum, 2017), so it is necessary to streamline the role of internal audit and external audit, especially in providing the effectiveness of public services, and law enforcement in order to reduce the potential for agency problems and support the realization of regional financial management and public service accountability in the subnational government.

Summary and conclusion
This study analyzes the extent to which public service, and government auditing are able to influence incumbent's electability by using a total of 291 sub-national elections in 2015 and 2017 in Indonesia. The findings indicate that audit findings, and follow-up on audit recommendations are important determinants for constituents to make political decisions. Specifically, the fewer the audit findings, the higher the chances of the incumbent being re-elected. Statistical testing also shows a positive effect between the number of follow-up on audit recommendations Dependent Variables for all models: INC t = Incumbent electability. Model 1C = The audit findings variable is measured by the number of audit findings related to weaknesses in the Internal Control System (FSPIt). Model 1D = Variable audit findings are measured by the total number of audit findings associated with non-compliance with statutory regulations (FLAWt). Model 1E = Variable of audit findings is measured by the number of audit findings related to non-compliance with laws and regulations and resulting in regional losses (FLOSSt). Model 1F = Variable of audit findings measured by the natural logarithm of the IDR value of audit findings related to non-compliance with statutory regulations (LnLAWt). Model 1G = Variable of audit findings is measured by natural logarithms of IDR value of audit findings related to non-compliance with laws and regulations resulting in regional losses (LnLOSSt).
Explanation of the operationalization of other variables in table 1. ***, **, * = Significant P-value 1%, 5%, 10%. and the electability of the incumbent in the second-term election. The results also found a relationship between public services and the output of electoral contestation. In this context, voters use non-financial information on the quality of public services as a basis for decision-making to determine their political preferences.
This study contributes to deepening insights regarding the mechanisms of information influence as a basis for decision-making in an electoral context. In addition, this study contributes an empirical analysis of the broad utility of auditing, which not only provides a normative function as a basis for government accountability and financial transparency, but also becomes a tool for constituents to make their political choices. The practical implication of this finding is an impetus for subnational executives to streamline their accounting and internal auditing functions to ensure the creation of quality financial reporting. For external auditors (BPK), these results have implications for the importance of carrying out a professional and independent audit process to improve the usefulness of financial reporting and reduce the potential for agency problems in regional financial management and public services. In addition, in the context of public management, these results suggest improving the delivery of quality public services at the local level, by encouraging services that are integrated, affordable, and transparent. In addition, the creation of public services that are free of corruption, effective and accountable is also recommended to provide satisfaction to the voting community and to increase the electability of the incumbent.
There are several limitations of this study that can be avenue for future research. First, the analysis in this study only focuses on the electoral output at the sub-national entities by examining available data from various agencies, so that personal and demographic analyzes are not examined. In addition, limitations also include the small variation in electoral votes that may be due to the absence of demographic representation as the basis for testing voter behavior and election results. Furthermore, this study does not examine partially the types of information in financial reporting and public services that are indicators of incumbent electability. Further research is suggested to analyze the extent to which personal and demographic factors of voters and candidates influence their electability. In addition, further research can further analyze the application of the government's accrual accounting system on the quality of public services in subnational governments in Indonesia.