Understanding the Indian Voter

The Bharatiya Janata Party’s (BJP’s) historic victory in India’s 2014 general election prompted declarations of a watershed in the behavior of the Indian voter. Upon closer inspection, the reality is more nuanced. On some parameters, such as voting based on economic and ethnic considerations, there were indeed discernible changes. However, the empirical evidence suggests these shifts were well under way before 2014. In other areas—namely, support for regional parties, dynastic politicians, and candidates associated with criminal activity—contemporary voters demonstrated much greater continuity with the past.

Continuity and Change in Voting Behavior

• Good economics can make for good politics in India. While parochial considerations have long been thought to play a central role in shaping voters’ choices, evidence from state and national elections suggests that macroeconomic realities are increasingly relevant.
   
• There has been much discussion of regional parties’ increasing influence in Indian politics. But recent electoral trends reveal a surprising degree of stability in the balance of power between national and regional parties.
   
• Dynastic politics may not be popular, but dynastic politicians are. At least one in five members of parliament elected in 2014 came from a political family.
   
• Indian voters have a long history of electing politicians who are the subject of ongoing criminal cases. The 2014 results demonstrate an underlying demand for politicians who can get things done—even if they are connected with wrongdoing.
   
• When voters cast their vote, they do not necessarily vote their caste. Social biases remain entrenched in India, but the transmission of those biases into the political domain is imperfect and may be weakening.

Broader Lessons for Indian Democracy

While the regional-national balance of power is steady, there has been an upheaval among national parties. For now, the ascendant BJP has replaced the foundering Indian National Congress as the pole around which political competition is organized.

Politicians who seek to gain strength using identity-based appeals alone have generally not fared well. While voters may harbor deep-seated social biases, identity-based concerns and economic evaluations are both in play. The most successful politicians have mastered the art of skillfully combining both types of appeals.

The composition of the candidate pool does not appear to be shifting, despite an increase in the absolute number of parties contesting elections. Voters have more choice than ever before, yet there is little qualitative change in the nature of the candidates themselves.

Introduction

The 2014 Indian general election was a historic event on multiple fronts.

It was exceptional, first and foremost, by virtue of its size. A staggering 834 million citizens were eligible to vote, of which nearly 554 million voters actually cast their ballots on Election Day at more than 900,000 polling stations throughout the country. This amounted to a voter turnout rate of 66.4 percent, the highest in India’s history and a full 8 percentage points higher than the turnout recorded in either the 2004 or 2009 national elections. When deciding whom to vote for, India’s voters had an expansive number of options to choose from: 8,251 candidates representing 464 political parties faced off in 543 parliamentary constituencies across 29 states and seven union territories.¹ Although the numbers are difficult to pin down, experts believe the 2014 poll was the second most expensive election ever held, following only the 2012 U.S. presidential election.²

The election’s outcome was also historic. Preelection polls had uniformly predicted that the opposition Bharatiya Janata Party (BJP), which held power from 1998 to 2004, would emerge as the single largest party in parliament. Yet few foresaw the size of the victory achieved by the BJP and, conversely, that of the defeat suffered by the ruling Indian National Congress (INC).³ Of the 543 seats in India’s lower house of parliament (Lok Sabha) on offer, the BJP claimed victory in 282, while its coalition allies notched wins in another 41 seats. Although it chose to form a government with its preelection alliance partners, the BJP was not compelled to do so; it had a clear majority in the Lok Sabha, the first time a single party has won such a majority since 1984 and the first time ever that a single party other than the Congress has achieved this feat.⁴

The Congress Party, on the other hand, saw its tally drop to a paltry 44 seats, a dramatic decline from the 206 seats it captured in 2009 (see figure 1) and its worst-ever performance in a general election. The party’s seat total was so small that it fell short of the minimal threshold necessary to nominate one of its own to serve as leader of the opposition in the lower house—a designation that requires a command over at least 10 percent of the body’s seats.

The BJP victory also appears to have redrawn the political map of India (see figure 2). After winning more than one-quarter (25.6 percent) of ballots cast and 182 seats in the 1998 general election, the BJP experienced a secular decline on both counts. In 2009, its tally dropped to 116 seats, limited to a few select pockets of the country. This smattering of seats grew by leaps and bounds in 2014, and now touches all four corners of the country, from Gujarat in the west to Assam in the east, and from Tamil Nadu in the south to Jammu & Kashmir in the north.

Of course, the election also garnered significant attention in no small measure thanks to the persona of the BJP leader, Narendra Modi. From October 2001 to May 2014, Modi had been the chief minister of Gujarat, a prosperous, midsize state of roughly 60 million people. In 2002, Gujarat was shaken by horrific ethnic violence in which over 1,000 people lost their lives, three-quarters of whom were from the minority Muslim community. In the minds of many Indian electors and election observers, Modi’s reputation was indelibly tarnished by the events of 2002 and their aftermath. Although it is hard to imagine now, just 1.9 percent of respondents to a national survey conducted after the 2009 general election said Modi was their choice to be India’s next prime minister.⁵

India’s ability to balance robust, participatory democracy with a diversity of ethnic groups—however imperfect—makes it the envy of the democratic world.

However, in the subsequent five years, Modi went from being a long-serving provincial official with a questionable biography to the most popular politician in all of India and, after the 2014 vote, the country’s prime minister. Central to his extraordinary rise in stature was the impressive economic record he had amassed in Gujarat over a decade, which provided a stark contrast to the flagging economy in the final years of the Congress-led United Progressive Alliance (UPA) government.⁶

But if one retreats from the immediate political context of the BJP victory, it is clear that India’s 2014 general election captured widespread attention for an even more fundamental reason. The magnitude of the BJP’s victory, and the manner in which it was achieved, raises the prospect that the very nature of the Indian voter is changing and that some of our bedrock assumptions about how she behaves are no longer accurate. For instance, given Modi’s constant emphasis on development and economic growth on the campaign trail, it is possible that economic issues matter more to Indian voters now than they have in the past. And, considering the utter defeat of the once hegemonic Congress Party, anchored by the Nehru-Gandhi political family, perhaps Indian voters are less enamored of dynastic politicians than in previous eras.

In light of the 2014 general election, it is worthwhile to look at what really makes Indian voters tick. The importance of this appraisal should not need elaboration. One in every six humans on earth resides in India, and the country has as many eligible voters as Europe and North America combined. In global terms, India is a rising power whose economic and geostrategic potential make it a significant player on the world stage. India’s ability to balance robust, participatory democracy with a diversity of ethnic groups—however imperfect—makes it the envy of the democratic world.

All things considered, the 2014 election presented a mixed picture. Economic factors played an unusually large role in shaping voting behavior. Traditional patterns of caste-based voting were much less evident, and regional parties, often thought to be gaining ground, suffered a setback. A slightly deeper look, however, reveals that these changes were not necessarily unique to the 2014 general election. There is evidence to suggest that many of these trends have been percolating beneath the surface for some time. What 2014 has done is to bring these trends to the fore of public consciousness.

However, on other dimensions, 2014 signaled more continuity than change. Despite widely held pronouncements that a young, aspirational, and increasingly urban India is no longer amenable to the idea that politics will be, to a significant measure, the province of a relatively small number of political families, dynastic politicians performed very well in the 2014 elections. The same can be said of politicians associated with criminal wrongdoing; as a percentage of India’s parliament, more politicians with criminal cases pending against them (including those of a serious nature) were elected than ever before.

These findings suggest that the Indian voter has neither fully entered a new era, nor stayed firmly in line with past behavior. For instance, while the aspirations of Indian voters may have changed insofar as they are placing greater emphasis on macroeconomic outcomes when deciding whom to vote for, there seem to be few changes in the personal characteristics of individual candidates on offer.

At the same time, social biases built around caste or ethnic identities remain deeply entrenched in politics, even as the way they are expressed might be changing. Cross-caste coalitions have been a hallmark of recent state and national elections, and appeals to identity politics are increasingly married with promises of economic development.

Finally, the regionalization of politics has stabilized in recent years, and may in fact have suffered a reversal. The share of the vote flowing to regional parties decreased in the 2014 general election, and many dominant regional parties saw their fortunes plummet even as a limited number of others held steady. The biggest change, however, was the composition of the vote accruing to the two national parties with truly pan-Indian appeal, the Congress and the BJP. In a sharp break with the prevailing trend in postindependence India, the BJP has now displaced the Congress as the “pole around which political competition is organized,” to borrow a phrase coined by Yogendra Yadav.⁷ In many states where the BJP had a marginal presence in past elections, the party profited at the expense of the Congress, whose local support greatly diminished.

The Commonly Held Beliefs

Since the dawn of Indian independence in 1947, there has been a large body of scholarship on the Indian voter. This collective understanding has been shaped and reshaped and—although consensus is fleeting when it comes to the study of any country as large and complex as India—in recent years a broadly accepted picture of Indians’ voting behavior has emerged. Five propositions in particular represent a significant portion of what is generally known about how Indian voters act.

To begin with, there is a belief in India that good economics does not make for good politics. On the contrary, the success of Indian politicians depends on how voters evaluate them on issues of a parochial nature, be it caste, religion, patronage, or clientelism.⁸ Unlike in many advanced democracies—such as the United States, for instance—in India, larger economic considerations are not thought to be at the top of voters’ minds on Election Day. Thus, when it comes to winning elections, the overall health of the economy is not perceived to be the preeminent factor.

There is also a general sense that regionalism is surging in India. The 1980s saw an onslaught of regional and caste-based parties that have since been thought to be chipping away at the once sacrosanct position of the two national parties.⁹ The end result of this regionalization of Indian politics is increased competition, heightened fragmentation, and a greater prevalence of coalition politics. National politics was thought to be little more than a sum of state-level dynamics.

Although family politics has been a fixture of India’s democracy since the early years of the republic (even reaching back into the colonial era), the conventional wisdom has suggested that the reign of dynastic politicians is gradually coming to an end. Thanks to oligarchic leadership, a lack of internal democracy, and limited upward mobility within parties, many parties resemble closed family firms rather than vibrant political organizations characterized by internal competition, contestation over ideas, and meritocratic selection.¹⁰ However, in recent years, some commentators have suggested that the mood among the electorate is shifting, and that voters now chafe at the political success of a privileged few who are seen as disconnected from ground realities.¹¹

Meanwhile, politicians associated with criminal activity have risen to national prominence, in spite of the fact that, at least since 2003, they must disclose their criminal records at the time of their nomination. Across all levels, from local to national politics, a significant number of politicians in India have thrived in elections even though they have had criminal cases—often of a serious nature—pending against them. Some researchers as well as representatives of civil society have argued that this is an outgrowth of the ignorance of India’s voters who, faced with a dearth of information about the true nature of their political candidates, unwittingly elect (and reelect) people linked to criminal activity.¹²

Lastly, voters do not so much cast their vote as vote their caste, or so the story goes.¹³ There is a wide body of scholarship across a range of social science disciplines that points to the commanding influence of caste on voting behavior, and to the value Indians place more broadly on identity-based considerations when encountering politics. Identity is also thought to shape the strategies that politicians use when seeking votes during their campaigns and the policies they pursue while in office.

