Research on decentralised fiscal responses to corruption is sparse in Indian federalism. This paper examines the effects of fiscal decentralisation on corruption in public service delivery process across states in India. Econometric estimations broadly suggest a varied impact of fiscal decentralisation on retrenchment of the incidences of corruption experienced by households while availing public services. Besides, a few socio-economic determinants enable people to raise their ‘voice’ against corruption. To correct the quality and quantum of transfers (formula-based and discretionary) that are characterised by adhocism and arbitrariness, is argued as an effective mechanism to reduce corruption in public service delivery process.
Keywords: corruption, fiscal decentralisation, intergovernmental transfers, public service delivery, citizen activism
JEL Codes: H77; D73
Introduction
Decentralisation is defined as a practice of the devolutions of political (functions), administrative (functionaries) and fiscal (finances) authorities, responsibilities and resources through deconcentration, delegation or devolution from central government to state governments to locally elected bodies in rural and urban areas for greater regional autonomy (Oates, 1972; Shah, 1994; World Bank, 2004). Decentralisation has lingered undefined at the applied sphere as it has grown as an adaptable practice across many emerging economies, and thereby, the overall decentralised public management has remained an ongoing contentious area. Empirical evidences regarding impacts of decentralisation on public service delivery greatly differ across and within countries. Fiscal decentralisation, which is a subset of decentralised governance where funds will be devolved after the devolutions of both functions and functionaries, is neither good nor bad for effective public service delivery.
Proponents of decentralisation ponder it as a solution, which leads to improve the public management in an economy (Shah, 1998) whereas opponents perceive that the impact of decentralisation on public management and service delivery will not produce the desired results unless: one, local level governments are given full authority to make decisions by their own as well as full access to resources (complete devolution of power); two, there is well understanding and healthy relationship among all the three tiers of governments; three, political healthy competition exists within each tier; and four, local politicians have abundant understanding of local needs (Prud’homme, 1995; Tanzi, 1996; Torres and Pachón, 2013). Downs (1957) claimed that in each tier of governments, healthy competition for elected office would lead the representatives to make decisions that are favoured by the median voter.
This paper assumes significance of fiscal decentralisation and corruption in public service delivery in Indian context for three important causes. The first reason is the legal fiat, i.e., the 73rd (Rural Local Bodies) and 74th (Urban Local Bodies) Constitutional Amendment Acts (CAAs) that had given recognition to the local self-governments as the third tier of governments. The second is the size of local level governments, where there are three tiers of the Panchayati Raj Institutions (PRIs), which are: Zilla Panchayat (or Zilla Parishad) at District level, Taluka Panchayat (or Panchayat Samiti) at Block level and Gram Panchayat at Village level. These tiers collectively constitute one of the biggest public self-governments in any democratic polity with as many as 259704 PRIs till 2017-18. The third reason is the asymmetries in transferring functions, functionaries and finances (3Fs) when the devolution of functions must be accompanied by concomitant functionaries followed by finances (Rangarajan, 2003). However, devolutions of functionaries and finances are sparse to adequately perform all the devolved 29 functions to the PRIs as per the Eleventh Schedule of Indian Constitution (Article 243G), and such asymmetry leads to the issue of ‘unfunded mandates’ (Chakraborty et al., 2016).
The 73rd and 74th CAAs theoretically enable the degree of ‘devolution’ of decentralisation to the local governments. The aim behind these two amendments was to remove the other two degrees of decentralisation (deconcentration and delegation), which are the representatives of partial form of decentralisation. However a considerable difference between what is declared in law and how it is acted upon is a serious concern. Although the benign intention of these two amendments was to improve service delivery at the lower level as a corollary of limited devolution of decision-making authority followed by corresponding functionaries and finances to the last tier of governments, the intergovernmental fiscal relation is at stress (Rao et al., 2011). This may lead to public sector corruption across and within the tiers of governments in India.
With such a high degree of the practice of political, followed by administrative and fiscal decentralisations, the level of corruption across states is certainly non-monolithic in nature. Corruption in each tier of governments in India – central, state and local bodies are adversely affecting the welfare making process. The misuse of public authority for private benefits in Indian federalism has been a serious cause of concern since independence (Bardhan, 1997). Corruption, although not confined to the public sector alone, is most commonly defined as “the use of public office for private gains” where agents entrusted with the responsibility of delivering a certain public service, engage in some kind of misappropriation or malfeasance for their private benefit. Nevertheless, attention needs to be drawn to the fact that developing economies are heavily dependent on their governments for the provision of basic and essential services.
This paper contributes to this strand of literature by empirically analysing the effects of fiscal decentralisation on corruption in public service delivery process across states in India. The paper is organised into five sections. Section 2 presents a succinct review of the literature on the impact of decentralised public management on the retrenchment of corruption. Section 3 critically discusses the issues in intergovernmental fiscal transfers, which may potentially create room for corruption in the process of public service provisioning. Section 4 discusses and presents the pattern of public sector corruption across states, and Section 5 describes the data and methodology adopted, section 6 presents the empirical results. Section 7 draws conclusions.
Effects of Decentralisation on Corruption: A Succinct Review of Literature
Empirical evidence on the impact of decentralised public management on the reduction of corruption in public service delivery varies across and within countries. A number of empirical studies across the globe found that fiscal decentralisation led to lower corruption (Huther and Shah, 1998; Fisman and Gatti, 1999, Arikan, 2004). Fiszbein (1997) found that political decentralisation in Colombia improved the quality of governance and reduced prevalence of corruption, while Kuncoro (2000) found that administrative decentralisation in Indonesia resulted in lowering the same. Boadway and Shah (2009) argued that decentralisation has a positive influence in controlling corruption by enhancing transparency, which consequently improves ‘voice’ mechanism and reduces the degree of ‘elite capture’ at the local level. Further, Lederman, Loayza and Soares (2005) found that accountability in political institutions in a decentralised setting is highly imperative in reducing the level of corruption. In contrast, Triesman (2000) found that countries with decentralisation and larger populations had higher perceived corruption. However, only what these literatures suggest is that the conceptual and empirical studies are inconclusive about the impact of decentralisation on corruption.
