“New public management and corruption: Empirical evidence of local governments in Indonesia”

This study is relevant because it examines the determinants of corruption in local governments that have a negative impact on the success of sustainable development. This study aims to examine the effect of New Public Management (NPM), as measured by fiscal decentralization, financial reporting quality and independent audits, on the level of corruption. The sample consisted of 433 local governments in Indonesia based on data from 2011–2017. PLS-SEM was used as a data analysis technique. The results test shows that fiscal decentralization positively affects corruption with a path coefficient of 0.19 and a p-value of 0.004. The quality of financial reporting has a negative effect on the level of corruption with a coefficient of –0.26 and a p-value < 0.001. Hypotheses testing results also show that audit finding positively affects corruption with a coefficient of 0.10 and a p-value < 0.10. On the other hand, follow-up audit results have no significant effect on corruption with a p-value > 0.10. This study concludes that the NPM mechanism in the form of fiscal decentralization positively affects corruption. These results imply that fiscal decentralization needs to be balanced with good governance, among others, by increasing the quality of financial reports and independent audits.


INTRODUCTION
Based on a survey conducted by the Association of Certified Fraud Examiners (ACFE), corruption ranks first among fraud schemes that occur in Asia Pacific with a percentage of 51% (ACFE, 2021).Meanwhile, the ACFE Indonesia Chapter in the 2019 Indonesia Fraud Survey Report also shows that the most common type of fraud in Indonesia is corruption with a percentage of 64.4% of cases (ACFE Indonesia, 2020).Based on the ACFE Indonesia 2020 report, 167 respondents, or 69.9%, stated that corruption was the most detrimental act in Indonesia.Government organizations are the institutions that suffer the most because of the fraud (ACFE Indonesia, 2020).To overcome the problem of corruption, governments in various countries have carried out public sector reforms known as New Public Management (NPM), including fiscal decentralization, the application of accrual-based accounting, and the implementation of independent audits.However, research results indicate that the effectiveness of the NPM mechanism is still a debate and an important research question (Changwony & Paterson, 2019).The practice of decentralization in Indonesia after 15 years of implementing regional autonomy and fiscal decentralization has caused many problems in the regions, one of which is the increasing level of corruption (Boolaky et al., 2018;Sylvia et al., 2018).
Previous research on the effect of fiscal decentralization on corruption have obtained mixed evidence.Studies that provide evidence of a negative effect of fiscal decentralization on the level of corruption include Fisman and Gatti (2002a), Fan et al. (2009), Ivanyna and Shah (2010), and Gurgur and Shah (2014).In contrast, the results of other studies actually provide empirical evidence that fiscal decentralization actually increases corruption (Triesman, 2000;Fisman & Gatti, 2002b;Saputra, 2012;Shon & Cho, 2019; Ulum et al., 2019).To explain the inconsistency of previous research, it is important to analyze other NPM mechanisms besides fiscal decentralization.

