Abstract
This article presents the findings of perceptions of electoral integrity survey for subnational elections in India between 2015 and 2017. The perceptions of electoral integrity survey is an attempt to provide a comprehensive, impartial and independent source of information derived from experts about elections conforming to internationally recognized principles and standards. The article then evaluates the factors that account for differences in the perceptions of electoral integrity in India. We show that higher-income states in India tend to have higher levels of electoral integrity. However, we argue that differences in the quality of governance best explain subnational variation in the perceptions of electoral integrity in India.
Introduction
The subject of electoral integrity 1 has gained currency with the increasing recognition of the limitations of a purely structural understanding of democracy. While the existence of a democratic system is no doubt important, it is not enough to justify fair and legitimate political processes. One of the defining features of democracy is the conduct of free and fair elections. Contested and flawed elections can pose a serious threat to the quality of democracy and political processes. As evidence from the World Values Survey suggests, public perceptions of electoral malpractices erode trust and confidence in elected authorities, discourage voter turnout, generate protests and even undermine regime stability (Norris et al., 2014).
Existing literature on elections highlights several reasons for flawed elections, ranging from failure of the administration of the electoral process (Kavanagh, 2015), the attitude of election administration (Elklit & Reynolds, 2002; James, 2013), counting of votes (Alvarez, et al., 2008), unequal access to the media and public resources, registration problems, vote-buying (Carreras & Irepoglu, 2013), disqualification of opponents, manipulation of media, unfair disparities in political campaign finance, violence (Norris, 2015), unjust electoral laws and management (Daxecker & Schneider, 2014) and so on. Comparing the various dimensions of electoral failures, Norris et al. (2013) highlight two levels of malpractices—first- and second-order—according to the severity of their potential consequences. First-order malpractices revolve around violations of generally accepted norms of freedoms and rights. Second-order malpractices concern more mundane issues of maladministration, lack of technical capacity or human error.
The literature on electoral integrity has examined the causes for the disparities in the quality of elections and explored between-country variations in outcomes. However, one of the most under-researched areas in the literature on electoral integrity pertains to within-country variations. This article focuses on India—a durable democracy with federal structure—to identify within-country variations in electoral integrity. Based on an original survey dataset of perceptions of electoral integrity (PEI) of state 2 assembly elections, the article shows that some subnational states have significantly higher levels of electoral integrity than others. While anecdotal evidence and case studies have pointed out differences in the quality of election across Indian states this, to the best of our knowledge, is the first structured comparison of the quality of elections.
On exploring the subnational electoral integrity, it becomes apparent that disparities in the socio-economic status of the different states can account for much of the observed variation. Broadly speaking, higher-income states have higher levels of electoral integrity. However, the survey reveals two important anomalies, namely a low-income state (i.e., Bihar) that has significantly higher levels of electoral integrity and a high-income state (i.e., West Bengal) that has very poor electoral integrity. The subsequent section of the article explains this important anomaly and argues that differences in the quality of governance provide a better explanation for the subnational variation in the PEI.
Literature Review
Given the importance of the quality of elections and their potential lapses, it is important to analyse, evaluate and investigate the conduct of elections. The assessment of elections is, however, challenging because of the concealed nature of electoral frauds and the complexity of the electoral processes. To meaningfully engage with the idea of the quality of elections, a number of studies and theoretical frameworks have been developed. Organizations such as National Democratic Institute, International Foundation for Electoral System and Electoral Assistance System of the United Nations have focussed extensively on elections and provide technical assistance to states to develop credible and sustainable electoral systems.
A survey of the literature on electoral integrity reveals that conceptualization and operationalization of electoral integrity have subtle but important differences (van Ham, 2015). Conceptually, electoral integrity has been defined both positively and negatively. The former perceives integrity in terms of the presence of certain conditions such as impartial election management bodies (Elklit & Svensson, 1997), while the latter highlights the absence of norms and criteria essential for free and fair elections (Birch, 2011). Intuitively, negative definitions emphasize on actors, intentionality and their consequences for election, while the positive definitions focus on outlining the norms that affect integrity (van Ham, 2015). The conceptualization of integrity also varies according to the choice of universal or particularistic criteria. A universal understanding defines integrity with reference to universal standards evolving from democratic theory (Norris et al., 2014). In contrast, the particularistic approaches define integrity through the opinion of the citizens and political parties foregrounding the importance of contextual evaluation (Pastor, 1999). The particularistic approach is apposite for case study but encounters limitations of macro theory building and comparative analysis. One can also note differences in conceptualization in approaching integrity as a process and as a concept. Concept-based approaches theorize integrity on ideal democratic standards (Norris, 2013), while process-based conceptualizations consider the electoral process as time based—before, during and after election day (Mozaffar & Schedler, 2002).