In this strand, sparse literature can be found in the Indian context. Wade (1997) found that over-centralised public management caused poor communications among the different tiers of governments and lack of consistent monitoring led to corruption and poor public provisioning of canal irrigation in South India. Crook and Manor (2000) found that democratic decentralisation in the late 1980s in Karnataka led to increased transparency and accountability, and reduced corruption in development funds diverted by the powerful people. The plausible magnitude of such severe issues can never be adequately addressed without research while this area has not been adequately probed as of now, especially in the context of India.
Issues in Fragmented Intergovernmental Transfers, and Room for Corruption
The traditional theories of intergovernmental transfers in fiscal federalism and its impact on service delivery; and the role and performance of politicians and bureaucrats in the transfers system have been classified as first-generation theories (FGT) and second-generation theories (SGT). Qian and Weingast (1997), for the first time, have dichotomised the theories as FGT and SGT. Later; Oates (2005) succinctly presented a survey of fiscal federalism literature with such irreconcilable difference. A good deal of mainstream literature (Samuelson, 1954; Tiebout, 1956; Musgrave, 1959; Arrow, 1970) has viewed the public sector as benign and, bureaucrats and politicians are always motivated to deliver services and do everything to preserve markets. While Samuelson supposed that expenditures are functioned at the central level, Tiebout considered the significance of local spending where citizens at the local level would be allowed to reveal their preferences and influence decisions through the ballot – a formulation of efficiency rationale for decentralisation.
The SGT of fiscal decentralisation emphasises on the political economy and institutional aspects of intertemporal budget constraints. It is often argued that excessive political and bureaucratic regimentation is unfavourable to democracy. Notably, the SGT emphasises on the public choice perspective of fiscal federalism, by focusing on the behaviour of agents, who are in the process of bringing services to the beneficiaries. In this regard, Niskanen (1968) set forth the possibility of bureaucrats’ budget maximising behaviour. On the basis of what the ‘flypaper effect’ notion and its empirical test have started gaining importance in the fiscal federalism literature.
Intergovernmental transfers have been broadly classified into two general forms: one is unconditional transfers (untied), and the other is conditional transfers (tied). The rationale of unconditional transfers is predominantly to offset vertical and horizontal imbalances; thus its efficiency gain principally depends on designing the suitable formula (Oates, 1972). On the other hand, tied funds are intended to provide fiscal incentives to intermediate and local governments for undertaking certain specific public service programmes or activities and to resolve inter-jurisdictional cost and benefit spillovers.
However, the discretionary tied grants in India are often criticised because these are generally skewed towards the states with higher income (Rao et al., 2011) since these are open to political bargaining (George, 1987). Thus, this type of transfers might not bring the anticipated results if political bias would appear between the centre and some of its states. Consequently, those particular favoured states might enjoy a greater amount of grants. In contrast, the formula-based tax devolution to the states, recommended by Union Finance Commission (UFC) is also not free from criticism. Some studies have highlighted three major debatable issues: one, successive changes in horizontal devolution formula (Lakdawala, 1987; Rangarajan and Srivastava, 2008); two, state governments generally levy conditions on UFC recommended tax devolution while giving to the local governments when it is explicitly meant to be flexible in practice[4] (Singh, 2008); and three, UFCs use population figures of the year 1971 for computing the inter se shares of states in tax devolution. The purpose of using this has been to penalise the states with higher rate of growth of population. However, the concern in this regard is, if some states face higher growth of population due to the influx of migrants, then should those states be penalised (Sen and Trebesch, 2003).
The conditional and unconditional transfers in the Indian federal fiscal system are being administered through three institutions, which are: Union Finance Commission, Planning Commission[5] and transfers administered by various central ministries and departments. However, apart from the mentioned issues relating to these two types of transfers, the overall decentralised mechanism has a few more important implications on the service delivery process. Cropped up from literature, there are four important causes, which are: one, lack of coordination amongst PC, UFC and various other central ministries and departments (Reddy, 2015); two, excessive state control over local tax regime (Rao et al., 2011); three, a proliferation of ‘discretionary’ tied grants with the corresponding reduction in ‘formula-based’ untied transfers (Bhadra, 2017); and four, the huge incongruity between states’ projections and UFC’s projections of fiscal variables (Fourteenth Finance Commission Report, 2015).
Aforementioned all the points insinuate that there is an enormous scope of discretion in the design of the overall decentralisation practice and fiscal transfer system in India, which drives to corruption. Besides, a few studies have put special emphasis on the centrally sponsored schemes (CSSs) and argued that the CSSs, which are discretionary tied grants, is one of the sizeable contributors to corruption (Saxena, 2007).
The concept of fiscal illusion in public choice theory predominantly arises from the political-bureaucratic nexus. The fiscal illusion is referred to as the situation where local taxpayers are not abundantly aware of intergovernmental transfers as the major source of local governments’ revenue receipts. Precisely, the fiscal illusion school is more alarmed about the problems of ‘government failure’ at the local level. It is widely acknowledged that bureaucracy is governed by certain rules and regulations. However, the bureaucrats are subjected to undertake decisions, which are outside the ambit of imposed norms that might bring discretionary part while practicing the governing rules and regulations. This will occur when bureaucrats could influence politicians in decision-making, while, by and large, bureaucrats are the implementers of the decisions made by the politicians. Besides, as bureaucrats own more information than other agents involved in the economy, they can create information asymmetry among politicians and citizens. However, due to these reasons, it is hard to empirically analyse the behaviour of bureaucrats and politicians that creates scope for public sector corruption in India.
Public Sector Corruption
Corruption in public sector hinders welfare making process of an economy (Banerjee, 1997) through promoting inefficiencies in utilisation of resources that distorts markets and public finance (Niehaus and Sukhtankar, 2013) and thereby, it impedes economic growth (Mauro, 1995; Gründler and Potrafke, 2019).