Of course, these beliefs are somewhat arbitrarily chosen; undoubtedly there are many other popular notions that are worth highlighting. However, these five issues are highly salient, and reasonably high-quality data about them are available.

Several of these propositions, it turns out, are worthy of revision; much of the thinking behind them is wrong, incomplete, or—at the very least—lacking in nuance. In some instances, it is not obvious that this conventional wisdom was ever fully backed by the available evidence.

However, the 2014 general election in and of itself does not necessarily represent a structural break in voter behavior. While this election offers a natural opportunity to pause and take stock, in many cases evidence militating against these widely accepted truths was apparent well before the 2014 poll was held.

And discarding old notions does not mean that constructing new understandings is equally easy; in several instances, gaps in knowledge make it difficult to build a new, complete picture of the Indian voter. In other words, while shortcomings in the conventional wisdom can be identified, a great deal is still unknown, making continued research essential.

Moving Toward “It’s the Economy, Stupid”

In most democracies, the overall health of the economy is perhaps the most important barometer of an incumbent government’s political popularity. When economic indicators are broadly positive, voters are more likely to approve of the job their representatives are doing and, hence, reelect them. The converse is also true. Although the precise relationship between economics and elections varies considerably by country and time period, the tight correlation between the two is regularly treated as received wisdom.¹⁴ As one scholar put it, “The notion that voters judge democratic governments by how well they manage the economy has taken on the ring of an incontrovertible social scientific fact.”¹⁵ Or, as political strategist James Carville summed it up in what became the defining motif of Bill Clinton’s 1992 U.S. presidential campaign, “It’s the economy, stupid.”

Indians have traditionally cared more about their own personal well-being than the country’s broader macroeconomic health.

Notwithstanding the popular consensus around economic voting, scholars have typically treated India as something of an outlier in this regard. Instead, it has been thought that good economics need not make for good politics in India because voters regularly prioritize other factors—namely patronage, populism, or parochialism—when selecting their representatives. The result is that even when governments perform in ways that objectively improve the state of their economies, voters often reject, rather than reward, them at the ballot box.

This description is slightly unfair insofar as it assumes Indians are irrationally unconcerned with their material well-being. Perhaps a more fair characterization is that Indians have traditionally cared more about their own personal well-being than the country’s broader macroeconomic health. This can then easily be linked to the distinction scholars have made between “egotropic” (pocketbook) and “sociotropic” (national, or macroeconomic) voting.¹⁶

If one assumes that parochial forms of voting have been driven by an underlying pocketbook logic, there does appear to be evidence of a shift in the voting patterns of Indians toward more macroeconomic concerns. However, the 2014 election, rather than signaling its arrival, provided a reaffirmation of a change that was already under way.

A Cautionary Tale

If there is one example analysts have repeatedly used to highlight the Indian voter’s relative inattention to economic performance, it is the demise of Chandrababu Naidu, the former chief minister of the state of Andhra Pradesh and head of the Telugu Desam Party (TDP).¹⁷ Naidu came to power in 1995 with the goal of revamping Andhra’s economic fortunes, quickly developing a reputation as a leader committed to good economic governance and the kinds of promarket reforms typically favored by international donors and multilateral financial institutions. Doubling down on this image, Naidu even took the unusual step of hiring Western consulting giant McKinsey to develop reform ideas, which were encapsulated in a manifesto entitled Andhra Pradesh: Vision 2020.¹⁸

Although Naidu won reelection in 1999, aided by an alliance with the BJP (which catapulted to power in that year’s national elections), his state’s voters dismissed his government at the next state election in 2004. Why the voters of Andhra Pradesh threw Naidu out of office is unclear; research suggests that a whole range of factors likely played a role, from alliance dynamics to caste calculations and perceptions that Naidu had a “procorporate” or “prorich” bias.¹⁹

Nevertheless, the notion that economic reforms were central to Naidu’s defeat quickly became popular lore, sending a warning shot to other politicians who hoped to institute similar structural changes. Indeed, the political scientist K. C. Suri commented at the time that Naidu’s defeat signaled that “India has not reached a stage where the people would prefer a CEO to a politician to run the government.”²⁰

The example of Chandrababu Naidu is simply one data point.²¹ However, at first glance, more systematic analyses seem to provide some empirical support for the contention that voters do not necessarily reward incumbents who demonstrate superior economic performance. A review of more than 120 state elections in India’s eighteen states from 1980 to 2012 reveals very little association between economics and elections (see figure 3).²² The absence of economic voting is striking because it appears to hold for numerous definitions of “electoral performance”—that is, whether one looks at the change in the percentage of seats or votes won by the incumbent, or simply considers a binary measure of whether the incumbent won reelection.

A Change Afoot?

Despite this apparent lack of connection, there is emerging evidence that the ground is shifting beneath politicians’ feet in India. Several analysts have noted that while Indian voters previously might not have evaluated their politicians on the basis of their economic performance in office, India’s political economy has undergone a structural break at least since the mid-2000s. Increasingly, they claim, Indian democracy too has come to resemble the standard model of retrospective economic voting widely documented in advanced industrial democracies. Reflecting on the growing correlation between positive economic governance and the electoral performance of incumbents, the economist Arvind Subramanian wrote in 2009: “Since independence, many Indian voters have reflexively ejected politicians from office even when they had compiled decent records in power. . . . Recently, though, Indian voters have started to reward good performance, especially in state-level politics.”²³

Politicians too seem to have picked up on this supposed shift. One of the first who articulated the notion that good economics increasingly made for good politics was Modi. Speaking to supporters in December 2012, after he won his third state election in Gujarat, Modi rejected the old narrative that higher rates of economic growth have no association with electoral outcomes. After all, he pronounced, for at least one decade, the people of Gujarat “have shown that [they support] good economics and good governance.”²⁴

Scholarly advocates of this new consensus cite several pieces of evidence in support of their assessments that economic variables have begun to shape the voting patterns of ordinary Indians on Election Day.

Economic variables have begun to shape the voting patterns of ordinary Indians on Election Day.

First, with respect to state elections, many ruling parties in recent years have succeeded in transcending narrow caste appeals and instead forging much broader rainbow coalitions of many discrete caste and religious communities across the social hierarchy. This shift toward stitching together wide-ranging social coalitions suggests that parties are gradually finding ways of broadening their repertoire beyond parochial concerns to include issues of a programmatic nature. The most common example cited in this regard is Bihar Chief Minister Nitish Kumar from the Janata Dal (United), or JD(U).

Kumar’s predecessor, the wily backward-caste politician Lalu Prasad Yadav, ruled Bihar (directly or by proxy) for fifteen years between 1990 and 2005, explicitly prioritizing social justice over growth and development, much to the detriment of governance in the state.

Kumar, another low-caste leader and former member of Lalu’s party, chose a different tack. Although he did not abandon identity politics altogether, Kumar campaigned on the idea of replacing “Jungle Raj,” the epithet often used to characterize the one and a half decades with Lalu (or his wife, Rabri Devi) at the helm, with “Vikas Raj” (vikas means development in Hindi). Kumar captured political power in 2005 on that pledge and, after successfully implementing his promise, won resounding reelection in 2010 with wide-ranging support from the upper, intermediate, and lower castes—as well as Muslims. Kumar’s electoral rout is seen as evidence that narrow, parochial interests are gradually giving way to more programmatic agendas, even in ethnically polarized states.

Reflecting on his reelection, Kumar stated that Bihar was witnessing a “nayi kahani” (new beginning) in which “voters had to choose between progress or antiquated politics. . . . The people whose strategies were based on caste and religion have been left disappointed.”²⁵

Similarly, the Samajwadi Party (SP) rode to a single-party majority victory in the 2012 Uttar Pradesh assembly elections on the back of a future-oriented campaign platform centered on promises to restore good governance and invest in development.²⁶ Like Nitish Kumar, the SP also assembled a broad-based coalition in support of its reform agenda, but—unlike Nitish during his first term—it has thus far failed to deliver on that promise.

Second, at least anecdotally, there are several recent examples—at the state as well as the national level—of governments being reelected at least in part on account of their perceived economic acumen. On the national stage, the UPA government won a resounding reelection in the 2009 general election, having delivered the fastest rates of economic growth in India’s history. At the state level, a crop of popular chief ministers has achieved lasting political power by combining good politics and successful economics. These leaders cut across party lines and represent a broad range of states; they include Raman Singh (from Chhattisgarh) and Shivraj Singh Chouhan (from Madhya Pradesh) of the BJP; Sheila Dikshit (who, until 2013, governed Delhi) of the Congress; and Naveen Patnaik (Odisha) and Kumar (Bihar), who lead regional parties. Singh and Chouhan have held power for eleven and nine years, respectively. Prior to her party’s defeat, Dikshit controlled the reins of power in Delhi for a decade and a half. In early 2015, Patnaik began his fifteenth year in office while Kumar has completed nine (not counting a nearly yearlong period during which he stepped aside after his party’s poor performance in the 2014 general election).

Finally, hard evidence is emerging that economic voting is now a reality. Analyzing the 2009 general election, Poonam Gupta and Arvind Panagariya found in a 2014 study that parliamentary candidates from ruling parties in states that exhibited better economic performance between 2004 and 2009 were significantly more likely to win election than candidates hailing from states which grew at a slower rate.²⁷ The authors found that 85 percent of the incumbent party’s candidates in “high-growth” states won reelection. In contrast, on average, candidates from state ruling parties in “medium-growth” and “low-growth” states won only 52 percent and 40 percent of the seats they contested, respectively.²⁸

Reconciling the Past and Present

These arguments, though compelling, stand in marked contrast to the finding that there has been no association between growth and electoral performance at the state level over the past several decades. But looking at the entire sweep of the past three decades obscures interesting variations beneath the surface. One way of reconciling this puzzle is to examine the relationship between growth and electoral outcomes by decade, rather than across the whole period, to explore whether it has been changing over time. Indeed, those who argue that economic assessments are shaping how Indian voters behave contend that this is a relatively recent phenomenon.

