Dept. of Political Science, the Ohio State University
Autumn, December 5, 2006
Background
The primary concern of this study is examining the correlation between political competition and level of corruption in Indonesia.[1] The country has long been positioned by various well-known indices as one of the most corrupt states in the globe. This year’s Corruption Perception Index of Transparency International (TI), for instance, places Indonesia, the world’s largest Muslim country, on its top-ten list (TI website).
This focus is led by two intriguing developments that have emerged since the country gained its democracy back in 1998. For over 40 years, the Indonesian democracy was suppressed under President Sukarno’s and Suharto’s authoritarian rules. In a stark contrast, today’s Indonesia has been experiencing a free-fight political competition induced by two milestones of the 1998 reform.
Political liberalization is the first one. After the parliament lifted the barrier of entry in establishing political party, the number has risen dramatically; from the three tightly restrained parties under Suharto’s regime, to 48 and 24 parties in the 1999 and 2004 general election, respectively.
The second force emerges from the decentralization policy implemented since 2001. The law mandates the central government to surrender a substantial amount of its authorities and to transfer a quarter of the national budget to 465 local governments, namely district (kabupaten) and municipality (kotamadya) (Liddle 2006). Not only does the decentralization affect the fiscal field, it has also changed the political game significantly. In the past’s authoritarian system, the governors, mayors, district heads, and local parliament members were in fact appointed by the central government. Formally, the local parliaments were required to approve the candidates proposed by the central government. Yet, in reality, the legislators played no more than a rubber-stamp role.
Today, the game runs differently. Local politics flourishes and does matter in determining the end-result. Moreover, while members of parliament compete through the open-list proportional representation electoral system, the decentralization law mandates heads of local governments to be directly elected by the people.[2] In brief, the two forces of reform have been pushing a significant increase in the degree of political competition at the Indonesian local scene.
Simultaneously, along with those developments, the country has also experienced an intriguing development in corruption. On one hand, the trend of corruption has been steadily decreasing since the political wave of reform hit the country in 1998. The TI’s corruption perception index scores the country from 1.7 in 1999 to 2.4 this year, with 1 indicates the most corrupt state. However, the most recent score is still lower than that of in Suharto’s era, 2.72 in 1997 (TI website). The International Country Risk Guide - Corruption Index shows a similar pattern. Indonesia’s score has worsened from 3.0 in 1997 to 1.0 during the era reform of 1999-2003 (World Bank website).
The ambiguity continues. Observers are frequently astonished seeing a dramatic increase of corruption trials against politicians, especially those have been occurring at the local stage. According to the Ministry of Home Affairs record, between 2004 and 2006 67 heads of local government or their deputies and 1,062 local parliament members has been indicted on corruption charges. Most interesting, in reaction to such developments, a group of legislators at the national level have petitioned government to stop the wave of trials against their local colleagues. They accuse that the prosecutions are merely the product of political fight and are skewed by the narrow interests (Komisi II-III DPR-RI 2006).
On the other hand, however, there is also an indication that the magnitude of local budget misuses have increased in this very period. The national coalition of anti-corruption NGOs reports that during 1999-2004 the corruption upon local government budgets in 18 provinces has reached Rp.454 billions or equivalent to $45.4 millions.[3]
These backgrounds then direct us to ask: What kind of role the current political liberalization has been playing on corruption? Does it create an impediment or impetus for fraudulent practices? To answer the puzzle, the research question of this study is formulated as follows: What is, if any, the relationship between the increasing political competition and the corruption level in Indonesia?
In line with that effort, this study aims to contribute its findings to fill a hiatus in the corruption study on Indonesia. As warned by Thee Kian Wie (2002), a prominent Indonesian economy historian, corruption—the primary cause of the collapse of Indonesia’s economy after the Asian financial crisis of 1997—has long been neglected by the country’s academia. More importantly, while numerous corruption studies has been conducted upon the context of the developed, the post-Communist or Latin American countries, the research on new democracies in Asia is scarce. In this regard, the twin backdrops we have seen, Indonesia as one of the most problematic country in corruption and as a new-born democracy, offer another contribution of this study to a further theory building that are currently rather underdeveloped.