Incorporating the conceptual and methodological distinctions, a consensus has emerged over the international conventions and global norms that constitute the foundations of electoral integrity. Electoral integrity has been identified with ‘international conventions and global norms, applying universally to all countries worldwide throughout the electoral cycle, including during the pre-electoral period, the campaign, the polling period and in the aftermath of polling’ (Norris et al., 2014).
The PEI index developed by Norris et al. (2014) builds on such an understanding of electoral integrity and provides a standard and uniform framework of evaluating the electoral process across the world. PEI is an expert survey on a range of issues across the spectrum of the electoral cycle to arrive at an evaluation about the quality of an election. As the normative basis of the survey is derived from the body of human rights treatises and international conventions, the framework is universal and precise. The emphasis on election sequence as well as specific dimensions of election allow an evaluation of the range of electoral malpractices from direct questions of rigging and violence to more indirect and nuanced manipulations of media, irregularities in electoral finances and so on. Since the focus is on the electoral processes, it allows a comparison between different regimes based on uniform characteristics without any a priori value judgement about the regimes, such as liberal democracies, anocracies and autocracies.
The empirical results of the PEI survey across the world have evoked some interesting issues; for example, while economically advanced democratic countries—such as Norway, Sweden, Iceland—are expectedly among the high PEI countries, the USA has been outperformed by some developing countries, such as Costa Rica, Rwanda, Mongolia and Micronesia (Norris et al., 2015). Such an outcome not only highlights the specific drawbacks of the US electoral system but also problematizes the simplistic and straightforward link that modernization theory tries to draw between development and democracy (e.g., Przeworski & Limongi, 1993).
The PEI index provides us with a range of factors which allow for engagement with the question of electoral legitimacy, without assuming that economic development guarantees a genuinely legitimate democratic process. Institutional theories (e.g., Barro, 1999; Boix & Stokes, 2003; Przeworski & Limongi, 1997) show that economic affluence has a general correlation to the necessary human, financial and technical development to strengthen democracy and the capacity of public sector management and hold elections meeting international standards. Amidst the backdrop of such understandings in both the traditional sociological and development literature—as well as in institutional theories—it is interesting to examine the divergences in electoral integrity, which allows us to problematize these conceptual understandings.
Elections and India
India is typically referred to as the world’s largest democracy; yet, it remains among the outliers in terms of stability of the electoral system among postcolonial states. With its high rates of poverty, illiteracy, low urbanization, pervasive ethnic, religious and linguistic divisions, weak capitalist class and a weak civil society, India has been a popular example defying modernization theory and institutional notions of democratic development (McMillan, 2008).
Every parliamentary election in India is hailed as the biggest in the history of democracy, involving a record number of electorate, engaging enormous resources and colossal organizational efforts. The latest parliamentary elections in 2019 had an electorate of about 900 million who decided the fate of 8,040 candidates for the 543 parliamentary seats in 7 phases (
Despite the overall success of the massive electoral exercise, several issues concerning electoral integrity have been noted such as violence and intimidation, lapses in voter registration and differential access to finance and media. In the 2019 parliamentary elections, concerns about the electoral voting machine with reports of malfunctioning and allegations of hacking, assertions of omissions in the electoral roll, flagrant violation of the model code of conduct and the severe imbalance in access to campaign finance dominated the discussions on electoral integrity (Khare, 2019; Tharoor, 2019). Many of these concerns are long-standing with news reports during elections replete with incidences of intimidation and vote-buying (Mahmood, 2015). The impact of voter buying is not limited to specific constituencies or election period but has wider ramifications. Kapur and Vaishnav (2011) have shown that in India, a distorted electoral incentive leads to institutional nexus between shady business and politics and unfavourable developmental outcomes.
Despite these shortcomings, the PEI index placed both the 2014 and 2019 general elections in India above the global average due to the favourable standing in the categories of election management bodies, laws, electoral procedures, counting and result declaration. On most counts, the position of India was higher than South Asian as well as global mean scores. Incidentally, most of these activities are under the purview of its election management body (EMB) (i.e., the Election Commission of India [ECI]), which is considered among the most efficient and independent public bodies in the country (Singh & Roy, 2019; Sridharan & Vaishnav, 2017).