Narayan et al. (2000) found that the poor segment across the globe often suffer from pervasive low-level corruption and lack of justice in public service delivery. Davis (2004) shows that widespread corruption in South Asia’s water and sanitation sector where contractors and politicians work together to obtain projects that raise their private benefit. Report card surveys in Bangalore revealed around one-third of people in Bangalore had to resort to bribery to avail a service or solve a service-oriented problem (Paul, 1993). Similarly, surveys conducted in rural and urban areas in Nepal, Bangladesh, Pakistan, Sri Lanka and Nepal to assess the incidence of corruption in public services of healthcare, education, land administration, power, taxation, police and judiciary (Cavill and Sohail, 2007). However, the scenario gets more disheartening as public sector corruption is not limited to bribery alone. There is a plethora of corruption that takes different forms, including embezzlement, favouritism, kickbacks and leakage, among others.
Huss et al. (2010) conducted a participatory and opportunistic evaluation design to analyse a case study of The Karnataka Lokyukta (KLA) – a public complaints agency to analyse the scope and level of poor governance in the health sector in Karnataka. Results show widespread corruption in various sub-sectors of health. In a survey conducted by Neihaus and Sukhtankar (2013) in selected districts in Odisha and Andhra Pradesh, comparing official records of disbursement of wages under Mahatma Gandhi National Rural Employment Guarantee Scheme (MGNREGS), revealed 70-80 percent embezzlement of funds before it reaches to its targeted beneficiaries. Two other important schemes, such as Sarva Sikhsha Abhiyan (SSA) and the National Rural Health Mission (NRHM) have also witnessed fund embezzlement. The CAG report (No. 15, 2006) revealed a loss of Rs. 472 crores by a total of 14 states through non-adjustment of advances, excess payment of contingent grants, diversion to non-SSA activities and distribution of goods to non-beneficiaries. In 2011, the NRHM scheme went into controversy since an audit by the CAG (2011) reported that the accounts of the State Health Society failed to show a transparent record of financial activities.
According to CMS-ICS 2018, 75 percent households perceive that corruption in the public sector has either increased or remained same, and 27 percent households have experienced corruption as demand for a bribe or they had to use contacts/middlemen while availing the public services. Sector-wise reasons and the average amount paid as a bribe for availing public services are presented in detail in Table A1.
However, the difference between perception and experience can stem from a bad reputation of government as well as from underreporting of the incidences of corruption. It has been observed from Figure 1 that this asymmetry is high in Andhra Pradesh, Gujrat, Punjab, Rajasthan and Uttar Pradesh, where the perception of increased corruption in public services is more than twice than its actual experience. Almost 77 percent of households across India have perceived corruption to be higher than their experience with the exception of Bihar, Karnataka and Telangana. The perceived increase in corruption is lowest for Maharashtra, followed by West Bengal and Delhi. Particularly in the case of Delhi, the difference in corruption perception and experience is minimum with 34 percent of households reported a perception of an increase in corruption, and 29 percent had experienced it.
On the other hand, in Andhra Pradesh, Rajasthan, Punjab and Tamil Nadu, more than 50 percent of the households perceive an increase in corruption in public services. A case of particular interest that draws attention is Telangana, where a substantial differential in perception and experience has been seen. In this state, while 73 percent of households have experienced corruption, it is surprising that only 13 percent of households perceive an increase in corruption in public service delivery.
It is a consensus that a high experience of corruption would act as a catalyst for protests and thereby implying a linearly increasing relationship between experienced corruption and citizens’ activism. Among Bihar, Telangana and Karnataka where households’ corruption experience is higher than their perception, citizen activism[6] is reported to be high only in Bihar and Telangana. Interestingly, Gujarat, Maharashtra, and Rajasthan have reportedly high citizen activism even though their corruption experience is less than their perception. The rest of the states, i.e. Uttar Pradesh, West Bengal, Andhra Pradesh, Madhya Pradesh, Delhi and Tamil Nadu are show low corruption experience compared to perception and display low citizen activism.
Figure 1: Households Experienced and their Perception about Corruption in Public Services
Source: CMS-ICS 2018
Corruption in the public service delivery process can be manifested through bribe payment or resorting to the services of a middleman in availing a certain public service. CMS 2018 report focuses on both these aspects and analyses the wedge between corruption experience and service delivery denial for the eleven covered public services.
It could be observed from Figure 2 that among the 11 sectors, transport and police record the highest corruption experienced by the households at 21 percent and 20 percent respectively. For police services, around 2 percent of households have been denied the service on account of inability to pay a bribe or arrange for a middleman and among the households who experienced corruption, an average annual bribe of Rs. 313 has been paid to file a complaint or FIR. In transport, the denial is faced by about 1 percent of households. Within the transport sector for the ones who availed the services, a maximum average annual bribe of Rs. 518 has been paid to renew or acquire a driving license and Rs. 327 for registration of a vehicle. The corruption experienced by the households is 10 percent in healthcare where an average annual bribe of Rs. 275 needed to pay for diagnostic or pathological services, and about 1 percent of households have been denied the service. In the case of the judiciary, water supply as well as PDS, 8 percent of households report corruption experience. For installation and maintenance of water supply, an average annual bribe of Rs. 333 has been paid by the households. Within PDS, the maximum average annual bribe paid is Rs. 256 to get a new ration card. In the case of the judiciary, the maximum bribe paid on average in a year is Rs. 314 to get a certified copy of the order. Among these three sectors, the maximum service delivery denial befell in PDS with almost 2 percent, followed by water supply and judiciary.
Figure 2: Households Experienced Corruption and Denied Service
Source: Same as Figure 1
In both school education and electricity sectors, 6 percent of households report corruption experience. In the case of school admission, an average annual bribe of Rs. 217 was paid by the households whereas the same stands at Rs. 367 to obtain a new electricity connection. In both these sectors, about 1 percent of households were denied from availing the services. Typically, often the beneficiaries, while trying to access a public service, face an ultimatum of paying a bribe or being ousted. Evidently, this part of the population is usually within low or middle-income categories, and thus, is heavily dependent on services from the public sector given their constraint to afford private care. This reflects a sufficiently high marginal utility of money which almost compels them to agree to bribe transaction to minimise the likelihood of service denial. Interestingly, in the banking sector, only 1 percent of households have experienced corruption, and nobody has been denied from availing the service but paid an average bribe of Rs. 5250 to take a loan, which is distinctly higher compared to bribe amounts paid in other sectors.