Only after disaggregating the relationship between economics and elections by decade does it become clear that a significant change has occurred in the period following the year 2000. Unlike in previous decades, the data reveal positive and significant electoral returns to governments able to generate faster growth rates in the 2000s (see figure 4). Statistical analyses, after controlling for a range of potentially confounding factors, reveal that a 1 percentage point improvement in a state’s growth rate in the 2000s is linked to a 7.5 percent increase in the likelihood that the incumbent party or alliance will be reelected, a 3.3 percentage point gain in seat share, and a positive (though statistically insignificant) rise in vote share.²⁹ To reiterate, there is no clear pattern of electoral rewards for higher growth in either the 1980s or 1990s; these effects are only evident in the most recent decade.

However, while the data show a clear relationship between economics and politics, there are several outliers. Certain states, such as Tamil Nadu, exhibit regular alternation between parties that is independent of the pace of economic growth. Nevertheless, on average the general shift toward rewarding or punishing incumbents on the basis of their economic performance is unmistakable.

Coming to 2014, scholars widely interpreted the BJP’s general election victory as conclusive evidence that good economics and good politics can proceed hand in hand. In the two years preceding the election, India’s economic growth slumped to below 5 percent, its worst showing in a quarter century. To compound matters, inflation remained elevated for a prolonged period, causing great stress to the finances of Indians across the socioeconomic spectrum.

Sensing an opportunity to capitalize on the economic success that his state of Gujarat had enjoyed, Modi focused extensively on the economy, making the rejuvenation of India’s growth story the centerpiece of his national political campaign at a time when the average Indian voter was seized with concerns over economic uncertainty. On the hustings, Modi’s campaign speeches were littered with a litany of pro-development catchphrases such as “sab ka saath, sab ka vikas” (together with all, development for all) and taglines like “maximum governance, minimum government.”

The strategy worked swimmingly; the available evidence suggests that the economy loomed large in the minds of Indian voters as they decided how to cast their ballots in the 2014 general election. An all-India preelection survey conducted in late 2013 by the Lok Foundation (for which I was a member of the research team) found that, among voters who expressed a preference, a plurality cited issues concerning the economy and development as among the most important factors shaping their vote choice (see figure 5).³⁰

What is intriguing about this finding is the surprising consistency of responses across both rural and urban areas. The discourse on Indian politics often paints a stark contrast between urban sections of the country, in which a more cosmopolitan citizenry emphasizes programmatic policies linked to governance and development, and rural areas, where voters’ decisions are supposedly driven by identity considerations and access to patronage rather than high-minded issues of economic growth. The Lok survey evidence suggests this divide may be somewhat artificial in today’s India. For both urban and rural residents, corruption, economic growth, and inflation were the three most pressing issues—although their relative importance varied at the margins.³¹

Voters’ preoccupation with the economy was further confirmed by a second survey, which allowed for open-ended responses and was carried out just a few days after voters had expressed their preferences at the ballot box. The 2014 National Election Study, conducted by the Center for the Study of Developing Societies (CSDS), also found that voters were highly motivated by the economy when casting their ballots, with 19 percent stating that price rise (inflation) was the most important election issue and nearly 11 percent citing the lack of development.³²

The evidence from state election results indicates that 2014 was an exclamation point on an existing trend rather than the start of something entirely novel. Indeed, the UPA government in New Delhi was arguably reelected in 2009 on the basis of the stellar growth record it amassed over the previous five years. Its failure during its second stint in office, some have argued, was to misinterpret this mandate as one in favor of greater welfare spending as opposed to economic growth.³³

While the data clearly support the contention that good economics can make for good politics in today’s India, what is less clear is why this shift has occurred and why in the past decade. This is an area where further research is necessary, but there are likely mutually reinforcing factors at work. Some have argued that the decade of the 2000s triggered a change in mind-set brought about by a new era of rapid growth, or what some have dubbed a “revolution in rising expectations.”³⁴ It is also possible that governments are doing a better job of ensuring that the effects of growth reach more segments of the population.

A second hypothesis has to do with the nature of political competition. After a politically tumultuous 1980s and 1990s, competition within most states began to stabilize around two parties or, in some cases, two opposing alliances, in the 2000s. Voters today have a much clearer picture of what the likely alternative ruling party or parties will look like and what programmatic priorities they espouse. This new format of party competition increases clarity of responsibility, thereby enhancing accountability.

A final hypothesis involves the simple passage of time. In the decades leading up to the 2000s, politics was dominated by identity considerations, as newly empowered social groups found greater political voice and began to challenge traditionally dominant castes and social groups by making popular appeals to social justice. Eventually, appeals made solely on the basis of social justice may have gradually lost their allure, as voters began linking social justice to larger issues of economic advancement.

Whatever the underlying drivers of the shift toward economic voting might be, the change in behavior appears positive from a governance perspective. After all, when voters hold their leaders accountable on the basis of actual performance criteria rather than narrow forms of parochialism alone, social welfare at large stands to gain.³⁵

The Arrested Rise of Regional Parties

The apparent rise of regional political parties is a recurring theme on the Indian political scene. Over the past two decades, India has witnessed a steep increase in the number of parties contesting parliamentary elections and a corresponding decline in the average margin of victory. Both of these data points are evidence of new power centers emerging from within the states. Indeed, aside from Modi and the duo atop the Congress Party, many of the most influential political leaders in India today hail from regional parties—such as Jayalalithaa Jayaram, the chief minister of Tamil Nadu; party leaders and onetime chief ministers of Uttar Pradesh Mulayam Singh Yadav and Mayawati; and the chief minister of West Bengal, Mamata Banerjee. Indeed the proliferation of new regional parties, whose appeal is often restricted to a single state or, even, a subregion of a single state, over the course of the past two decades has remade India’s electoral system.

But whether regional parties will be able to accumulate greater influence over governance in New Delhi and in India’s states remains an open question. By any measure, the 2014 election results were a repudiation of the conventional narrative that treats regional parties as constantly on the rise, acquiring greater political space. In fact, talk of the perpetual rise of regional parties is at odds with several recent empirical trends.

A Steady Balance of Power

The data from the past several election cycles reveal that there is a surprising degree of stability in the relative appeal of regional parties, notwithstanding the hype to the contrary (see figure 6). In national elections, the vote share earned by parties other than the BJP and the Congress has hovered around the 50 percent mark since 1996. The remaining share of the all-India vote, the other 50 percent, has gone to the Congress and the BJP. This balance of power suggests that while regional parties have been increasing in number, they appear to be dividing a roughly constant vote share among themselves, leaving the vote share going to national parties relatively untouched.

To be fair, it is an oversimplification to neatly partition vote share in this manner, given that national parties have regularly contested elections over the past several election cycles as part of broader alliance structures featuring multiple regional parties. Therefore, in any given state where either the Congress or the BJP has a local partner with which it has a seat-sharing arrangement, there is some vote share that would ordinarily go to its partner that gets transferred to it instead, and vice versa. In Maharashtra, for example, the BJP and its principal local ally, the Shiv Sena, have a seat-sharing arrangement in which the two divvy up constituencies based on relative strength and local realities. The BJP’s vote share in the state is thus an amalgamation of what it would have earned on its own plus what the Shiv Sena transfers to it—and the opposite is true for the Shiv Sena.³⁶

The give-and-take that transpires between national and regional parties as part and parcel of broader coalition realities is not new to Indian politics. What changed significantly in 2014 was not the distribution of votes between national and regional parties, but the distribution of the national party vote. Essentially, without greatly changing their aggregate vote share, the two national parties switched places in the five years between 2009 and 2014. The Congress, which racked up a 28.5 percent vote share in 2009, saw this drop to 19.5 percent five years later. Conversely, the BJP grabbed 31.3 percent of the vote in 2014, a sizeable increase from just 18.8 percent in the previous election and slightly higher than where the Congress stood five years prior.

The same trend appears to hold when one calculates the share of the vote that went to alliances in which the two parties participated. The BJP-led National Democratic Alliance (NDA) won 38.1 percent of the all-India vote in 2014 compared to just 23.3 percent for the Congress-led UPA.³⁷ In the 2009 polls, the NDA garnered 24.2 percent of the vote compared to UPA’s 36.5 percent.³⁸

This changing composition of the vote share, however, does not neatly translate into proportional changes in parliamentary seats. This is an artifact of India’s first-past-the-post electoral system, which is known to create disproportional outcomes due to its winner-take-all nature. For instance, in 1952 the Congress won 45 percent of the all-India vote but nearly three-quarters of seats in the Lok Sabha (74.4 percent). In 1977, when the party was thrown out of office following a nearly two-year period of emergency rule, 34.5 percent of the vote translated into a measly 28.4 percent of seats. In 2014, the Congress enjoyed its least efficient performance on record, winning 19.5 percent of the vote but ending up with only 8.1 percent of the seats. The BJP, on the other hand, enjoyed the opposite fate: its 31.3 percent of the vote translated into more than half (51.2 percent) of the seats in the lower house in 2014.

Do Regional Parties Undermine National Parties?

Given the explosive rise of regional parties and their rapidly increasing share of the vote in the 1980s and early 1990s, many assumed that their further growth was both inevitable and, by extension, detrimental to the fortunes of the national parties.

One reason that the pessimistic forecast for national players has not come to pass is that many regional parties threaten other regional parties as much as they do national parties. Again, India’s first-past-the-post electoral system—in which a party can win a constituency with a small minority of the total vote—lends itself to intense fragmentation, with a larger number of parties fighting for a fixed vote share in a given electoral contest. This is especially the case when there are more than two viable parties in the fray, which is a common occurrence in India (unlike in the United States, which has similar electoral rules). In 2014, for instance, only 36.8 percent of electoral districts were won with a majority of votes, which was actually a steep increase from 22.1 percent in 2009. The net result has often been that regional parties have crowded out rival regional parties. To some extent, fragmentation can be mitigated though coordination and preelectoral alliances, in which parties work out detailed seat-sharing arrangements to avoid cutting into one another’s votes. But the intense factionalism in Indian politics and underlying diversity of society place limits on this.