Election Management Body: The Election Commission of India
Electoral management bodies administer and regulate the multiple dimensions of electoral contests from issuing detailed administrative guidelines governing the conduct of staff, candidates and political parties; validating and registering the nomination of political parties and candidates for elected office; determining qualifications and maintaining electoral registration databases; regulating procedures and facilities for polling; monitoring the vote tabulation process and compiling nation-wide results; and authorizing the official declaration of the vote and seat distribution (Norris, 2015).
Intuitively, the performance of these activities is a reflection of the wider nature of governance in a state. Scholars like Birch (2002) and McAllister and White (2015) have found that the de jure autonomy of EMBs was unrelated to public confidence in electoral processes. Public perceptions of electoral unfairness have their roots in seeing widespread corruption among public officials. Such findings point to the importance of governance in shaping perceptions of integrity due to the possible or actual breach between the everyday and the formal.
In India, the Election Commission is the constitutional body responsible for the conduct of elections to the national parliament and the state legislature and the offices of president and vice-president. Under Article 324 of the Indian Constitution, the Election Commission is empowered to ‘superintend, direct and control’ every aspect of elections from the announcement of elections to the declaration of results. During this period, the Commission can direct governments to follow guidelines to prevent misuse of state power for elections, manage transfers and postings, monitor code of conduct and exercise control over government staff associated with the election process. The performance of such wide range of activities has led to the varied characterization of ECI from a ‘regulatory body enforcing rules’ to ‘a prominent player in the electoral game’ to even a case of ‘administrative overreach’ (Singh & Roy, 2019, p. 7).
Institutionally, ECI can be categorized under the agency model as it enjoys de jure autonomy to prevent political interference and potential conflicts of interest. Over the years, ECI has asserted its autonomy in determining the timing and conduct of elections, control of the administration and the regulation of parties and candidates during the election (Katju, 2016). Instances such as the adoption of model code of conduct, declaration of criminal records, family financial assets and educational qualifications of candidates and vulnerability mapping to determine the likelihood of violence have reinforced the autonomy and efficiency of ECI (Sridharan & Vaishnav, 2017).
Nevertheless, as a bureaucratic apparatus of the state, ECI is embedded in the wider institutional framework, and its functioning is shaped by the nature of governance and political dynamics in the state. It is worth noting that the assertiveness of ECI occurred in the 1990s with the decline of the one-party dominant system and the ‘second democratic upsurge’. It is the wider political context of plural and competitive electoral system and the emergence of the regulatory state that facilitated the activism of ECI (Singh & Roy, 2019). There have also been attempts by the executive and political parties to influence the activities of ECI. Sridharan and Vaishnav (2017) express concerns about the autonomy of ECI due to the possibility of partisan appointments in the absence of precise provisions on appointment of the Chief Election Commissioner (CEC).
The embeddedness of electoral management bodies has been recognized in literature. As Grofman et al. (1999) argue that like all democratic institutions, electoral governance institutions are embedded institutions in historical configurations of institutional, cultural, and socio-economic factors. Mozaffar and Schedler (2002) point out that electoral governance involves rulemaking, rule application and rule adjudication to provide procedural certainty but remain socially and institutionally embedded, which matter most during the indeterminate conditions. Understandably, the prevailing political system and the dynamics of political competition not only influence the choice of rules of electoral governance but also their operational effectiveness and political outcomes. In this context, the overall condition of governance may impact the operation and effectiveness of EMB.
Even though the performance of EMBs in India is commendable, subnational differences in performance and efficiency have remained relatively unexplored. Intuitively, it can be argued that the operation of the election commission at the ground level is related to the overall governance in the subnational states. The federal framework of India has 29 subnational units and 7 union territories, and the elections to the national parliament, as well as state legislatures, are conducted by the same body, ECI. The federal structure and electoral system provides a natural experiment to analyse the importance of the institutional dimensions of election (i.e., EMBs) vis-à-vis the non-material cultural dimensions of administration, that is, administrative ethos.
Although the institutional framework of ECI corroborates the general prescriptions of autonomy and impartiality, the members of ECI are appointed from the civil service, and in the exercise of electoral functions, the commission has to rely on governmental staff. This provides scope for some potential de facto external interference. During elections, ECI calls upon the resources of the public sector, particularly the permanent pool of skilled, experienced and trained full-time employees at national and local levels to administer contests. Electoral authorities also use the existing technological and communication infrastructure of national and local governments, reducing overhead costs and the duplication of services. Given the scope of indirect intervention of government, it is possible that formal legal independence and impartiality may be compromised at the level of practice.