It is important to note that corruption can be different at the ‘intensive’ and ‘extensive’ margins. While corruption can be lower for a few public services at the extensive margin, captured by the proportion of households experiencing the same, it can be much higher at the intensive margin captured by the bribe amount paid to avail services. Figure 3 presents a relationship between the households’ experienced corruption and the average bribe money paid across sectors. For a cluster of sectors including healthcare[7], judiciary, PDS, water supply, electricity and school education, the average bribe paid annually across these sectors ranging from Rs. 190 to Rs. 300. In these sectors, the corruption experienced by the households rests in between 6-10 percent. For another cluster of sectors including land, police[8] and transport the corruption experienced by the households falls between 16-21 percent with average annual bribe paid to range from Rs. 300 to Rs. 450.
Two particular sectors that draw attention are transport and banking. Corruption experience in the banking sector is lowest at only 1 percent whereas it is higher at 21 percent for the transport sector. It suffices to say that corruption is much lower at the extensive margin in the banking sector vis-à-vis transport sector. However, at the same time, the average annual amount of bribe paid in the transport sector is Rs. 423 whereas in the banking sector it stands at Rs. 1942, which is substantially higher in comparison with all other sectors. This merely reflects higher corruption prevalence at the intensive margin in the banking sector as compared to the transport sector.
Figure 3: Relationship between Sector-wise Households Experienced Corruption and Bribe Paid
Source: Same as Figure 1
Interpreting Data and Methodology
To examine how the perception and experience of corruption in public service delivery is influenced by fiscal decentralisation, we perform an empirical analysis using cross-section data for several variables comprising fiscal, socio-economic and corruption categories. We introduce six key variables including Corruption in Public Service Delivery, Citizens’ Activism against Corruption, Fiscal Indicators at State Level, Fiscal Transfers from Centre to States, Fiscal Decentralisation to the PRIs, and Political and Administrative Decentralisation.
The definitions, categories and sources of the variables used in the study are reported in detail in Table 1, and descriptive statistics of the variables are presented in Table A2. Concerning corruption variables, there is a dearth of data at the State level. Transparency International publishes a corruption perception index (CPI) every year since 1995, which ranks countries by perception-based surveys on public sector corruption. With the CPI score[9] of 41, India’s rank is 78 out of a total of 176 countries in 2018. Lack of detailed data in a time-series manner on public sector corruption across states in India limits the scope of empirical analysis. In this regard, the Centre for Media Studies conducted India Corruption Survey in 2018 (CMS-ICS 2018) and presented data on public sector corruption covering both rural and urban areas of 13 states and 11 public services. The survey report predominantly provides two types of corruption data: one, households’ perception about corruption in public sectors/services; and two, households’ experience with corruption while accessing public services. This is a 12th round of CMS-India Corruption Study. The data collection was conducted during February-March 2018. It covered a sample of 160 households from two districts in each state (one of the districts covered has been the state capital) spread across 10-12 locations (rural and urban). The survey report claims that the sample is fair enough “because of accumulated survey data of over 15 years indicate consistency in findings and in consonance with expert’s opinion”.
The empirical analysis is done at two levels. At the first level, we run an OLS regression to analyse the decentralised fiscal response to the corruption where Experienced Corruption in Public Services has been considered as the dependent variable. At the second level, we perform a Probit estimation with citizens’ activism against corruption as the dependent variable to observe how some socio-economic determinants, corruption (both perceived and experienced) influence the activism against corruption. It is important to mention that a possible presence of endogeneity due to the simultaneity bias is a common problem with almost all corruption analyses. However, a lack of time series data on corruption prevents further investigation. Further, we use the Breusch-Pagan / Cook-Weisberg test for heteroscedasticity where results showed constant variance (no heteroscedasticity). The total number of observations in both the regression models is 12. Since CMS-ICS 2018 provided data for 13 states, however, we did not include Delhi in the analysis since it is devoid of PRIs.
Empirical Results
The varied levels of perceived and experienced corruption in public service delivery by households across states in India point to examine how the decentralised public finance is effective to combat these levels of corruptions. In order to empirically examine this, this paper especially examines two aspects, which are: one, what is the decentralised fiscal response to the corruption experienced in public service provisioning; and two, how various socio-economic determinants, and perceived and experienced corruption influence citizens’ activism against corruption.
Table 1: Variables used in the Study
| Variable | Category | Definition | Data Source |
| perception | Corruption in Public Service Delivery | CMS-ICS 2018 | |
| experienced | Percent of households experienced corruption (among those households who interacted with at least one public service during last one year) | ||
| activism | |||
| owntax | Fiscal Indicators at State Level | States’ own tax revenue as a percent to GSDP | State Budgets 2019-20 |
| taxdev | Fiscal Transfers from Centre to States | States’ tax devolution (share in central taxes) as percent to GSDP | |
| gia | States’ grants-in-aid as percent to GSDP | ||
| devfin | Fiscal Decentralisation to the PRIs | Percentage devolution of finances pertaining to the devolution of functions (29 Subjects as per Eleventh Schedule of Constitution – Article 243G) | Ministry of Panchayati Raj |
| n_pri | Political and Administrative Decentralisation | State-wise total number of PRIs | |
| pci | Socio-economic Determinants | Per capita Nominal GSDP for 2017-18 | State Budgets 2019-20 |
| litrate | Percent of total literate population across states | Census 2011 | |
| womenlit | Percent of women in the age 15-49 years who are literate | NFHS-4 in 2015-16 | |
| menlit | Percent of men in the age 15-49 years who are literate | ||
| hh_internet | Percentage of total households possessing internet facility |
The descriptive statistics broadly show a lack of a predictable pattern on the link between fiscal variables and corruption categories that forbids us from arriving at any ubiquitous inferences. For instance, for the state’s tax devolution, categories A and C reflect a lower average compared to the overall average, and for states’ own tax revenue, categories A, C and D are higher vis-à-vis the overall average value.
In some states, perception about corruption in public service delivery is much higher than its actual experience. It may ensue essentially because of the fact that the reputation of the respective state governments’ service delivery is worse than its actual experience while availing services. Keeping such asymmetry in mind, we chose the households’ experienced corruption in public services as the dependent variable to see the impact of decentralised public finance and various socio-economic determinants on it. The results presented in Table 2 (Model I) report statistical significance at one percent for all explanatory variables.