Indeed, the data from recent Lok Sabha elections point to the fact that while regional parties’ share of the vote has remained healthy, their share of seats has suffered (see figure 7). The share of the popular vote won by regional parties stood at 51.3 percent in 2004, hit an all-time high of 52.6 percent in 2009, and then dipped to 48.6 percent in 2014—on par with their vote performance in the 1998 elections. Their seat share, however, has had a very different trajectory. Regional parties held 47.9 percent of seats after the 2004 vote, but that dipped to 40 percent in 2014, even as their share of the vote increased, due to intense fragmentation. This suggests that the proliferation of regional parties may lead them to cannibalize the nonnational party vote share, rather than chip away at the power of national parties.

Another possible threat posed by regional parties was that, as their popularity grew, they would eventually morph into national parties themselves by virtue of their expanding reach. While it is true that several regional parties officially made that jump—although none can claim the extensive reach the Congress and BJP boast—it turns out this fear was overblown. Most regional parties have struggled mightily to translate their regional standing into a platform for success at a broader, pan-Indian level—despite being accorded “national party” status by the Election Commission on the basis of their rather limited presence in multiple states.³⁹

For instance, in the 2014 general election, Mayawati’s Bahujan Samaj Party (BSP) experienced an utter collapse. The BSP fielded candidates in 503 constituencies across India, more than any other party. Yet the party took home exactly zero seats. In fact, the BSP was not even a contender in the vast majority of constituencies in which it entered the fray; its candidates finished among the top two in just 36 constituencies. An astonishing 89.1 percent of the party’s candidates forfeited their deposits—a fate accorded to candidates who fail to win at least one-sixth of the constituency vote.

Contrast this with the BJP, which contested 428 seats, won 282, and finished second in another 54 seats around the country. The Congress bagged just 44 seats but finished second in another 224. Although this was a monumentally bad performance for the Congress, it still earned more votes than the next five largest vote-getting parties combined—BSP, Trinamool, SP, the Communist Party of India (Marxist), or CPI(M), and All India Anna Dravida Munnetra Kazhagam (AIADMK).

The two major leftist parties—the Communist Party of India and CPI(M)—which were once perceived as significant players in several states across the country, have also fallen on hard times (see figure 8). Together, the two parties held 9.8 percent of the all-India vote in 1971. The Left has failed to recapture those heights in the intervening years, stagnating at around 7 percent over the last few election cycles before dropping precipitously to 4.1 percent in 2014. Thanks to the regionally concentrated nature of its supporters, the Left had much better luck in terms of winning seats: in 2004, it won 53 seats in parliament—a record high (see figure 9). By 2009, however, that total had been more than cut in half (20 seats), and it was reduced by half once again in 2014 (10 seats).

The final regional-turned-national player, the Nationalist Congress Party (NCP), also experienced dimming fortunes in 2014. A primarily Maharashtra-based faction of the Congress Party, which then became a Congress coalition partner, the NCP witnessed its vote share slump from 2 to 1.6 percent (and its seat total dip from nine to six).

Ruling the Regions

Evaluating the performance of regional parties in national elections only tells half the story, because it is in India’s regions (that is, states) that these parties got their start—and from which they continue to derive much of their grassroots support. An examination of data from India’s states, however, underscores the real limits to the influence of regional forces.

As of mid-2015, chief ministers from national parties controlled the reins of power in more than half (54.8 percent) of all state assemblies—seventeen of 31 to be exact—while regional parties controlled fourteen states (45.2 percent). If one looks at the late 1990s, however, the numbers told a nearly opposite story; in 1999, which represented regional parties’ peak performance at the state level, they controlled nearly two-thirds (63 percent) of Indian states. Since that high-water mark, the share of states run by regional parties has fallen off considerably, although it rebounded somewhat in recent years (see figure 10).

Regional parties are ascendant in many of India’s most populous and, hence, electorally consequential states. Three states—Bihar, Uttar Pradesh, and West Bengal—are home to 400 million Indians, or nearly one-third of India’s entire population, and they are all run by regional parties. But even when judged by the population residing in states ruled by regional versus national parties, the latter still come out narrowly ahead: as of the end of 2014, a majority of Indians lived in places with a Congress or BJP chief minister.

Are Regional Parties Transforming Governance?

While the numbers demonstrate that regional parties may not be as dominant as some analysts expected, there are more qualitative arguments to consider. For instance, some observers have remarked that regional parties have been crucial to the innovations in governance that India has seen in recent years. To justify this claim, many point to the rise of a new cadre of state leaders—Biju Janata Dal’s Naveen Patnaik of Odisha or Nitish Kumar of Janata Dal (United) from Bihar and, now, Arvind Kejriwal of Delhi’s Aam Aadmi Party (AAP) are among the most commonly cited examples—who have demonstrated to their respective electorates that they are willing to experiment with fresh thinking on development and governance. Morgan Stanley economist Ruchir Sharma has often argued this case, writing that, “as a rising force, the regional parties represent hope: they are young, energetic, focused on economic development, and . . . in sync with the practical aspirations of the youthful majority.”⁴⁰

Sharma is on strong ground when he claims that regional parties have given politicians from national parties a run for their money, introducing a new dynamic of “competitive federalism” in which states vie with one another for capital, both financial and human. Coupled with a move in the post-1991 period toward greater decentralization, this new constellation of forces has no doubt introduced much greater variation in state-level policies.

Yet, there are two limitations to this thesis. The first is that regional parties are far from unique when it comes to providing their citizens with hope. While it is certainly true that there is a new crop of state chief ministers who have found ways to buck the underlying trend of anti-incumbency in Indian elections, a number of chief ministers hailing from the Congress and the BJP have also managed to connect with voters in ways that have been politically gratifying (see figure 11).

The second shortcoming of the thesis involves the large variation within the categories of regional and national party leaders. Indeed, for every reformist chief minister from a regional party, there is an Akhilesh Yadav, a regional leader who came to power with a historic mandate and a promise to break with the past—but who has largely failed to deliver on that pledge as chief executive of India’s most populous state.

From 2003 to 2007, the Samajwadi Party (under the leadership of Akhilesh Yadav’s father, Mulayam Singh Yadav) ruled Uttar Pradesh and along the way developed a less-than-stellar reputation for its poor governance. The SP’s stint in power—dubbed “Goonda Raj” (literally “thug rule”) by its opponents—was characterized by brazen corruption, a breakdown in law and order, and poor developmental outcomes. The party’s weak performance resulted in a decisive electoral defeat in 2007 at the hands of Mayawati’s BSP, which would govern the state for the next five years.

In 2012, at the ripe age of thirty-eight, Yadav rallied the SP back to power with a single-party majority—an impressive feat in a diverse state of 200 million. As the face of the party’s campaign and the projected chief ministerial candidate, Yadav garnered widespread praise for engineering a historic comeback.

In the wake of a lackluster performance by the BSP, voters had extremely high expectations for Yadav, underpinned by the implicit promise that the new leader, guided by a more modern outlook, Western education, and younger age, would reject the retrograde politics of his father and his party elders. But by mid-2015, Yadav had failed to offer concrete signs of a new era of transformational governance. To the contrary, the SP’s recent time in power has earned it many comparisons to the previous era of Goonda Raj, and the party was roundly trounced in the 2014 national elections. The BJP (with its ally, Apna Dal) won 73 of the state’s 80 parliamentary seats, leaving just five for the SP, a huge reversal from the 23 seats it had won in 2009.

Even the celebrated Nitish Kumar, the poster child of India’s next generation of progressive, reform-minded regional leaders, has come into hard times. During his first term in office, Kumar won praise for implementing sweeping reform in one of India’s poorest and historically worst-governed states. After breaking his partnership with the BJP in June 2013 due to his opposition to Modi’s prime ministerial candidacy, Kumar joined forces with his onetime nemesis, Lalu Prasad Yadav, in order to forestall the BJP’s rise. This shifted his political ethos from one focused on governance and development to one grounded in caste and communal calculations. The BJP badly routed the Kumar-Yadav partnership in the 2014 Lok Sabha elections, leading to Kumar’s resignation. Although he assumed the chief ministership of Bihar once more in February 2015, the shine has come off the Nitish Kumar brand.

The variation in the performance of regional parties amply demonstrates that there is nothing inherently transformative about regional party rule. And even reform-minded leaders have fallen into a trap that parties of all stripes in India have too often experienced. Most regional parties have not invested in long-term institution building within their own party organizations, which raises serious doubts about their ability to transform governance over the long term. Not only have few regional leaders invested in building sound institutional structures, but they have, to the contrary, actively worked to undermine whatever semblance of internal party democracy might have once existed. What this means in practice is that most critical decisions of policy and personnel are ceded to the supreme leader, with very limited information flowing upward (or back down). Party leaders take decisions and then make pronouncements.⁴¹

Any honest assessment of the health of democracy in India must recognize that the infirmities of state parties are often replicated when it comes to their national counterparts, the Congress and the BJP. Nevertheless, the situation facing regional parties is arguably more dire given that they rarely have a second-tier leadership beyond the party president.

When Tamil Nadu chief minister Jayalalithaa was convicted on corruption charges in September 2014, she was disqualified from holding public office. In her absence, the party lacked a cadre of second-rung leaders with the legitimacy or authority to fill in behind her. Although her conviction was thrown out in May 2015, setting the stage for her return as chief minister, during the entirety of the period when she was out of office, no major decisions of governance could be taken without first consulting her, introducing major challenges in what has traditionally been thought of as one of India’s better-run states. During her brief interregnum, Jayalalithaa’s office remained unoccupied, and her seat in the Tamil Nadu state assembly sat vacant—a testament to the dominance she exudes in the state.⁴²

Even India’s newest regional chieftain, Arvind Kejriwal of the AAP, has surprised observers by the extent to which he has taken steps to sideline internal party dissent. Prior to taking office, Kejriwal preached the gospel of internal party democracy, transparency, and accountability. But after the party’s resounding electoral victory in Delhi’s February 2015 state elections, Kejriwal was made chief minister and quickly moved to consolidate his position within the party, eventually expelling two of its leading intellectuals who had questioned many of the party’s decisions leading up to the election.

A New Chapter?

The regional revolution is arguably one of the most important and interesting developments to have occurred over the last six and a half decades of Indian electoral politics. Today, regional parties command significant influence, power, and financial resources; however, their rise should not be overstated.