Case Study and Methodology
This article interrogates whether the framework of electoral integrity that exists in national parliamentary elections also translates to India’s state legislative assembly elections. It focuses on electoral integrity at the subnational level and explores the variations in integrity. It offers a range of explanations that account for any observed variation in the perception of electoral integrity. Given the similarity in the structural institutional framework provided by ECI, subnational variation in electoral integrity, if any, could be attributed to state-specific socio-economic factors and political processes. The broader institutional frameworks are largely similar across the states, and ECI conducts state assembly elections following uniform regulations and procedures. As such, the perceived difference in electoral integrity directs attention to subnational variation in socio-economic status and the quality of governance in different states.
To evaluate the subnational electoral integrity in India, we conducted an original (expert) opinion survey of electoral integrity in India. This was the first effort ever in which the electoral integrity framework was applied to study Indian elections. Between 2015 and 2017, the PEI project evaluated nine subnational elections. The survey of electoral integrity was conducted across Assam, Bihar, Goa, Kerala, Manipur, Punjab, Tamil Nadu, Uttar Pradesh and West Bengal. Some of these states have very large populations (e.g., Uttar Pradesh has a population of 204 million people), while others have a comparatively small population (e.g., Goa has a population of 1.8 million people). Overall, these nine states account for 43 per cent of India’s total population.
The focus on subnational electoral integrity stems from the recognition of the increased prominence of subnational states in policymaking and administration, particularly after economic liberalization (Saez & Mahmood, 2016). It is also due to the recognition that politics in India has largely been played out at the level of subnational states (Yadav, 1996). It is important to note that turnout rates in state-level elections have overtaken turnouts in national-level elections; for instance, the average voter turnout in the period 2003–2007 for national elections was 60 per cent, while the turnout for state elections in the same period was close to 65 per cent. In the period 2008–2012, average voter turnout for national elections reduced to 59 per cent, while for state elections, it increased to almost 69 per cent (Vaishnav, 2013).
The PEI index measures expert perceptions and evaluations of electoral integrity much like the surveys on human rights, corruption and democracy such as Freedom House and Transparency International. The survey provides proxy indicators of 11 relevant dimensions of electoral practices over the election period. The PEI consists of election laws, constituency boundaries, party and candidate registration, campaign finance, vote count, election procedure, campaign media, voter registration, voting process, result transmission and election management bodies. It deconstructs the larger electoral process into the three electoral cycles of pre-election stage, election day and post-election period. The criteria to conceptualize electoral integrity are explicit, can be validly measured and comparable across cases. Based on expert evaluation, the PEI survey produces an integrity score from 0 to 100, which enables comparison across countries as well as different dimensions of election.
The survey of experts to evaluate electoral integrity is apposite due to the furtive nature of election fraud and the complexity of electoral processes. It is preferable to reports which are often impressionistic, focused on the election day and consider election quality largely in bimodal terms (Elklit & Andrew Reynolds, 2002). On the contrary, the PEI expert survey is attentive to different dimensions of the electoral process and elicits opinion of experts with knowledge of the specific election (Norris et al., 2014).
Opinion is elicited from experts who are political or social scientists and have published on or have other demonstrated knowledge of the electoral process in a particular state. Demonstrated knowledge is evaluated by the following criteria: (a) membership of a relevant research group, a professional network; (b) existing publications on electoral or other state-specific topics in books, academic journals or conference papers; and (3) employment at a university or college as a teacher (Norris et al., 2014). To ensure that biases in the expert evaluation are minimized, the survey design collates background information on gender, age, education, employment status, political orientation and partisan support.
To find experts, we reached out to organizations like Centre for the Study of Developing Societies, New Delhi and various universities to find scholars with concentration and publication in electoral issues. The membership list of the Indian Political Science Association was also useful to identify state-level experts. This pool was, however, not adequate (40 experts in each state) in some cases, and we broadened the academic focus of the experts to include state politics in general. In hindsight, broadening the topics on which academics conducted research beyond electoral politics arguably reduced the possibility of biasing our sample. In some states, the aforementioned criteria did not generate the required number of experts, and we had to resort to snowballing by inviting experts in our contact to suggest others with relevant state- and election-specific knowledge. Once we had the required number of experts, we sent an email invitation to complete the online survey. Two reminder emails were sent to non-responders 1 and 2 weeks after the initial invitation.
Subnational Electoral Integrity
The survey data provide for the first time an opportunity to not only examine the quality of subnational elections but also the role of sociopolitical variables in determining electoral integrity, given the broad structural institutional similarities. Figure 1 shows the aggregate results for the PEI index across nine states in India. As Figure 1 illustrates, there are significant variations in the overall perception of electoral integrity in the chosen subnational states of India (see Figure 1).