It could be seen from the ordinary least square estimates reported in Model I that, with the increase in states’ tax mobilisation from own sources, households’ experienced corruption in public services would also significantly increase. Intuitively, a plausible implication of such a result could be because own revenue is sometimes used by the politicians for their political survival. However, the paucity of information regarding how state government utilises its own revenue and intergovernmental transfers prohibits us from arriving at a concrete conclusion. Regardless, a positive and significant impact of states’ own tax revenue on corruption experience shows an underlying lack of accountability. A priori, we expect negative signs for literacy rate and per capita income, implying that an increase in these two socio-economic determinants would significantly reduce the level of experienced corruption in public services. However, per capita income emerged to be positively and significantly associated with households’ experienced corruption in public services.
Conventional wisdom and literature (see Svensson, 2005) argues that the level of corruption goes down as a country’s income goes up. Thus, a similar argument can be put forward at state level as well. However, a plausible counter argument is richer states have more resources and hence a potential for corruption. However, this argument has little room when the kind of corruption is petty and associated with routine public service delivery. To comment on the unexpected sign of per capita income, much deeper analysis is highly imperative. We consider this result to be one of our caveats. In contrast, the results reveal that improvement in literacy rate would significantly reduce the experience of corruption in public services. This suggests that illiterate persons are perhaps more susceptible and likely to experience corruption while availing public services since the basic cognitive ability is restricted.
The main argument of our paper was that the fiscal decentralisation has a key role to play in deciding the corruption that has been experienced by the households while availing public services. In this direction, the results, regarding the impact of fiscal decentralisation on households’ experienced corruption while availing public services, have emerged to be somewhat ambiguous. Similar to own tax revenue, the positive and significant coefficient of tax devolution from the centre to states implies that due to its increase, households’ experience of corruption while availing public services will also significantly increase. Arbitrariness and adhocism in tax devolution is a contributing factor behind such an outcome. In contrast, the negative and significant coefficient of devolution of finances to the PRIs pertaining to the functions signifies that the increase in more finances pertaining to the functions would significantly reduce the level of experienced corruption by the households while availing public services. Therefore, it is highly imperative for the states to considerably reduce the magnitude of ‘unfunded mandates’ to their PRIs. Surprisingly, although severely criticised the grounds of political bargaining, grants-in-aid, the discretionary tied transfers from the centre to states, have appeared as statistically significant in reducing corruption experience.
These results however appear perverse at first, since the impacts of all the three variables of fiscal decentralisation viz. tax devolution, grants-in-aid and devolution of finances to PRIs pertaining to functions on the experienced corruption are greatly differing from each other. The number of PRIs, a reflective variable of political and administrative decentralisations, has shown that with its increase, the level of experienced corruption while availing public services by the households would significantly come.
As the second level of analysis, to explore whether perception about, and experience of corruption, and other socio-economic characteristics of the states are useful to explain the citizens’ activism against corruption, a probit analysis has been done. The results are also reported in Table 2 (Model II). All the explanatory variables have emerged statistically significant in explaining the citizens’ activism at one percent level of significance. It is important to note that the citizens’ activism against corruption significantly increases if the perceived level of corruption and experienced corruption in public service provisioning increases. Besides, it has been observed that the citizens’ activism significantly goes up with the increase in households’ use of internet service.
It is important to note that there can be an issue of reverse causality since it may also happen that with an increase in activism the perception of corruption is revised and at the same time a higher reporting of corruption is encouraged, both of which would lead to an increase in perception and experience of corruption. However, this remains another caveat in our analysis as a lack of sufficient time-series data does not allow us to explore the outcomes further.
The effects of women’s and men’s literacy rates have surprisingly put up a contrasting result. The increase in literacy rate of men would significantly increase the level of citizens’ activism against corruption while surprisingly an increase in female literacy leads to a significant fall in activism. The latter result could plausibly be due to a higher opportunity cost in terms of physical and financial capabilities for a literate woman compared to a semi-literate or illiterate woman. However, it might not be the case and the result could be interpreted as a fact that literate women also can have engagement in domestic activities, or simply have different interests. Nevertheless, what the results of this model suggest is that the citizens’ activism against corruption is not only statistically significantly responsive to the perceived and experienced corruption but is significantly responsive to the socio-economic determinants as well.
Table 2: Decentralised Fiscal Response to Corruption in Public Service Delivery
| Regressors | Experienced Corruption in Public Services | Citizens’ Activism against Corruption |
| Model I: OLS Estimation | Model II: Probit Estimation[10] | |
| owntax | 13.335***
(18.45) |
|
| taxdev | 5.93***
(13.69) |
|
| gia | -7.25***
(-6.17) |
|
| devfin | -0.27***
(-7.52) |
|
| n_pri | -0.00067***
(-9.97) |
|
| pci | 0.00027***
(8.25) |
|
| litrate | -1.51***
(-7.71) |
|
| perception | -10.25***
(-34.99) |
|
| experienced | -0.240***
(-36.26) |
|
| womenlit | 2.971***
(64.09) |
|
| menlit | -6.663***
(-63.41) |
|
| hh_internet | -0.627***
(-41.76) |
|
| _cons | 25.04
(1.74) |
387.75***
(62.35) |
| N | 12 | 12 |
| R2 | 0.9950 | |
| Prob.>F | 0.0001 | |
| Root MSE | 2.0014 | |
| Wald Chi2 | 5851.65*** | |
| Pseudo R2 | 1.0000 |
Notes: 1) Figures in parentheses are t-statistics
2) *** p ≤ 0.01, ** 0.01 < p ≤ 0.05, * 0.05 < p ≤ 0.10
Source: Authors’ computations based on sources mentioned in Table 1
Conclusions and Implications
The analyses in this study identify that sector-wise corruption may differ at the ‘intensive’ and ‘extensive’ margins in terms of the amount of bribe paid and experienced corruption, respectively. Our preliminary results suggest that although in varying margins, corruption prevails in the sectors that are of primary importance to long-run economic growth and development. In this regard, the role of decentralised governance is imperative, and therefore, this paper empirically examines the decentralised fiscal responses to corruption. Research on this strand of literature is sparse in Indian federalism. In this direction, the contributions of this paper are two-pronged – one has critically discussed the fiscal federalism literature focussing on the issues in intergovernmental fiscal transfers that creates room for corruption, many aspects of which could gain attention for future research. The other contribution emerged through an empirical probing of decentralised fiscal response to corruption experienced by the households across the country while availing public services.