Theoretical Framework
In this study corruption is delineated along the World Bank’s parsimonious definition, i.e. the use of public office for private gain.
Thus far we have observed the indication of how the current political competition and decentralization might affect the level of corruption in Indonesia. A substantial number of scholarly works done in this particular area have also underlined the significant impact of those two factors on corruption. However, the prevailing theorems are far from reaching an agreement.
First, let us consider the relationship between political competition and corruption. Scholars have found that the strength of political competition tends to reduce the likelihood of corruption by raising the risk of detection and electoral cost. Competitive political environment will constrain politicians’ choices in acting against constituent’s interest on the presence of less corrupt circumstances, whereas the absence of competition will provide a space for political actors to increase the magnitude of extortion by launching a threat of proposing damaging legislations (Rose-Ackerman 1999).
In line with that argument, examining 83 countries by regressing the TI Corruption Perception Index against the index of political competitiveness derived from the Polity III database, Michael Johnston (2002) finds a significant negative correlation between those two variables. Doig offers County Durham of England as another example where the most entrenched corruption is found in the locality with low political competition.
The impact of cartel-party system is the next proof of this point of view. Mair and Kartz have warned that inter-party collusion would prevent the participated parties to check each other’s abuses and could establish an auction of bribery that would result in a higher pay-off (in Johnston 2002). Another confirmation comes from the mainland China. Investigating the failure of country’s anti-corruption campaign, Melanie Manion (2004) concludes that the major impediment should be attributed to the Chinese monolithic party system. In Indonesian context, observers tend to believe that today’s political reform has been enabling ordinary citizen to sanction political actors at some degree and hence it has been increasing politician’s risk to engage in corruption practices (Hamilton-Hart 2001).
Other scholars argue the reverse. The rise of political competition might contribute in increasing corrupt activities and donors’ influence to political actors and parties. Examining corruption in Italy, Golden and Chang (2001) found a strong positive connection between competition among Christian Democratic MPs in haunting campaign funds and the degree of political corruption. However they fail to obtain a significant contribution of the inter-party competition variable.
Susan Rose-Ackerman (1999) proposes a middle way. She argues that the theoretical dispute reveals the importance for researchers to be more precise in measuring the degree of competitiveness to be able to capture its true impact on corruption. She argues that the correlation between the variables is based on what she terms “paradox of stability;” the too-stable political environment could increase the likelihood of corruption, but a very volatile circumstance would also produce a similar result.
Type of electoral system also matters. Scholars have argued that certain types of electoral rules tend to force elected officials more accountable than others and hence to reduce the incentive of corruption. They argue that party list mechanism leads to more corruption than plurality or other political systems in which individuals compete against each other. However, the opposite argument contends that the latter system is prone to more corruption, higher rents and less accountable. This finding is relevant to be considered since Indonesia is currently adopting the open-list proportional representation system in which previous works have detected the existence of a significant correlation between intra-party competition and level of corruption.
The second important explanatory factor we need to discuss is decentralization. Thus far, like political competition, most scholars have reach agreement about a close linkage between decentralization and corruption. However, there is still much disagreement of how the former would predict the latter. Some argues that decentralization can not be generalized across countries due to its nature that is highly context-dependent.
At the optimistic camp, scholars believe that decentralization brings check and balance mechanism closer to citizens and hence it enables them to closely evaluate and then reward or sanction elected officials based on their performances. Put another way, as local democracy flourishes corruption shrinks. Inter-jurisdictional competition among localities provides another force toward that direction. Others at the opposite side argue that decentralization promotes corruption. The power diffusion to the local level might also be followed by the distribution and multiplication of fraudulent practices in a less controllable manner. Less professional local officials with a much larger discretion of authorities then only exacerbates the problem (Bardhan and Mookherjee 2005).
These two different views have been supported by evidence provided by the respective studies. Fisman and Gatti (1999), for instance, find a strong negative correlation between fiscal decentralization and corruption level across countries. They run a statistical analysis over the aggregate data analysis derived from the International Monetary Fund’s Government Finance Statistics in the period of 1980-95.