In absolute terms, the all-India average PEI score for these 9 states was 58. In terms of absolute scores, the state of Bihar scores the highest with a PEI score of 68.9, followed by Kerala with 65.4 and Tamil Nadu with 61.6. On the other hand, states like Goa and West Bengal trailed behind other states with a PEI score of 45 and 52.9, respectively. As Figure 1 reveals, some states—like Bihar, Kerala, Punjab, Tamil Nadu—have a PEI score above the all-India average, indicating relatively free and fair elections there. In contrast, some states like Goa and West Bengal fall below the India score, indicating below par elections.
The PEI scores and the subnational variations are counter-intuitive and raise a number of serious questions. In terms of major socio-economic indicators, Tamil Nadu and Kerala are considered among the developed states of India and expectedly corroborate the better quality of elections. However, the position of Bihar, West Bengal and Goa challenge the direct correlation between development and quality of elections. In most relevant indicators, the state of Bihar has the lowest position among the states considered.
The PEI score is the aggregate measure of the various dimensions of election and interrogating the specific dimensions can illuminate the underlying processes resulting in such counter-intuitive outcome. To interrogate the relationship between overall development and electoral integrity further, the dimensions of elections are considered in a disaggregated manner. As mentioned, these parameters include fairness of the media, electoral laws, voter registration procedures, party registration procedures, drawing of state boundaries, electoral finance, the role of electoral management bodies, voting day situation, counting of votes, results and finally, electoral procedures. The results for the different dimensions of the electoral process are reported in Table 1. To identify the specific dimensions of election that contribute to the interstate variation in electoral integrity, we measure the standard deviation of each specific category. Together with the PEI scores of India, such a measure can highlight the more crucial components of the election that require greater attention (see Table 1).
Measures of the Various Dimensions of Electoral Integrity
Certain distinct features of electoral integrity can be identified from the data in Table 1.
First, campaign finance, voter registration and media fairness show the greatest inherent weaknesses of the election system in India. The weakness of campaign finance and media bias in India has been highlighted by practitioners and scholars (Mahmood, 2015).
Second, in terms of voter registration as well as party registration, the subnational states perform better than the Indian average. Voter registration which involves enlisting of voters and preparation of electoral list can be expected to be better in state elections due to the smaller size of constituencies, and the fact of elections being closer to the ground compared to parliamentary elections.
Third, the quality and impartiality of election laws are very high across the states. In terms of media coverage, although the scores are low, most of the states have similar scores. Biased coverage and partisan reporting are endemic and somewhat similar across the states.
Finally, election procedure, result declaration, vote count and EMBs were the four dimensions of the election that exhibited the most variation in response. Although the subnational performance on these dimensions is relatively high, they have maximum divergence across states. Interestingly, EMB, which is institutionally similar, that is, ECI, has a crucial role in the performance of these functions.
Explaining Subnational Variation
The PEI survey outcomes thus present two interrelated but distinct questions, namely, how can one account for the subnational variation in electoral integrity, given the similarity in institutional factors that govern elections in India? Secondly, how can one explain the high level of electoral integrity in Bihar (a low-income state) and low integrity in West Bengal (a high-income state)?
As a possible answer to the questions raised, the article considers the subnational variation in socio-economic status and the quality of governance in different states. These two broad explanatory variables—socio-economic development and governance—have been highlighted in the literature as responsible for variations in democracy and electoral integrity.
The literature on variations of the quality of democracy hinges around a number of overlapping themes, typically stressing institutional or socio-economic factors that account for variations in the quality of democracy (Diamond & Morlino, 2004). The main thrust of this literature is that the core structural dimensions of democracy (e.g., freedom, rule of law, accountability, participation) are interrelated. In this context, variations in economic equality can affect within-country or between-country differences in the quality of all dimensions of democracy (Rueschemeyer, 2004). One of the most important empirical insights in this literature is that there is a positive relationship between higher levels of income and democracy. Countries with higher income are more likely to have more successful transitions to democracy, while their process of democratic consolidation will be more stable (Barro, 1999; Boix & Stokes, 2003; Przeworski & Limongi, 1997).