Three important variables have been considered to represent fiscal decentralisation in India, which are: one, devolution of finances to the PRIs pertaining to 29 functions enlisted in the Eleventh Schedule of Constitution; two, tax devolution, which is an untied and formula-based transfer to the states; and three, grants-in-aid to the states, which is a tied and discretionary form of transfer. The analyses suggest mixed and varied results regarding the effects of fiscal decentralisation on retrenchment of corruption experienced by the households across the country while availing public services.
Such ambiguity in results suggests that the impact of overall fiscal decentralisation on the reduction of incidences of corruption experienced by the households while availing public services are somewhat inconclusive. On the other hand, political and administrative decentralisations (through the variable of number of PRIs across states) have significant impact on a decrease in the level of experienced corruption while availing public services by the households. In sum, different types of decentralisation have different effects on corruption. As is expected, it has emerged from our analyses that education is a significant determinant that enables people to raise their ‘voice’ against corruption through several activities, and thereby, the degree of their experiencing of corruption while availing of public services goes substantially down.
To sum up, even when, political decentralisation is taking place through regular panchayat elections across many states, the transfers of finances and functionaries are still lagging behind the transfer of functions. A lower level of states’ fiscal autonomy and quality and quantum of transfers (formula-based and discretionary) characterised by adhocism and arbitrariness are the major challenges in the provisioning of corruption-free public services at the intermediate and local level. It is argued that increasing share of untied formula-based transfers would be an effective way to rectify the anomalies in the intergovernmental transfer mechanism, which would help to reduce the level of corruption in public service delivery process.
The results, however, need to be considered with a few caveats. The CMS-ICS 2018 covered only 13 states, while more than 50 percent of the states and Union Teritorries (UTs) have remained uncovered. Had the study covered all the states and UTs, the result could have been different. Perpetual policy commitment through budgets is inter-alia, an important way to combat against corruption in various public sectors. Another key constraint towards rigorous policy commitment is the limited availability of disaggregated fiscal data for the PRIs of each state, which to some degree inhibits the scope of rigorous analysis regarding the effects of fiscal decentralisation in order to address corruption at the panchayat level, which could accumulate to the state level. Such a restricted scope of profound analysis is further compounded by the fact of unavailability of corruption data in a time-series manner. Based on these cautions and limitations of data, it would be useful to take these results into justification and to take up further research in this strand of literature from our inferences.
[1] In this regard, the commissioning of SFC has been an important step towards lessening the adhocism and arbitrariness in the devolution of funds to the local level governments from their respective state governments. However, the situation remains compounded with the states’ ignorant behaviour towards their respective Finance Commission’s recommendations. While the states are not legally bound to comply with the recommendations of their respective State Finance Commission, but paying no attention to this institution makes it somewhat irrelevant (Rao et al., 2011). Therefore, a cause of worry on this aspect is the magnitude of control exerted by the state governments over their respective local units about the utilisation of such untied transfers.
[2] Planning commission has been replaced with a new name i.e. NITI Aayog in 2014, and it does not have the power to recommend transfers.
[3] The CMS-ICS 2018 report defined citizen activism using four indicators, which are: one, ever complained about corruption in public services; two, ever participated in public meetings/protest rally to raise concern about prevailing corruption in public services; (three), awareness and usage of RTI Act; and (four), online complaint registration and sought information.
[4] In the healthcare sector, a nominal amount of Rs. 126 has been paid as an average annual bribe to avail preventive healthcare as out-patient service. While such a petty annual amount does not seem to entail much botheration in terms of corruption, the fact that seems to be more alarming is the regularisation of such an action. This is captured by the fact that almost 50 percent of households paid bribes while availing the service. This, in turn, reflects a society where corruption has been internalised and accepted as a part of routine activities, and certainly cripples the overall operation of public service delivery.
[5] 34 percent of households have paid Rs. 234 as an average annual bribe in the police sector to avoid a punishment due to traffic rule violation and 19 percent have paid an average of Rs. 460 to remove their names as accused. While these types of corruption do not involve a ‘delivery of public service’ per se, but due to the very infectious nature of corruption, it forms a skewed service delivery system.
[6] The index scores on a scale of zero (highly corrupt) to 100 (very clean).
[7] The marginal effects for Probit analysis generated similar estimates and hence we have refrained from reporting the same to avoid repetition.
References
Arikan, Gulsun, (2004). Fiscal Decentralization: A Remedy for Corruption? International Tax and Public Finance, 11(2), 175–195.
Arrow, Kenneth, 1970, The Organization of Economic Activity: Issues Pertinent to the Choice of Market Versus Non-Market Allocation. In Joint Economic Committee, The Analysis and Evaluation of Public Expenditures: The PPB System, Vol. I. Washington, D.C.: U.S. GPO.
Banerjee, Abhijit, 1997, A Theory of Misgovernance. Quarterly Journal of Economics, 112(4), 1289-1332.
Bardhan, Pranab, 1997, Corruption and Development: A Review of Issues. Journal of Economic Literature, 35, 1320-1346.
Bhadra, Kaushik, 2017, Fragmented Inter-Governmental Fiscal Transfer System in India under Reforms: Issues and Challenges. In A Saga of Economic Reforms in India. Edited by Neelakanta N. T. et al. Madhav Books (P) Ltd., New Delhi.
Boadway, Robin, and Shah, Anwar, 2009, Fiscal Federalism. Cambridge Books. Cambridge University Press, January.
Cavill, Sue, and Sohail, M., 2007, Accoutnability Arrangements to Combat Corruption: Literature Review. Loughborough: WEDC, Loughborough University.