The electoral success of the BJP in the 2014 general election holds out the distinct possibility that the pendulum will swing back toward national parties in India’s states, although the setback suffered by the BJP in the February 2015 Delhi assembly elections at the hands of the AAP marked a formal end to Modi’s honeymoon period. Indeed, with the Congress in decline, there is a new opportunity for aspirant regional players such as the AAP to provide a pan-Indian alternative to the BJP. But even if regional parties can recover, their rise is certainly not steady or irreversible.⁴³

Popular Mandate for Dynastic Politicians

Another oft-discussed attribute of India’s electoral politics is the continued dominance of political families or dynasties. In particular, a great deal of attention has been focused on one of the most storied political dynasties anywhere in the world, the Nehru-Gandhi family. The dynasty reaches back nearly a century, beginning with Motilal Nehru, a lawyer and activist in the Indian nationalist movement who twice served as president of the Congress Party, from 1919 to 1920 and again from 1928 to 1929. For thirty-seven of the nearly sixty-eight years that India has been independent, the prime minister has come from this family: Jawaharlal Nehru, from 1947 to 1964; daughter Indira Gandhi, from 1966 to 1977 and 1980 to 1984; and her son, Rajiv Gandhi, from 1984 to 1989. A Nehru-Gandhi has not been prime minister in subsequent years, but Rajiv’s widow, Sonia, has been Congress Party president since 1998 and was an influential power center during Congress rule from 2004 to 2014. Her son, party vice president Rahul Gandhi, is the heir apparent to the family business.

Dynastic politics is not unique to India. According to a 2000 study, 8.7 percent of sitting U.S. legislators had a relative who had also served in Congress, while the share reached roughly 9 percent in the United Kingdom in 2014.⁴⁴ In Japan in the late 2000s, some estimates placed the share of such “legacy” members of parliament as high as 20 percent, and many smaller democracies ranging from Iceland to Ireland report significant numbers of dynastic politicians.⁴⁵

But the issue is of particular interest in the Indian context because dynastic politics prevails despite widespread anti-incumbency and electoral volatility. Unlike the United States, for instance, where incumbents enjoy a massive advantage, India’s elected politicians are at least as likely to be voted out of office as they are to be voted back in.⁴⁶ Furthermore, due to the long-standing dominance of the Congress Party, dynastic politics was front and center in the political scene for much of the country’s postindependence era.

Many observers interpreted the 2014 electoral verdict as a repudiation of dynastic politics. Modi himself argued that a BJP victory in 2014 would spell the end of this scourge in India⁴⁷ Indeed, Finance Minister Arun Jaitley declared in January 2015 that the BJP’s general election victory “has put an end to dynasty politics” in India.⁴⁸

The BJP leaders’ rhetoric aside, the reality is far more complicated. The 2014 elections certainly delivered a blow to the Nehru-Gandhi dynasty writ large. But both serving dynasts, Rahul and Sonia Gandhi, won their seats in parliament, and the underlying tendency toward dynastic politicians does not show clear signs of easing.

Widespread Dynasticism

For starters, and contrary to the pronouncements of Modi and Jaitley, the practice of dynastic politics is hardly the monopoly of the Congress Party (see figure 12). Data collected by Kanchan Chandra reveal that 20 percent of Lok Sabha members of parliament elected in 2004 came from political families. This number rose even further, to 29 percent, in the 2009 national election. In 2014, the supposed antidynasty election, the share of parliamentarians hailing from political families stood around 21 percent—a decline, but by no means a wipeout.⁴⁹

Data from the 2014 election clearly indicate that parties across the political spectrum engage in dynastic politics (see figure 13). Looking at parties with five or more members of parliament and the share of those delegations that belongs to political families, the Samajwadi Party tops the list; all five of its parliamentarians are members of the same family. Mulayam Singh Yadav, the president of the party and a member of parliament from Azamgarh, is the family patriarch. In the Lok Sabha he is joined by two of his nephews, his daughter-in-law, and a grandnephew. The list does not stop there: Mulayam’s son, Akhilesh, is the chief minister of Uttar Pradesh and his younger brother is a cabinet minister. Mulayam’s cousin also represents the SP in the Rajya Sabha, India’s upper house of parliament.⁵⁰ The Congress actually is located in the middle of the pack; roughly 41 percent of its members of parliament are dynasts. The BJP, on the other hand, is near the lower end of the spectrum with 16.7 percent of its delegation from political families.

Stepping back from counting individuals who hail from political families, many parties themselves remain the province of a single political family. This pattern is discernible among families in many states representing a wide diversity of parties, including the Abdullahs (Jammu & Kashmir, Jammu & Kashmir National Conference), Badals (Punjab, Shiromani Akali Dal), Karunanidhis (Tamil Nadu, Dravida Munnetra Kazhagam), Hoodas (Haryana, INC), Paswans (Bihar, Lok Jan Shakti Party), Patnaiks (Odisha, Biju Janata Dal), Pawars (Maharashtra, NCP), Reddys (Seemandhra, YSR Congress Party), Scindias (Rajasthan/Madhya Pradesh, INC/BJP), Thackerays (Maharashtra, Shiv Sena), and Yadavs (Bihar, Rashtriya Janata Dal).

Sources of Support

But what accounts for the continued political support for hereditary politicians? There seems to be something of a paradox: the conventional wisdom is that dynastic politics is unpopular, yet individual dynastic politicians are clearly popular enough to get elected and reelected.

Curiously, the literature on dynastic politics has not empirically evaluated the appeal of family-backed politicians, but there are several theoretical reasons why voters might support them. For instance, dynastic politicians might thrive in elections simply because of enhanced name recognition; candidates who have the same last name as popular incumbents have a natural advantage during campaigns. A second possible explanation is that dynastic politicians are simply better at politics given their genetic makeup and family experience. A third, related though distinct, possibility holds that dynastic politicians excel because they have greater exposure to politics, perhaps grounded in political networks and connections. A final hypothesis focuses on an enhanced capacity to deliver services to constituents. These lines of argument are not mutually exclusive, but they do have unique points of emphasis.

To test these various theoretical claims, the Lok Foundation 2013 preelection survey asked respondents whether, in the upcoming 2014 parliamentary election, they would prefer voting for a candidate who comes from a political family. In an election in which the dynastic control of the incumbent Congress Party became a major election issue, 46 percent of respondents answered in the affirmative.

Respondents who preferred dynastic candidates were asked why and presented with three options to explain their support. Forty-five percent said they preferred dynastic candidates due to the fact that they are better at doing politics “because it is their occupation,” which suggests that the practice of politics is seen as hardwired into their DNA. Forty percent answered that they preferred candidates from political families because they are likely to succeed due to their greater exposure to politics. This suggests that Indians believe family connections give politicians a leg up due both to nature (they are inherently better at politics) as well as to nurture (they are more exposed to politics during their formative years). Only 15 percent said they support hereditary politicians because their family ties make it easier to deliver services to constituents.

This finding may help explain why dynasticism is especially prevalent among younger politicians. Data from the Hindu newspaper on the age breakdown of members of parliament with family connections show that the younger the parliamentarian, the more likely he or she is to be from a political family (see figure 14).⁵¹ In 2014, for instance, 81 percent of members of parliament aged thirty or below and half of lawmakers between thirty and thirty-nine were dynasts. This share generally decreased as the age of members of parliament increased.⁵²

Implications

The 2014 result does not validate the hypothesis that Indian voters are fed up with dynastic politicians. While the overall share of politicians from political families did decline between 2009 and 2014, it is still on a par with the levels witnessed in 2004. Furthermore, given the fact that dynasticism is most prevalent among younger members of parliament, the arrival of a new generation of politicians might militate against any decline.

Even for the ascendant BJP, which exhibits a relatively lower share of dynastic politicians, dynasticism is more prevalent among younger members of parliament. The average age of the party’s dynastic parliamentarians is forty-eight, compared to fifty-five for the party as a whole; and 50 percent of the party’s members of parliament under the age of forty have a family connection in politics, compared to just 12.5 percent for those above forty years old.

The well-known lack of internal party democracy in India, if left untouched, will likely continue to reproduce dynastic politics. Political parties remain largely top-down affairs, and it can be difficult for those without connections to the party elite to get their foot in the door. Hence, as long as there are significant barriers to entry in India’s leading political parties, one should not expect drastic change on this front. But the issue is not solely one of supply; there is also a demand-side logic from the perspective of the Indian voter. Insofar as politics is viewed as an occupation whose skills are passed down through family channels, it is difficult to imagine that family politics will disappear from the scene.

Support for Candidates Tied to Crime

One of the most interesting—and confounding—developments in India’s domestic politics in recent memory has been the large number of politicians seeking, and winning, elected office who are under criminal scrutiny. Since 2003, thanks to the intervention of the Supreme Court and the oversight of the Election Commission of India, candidates seeking election to state and national office are required to publicly furnish details of their criminal, educational, and financial records. One of the underlying premises motivating this shift was the belief that voters, once exposed to credible information about the attributes of political candidates, might withdraw popular support from those who are seen as tainted.⁵³

In the 2014 elections, more attention than ever before—on the part of the media, civil society, and concerned activists—was devoted to publicizing the biographical details of parliamentary aspirants. In fact, anyone with a mobile phone could send a text message to a service set up by a good governance watchdog group and instantly receive details on the personal characteristics of candidates competing in their constituency. Notwithstanding this publicity, a record number of candidates involved in criminal cases, including those facing charges considered serious, won election to the Lok Sabha.

The remarkable staying power of this phenomenon, not to mention its growth over time, raises the question of why—with the biographical profiles of all candidates readily available for the past decade—the number of criminally suspect legislators has gone up, rather than down. The short answer is that a lack of information is not the primary driver of voter support for politicians suspected of criminal activity; rather, as the 2014 election results demonstrate, there is an underlying demand for politicians who can get things done, even if (indeed, especially if) they are connected with wrongdoing.

An Entrenched Dynamic

The presence of large numbers of parliamentarians facing criminal cases is not a new development, but there is evidence to suggest the problem is getting worse rather than better (see figure 15). In 2004, the first year for which candidates were required to submit disclosures, 24 percent of members of parliament declared that they faced pending criminal cases at the time of their election. To be fair, these were cases—rather than convictions—but they were more than simple allegations. Under the disclosure regime, candidates are only required to report cases in which a judicial process has already commenced. Keeping this in mind, what is striking is the degree of alleged serious crime. In 2004, 12 percent of members of parliament reported that they were facing serious cases before the courts—that is, cases that involve charges such as murder, attempted murder, kidnapping, and various crimes against women. Over the next ten years, those numbers steadily increased to the point where one-third (34 percent) of all members of parliament in 2014 declared that they were involved in criminal cases, including 21 percent who disclosed ongoing cases of a serious nature.⁵⁴

The situation among India’s 4,120 state legislators is roughly similar, although it often takes a backseat to the state of affairs in the Lok Sabha. As of 2014, about one-third of state assembly representatives (31 percent) faced at least one criminal case. Again about half of those, or roughly 15 percent, reported cases in which they stood accused of committing serious crimes.⁵⁵ Very little is known about the extent of suspected criminality in local-level politics because there has been no systematic analysis of the situation in panchayat (village governments) and urban local governing bodies. However, there is some evidence to suggest that even these local tiers of governance are not free of politicians linked to crime. Data compiled by the Association for Democratic Reforms show that 17 percent and 21 percent of municipal corporators in Mumbai and Delhi, respectively, declared involvement in criminal cases.⁵⁶ Similarly, 16 percent of winners from Navi Mumbai’s 2015 municipal elections faced pending criminal cases at the time of their nomination.⁵⁷

Credibility or Ignorance?