In contrast, reviewing previous studies on this subject, Bardhan and Mookherjee (2005) conclude that most cross-country empirical studies fail to show strong evidence of the significant impact of decentralization on corruption reduction. The most widely cited study in this line of argument was conducted by Shleifer and Vishny (1993). They conclude that centralized polities are less prone to corruption than systems where power and authority are more diffused. In parallel with that, the 1993 report of Great Britain Audit Commission indicates that the country’s decentralization policies “have increased the risk of fraud and corruption occurring” (in Klitgaard et al. 2000, 10).
In the case of Indonesia, Ari Kuncoro surveyed 1800 corporations across the country and he found a decline of the level of bribery (as a percentage to total production cost) reported by the local business leaders after the decentralization law enacted (2002). However, he notes that the finding should be reserved and need to be further examined due to a suspicion that the decline of bribery might related to a drop in corporation’s profit due to the crawling economy following the 1997 Asian financial crisis. It should be noted that another Kuncoro’s finding shows the impact of decentralization on the increase of business uncertainty. In regard that this study will be focusing on political competition as its independent variable, I take decentralization into account as a control variable, explained later.
Given the theoretical disputes we have seen, this study would make a contribution in a theory testing on the interaction among those variables in the context of Indonesia, one of the most corrupt countries in the world as indicated previously.
Hypotheses
Based on the background and theoretical framework elaborated above, I state the hypotheses as follow:
a. The increase of political competition induces a net effect of the decrease of corruption at local level.
b. The correlation stated on (a) occurs only when decentralization taken into account;
c. Considering (a) and (b), the largest decrease of corruption would present at the district/municipality level.
Unit of analysis
There are two units of analysis of this study, i.e.:
a. For the aggregate data analysis: The entire 33 provinces and 465 districts and municipalities in Indonesia. District and municipality are the same level government. They differ in term of type of area coverage. The former pertains to the rural area while the latter to the urban. The decentralization law renders autonomy to the municipality and district and positions provincial government as the arm of central government in coordinating and monitoring the implementation of local governance. However, this study takes province into account since the political competition and the dynamics of corruption at this level also shows a significant development. Moreover, a comparison between the dynamics in provinces (that functions as the central government’s arm) and municipality/districts (as the unit of decentralization) would provide a useful insight to reveal the net effect of decentralization.
b. For the survey: Corporations in 33 provinces. These corporations will be selected with the systematic random sampling method from a list provided by the company index provided by the Ministry of Justice. In Indonesia, every corporation entity should be legally registered in this department.
Period of study
The time span of this study is a period from 1999 to 2009 that covers three elections of 1999, 2004, and 2008 after reformasi, the era reform marked by the fall of the authoritarian Suharto’s regime in 1998.[4] This period encompasses the two major shifts in Indonesian political system this study concerns; the political liberalization in 1999 and the decentralization policy in 2001. The problem of data availability prevents this study to include the era of Suharto’s regime. However, I would also take this period into account when I employ a historical analysis to shed light on the current pattern.
Variables and Source of Data
a. Independent Variable:
Political competition, as measured by two following estimators:
a.1. VOTE-INTRA: This variable is employed to capture the intra-party competition that has a significant impact on corruption. Indonesia’s recent political system is multi-party presidentialism. In term of electoral system the country, as I have mentioned previously, is adopting the open-list proportional representation in which theories has indicated that intra-party competition might play a significant effect on level of corruption. For this purpose I will be examining the candidate lists in the three general elections and will select three top MP candidates from each three top parties in every locality, including at the provincial level. The degree of competitiveness will be judged by the number of votes they obtained. The closer number of votes obtained by each candidate the higher level of intra-party competition. Such judgment will be coded into an index of competitiveness (coded 1 to 5 in which 1 indicates the lowest level of intra-party competition).