Chakraborty, Pinaki, Chakraborty, Lekha and Mukherjee, Anit, 2016, Social Sector in a Decentralized Economy: India in the Era of Globalization. Cambridge University Press, New Delhi.
Crook, Richard, and Manor, James, 2000, Democratic Decentralization. OED Working Paper Series, The World Bank, Washington, D. C.
Davis, Jennifer, 2004, Corruption in Public Service Delivery: Experience from South Asia’s Water and Sanitation Sector. World Development, 32(1), 53-71.
Downs, Anthony, 1957, An Economic Theory of Democracy. New York: Harper & Row.
Fisman, Raymond, and Gatti, Roberta, 2002, Decentralization and corruption: evidence across countries. Journal of Public Economics, 83, 325-345.
Fiszbein, Ariel, 1997, The Emergence of Local Capacity: Lessons From Colombia, World Development, 25(7), 1029-1043.
George, K. K, 1987, Approach of the Finance Commissions In Centre-State Budgetary Transfers. Edited by I. S. Gulati., 247-264. Sameeksha Trust: Oxford University Press.
Government of India, 2006, Performance Audit of Sarva Shiksha Abhiyan. Comptroller and Auditor General India. Report No. 15.
Government of India, 2011, Performance Audit on National Rural Health Mission of Government of Uttar Pradesh. Comptroller and Auditor General India.
Government of India, 2015, Report of the Fourteenth Finance Commission, New Delhi.
Gründler, Klaus, and Potrafke, Niklas, 2019, Corruption and Economic Growth: New Empirical Evidence. CESifo Working Paper No. 7816.
Gupta, Sanjeev, Davoodi, Hamid, and Alonso-Terme, Rosa, 1998, Does Corruption Affect Income Equality and Poverty. IMF Working Paper No. 98/76. IMF: Washington D. C.
Huss, R. et al., 2011, Good Governance and Corruption in the Health Sector: Lessons from the Karnataka Experience. Health and Policy Planning, 26, 471-484.
Huther, Jeff, and Shah, Anwar, 1998, Applying a Simple Measure of Good Governance to the Debate on Fiscal Decentralization. Policy Research Working Paper Number 1894: World Bank, Washington, D. C.
Kuncoro, Ari, 2000, The Impact of Licensing Decentralization on Firm Location Choice: the Case of Indonesia. Paper presented at the Annual Bank Conference on Development Economics.
Lakdawala, D. T, 1987, Eighth Finance Commission’s Recommendations. In Centre-State Budgetary Transfers, ed. I. S. Gulati., 195-207.Sameeksha Trust: Oxford University Press.
Lederman, Daniel, Loayza, Norman, and Soares, R. R, 2005, Accountability and corruption: Political institutions matter. Economics & Politics, 17, 1-35.
Mauro, Paolo, 1995, Corruption and growth. Quarterly Journal of Economics, 110, 681-712.
Narayan, Deepa, Chambers, Robert, Shah, Meera, and Petesch, Patti, 2000, Voices of the Poor: Crying out for Change. Oxford University Press: Oxford.
Niehaus, Paul, and Sukhtankar, Sandip, 2013, The Marginal Rate of Corruption in Public Programs: Evidence from India. Journal of Public Economics, 104, 52-64.
Niskanen, William, 1968, Nonmarket decision making: The peculiar economics of bureaucracy. Applied Economic Review (Supplement), 58, 293-305.
Oates, Wallace, 1972, Fiscal Federalism.2nd ed. New York: Harcourt, Brace and Jovanovich.
Oates, Wallace, 2005, Toward a Second Generation Theory of Fiscal Federalism. International Tax and Public Finance, 12, 349-373.
Paul, Samuelson, 1993, Bangalore’s Public Services: A Report Card. Economics and Political Weekly, 28(52), 2901-2909.
Prud’homme, Remy, 1995, On the Dangers of Decentralization. World Bank Research Observer, 10(2), 201-220.
Qian, Yingyi and Weingast, Barry, 1997, Federalism as a Commitment to Preserving Market Incentives. Journal of Economic Perspectives, 11(4), 83-92.
Rangarajan, Chakravarthi, 2003, Fiscal decentralization. Seminar on Rural Fiscal Decentralization and Role of Central Finance Commission, National Institute of Public Finance and Policy, New Delhi, September 3, 1-6.
Rangarajan, Chakravarthi, and Srivastava, D. K, 2008, Reforming India’s Fiscal Transfer System: Resolving Vertical and Horizontal Imbalances. Economic and Political Weekly, 43(23), 47-60.
Rao, M. G., et al., 2011, Fiscal Decentralization to Rural Local Governments in India: Selected Issues and Reform Options. National Institute of Public Finance and Policy, Mimeo.
Reddy, Venugopal, 2015, A Tale of Two Commissions and the Missing Links. Economic and Political Weekly, 50(4), 32-38.
Samuelson, Paul, 1954, The Pure Theory of Public Expenditure. Review of Economics and Statistics, 36, 387-389.
Saxena, N. C., 2007, Rural poverty reduction through centrally sponsored schemes, Indian J Med Res, 126, 381-389.
Sen, Tapas, and Trebesch, Christoph, 2003, The Use of Socio-Economic Criteria for Intergovernmental Transfers: The Case of India. Paper presented at a conference on Issues before the Twelfth Finance Commission.
Shah, Anwar, 1994, The Reform of Intergovernmental Fiscal Relations in Developing and Emerging Market Economies. World Bank, Washington, D. C.
Shah, Anwar, 1998, Fiscal Federalism and Macroeconomic Governance – For Better or For Worse? Policy Research Working Paper No. 2005. World Bank, Washington, D. C.
Singh, Nirvikar, 2008, Decentralization and Public Delivery of Health Care Services in India. Health Affairs, 27(4), 991-1001.
Svensson, Jacob, 2008, Eight questions about corruption. Journal of Economic Perspectives, 19(3), 19-42.
Tanzi, Vito, 1996, Fiscal federalism and decentralization: A review of some efficiency and macroeconomic aspects. Annual World Bank conference on development economics: 295-316. World Bank, Washington, D. C.
Tiebout, Charles, 1956, A Pure Theory of Local Expenditures. The Journal of Political Economy, 64(5), 416-424.