Broadly speaking, there are two views regarding the proclivity of Indian voters to support politicians involved with serious criminal cases.⁵⁸ One draws its inspiration from a venerable political economy literature, which argues that limited information can often hinder citizens’ ability to hold their governments accountable.⁵⁹ Hence, if it is the case that voters in India lack access to credible sources of information about politicians, they will struggle to distinguish between “low-quality” and “high-quality” candidates.⁶⁰ According to this logic, voters may not willingly support politicians linked with criminality but might do so out of ignorance. Once presented with credible information, however, voters would change their minds and switch their support to “cleaner” politicians. Indeed, this is the implicit—or, in some cases, explicit—view many good government groups adopt in arguing, to borrow an oft-quoted phrase, that “sunlight is the best disinfectant.”

An alternative narrative suggests a very different explanation centered on the credibility of politicians with criminal reputations. According to this line of thought, voters may possess information about the biographical details of politicians, but still choose to support them because they are perceived to be effective representatives who are capable of “getting things done.”⁶¹ Candidates with criminal reputations often campaign or defend their activities precisely on these grounds. As one member of parliament from the state of Maharashtra who faced at least fourteen criminal cases winding their way through the courts explained in 2014: “The cases against me pertain to issues related to the public for whom I have been fighting all along. One case is of assaulting an engineer, which was over a delay in flyover work. It was not a personal issue but for the people.”⁶²

Research from India has shown that this link between credibility and criminality is typically mediated by identity or ethnic politics.⁶³ In other words, candidates involved in criminal cases often manipulate the deep social divisions that are prevalent in Indian society, using them as a cleavage along which they can mobilize support and deliver constituent services. If the latter view accurately characterizes the reality on the ground, it could help explain why Indian voters—seemingly outraged about malfeasance and misgovernance—would at the same time back criminally suspect politicians in such great numbers.

The Lok preelection survey provides an opportunity to adjudicate between these two competing explanations. The survey asked respondents a simple, direct question: “Would you vote for a candidate who delivers benefits to you even if s/he faces serious criminal cases?” In response, 26 percent said that they would vote for a candidate who gets things done but also faces serious pending criminal cases. In other words, one out of four Indians surveyed was willing to openly admit that he/she would vote for a candidate who is involved in a criminal case but is perceived to be an effective representative for the constituency. This seems to support the credibility hypothesis; voters can be well informed and support candidates with criminal reputations.

However, the credibility thesis is about more than getting things done. To test the latter dimension of the hypothesis—that credibility is tied to the salience of ethnic or identity considerations—one can link respondents’ views on criminality to another question from the Lok survey on caste sentiment. The survey asked respondents: “Is it important to you that a [Upper Caste/Other Backward Class (OBC)/Scheduled Caste (SC)/your community] candidate wins the election in your constituency?”⁶⁴ The caste group mentioned in the prompt varied according to the identity of the respondent. For example, if a respondent belonged to an upper-caste community, he/she would be shown a prompt for upper caste.

The survey found that 46 percent of respondents indicated that it was important that a candidate from their own community won the election. This is a fairly direct measure of positive ethnic bias or co-ethnic affinity. To bring these two pieces together, a 2015 study I co-authored with Neelanjan Sircar looks at the relationship between the criminality and positive ethnic bias questions (see figure 16).⁶⁵ Our preliminary findings suggest that there is a very strong connection between the percentage of respondents in a given state who would support candidates who are competent but face serious criminal cases and the share of those who favor a candidate from their own community when making political choices. This fits very well with the credibility hypothesis.

Good Governance Election

The 2014 polls were widely heralded as India’s “good governance” election. Modi ran explicitly on this theme, delivering repeated assaults on the misgovernance and malfeasance of the incumbent UPA government while continually touting the “Gujarat model” of development, and the stewardship of the economy and the administration of the nation’s affairs were clearly key drivers of the Modi wave. How does one reconcile that development with the seemingly contradictory support for candidates who are involved in criminal cases? The same election that produced a wave for Modi also saw the election of 281 other members of parliament from the BJP, of whom 35 percent faced criminal cases and 22 percent faced serious cases.⁶⁶

For starters, Indian voters might value leaders who are honest, upstanding, and effective, but at the end of the day they are selecting local candidates who will be charged with looking after their interests. Given the realities of state weakness and the real and perceived distance between the state or national capital and the places where most Indians actually reside, within their own constituencies voters might be inclined to put their faith in someone who is willing to use whatever means are at his disposal to get things done.

Secondly, voters are not rewarding politicians with criminal reputations because they have an interest in rewarding misgovernance; rather, they are doing so precisely because they perceive, rightly or wrongly, that such candidates can actually use their no-holds-barred reputations and skills to provide the very governance that has been absent. Until the quality of the state improves or “clean” politicians can convincingly demonstrate that they can deliver, even well-informed voters might have good reasons to seek the assistance of candidates who try to pass off their criminal records as signs of their competence.⁶⁷

Messy Realities of Ethnic Voting

A final notion about Indian politics that is worthy of closer scrutiny is the proposition that Indian voters do not cast their vote so much as they vote their caste. The Lok survey data on positive ethnic bias plainly demonstrate that the salience of identity politics remains high in India. In addition to providing evidence of positive ethnic bias, the answers to a related question from the Lok survey also uncovered high levels of negative ethnic bias—that is, the extent to which voters discriminate against candidates from other communities. Survey participants were asked whether they would be troubled if a candidate from a community other than their own won the election; 38 percent of respondents answered in the affirmative.

Considering the data on negative and positive ethnic bias together, it is hard to conclude that the conventional understanding of caste is wildly off base. But recent research points to at least two possibilities. The first is that the conventional understanding of the role caste plays in shaping political behavior is oversimplified and the relationship is much more nuanced than is commonly recognized. The second is that the dynamics of caste are changing in a way that is diminishing its influence. There are signs pointing in both directions, and support for these initial inclinations in the 2014 election results.

Co-Ethnic Voting

One difficulty with the simple vote-your-caste logic is that it is not clear that voters do regularly and literally vote their caste. Survey evidence I collected from Bihar in conjunction with the Center for the Study of Developing Societies illustrates this phenomenon. Just days after Bihar’s 2010 state assembly election, we surveyed more than 2,000 voters about their voting behavior. Survey enumerators conducted a simulated poll to discern respondents’ stated vote choice. The research team separately collected data on the caste affiliations of the candidates in order to determine the extent of co-ethnic voting. We found that only 14.2 percent of respondents voted for a candidate who was from the same jati, or specific caste category(see figure 17).⁶⁸ Jati is a pretty restrictive category, but even if one looks at the rate at which voters said they had selected candidates from their own umbrella caste group (e.g., Upper Caste, OBC, SC), it is just under 40 percent. This finding is especially surprising in a state like Bihar, where caste conflict and identity-based mobilization are ingrained in the history of electoral politics.⁶⁹

One reason the rates of voting for someone of the same caste category are so low could be the fact that voters often do not have the opportunity to vote for a co-ethnic because one is not on the ballot. Given the multiplicity of jatis in any given constituency, this is not a trivial consideration. In the absence of a co-ethnic candidate, the relevance of ethnicity is not negated but instead changes form. For instance, ethnicity could still be a salient consideration if voters engage in negative voting, or voting against a specific group, or even if they vote in favor a given group that might be considered an ally.⁷⁰

A second possibility is that voters value the ethnic “brand” of a given party label, not that of the individual candidate. That is, a low-caste Dalit voter might vote for the pro-Dalit BSP not because the BSP is running a Dalit candidate, but because, ideologically, the BSP is perceived to be a party dedicated to the cause of Dalit advancement. While this is a distinct possibility to be considered, it is also the case that the ethnic brands of many parties are being diluted in the current era of Indian politics, in which many parties or alliances are winning election on the backs of rainbow coalitions of diverse communities. In the 2010 Bihar elections, for instance, postelection data collected by the CSDS indicate that the victorious National Democratic Alliance—a coalition of the BJP and JD(U)—cobbled together an impressive, broad-ranging collection of voters (see table 1).⁷¹ The NDA earned a majority, or near majority, of votes from several groups across the social hierarchy, from upper-caste Brahmins and Rajputs to backward-caste Kurmi-Koeris, other SCs, and non-Muslim religious minorities.

Analyses of the BJP’s performance in the Hindi heartland of north India in the 2014 general election, drawing on CSDS postpoll survey data, suggest that it too cobbled together disparate sources of support across the social spectrum—the glaring exception being the minority Muslim community (see figure 18). In caste-conscious Uttar Pradesh, for instance, the CSDS postelection poll suggests that the BJP won 45 percent of the non-Jatav Dalit vote (the Jatavs being the core Dalit supporters of the rival BSP).⁷² Similarly, in neighboring Bihar, the BJP performed very well among sections of the OBCs that had previously voted for the BJP’s erstwhile ally in the state, the Janata Dal (United).

One obvious driver of these results was a polarization of Hindu-Muslim votes due to the Hindu majoritarian rhetoric of the BJP and the Sangh Parivar, a collection of Hindu nationalist groups, and to their tactics of mobilizing along communal lines. Clearly this sort of polarization was an important aspect of the BJP’s election campaign, even if it was not necessarily its cornerstone on a pan-Indian basis. It is very hard to quantify the precise impact of this approach, however, because election surveys are a blunt instrument with which to detect such effects (i.e., many voters might not willingly disclose the religious or communal motivations underlying their vote choice). But even assuming religious polarization was an important driver of Hindu vote consolidation, the dilution of clear caste-based voting is still striking given that north India is the region of the country where scholars believe caste calculations are most salient.