Data source: KPU (The National Election Commission).
a.2. VOTE-INTER: Vote share at provincial and district/municipality level per election cycle. This variable aims to capture the dynamics of inter-party competition. In every region, only three top parties at the respective parliaments are taken into account. A close difference of the proportion of vote share would determine the higher level of competitiveness. This variable will also be transformed into an index with a similar method with the one of VOTE-INTRA.[5]
Data source: KPU.
a.3. TURNOVER: Turnover rate of local government head (governor, mayor/district head) in the respective localities.[6] This is defined as number of changes of local government heads accumulated in the period of study, from 1999 - 2009 (adapted from Przeworski et al. 2000).
Data source: KPU.
b. Dependent Variable:
The level of corruption at the provincial and municipal or district level, as measured by two following estimators:
b.1. INDICTMENTS: Number of indictments on corruption charges against local government heads and parliament members during the period of study. This includes any charges relating to the anti-corruption law and/or to the anti-corruption provisions in the criminal penal code.[7] For these study purposes, the number of indictments is preferable rather than the number of verdicts due to the latter is more likely distorted by the high level corruption in the Indonesian judiciary system.
Data source: Ministry of Home Affairs and the national coalition of anti-corruption NGOs (GERAK).
b.2. CORRUPTED BUDGET: The amount of local budget loss in corruption or other malfeasance practices as concluded in the audit report of BPK (the Supreme Audit Agency).
Data source: BPK’s quarterly publications.
b.3. BRIBERY: The amount of illicit payment and/or gifts rendered by corporations to the bureaucracy, member of parliaments, and/or the political parties.
Data source: Survey.
b.4. PERCEPTION: The corporate executive’s perception toward the level of corruption in relations with the level of political competition and decentralization. Gunnar Myrdal (2002) underlines the importance of public perception in anti-corruption drives. Examining corruption in South Asia, Myrdal argues that “the folklore of corruption,” people’s belief of how corruption can or can not be combated, significantly determines the level of corruption and the success of anti-corruption campaign.
Data source: Survey.
c. Control Variables:
c.1. DECENTRALIZATION: Total expenditure of local government (province and district/municipal) over total spending of all levels of government (national, province, district/municipality) per annum. I choose expenditure rather than revenue, since the former could better capture corruption than the latter. This variable is adopted from Fisman and Gatti’s study (1999).
Data source: BPS (the National Statistics Agency).
c.2. ECONOMY: The level of economic development and affluence of each locality measured by the respective income percapita. This is an important control variable since most prior studies have reached an agreement that the economic wealth is a strong predictor for a less corrupt society. The importance to control level of economic development has been shown by Johnston (2002). His study displays that by including GDP percapita into the regression-equation, the GDP becomes a strong negative predictor of corruption and changes the political competition variable as a not significant explanatory variable.
Data source: BPS.
c.3. EDUCATION: Literacy rate at each localities as determined by the national demographic survey.
Data source: BPS.
c.4. CIVIL ASSOCIATIONS: Number of civil society associations (NGOs, religious and other secondary organizations). Dinissa Duvanova (n.d.), for instance, has shown the positive correlation between level of bureaucratic corruption and business association formation in 25 post-communist countries.
Data source: The Ministry of Justice.
c.5. RELIGIOSITY: Level of religiosity of each locality’s population. This study will borrow the findings of LSI and other’s surveys in Indonesia (Mujani and Liddle, forthcoming)
Data source: LSI survey (Indonesia Survey Institute).
c.6. REGION: This is the dummy variable to distinguish localities in Java, coded by 1, and they are in outer Java (0). Kuncoro’s study (2002) points out the importance to capture the Java-outer Java variances due to a wide socio-economic disparity between two regions. Examining the correlation between corruption and decentralization in Indonesia, he found the significant differences between two regions.
Methodology
Data validity and bias of measurement are the two common major challenges of study on corruption in regard to its secretive and illicit nature. Moreover, due to the institutionalized nature of corruption in Indonesia, such threat is even more prominent. These circumstances are recognized as the major limitation of this study. To overcome those obstacles I will generate four different methods of data collections. This multi-stage method would provide an advantage in which the validity of each resulted data can be checked one against to another. By doing this I expect to be able to mitigate the threat to a minimum level.