While discussing the subnational variation in the outcomes of electoral integrity, one finds that—for the most part—subnational divergence in the institutional and socio-economic status corresponds to electoral integrity. To elucidate, the nine states are culturally and geographically diverse with Assam and Manipur in the north-east, Punjab in the north-west, Uttar Pradesh in the mid-west, Bihar and West Bengal in the east, and Tamil Nadu and Kerala in the south of India. The nature of political contestation is also very diverse across the states, with electoral competition between regional parties in Tamil Nadu, the contest between regional and national parties/alliances in Bihar and West Bengal and the contest between two national parties in Assam and Manipur and between two stable alliances in Kerala. In Uttar Pradesh and Punjab, the electoral competition was between three and four parties and included both regional and national parties. The subnational states are also marked by considerable differences in socio-economic status such as rates of literacy, economic growth and urbanization, all factors that have consequences for the quality of elections.
In general, higher-income states in India appear to have higher levels of electoral integrity. To interrogate the positive relationship between income and electoral integrity, we focus on a number of illustrative socio-economic variables. The first indicator is the state’s net state domestic product (NSDP) per capita. NSDP is an economic indicator equivalent to GDP at the national level. Figure 2 shows the per capita NSDP for the selected nine states in India (see Figure 2).
As the data in Figure 2 show, for the year 2013–2014, the state of Goa has the highest per capita income. Tamil Nadu, Punjab and Kerala are among the high per capita income states of India, while Assam, Bihar, Manipur and Uttar Pradesh are among the low-income states. The position of West Bengal is intermediate. At absolute terms, the per capita NSDP of Bihar is lowest at ₹31,199, and Goa is highest with ₹224,138 followed by Tamil Nadu with ₹112,664 and Kerala with ₹103,820. The per capita NSDP of West Bengal with ₹70,059 is near the India average, while Assam with ₹44,263, Manipur with ₹41,573, and Uttar Pradesh with ₹36,250 are well below.
The disparities between the selected states are also evident in terms of absolute poverty levels. The Government of India measures poverty levels based on household consumer expenditure surveys undertaken by the National Sample Survey Office (NSS), Ministry of Statistics and Programme Implementation. Figure 3 shows the share of households in a given state’s population living below the official poverty line (see Figure 3).
As Figure 3 shows, some states have a large percentage of its population living under the poverty line. Manipur is the state with the maximum proportion of the population living below poverty lines at 36.9 per cent closely followed by Bihar at 33.7 and Assam at 32 per cent. In contrast, Goa has the least share of population living below the poverty line at 5.1 per cent followed by Kerala at 7.1 and Punjab at 8.3 per cent (Planning Commission of India, 2013). The divergence in economic condition among the states broadly corresponds to other related socio-economic variables. The rate of literacy and urbanization across the states broadly mirror the macroeconomic situation. In terms of literacy, Bihar has the lowest literacy rate at 61.8 per cent as compared to 94 per cent in Kerala, 88.7 per cent in Goa, 80.09 per cent in Tamil Nadu, 76.9 in Manipur 76.26 per cent in West Bengal, 75.8 in Punjab 72.19 per cent in Assam and 67.7 in Uttar Pradesh (Census, 2011).
It is well recognized that social and development indicators of the states of Kerala, Goa are very high compared to the rest of India. The states also vary significantly in terms of urbanization. The Government of India calculates the percentage of the population living in an urban setting based on minimum population size and density, plus the proportion of the male working population engaged in non-agricultural pursuits. The 2011 census of India calculated that 31.1 per cent of India’s population lived in urban areas. Given this context, the share of urban population is very high in the states of Goa with 62.17 per cent of the population living in urban areas closely followed by Kerala and Tamil Nadu with nearly 50 per cent of the population living in urban areas. As expected, based on the socio-economic data provided in Figures 2 and 3, the poorer states of Assam, Uttar Pradesh and Bihar have very low levels of urbanization (see Figure 4).
The data in Figures 2–4 show a consistent pattern: high-income states have lower levels of poverty and higher levels of urbanization. The reverse is true for low-income states in India. These socio-economic patterns are significant because it is well recognized that economic condition, poverty, literacy and urbanization have implications for the quality of elections (Norris, 2015). Intuitively Bihar, Uttar Pradesh and Assam should have lower levels of electoral integrity, with Goa, Kerala and Tamil Nadu positioned along the high end of integrity spectrum. How these differences play out and affect electoral integrity becomes an interesting case for study.
We could infer that differences in the electoral integrity of different states stem from the subnational variation in their socio-economic status. However, these socio-economic variables cannot adequately explain why a poor state like Bihar has a higher degree of electoral integrity than a high-income state like West Bengal. 3 To explain these anomalies, we examine whether political processes, especially the quality of governance, can better explain the outcomes of electoral integrity.