Clearly, the salience of caste as a useful social category for political mobilization is by no means dead. However, there are incipient signs that the influence caste once commanded may be waning. In the 2014 election, there was a clear divergence in the performance of key regional parties that can be tied to caste. As K. K. Kailash has argued, “regionally located” parties performed far worse than “regionalist” parties, the key difference being that the former are typically motivated by underlying caste and community considerations while the latter represent the interests of a particular state, regional culture, or language.⁷³ Indeed, the top performing regional parties in 2014—AIADMK, Trinamool Congress, the Biju Janata Dal (BJD)—which collectively won 91 seats, all conform to the “regionalist” conception.

Caste Identification

The 2010 Bihar survey, which found relatively low levels of co-ethnic voting, also points to a second puzzle. The survey asked respondents whether they could identify the caste (jati) of the candidates for the state legislature that they had voted for just a few days earlier.⁷⁴ Interestingly, 29.2 percent (or just under one-third) misidentified the jati of their preferred candidate. To be fair, when voters got it wrong, they were not very far off base. Closer to 90 percent managed to place their candidate within the correct umbrella caste grouping (Upper Caste, OBC, etc.). But the data reveal considerable structure to the errors. For instance, voters who are co-ethnics (i.e., they share the same jati as the candidate) did markedly better, while those with higher levels of education actually did less well at making accurate identifications.

Curiously, Muslims are an outlier when it comes to identifiability. Based on the results, Muslim candidates are much more readily identifiable than non-Muslim candidates. To some extent, this is not a surprise, as a candidate’s last name alone might be enough to signal his or her religious identity. What is more curious is that Muslim voters are much better than non-Muslim voters at identifying candidates’ ethnic backgrounds. This finding holds even when Muslims are voting for non-Muslim candidates, which is the case the majority of the time. What these results suggest, therefore, is that voters’ ability to readily identify the ethnic or caste affiliation of politicians varies far more than what has come to be expected.

There is some evidence to suggest that errors may flow the other way as well. Surveys conducted by Mark Schneider in Rajasthan in 2013 found that village-level politicians performed relatively poorly when asked how actual voters in their respective panchayat would vote.⁷⁵ In an experimental study, sarpanch (village presidents) correctly guessed the way villagers said they had voted in the previous election less than two-thirds of the time (65.5 percent); this was 1 percentage point better than their predictions of voters’ intentions in the next election. While village politicians performed better than random chance (50 percent, or a coin toss), they fared no better than the results that could have been achieved by using a simple benchmark based on highly visible demographic criteria or guesses informed by publicly available election survey data.

A Mixed Assessment

No one argues that ethnic identity, namely caste, does not play a role in shaping voter behavior. However, recent empirical evidence and data drawn from election surveys suggest that there is considerably more nuance than conventional identity-based models of voting behavior often convey. How then does one square this circle? There are at least two possibilities.

First, it is plausible that social bias, however deeply felt, does not necessarily translate into political choices. This could be, for instance, because voters hold certain partisan attachments that transcend caste. A BJP supporter could vote BJP even if it goes against her narrow sectarian interest. Or perhaps voters are often making a negative decision on Election Day, choosing to vote against a candidate of a particular group rather than in favor of someone.

A second possibility is that identity does matter, but it matters in conjunction with other factors such as the economy. As Neelanjan Sircar has pointed out, identity is relatively fixed but there is a significant degree of electoral volatility in India.⁷⁶ This does not mean that ethnic voting is not operative; indeed, one does see a strong relationship, for instance, between identity groups and support for either NDA or UPA in the 2014 election. This relationship in 2014, however, looked very different from how it appeared in 2009. For example, the increase in OBC support for the NDA between 2009 and 2014 might have been driven by factors that went well beyond identity alone, such as economic performance or the perceived leadership advantage of the Modi-led BJP.

Conclusion

When it comes to the Indian voter in 2015, recent evidence presents a mixed picture.

On the one hand, it is clear that there are important changes afoot, necessitating revision, if not full-scale replacement, of many commonly held notions. While the 2014 election was a watershed in terms of the outcome it produced, many of the changes it appeared to herald were actually in the works well in advance of the landmark polls. The empirical evidence suggests that nuanced changes in the nature of economic and ethnic voting had already been under way. What the 2014 election may have accomplished was to crystallize these changes.

On the other hand, the case for change has been overhyped when one considers other important dimensions. For example, while the Indian voter in 2014 did seem mobilized against the dynastically led Congress Party, there is no evidence of a broader backlash against hereditary politicians. And even this sentiment had its limits; after all, both Rahul and Sonia Gandhi handily won reelection to their seats in parliament. Two other members of the extended family, Maneka Gandhi and her son, Varun, won reelection on BJP tickets. With regard to preferences for politicians implicated in criminal activity, the 2014 result shows the nexus of crime and politics is more entrenched than ever. And the arrested rise of regional parties demonstrates that, although they remain key power brokers in Indian politics, there is nothing linear or preordained about their rise.

More broadly, the 2014 election reveals much about the state of India’s democracy.

First, while voters may harbor deep-seated social biases, political appeals to today’s Indian voter based on identity grounds rarely seem effective on their own. In 2014, the BJP managed to forge a genuine cross-caste political coalition that proved to be electoral dynamite.⁷⁷ Skeptics rightly point out that such intercaste bonhomie was only possible because of the BJP’s appeals to Hindu nationalism, which polarized the votes of Hindus (in favor of the BJP) and Muslims (largely against the party, given that 8.5 percent of this community voted BJP). While Hindu nationalism did have a role to play in the campaign, it alone was not powerful enough to create a pan-Indian appeal for the BJP. This religious solidarity had to be paired with a plank of governance and development.

While the motivations of voters might be shifting, the makeup of the candidate pool they have to choose from is not.

On this score, it is encouraging to see voters demonstrating an increased propensity to consider broad indicators of their government’s performance in managing the economy when making their electoral calculations. Viewed narrowly through the prism of governance and democratic accountability, this is a positive development. For the moment, identity-based concerns and economic or programmatic evaluations are both in play; it is impossible to conclude that one has clearly overtaken the other. However, it is notable that politicians who seek to gain strength on the basis of identity-based appeals alone have generally experienced hard times. The politicians who have fared better have found credible ways of marrying traditional appeals on the basis of identity with a forward-looking, aspirational agenda.

Second, it is striking that while the motivations of voters might be shifting, the makeup of the candidate pool they have to choose from is not. In some sense, voters have more choice than ever before, as evidenced by the increase in the absolute number of parties contesting elections. Yet, there is little qualitative change in the nature of candidates themselves. Dynastic politicians and those with criminal records remain well ensconced in state and national politics. Furthermore, politics continues to be a heavily male-dominated affair. In 1957, the first general election for which data on gender are available, only 3 percent of parliamentary candidates and 4.5 percent of winners were women. While those numbers have gradually increased—in 2014 women accounted for 8.1 percent of the overall candidate pool and 11.2 percent of winners—women remain massively underrepresented in parliament relative to their share of the general population.

Finally, while the overall balance of power between regional and national parties seems to have reached a steady equilibrium, the relative fortunes of the Congress and the BJP have seen significant shifts. For the first four decades of the postindependence period, the Congress was the pole around which Indian politics was organized. That designation was up for grabs following the landmark election of 1989, which marked the dawn of the era of coalition politics in Delhi, when there was no clear center of gravity in national politics.

By virtue of its performance in the 2014 general election and state assembly elections held before and after the national poll, the BJP has filled that vacuum—and its electoral gains since 2013 have come at the expense of the Congress, which followed up its worst-ever showing in a general election by delivering poor performances in several state elections. The continued decline of the Congress could mean that the BJP will lack a true national-level competitor within a few years. While the designation of India’s dominant political formation has now been passed on, it is difficult to foresee whether this shift is temporary or will be sustained over time.

In some ways, from the BJP’s perspective, this changing of the guard is both a blessing and a curse.⁷⁸ In states where the BJP and Congress are the only game in town, the Congress Party’s loss will mechanically be the BJP’s gain. But the situation is more complex in states where the BJP and the Congress are vying for space with one or more regional parties. In these states, such as Bihar and West Bengal, a rapid decline of the Congress could lead to the consolidation of an anti-BJP front, especially if the Congress forsakes contesting elections on its own and develops new regional allies. As the BJP might soon learn, this is one of the unfortunate consequences of being the pole around which politics is organized.

***
Some of the findings discussed in this paper are a result of joint work with Devesh Kapur, Megan Reed, and Neelanjan Sircar of the Center for the Advanced Study of India at the University of Pennsylvania, Danielle Smogard, and Reedy Swanson of the University of Virginia Law School. I am grateful to them for their collaboration. Kapur, Reed, Sircar, and I also collaborated with the Lok Foundation on the 2013 preelection survey cited here. Several sections of this paper draw on material from articles first written for the Carnegie Endowment’s India Decides 2014 initiative. I would like to thank seminar participants at Harvard, the University of Pennsylvania, Yale-National University of Singapore, and Yale University for comments. William Hayes provided excellent research assistance. I am grateful to Rebecca White and Holly Yeager for editorial comments. All errors are my own.

Notes

¹ Government of India, Press Information Bureau, “2014 General Elections,” http://pib.nic.in/elections2014/elemain.aspx.

² Sruthi Gottipati and Rajesh Kumar Singh, “India Set to Challenge U.S. for Election-Spending Record,” Reuters, March 9, 2014, http://in.reuters.com/article/2014/03/09/india-election-spending-idINDEEA2804B20140309.

³ E. Sridharan, “India’s Watershed Vote: Behind Modi’s Victory,” Journal of Democracy 25, no. 4 (2014): 20–33.

⁴ Milan Vaishnav and Danielle Smogard, “A New Era in Indian Politics?” Carnegie Endowment for International Peace, June 10, 2014, http://carnegieendowment.org/2014/06/10/new-era-in-indian-politics.

⁵ Center for the Study of Developing Societies, India National Election Study 2009 (New Delhi: CSDS, 2009).

⁶ Milan Vaishnav, “The Modi Debate Worth Having in India,” Carnegie Endowment for International Peace, October 9, 2013, http://carnegieendowment.org/2013/10/09/modi-debate-worth-having-in-india.

⁷ Yogendra Yadav, “Electoral Politics in the Time of Change: India’s Third Electoral System, 1989–99,” Economic and Political Weekly 34, no. 34/35 (August/September 1999): 2393–99.

⁸ Kanchan Chandra, Why Ethnic Parties Succeed: Patronage and Ethnic Head Counts in India (New York: Cambridge University Press, 2004).

⁹ Yogendra Yadav, “Understanding the Second Democratic Upsurge,” in Transforming India: Social and Political Dynamics of Democracy, ed. Francine Frankel, Zoya Hasan Rajeev Bhargava, and Balveer Arora (New Delhi: Oxford University Press, 2002).