I will conduct the data collections with the following steps:
a. In-depth interview.
I will initiate the whole process of data collection with in-depth interviews with the key national figures in this field. For this purpose the semi-structured interview technique will be utilized. The sources are including, but not limited to: major political party leaders and legislatures, the commissioner and commissioner deputies of the Corruption Eradication Commission (KPK), head and staff of the Presidential Unit for Reform Management, the attorney general and his staff, the chief of police force and his staff, and leaders of corruption-watch NGOs, such as Transparency International - Indonesia. The main purpose is to obtain an accurate picture of the variables and their interactions that will then be utilized to examine and refine this research design and to determine the direction of the aggregate data analysis and survey.
b. Aggregate data collection pertaining to the relevant variables as described above.
c. Nation-wide survey.
I will be focusing the survey on local business leaders across the country. The corporations are selected as the unit of analysis relating to the three primary objectives of the survey. First, it tries to estimate the magnitude of corruption with measuring the amount of illicit payment to local bureaucrats, parliament members, and political party leaders. I suspect they play two pivotal roles in the scope of corruption concerned by this study; as the major bribe providers and also the victims of political extortion. I expect that surveying the giving side of corruption would result in a greater reliability than asking those in the taking side.[8] Secondly, the survey will try to measure the perception towards the current level of corruption and its relations to political competition and decentralization. The two areas of findings can not be captured by the aggregate data analysis. More importantly, the survey will also function as a validity-check to the aggregate data that are threatened by corruption within the national audit agency and judiciary system.
The survey will be conducted through face to face interviews with one executive or their representative in each firm. Corporations surveyed here are the ones that fall into two size categories defined in Kuncoro’s study (2002), i.e. medium corporations for those having annual sale amount of Rp.1-10 billion and large corporations for those with the annual sale of more than Rp.10 billion. Small firms are not included since they are not significant players in regard the political corruption this study concerns.
d. Contextual study.
At the final stage, I will employ this type of study in two districts and municipalities, each in Java and outer Java. To avoid the threat of selection bias, the two localities are selected randomly from those having high number of corruption indictments and budgetary irregularities. In-depth interviews with political and business actors will be employed to capture the dynamics among the variables of this study that can not be entirely portrayed by the two preceding methods.
To examine the net effect of political competition on the corruption level, I will conduct the following analysis:
a. Statistical analysis upon aggregate and survey data.
This analysis will be performing three main tasks. First, it will compare the variances inter municipality/district data to examine the differences between municipalities/districts. Secondly, it will investigate any contrast between data obtained in the provincial level versus the ones in municipalities/districts. This comparison would provide a benefit in investigating the individual impact of decentralization. As explained before, the decentralization law is implemented at the district/municipality level, not at the provinces. Therefore, by comparing the data of the latter against those of the former, we could inspect how the decentralization intervene, or not, the correlation between political competition and corruption level. And the third is to measure the magnitude of corruption using the aggregate and survey data.
b. Qualitative analysis.
I employ this type of analysis upon the data resulted in the in-depth interview and contextual study in order to get insights on the dynamics shaping the correlation, if any, among corruption, political competition, and other relevant variables. Including this type of analysis is the historical analysis on the path of Indonesian corruption.
* * *
References
Bardhan, Pranab, and Dilip Mookherjee. 2005. “Decentralization, Corruption and Government Accountability: An Overview.” Revised draft for Susan Rose-Ackerman and Edward Elgar, ed. Handbook of Economic Corruption. Forthcoming.
Duvanova, Dinissa. n.d. “Bureaucratic Corruption and Collective Action: Business Associations in the Post-Communist Transition.” Department of Political Science, the Ohio State University.
Fisman, Raymond, and Roberta Gatti. 1999. “Decentralization and Corruption: Cross-Country and Cross-State Evidence. Washington, D.C.: The World Bank.
Golden, Miriam A., and Eric C.C. Chang. 2001. “Competitive Corruption: Factional Conflict and Political Malfeasance in Postwar Italian Christian Democracy.” World Politics, 53, 588-622.