Subnational Divergence in the Quality of Governance
Based on a large-scale quantitative analysis, Norris (2015) has found that the quality of elections is influenced by the general effectiveness of governments, such as the quality of the civil service and public services, the independence of public sector agencies from political pressures, and the quality of policy formulation and implementation. Where public administration is well-trained, experienced and enjoys procedural autonomy, then electoral authorities can call on these resources to manage effective and efficient contests. Where these qualities are lacking, the public sector is undermined by partisan politics, clientelistic favours and corrupt practices.
The incongruity between socio-economic development and electoral integrity across some of the subnational cases brings into focus the role of governance. Norris (2015) has found a robust relation between the PEI index scores and quality of the public sector and government effectiveness based on the World Bank Quality of Governance survey, and the World Bank Institute’s ‘good’ governance indices have been developed by Kaufmann et al. (2009). In the context of the selected cases also, it is possible to highlight the relationship between governance and electoral integrity, especially the performance and effectiveness of EMB.
The issue of governance at the subnational level has received some attention since economic liberalization as the states emerged as key players in the process of reforms (Saez & Mahmood, 2016). Despite the rhetorical and policy importance, there have been only a few attempts to measure or qualify governance in subnational India. The most prominent and important measure provided is by Mundle et al. (2012). Based on an assessment of the three branches of a developmental state—namely the executive, the judiciary and the legislature—they measure the delivery of infrastructure services, delivery of social services, fiscal performance and maintenance of law and order, delivery of legal services under the judicial pillar and the quality of the legislature under the legislative pillar. These six dimensions of governance are widely agreed dimensions of governance in literature, such as Besley and Persson (2011a) and Kaufmann et al. (2009).
Importantly, the study by Mundle et al. (2012) not only provides a measure of governance indicator for 17 major states of India, but it also disentangles the impact of ‘development clusters’. According to Besley and Persson (2011b), there exists a strong correlation between good governance and development. This is partly because of their dependence on the same underlying drivers and also because of their mutual inter-dependence, giving rise to development clusters. This can give rise to a governance rating bias in favour of the more developed states (Mundle et al., 2012).
In order to isolate the quality of governance, independent of the level of development, Mundle et al. (2012) transform the raw indicators using principal component analysis (PCA) and the Borda ranking or the average of the sum of ranks. PCA summarizes and captures the variation in the data in the form of uncorrelated components (vectors) called principal components. However, this method for summarizing multidimensional data involves a considerable loss of information on the absolute scale of variance of some indicators compared to others as it equalizes mean and variance of all the indicators and gives higher weights to those indicators that have higher correlations with other selected indicators, though such weighting may not be justified by the intrinsic importance of the indicator (Mundle et al., 2012).
In contrast, the Borda method or average of the sum of ranks arranges the states according to ranking in each individual indicator of governance and assigns a rank score. The rank score received by the states under each dimension is added up to give us the sub-index of quality of governance for that dimension. The final index of quality of governance is obtained by adding up the sub-indices and then dividing the number of dimensions of governance by six. Thus, it does not privilege any one dimension of governance over another as there is no a priori reason to consider a particular branch of governance as being more important than another (Mundle et al., 2012).
Table 2 reports the PCA and Borda ranking for the nine states that we have examined in this article. The last column reports the PEI scores for these selected states (see Table 2).
Evident from Table 2 is the strong positive relationship between the quality of governance rankings and the electoral integrity scores. In this context, the Borda ranking has greater proximity to PEI scores
Governance Ranking and PEI
compared to PCA ranking, primarily due to methodological issues. Evidently, states with better governance have better electoral integrity, although the correlation is not absolute. Notably, the governance rankings are at the All India level, while PEI scores are available for the selected states only. It will be interesting to note the relation once PEI data for 17 major states are available.
Anecdotal evidence corroborates the quality of governance across the subnational states. Since 2005, the government of Bihar, under Chief Minister Nitish Kumar, instituted several legislative acts—like the Right to Public Service Act 2011 and the Bihar Police Act 2007—to redefine the role of the bureaucracy and police, by promoting transparency in dealings with civil servants and how the police deals with crime (Mukherji & Mukherji, 2012). Improved governance is reflected in the enrolment rate in primary schools, which increased from 86.5 per cent in 2005 to 97 per cent in 2012. The state’ bicycle scheme started by the government in 2006, which provides ₹2,500 to every child enrolled in Grade 9 to purchase a bicycle, contributed to the hike in enrolment in secondary schools. These seemingly scattered facts show a structured attempt at improving governance in Bihar, in both social and economic terms which Mathew and Moore (2011) characterize as a ‘governance miracle’. While in absolute terms Bihar may be lagging behind other states in India—due to Bihar’s history of economic backwardness—there has been considerable economic development in that state. We argue that such structured development, which brings governance into the everyday lives of the people, can create a positive impact about the perception about the quality of governance in the state, which lends credibility to the state’s institutional arrangements.