LITERATURE REVIEW
Agency theory describes a contract between a principal that asks an agent to do work on his behalf, which includes the giving of some decision-making authority to the agent (Jensen & Meckling, 1976).Local government in the context of fiscal decentralization is basically a contractual relationship between agents and principals (Zimmerman, 1977).This contractual relationship is manifestly embodied in various regulations in the government sector, including in Law Number 32 of 2004 and 33 of 2004.These two laws and their elaboration or regulatory derivatives to the technical level are the basis of legitimacy for fiscal decentralization for local governments in Indonesia.
The NPM mechanism is designed to fulfill the principles of good governance in agency relations in government organizations (Fisman & Gatti, 2002a Government regulations related to fiscal decentralization are the legal basis for a contractual relationship between principals and agents (Zimmerman, 1977;Malagueno, 2010;Liu & Lin, 2012;Atuilik, 2016;Jeppesen, 2019).In this contractual relationship, the executive is given the authority to make fiscal decisions by the principal to carry out the actions or activities necessary to fulfill the wishes of the principal (legislative / people).This authority makes executives have discretionary power or freedom of action.Therefore, the activities carried out by the executive cannot always be observed by the legislative / the people (as the principal), resulting in information asymmetry.By assuming that both parties will maximize their utility, it can be predicted that the executive will not always behave in the interests of the principal (Jensen & Meckling, 1976).
Apart from being able to be explained by agency theory, the effect of fiscal decentralization on corruption can also be approached by using the the-ory of fraud triangle theory (Cressey, 1953), which consists of pressure, opportunity, and rationalization.High political costs resulting in high-cost recovery are a strong pressure to commit corruption.
In terms of the fraud triangle theory, the opportunistic behavior of the executive and legislature can be explained through three points of view such as pressure, perceived opportunity, and rationalization (Cressey, 1953).First, the pressure point of view.It is common knowledge that in the context of politics in Indonesia, the executive requires high political costs.Therefore, cost recovery by the executive and legislature is a necessity and a strong pressure to commit corruption through the budget.Second, the perspective of perceived opportunity.The regional head and all the bureaucracy have the authority to compile and implement the budget based on laws and regulations governing regional finances in Indonesia.Therefore, executives become very aware (have general information) and also have expertise (technical skills) regarding public sector budgets that are a requirement for perceived opportunities.Two conditions that are parts of the opportunity in the fraud triangle have been fulfilled, so that the opportunity to commit corruption becomes an easy thing.Third, the point of view of rationalization.In the pre-decentralization era, corruption occurred at the level of the central elite, and when fiscal decentralization occurred, the corrupt behavior of the central elite became a justification for the local elites to take the same action.In the end, this corrupt behavior becomes a kind of snowball, which inspires legislators and executive officials in other regions to do the same.Corruption is an executive opportunistic act in local governments that is detrimental to the principal (the community), therefore, the principal is involved in various forms of supervision of their agents.The literature on delegation of authority identifies four main measures by which principals can limit or reduce agency loss: contract design; screening and selection mechanisms; monitoring and reporting requirements; and institutional checks (Malagueno,  According to the State Financial Audit Standards, auditors must examine the internal control and design audit programs (Ulum et al., 2019).The auditing standard also requires an examiner to report fraud and deviation from statutory provisions to the competent authority in accordance with the applicable provisions of BPK.
Audit reports by the Supreme Audit Board are more precisely the findings of an examination of financial reporting weaknesses.Therefore, the more findings or detection of irregularities, the greater the incidence of corruption.The main objective in the audit of the local government sector is to test whether budget management is in accordance with regulations and conclude whether there are material misstatements in financial reporting (Malagueno et Kurniawati and Pratama (2021) show that irregularities detection positively affects corruption.
The number of irregularities detected in government auditing is a reflection of how many violations there are in the local government bureaucracy.However, an audit report that exposes this impropriety is not effective enough to deter corruption.Therefore, detecting irregularities in government auditing is only the first step that results in several recommendations for improvement by the auditors and a very important second step is to hold local governments accountable for making corrections and implementing improvements as recommended by the auditors (Aikins, 2012; Liu & Lin, 2012).
Corruption is an opportunistic behavior of local government executives, therefore, from the perspective of agency theory to reduce this, it is monitored by auditors of BPK.The auditee is obliged to follow up on the recommendations as stated in the management letter.Furthermore, they are also responsible for reporting on the status of the follow-up to the recommendations from the BPK audit results.Government audits will only work as a deterrent to corruption if recommendations for follow-up corrections to abuse in the process of receiving and disbursing funds are carried out entirely (Malagueno et

AIMS AND HYPOTHESES
This study aims to analyze the effect of the New Public Management (NPM) mechanism as measured by fiscal decentralization, financial reporting, and the implementation of independent audits on the level of corruption of Indonesian local governments.
Based on the literature review, this study proposed hypotheses: H1: Fiscal decentralization positively affects the level of corruption at local governments.
H2: Financial reporting quality negatively affects the level of corruption at local governments.
H3: The finding of the audit positively affects the level of corruption at local governments.
H4: Follow-up on the recommendations of the audit results negatively affects the level of corruption at local governments.

METHODS
All local governments in Indonesia were the population of this study.Sampling was carried out purposively between 2011 and 2017.The dependent/ endogenous variable is the level of corruption.The measurement of corruption (CORP) in this study refers to the real measurement that has been done.The corruption level in this study uses real losses due to corruption cases according to the final decision of the court.The formula for measuring the fiscal decentralization (FD) variable is developed into three indicators, namely fiscal decentralization (DFORI), fiscal decentralization per population (DFPOP), and fiscal decentralization per area (DFAREA).
Financial statement quality is measured by the Indonesia Supreme Audit Agency data.Audit findings are measured by using two indicators, namely findings of systems weaknesses and the number noncompliance findings.Follow-up recommendations for audit results are measured by the total rupiah value of findings that have been followed up by the regional government in audits conducted by the BPK for periods t and t-1.