Torres, Sanchez, and Pachón, Monica, 2013, Decentralization, Fiscal Effort and Social Progress in Colombia at the Municipal Level, 1994 – 2009: Why Does National Politics Matter?. IDB Working Paper Series.
Triesman, Daniel, 2000, Decentralization and the Quality of Government. Department of Political Science Working Paper. University of California, Los Angeles.
Wade, Robert, 1997, How Infrastructure Agencies Motivate Staff: Canal Irrigation in India and the Republic of Korea. In Infrastructure Strategies in East Asia. Ed. A. Mody. The World Bank, Washington, D. C.
World Bank, 2004, World Development Report 2004: Making Services Work for Poor People. New York: Oxford University Press.
Annexures
Table A1: Sector-wise Reasons and Bribe Paid for Availing Public Services
| Public Services | Reasons for Paying Bribe | % HHs paid bribe | Average Bribe Paid in a year (In Rs.) |
| School Education | School Admission | 45 | 217 |
| Issuance of certificate | 12 | 271 | |
| Low Attendance/promotion to higher class | 19 | 311 | |
| Application for scholarship | 22 | 329 | |
| Healthcare | Getting the prescribed medicine | 18 | 160 |
| For diagnostic services/Pathological tests | 10 | 275 | |
| As out-patient | 49 | 126 | |
| As in-patient/ for getting bed/services | 15 | 207 | |
| Water Supply | Installation of new water connection | 36 | 278 |
| Installation/Maintenance of water supply | 70 | 333 | |
| PDS | To get new ration card | 39 | 256 |
| Deletion/Addition of name in ration card | 20 | 220 | |
| For taking monthly ration | 28 | 202 | |
| Land and Housing | To get plots in auction/Transfer of ownership/Mutation | 27 | 212 |
| Purchase of land/Stamp paper/obtaining land and property document | 54 | 327 | |
| Building approval/get house plan sanctioned/new construction | 13 | 712 | |
| Electricity | New connection | 25 | 367 |
| Correction of the faulty meter/inflated bills | 56 | 357 | |
| Load enhancement | 11 | 245 | |
| Change/Correction of name/address and bill | 6 | 204 | |
| Police | Avoid Challan for Violation of traffic rule | 34 | 234 |
| Get the complaint/FIR Registered | 29 | 313 | |
| Remove name as an accused/witness | 19 | 460 | |
| Judiciary | To get hearing date of choice | 52 | 220 |
| To get certified copy of the order | 43 | 314 | |
| Transport | Registration of Vehicle | 12 | 327 |
| To get new/renew driving license | 83 | 518 | |
| Banking | To open account/documentation process | 40 | 225 |
| To get pension/scholarship | 20 | 350 | |
| To take loan | 30 | 5250 |
Source: Same as Figure 1
Table A2: Descriptive Statistics of the Variables Used in the Study
| State | devfin | owntax | taxdev | gia | n_pri | pci | litrate | womenlit | menlit | hh_internet |
| Andhra Pradesh | 17.24 | 6.20 | 3.57 | 2.83 | 12920 | 89622 | 67.0 | 62.9 | 79.4 | 3.8 |
| Bihar | 27.59 | 4.74 | 13.35 | 5.27 | 9142 | 45919 | 61.8 | 49.6 | 77.8 | 6.1 |
| Gujarat | 51.72 | 5.42 | 1.57 | 1.20 | 14066 | 205599 | 78.0 | 72.9 | 89.6 | 4.0 |
| Karnataka | 100.00 | 7.12 | 2.42 | 1.17 | 5654 | 206649 | 75.4 | 71.7 | 85.1 | 11.3 |
| Madhya Pradesh | 34.48 | 6.72 | 7.37 | 4.11 | 22827 | 86209 | 69.3 | 59.4 | 81.8 | 13.0 |
| Maharashtra | 62.07 | 6.73 | 1.49 | 0.87 | 27851 | 203090 | 82.3 | 80.3 | 92.8 | 16.0 |
| Punjab | 44.83 | 6.47 | 2.26 | 1.63 | 11758 | 158696 | 75.8 | 81.4 | 87.5 | 40.3 |
| Rajasthan | 62.07 | 6.15 | 4.50 | 2.91 | 9888 | 109942 | 66.1 | 56.5 | 85.4 | 11.4 |
| Tamil Nadu | 9.00 | 6.46 | 1.87 | 1.01 | 12525 | 207151 | 80.1 | 79.4 | 89.1 | 7.9 |
| Telangana | 17.24 | 7.71 | 2.24 | 1.10 | 8695 | 182558 | 67.0 | 65.2 | 83.4 | 6.1 |
| Uttar Pradesh | 41.38 | 7.10 | 8.82 | 2.96 | 58895 | 61012 | 67.7 | 61.0 | 82.4 | 7.7 |
| West Bengal | 41.38 | 5.65 | 4.34 | 2.56 | 3357 | 127456 | 76.3 | 70.9 | 81.1 | 7.2 |
| Average (Overall) | 42.42 | 6.37 | 4.48 | 2.30 | 16465 | 140325 | 72.23 | 67.60 | 84.62 | 11.23 |
| Average: Category A | 42.77 | 6.56 | 2.53 | 1.66 | 10714 | 165530 | 74.58 | 73.85 | 85.28 | 15.83 |
| Average: Category B | 39.66 | 6.01 | 5.41 | 2.62 | 10448 | 136005 | 68.23 | 61.05 | 84.05 | 6.90 |
| Average: Category C | 62.07 | 6.73 | 1.49 | 0.87 | 27851 | 203090 | 82.30 | 80.30 | 92.80 | 16.00 |
| Average: Category D | 39.08 | 6.49 | 6.84 | 3.21 | 28360 | 91559 | 71.10 | 63.77 | 81.77 | 9.30 |
Note: Various categorical averages are computed based on the perception of corruption (high and low) and citizens’ activism against corruption (high and low). Category “A” stands for High Perception & Low Activism; “B” stands for High Perception & High Activism; “C” stands for Low Perception & High Activism; and “D” stands for Low Perception & Low Activism
Source: Same as Table 2