In contrast, governance deficit has been a recurrent theme concerning evaluations of West Bengal. Scholars like Ray (2011) point out that poor work culture, political interference and failure of governance are characteristics of West Bengal. The change in government in 2011 has led to some improvement in governance indicators like school dropout rates and child marriage, but the period has been marked by deteriorating law and order management with incidences of violence and extortion (Mitra & Bhattacharyya, 2018). In a recent survey on ‘Governance issues and Voting Behaviour’, the state performed poorly in almost all governance issues such as employment generation, agriculture loan and price realization for farm products. The state’s performance was only average in the field of empowerment of women and electricity for domestic use (PTI, 2016).
The subnational PEI index of electoral integrity suggests variation in the expert evaluation of EMB fairness and competence. This divergence can be correlated to the differences in quality of governance encapsulating functional effectiveness and state capacity. As Norris (2015) points out, the delivery of services and administrative ethos are the two dimensions that exert influence on electoral administration. A well-functioning bureaucratic apparatus characterized by a culture of impartiality, effectiveness, legality, efficiency, transparency, and integrity is necessary for well-managed elections. Electoral integrity at the sub national level corroborates that states with improved governance not only had better election but the EMBs had greater credibility in the PEI index.
Conclusion
Electoral integrity has been an area of concern for both election practitioners and scholars emphasizing on substantive democracy. The evaluation of elections at the subnational level is an innovative exercise in the context of India. The enormity of the elections in India is reflected in the subnational states as many of them have population greater than most developed countries. The federal structural with significant heterogeneity in socio-economic development and political dynamics makes India an important natural case study.
At a specific level, elections in India, both at the national and subnational levels, are beset with problems of voter registration, media access and campaign finance. All the states perform poorly in these regards, which require intervention if elections are to become free and fair. In recent years, voter registration has got entangled in the citizenship debate as the preparation of electoral roll has coalesced with identification of citizens (Singh & Roy, 2019). Interestingly, in the context of demonetization of currency in India, the issue of campaign finance regulation has been carefully bypassed by the government, and the introduction of electoral bonds has skewed campaign finance.
The subnational comparative evaluation of electoral integrity highlights not only the quality of elections but also the potential factors that influence elections. The PEI model with a uniform scheme allows meaningful comparison between states which differ culturally, geographically and politically. Even though the integrity scores of a perception survey are only indicative, the results show significant variations in the quality of elections across the states. Bihar, Kerala, Tamil Nadu have PEI score above India, indicating relatively free and fair elections. In contrast, states like West Bengal and Assam fall below the India score, indicating below par elections.
As expected, the main findings suggest that electoral integrity is usually stronger in states characterized by effective governance. All the selected cases are characterized by the agency model of EMB with de jure independence. Even if we cannot claim, the irrelevance of de jure EMB, based on the case study, it is clear that quality of elections depends not on the de jure legal independence but state capacity and bureaucratic culture. Following the categorization of Norris (2013) between the first- and second-order malpractices, it is evident that ‘second-order’ administrative issues influence the quality of elections across subnational states of India.
The results of the PEI survey from across the world and subnational India show governance has a critical role in shaping electoral integrity. Positive governance such as regulation of electoral laws, control of campaign finance, ease of voter registration, proper counting and announcement of results and the role of electoral management bodies can also lend credibility to the electoral process despite the presence of structural hindrances, as seen in the case of Bihar. In contrast, despite better socio-economic indicators, the evaluation of elections in West Bengal is poor due to governance deficit. Electoral governance institutions are embedded in historical configurations of institutional, cultural and socio-economic factors and are strongly affected by the political system and the dynamics of political competition. The overall condition of governance impacts the operation and effectiveness of EMB.
Finally, the outcome also highlights the limitations of orthodox modernization theory that emphasizes on prerequisites for a successful democracy, particularly between democracy and development. Elections are a vital structural and substantive component of democracy, and PEI data show that relation is more nuanced and complex. While there is no doubt that socio-economic development and democratic institutions are critical to creating positive conditions for democracy, they are only a part of the necessary condition. The quality of governance is indispensable for democratic processes and perception of legitimacy that cannot be entirely subsumed under economic development. Economic development does not necessarily guarantee a genuinely legitimate democratic process, even if economic affluence has a positive correlation to democracy.
Footnotes
Declaration of Conflicting Interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author received no financial support for the research, authorship, and/or publication of this article.