RESULTS
The study population consisted of 514 local governments in Indonesia.The final sample consisted of 433 local governments.Table 1 presents descriptive statistics for each research variable.On average, the value of corruption is at a moderate level with an average loss value of IDR 1,723,966,769.19.Overall, the level of fiscal decentralization in Indonesia is relatively large, as indicated by the average magnitude of the three indicators.Descriptive statistics also show that there is still a fair amount of evidence about internal control weaknesses and non-compliance with laws and regulations, as well as the audit results of BPK.Meanwhile, the follow-up examination results were still mixed for periods t and t-1.The financial statements of local governments (untabulated) show that most are given unqualified (65.6%), qualified (15.8%), unqualified with explanatory paragraphs (6.6%), adverse (0.5%), and disclaimer opinion (11.5%).
This study used latent/unobserved variables and was measured using formative/index indicators.
Analysis of the measurement model can be uti-lized from the feasibility of the formative indicator by looking at the significance value of weight and co-linearity (variance inflation factor/VIF).The result of the measurement model using WarpPLS 7.0 in Table 2 shows that p-value for weight significance of all indicators of formative variables < 0.001 and co-linearity of all formative indicators shows VIF value < 3.3.Thus, the measurement model has fulfilled the criteria for formative constructs.Table 3 shows the hypothesis testing results based on the structural model.The result of this study provides empirical evidence that H1 is accepted, with the coefficient 0.19 and p-value 0.004.H2 is accepted with the coefficient of quality of financial reporting -0.26 and significant with p-value < 0.001.H3 is also accepted with the coefficient of audit findings 0.10 and p-value 0.095 (< 0.10).Meanwhile H4 is not accepted due to p-value 0.435 (> 0.10).

DISCUSSION
The result of this study provides empirical evidence that Hypothesis 1 is supported.Test results of Hypothesis 1 in this study support the agency theory.The results of the study support Prud'homme's (1995) argument that corruption in developing countries is more common at the local government level than at the national level.This is because executives in local governments are often pressured to comply with the demands of their political groups.With the existence of greater public financial management authority after the era of fiscal decentralization, the behavior of economic rent will increase.The findings of this study also support Cressey's (1953)  Furthermore, this finding also supports the fraud triangle theory that fraud can be reduced by limiting opportunities, among others, by practicing transparency and accountability in financial statements that are presented in accordance with accounting standards.The empirical evidence from the results of this study also shows the importance of the role of financial audits, indicated by their negative influence on the level of corruption.The results of this study also support the argument of Liu and Lin (2012), which states that the fundamental purpose of financial auditing is to supervise, guarantee, and support government accountability, which is an important institutional part of modern government governance.
The test results show support for Hypothesis 3 that the number of audit findings has a positive effect on the level of corruption.The more the financial audit findings in the form of weaknesses in the internal control system and non-compliance with laws, the greater the level of corruption.These results support the fraud triangle theory argument that weaknesses in internal control are an opportunity that can lead to fraud.In addition, this result supports Liu and Lin (2012)

CONCLUSION
The purpose of this study is to examine the impact of the New Public Management variables, namely fiscal decentralization, quality of financial statement, audit findings, and audit follow-up on the level of corruption.The results of this study provide empirical evidence that fiscal decentralization positively affects corruption.Financial audits in the form of providing opinions on financial reports have a negative effect on the level of corruption.Meanwhile, weaknesses in the internal control system and non-compliance with laws and regulations tend to increase the level of corruption.These results provide support for the agency theory and the fraud triangle theory.According to the fraud triangle theory, fiscal decentralization provides an opportunity for corruption in local governments in Indonesia.Meanwhile, financial statement and audits act as deterrent factors that reduce opportunities for corruption.Furthermore, the results of this study show that the NPM mechanism in Indonesia in the form of fiscal decentralization can actually increase corruption.The results of this study imply that fiscal decentralization in local governments in Indonesia needs to be controlled by other NPM mechanisms, namely by improving the quality of financial reports and auditing financial management through independent audits.
This study has limitations in measuring fiscal decentralization and corruption based on various proxies.In this study, an attempt was made to measure corruption using the best approach, namely the total level of losses caused by fraud.Corruption is a complex phenomenon that can provide future research opportunities to add variables such as political costs, culture, financial performance, politics variables, etc.When measuring variables, of course, the availability of data must be taken into account.
Several previous studies have shown empirical evidence that fiscal decentralization positively affects corruption.Saputra (2012) and Ulum et al. (2019) provide empirical evidence that decentralization will increase the level of corruption in local government in Indonesia.Albornoz and Cabrales (2013) and Shon and Cho (2019) also show that decentralization positively affects the level of corruption.In the sample of local government in Indonesia, Maria et al. (2021) show that independence in financial management increases the amount of corruption.

Figure 1
Figure 1 presents the WarpPLS 7.0 output for the structural model test results.

Figure 1 .
Figure 1.Structural model results 2010; Liu & Lin, 2012; Atuilik, 2016; Jeppesen, 2019; Furqan et al., 2020; Hamed-Shidom et al., 2022; Lino et al., 2022).Monitoring and reporting force executives to share with the legislature and the public information they may not have obtained.Law Number 32 of 2004 requires regional heads to provide accountability reports to the legislative, which have been examined by the Audit Board of the Republic of Indonesia (in the Indonesian context, it is known as the Badan Pemeriksa Keuangan, hereinafter abbreviated as BPK).

Table 2 .
Measurement model evaluation

Table 3 .
Hypothesis testing results