Hamilton-Hart, Natasha. 2001. “Anti-Corruption Strategies in Indonesia.” Bulletin of Indonesian Economic Studies, 37(1), 65-82.
Johnston, Michael. 2002. “Party Systems, Competition, and Political Checks.” In Arnold J. Heidenheimer, and Michael Johnston, eds. Political Corruption: Concepts & Contexts, 3rd ed. New Brunswick and London: Transaction Publishers, 777-794.
Klitgaard, Robert. 1988. Controlling Corruption. Berkeley, Los Angeles, London: University of California Press.
Klitgaard, Robert, Ronald Maclean-Abaroa, and H. Lindsey Parris. 2000. Corrupt Cities: A Practical Guide to Cure and Prevention. Oakland, California: CS Press.
Komisi II-III DPR RI (Commission II and III, the House of Representatives). October 3, 2006. Laporan dan Rekomendasi Panja Penegakan Hukum dan Pemerintahan Daerah (Report and Recommendation of Law Enforcement and Local Government Working Committee). Jakarta: The House of Representatives of the Republic of Indonesia.
KPPOD website (Komisi Pemantauan Pelaksanaan Otonomi Daerah, Monitoring Commission on the Regional Autonomy Implementation). Database Perda (Local Regulation Database).
Kuncoro, Ari. 2002. “The New Laws of Decentralization and Corruption in Indonesia: Examination of Provincial and District Data.” Presented at the IMF Research Department, Washington, D.C.
Liddle, R. William. 2006. Indonesia. In Shively, ed. Comparative Governance: Political Structure and Diversity Across the Globe. USA: McGraw-Hill Primis.
Manion, Melanie. 2004. Corruption by Design: Building Clean Government in Mainland China and Hong Kong. Cambridge and London: Harvard University Press.
Mujani, Saiful, and R. William Liddle. “Leadership, Party, and Religion: Explaning Voting Behavior in Indonesia.” Comparative Political Studies. Forthcoming.
Myrdal, Gunnar. 2002. “Corruption as a Hindrance to Modernization in South Asia.” In Arnold J. Heidenheimer, and Michael Johnston, eds. Political Corruption: Concepts & Contexts, 3rd ed. New Brunswick and London: Transaction Publishers, 253-264.
Przeworski, Adam, Michael E. Alvarez, Josẻ Antonio Cheibub, and Fernando Lumongi. 2000. Democracy and Development: Political Institutions and Well-Being in the World, 1950-1990. Cambridge, New York: Cambridge University Press.
Rose-Ackerman, Susan. 1999. Corruption and Government: Causes, Consequences, and Reform. New York: Cambridge University Press.
Shleifer, Andrei, and Robert M. Vishny. 1993. “Corruption.” Quarterly Journal of Economics, 108(3), 599-617.
Thee, Kian Wie. 2002. “The Soeharto Era and After: Stability, Development and Crisis, 1966-2000.” In Howard Dick, et.al., eds. The Emergence of a National Economy: An Economic History ofIndonesia, 1800-2000. Australia: Allen and Unwin.
TI (Transparency International) website. TI Corruption Perception Index.
UNODC (United Nations Office on Drugs and Crime). Why a Global Programme Against Corruption?
World Bank website. Worldwide Governance Indicators Country Snapshot.
End Notes:
[1] I thank Prof. Richard Gunther, Prof. R. William Liddle, and Dinissa Duvanova for their helpful insight and advices.
[2] Law No. 32/2004 on Local Government.
[3] Korupsi Dana APBD Rp. 454 Miliar (Corruption of Local Budget Reaches 454 Rupiahs). Sept 18, 2004. Jawa Pos.
[4] This study is intended as a draft of my dissertation proposal. The end of the period of study, year 2009, is determined by the time when the field-study will be started.
[5] The index has not been precisely constructed and will be developed later by adapting the Ranney’s index into Indonesian context.
[6] Prof. Richard Gunther suggested this estimator.
[7] Law no. 31/1999 on Corruption Eradication as amended by law no. 20/2001.
[8] Personal communication with Dinissa Duvanova, November 22, 2006.
No comments:
Post a Comment