¹⁰ Pratap Bhanu Mehta, “Reform Political Parties First,” Seminar 497 (January 2001), www.india-seminar.com/2001/497/497%20pratap%20bhanu%20mehta.htm.

¹¹ Soutik Biswas, “Is India’s Politics Becoming Less Dynastic?” BBC News, July 27, 2014, www.bbc.com/news/world-asia-india-28478544.

¹² Abhijit V. Banerjee, Donald Green, Jeffrey McManus, and Rohini Pande, “Are Poor Voters Indifferent to Whether Elected Leaders are Criminal or Corrupt? A Vignette Experiment in Rural India,” mimeo, Massachusetts Institute of Technology, 2012.

¹³ Christophe Jaffrelot, “Caste and Political Parties in India: Do Indians Vote Their Caste—While Casting Their Vote?” in Routledge Handbook of Indian Politics, ed. Atul Kohli and Prerna Singh (London: Routledge, 2012).

¹⁴ Michael S. Lewis-Beck and Mary Stegmaier, “Economic Determinants of Electoral Outcomes,” Annual Review of Political Science 3 (2000): 183–219.

¹⁵ Christopher J. Anderson, “The End of Economic Voting? Contingency Dilemmas and the Limits of Democratic Accountability,” Annual Review of Political Science 10 (2007): 271–96.

¹⁶ Ibid.

¹⁷ “Learning From the Pioneer,” Economist, May 30, 2015.

¹⁸ McKinsey and Company, “Andhra Pradesh Vision 2020,” 1999, www.scribd.com/doc/123915459/McKinsey-Andhra-Pradesh-Vision-2020-1999#scribd.

¹⁹ Dehejia and Subramanya argue that the “most plausible explanation for TDP’s [Naidu’s Telugu Desam Party] defeat in 2004 is that it was a victim of circumstance and of its own hype.” See Vivek Dehejia and Rupa Subramanya, “Why No Centre-Right Political Party in India Today?” Business Standard, February 16, 2014, www.business-standard.com/article/opinion/vivek-dehejia-rupa-subramanya-why-no-centre-right-political-party-in-india-today-114021600697_1.html.

²⁰ K. C. Suri, “Andhra Pradesh: Fall of the CEO in Arena of Democracy,” Economic and Political Weekly 39, no. 51 (December 2004): 5404–11.

²¹ Milan Vaishnav and Reedy Swanson, “Does Good Economics Make for Good Politics? Evidence From Indian States,” India Review 14, no. 3 (2015).

²² Ibid.

²³ Arvind Subramanian, “The India Vote,” Wall Street Journal Online, April 17, 2009, http://online.wsj.com/article/SB123990409247725869.html.

²⁴ “Narendra Modi Thanks Gujarat for Hat-Trick: Highlights of Speech,” NDTV, December 20, 2012, www.ndtv.com/assembly/narendra-modi-thanks-gujarat-for-hat-trick-highlights-of-speech-508126.

²⁵ “Nitish’s Bihar: From Goonda Raj to Vikas Raj,” NDTV, November 24, 2010, www.ndtv.com/assembly/nitishs-bihar-from-goonda-raj-to-vikas-raj-440147.

²⁶ Pratap Bhanu Mehta, “Mandate for a Dream,” Indian Express, March 7, 2012, http://indianexpress.com/article/opinion/columns/mandate-for-a-dream.

²⁷ Poonam Gupta and Arvind Panagariya, “Growth and Election Outcomes in a Developing Country,” Economics and Politics 26, no. 2 (July 2014): 332–54, http://onlinelibrary.wiley.com/doi/10.1111/ecpo.12038/abstract.

²⁸ Poonam Gupta and Arvind Panagariya, “What Voters Reward: Evidence From the 2009 Indian Parliamentary Elections,” Vox, March 17, 2014, www.voxeu.org/article/voters-reward-performance-evidence-india.

²⁹ For details, see Vaishnav and Swanson, “Does Good Economics Make for Good Politics?”

³⁰ The Center for the Advanced Study of India at the University of Pennsylvania, in conjunction with the Carnegie Endowment, partnered with the Lok Foundation to conduct a survey of a cross section of the population. The survey involved face-to-face interviews of 68,500 randomly selected Indians across 24 states and union territories between September and December 2013. For more detail, see http://indiaintransition.com/lok-pre-election-survey.

³¹ There could be several reasons for this blurring between urban and rural, including rural-urban migration and the growing importance of nonfarm income in rural India; the growth of technology like the Internet and mobile phones and better connectivity; and the psychological “revolution in rising expectations” brought about by unprecedented levels of economic growth in the 2000s. See Neelanjan Sircar and Milan Vaishnav, “Guest Post: The Rural/Urban Divide Dies Out,” Beyond Brics (blog), Financial Times, April 3, 2014, http://blogs.ft.com/beyond-brics/2014/04/03/india-elections-the-ruralurban-divide-dies-out.

³² According to the CSDS survey, voters were most exercised by inflation, rather than growth, with corruption occupying the number two spot. One reason for the discrepancy between these two surveys, other than their timing, could be the structure of the survey questions. Whereas the Lok survey offered respondents a list of eight issues and asked them to identify which one was the most important to them, the CSDS survey allowed for open-ended responses. It is possible that growth comprises several components, which are captured separately in the CSDS survey. See Center for the Study of Developing Societies, India National Election Study 2014 (New Delhi: CSDS, 2014).

³³ Suyash Rai and Milan Vaishnav, “The Politics and Plumbing of Reforms,” Mint, July 28, 2014, www.livemint.com/Opinion/TKFuvmJ7J8ziPDwru0TzeM/The-politics-and-plumbing-of-reforms.html.

³⁴ As the economist Arvind Panagariya noted, “People at all levels, especially the poor, had discovered that unlike during the first four decades of independence, rapid improvement in the economic fortune was possible.” See Arvind Panagariya, “Is Anti-Incumbency Really Pass?” Economic Times, May 28, 2009, http://articles.economictimes.indiatimes.com/2009-05-28/news/29403788_1_anti-incumbency-parliamentary-elections-seats.

³⁵ For more on this topic, see Milan Vaishnav and Reedy Swanson, “Does Good Economics Mean Good Politics in India?” Carnegie Endowment for International Peace, December 19, 2013, http://carnegieendowment.org/2013/12/19/does-good-economics-mean-good-politics-in-india.

³⁶ In the 2014 general election, the BJP alliance in the state of Maharashtra also included three smaller parties—Swabhimani Paksha, Republican Party of India (Athvale), and Rashtriya Samaj Paksha—in addition to the Shiv Sena. The BJP contested 24 (of 48 seats), the Shiv Sena 20 seats, and the smaller parties the remaining four.

³⁷ Adnan Farooqui and E. Sridharan, “Is the Coalition Era Over in Indian Politics?” Round Table: The Commonwealth Journal of International Affairs, December 2014, https://casi.sas.upenn.edu/sites/casi.sas.upenn.edu/files/upiasi/Farooqui-Sridharan,%20Round%20Table%20article.pdf.

³⁸ Data sourced from www.indiavotes.com.

³⁹ The Election Commission has rather idiosyncratic rules governing the designation of “national party” status, which I ignore here. Their criteria can be found at Press Information Bureau,
Government of India, Election Commission, “Dynamics of Elevation of Political Parties to State or National Party,” March 8, 2014, http://pib.nic.in/newsite/PrintRelease.aspx?relid=104537.

⁴⁰ Ruchir Sharma, “The Rise of the Rest of India: How States Have Become the Engines of Growth,” Foreign Affairs, September/October 2013, www.foreignaffairs.com/articles/139646/ruchir-sharma/the-rise-of-the-rest-of-india.

⁴¹ Mehta, “Reform Political Parties First.”

⁴² Annie Gowen, “Ex-Movie Star and Convicted Politician Still Running Her Indian State,” Washington Post, February 16, 2015, www.washingtonpost.com/world/asia_pacific/ex-movie-star-and-convicted-politician-still-running-her-indian-state/2015/02/16/bc2f48d6-abc2-11e4-8876-460b1144cbc1_story.html.

⁴³ For more on this topic, see Milan Vaishnav, “The Complicated Rise of India’s Regional Parties,” Carnegie Endowment for International Peace, November 13, 2013, http://carnegieendowment.org/2013/11/13/complicated-rise-of-india-s-regional-parties; and Vaishnav and Smogard, “A New Era in Indian Politics?”

⁴⁴ Ernesto Dal Bó, Pedro Dal Bó, and Jason Snyder, “Political Dynasties,” Review of Economic Studies 76, no. 1 (2009); House of Commons Library, “Current MPs Related to Other Current or Former Members of the House of Commons,” www.parliament.uk/briefing-papers/SN04809/current-members-related-to-other-current-or-former-members.

⁴⁵ Daniel M. Smith, “Succeeding in Politics: Dynasties in Democracies,” Dissertation, Department of Political Science, University of California–San Diego, 2012.

⁴⁶ Yogesh Uppal, “The Disadvantaged Incumbents: Estimating Incumbency Effects in Indian State Legislatures,” Public Choice 138 (2009): 9–27; Leigh L. Linden, “Are Incumbents Really Advantaged? The Preferences for Non-Incumbents in Indian National Elections,” mimeo, University of Texas–Austin, 2004.

⁴⁷ Dhaval Kulkarni, “Vote for India, Free It of Dynasty Rule: Narendra Modi,” DNA India, December 23, 2013, www.dnaindia.com/mumbai/report-vote-for-india-free-it-of-dynasty-rule-narendra-modi-1939374.

⁴⁸ Kritika Sharma Sebastian, “BJP Has Ended Dynasty Politics,” Hindu, January 31, 2015, www.thehindu.com/news/cities/Delhi/bjp-has-ended-dynasty-politics/article6841190.ece.

⁴⁹ Kanchan Chandra, “Hardly the End of Dynastic Rule,” Economic and Political Weekly 49, no. 28 (July 2014): 25–28.

⁵⁰ Deepak Gidwani, “After Tipu, Teju Pratap Ninth of Yadav Clan to Enter UP Politics,” DNA India, September 17, 2014, www.dnaindia.com/india/report-after-tipu-teju-pratap-ninth-of-yadav-clan-to-enter-up-politics-2019167.

⁵¹ Rukmini S., “They Have Politics in Their DNA,” Hindu, June 9, 2014, www.thehindu.com/news/national/they-have-politics-in-their-dna/article6095560.ece.

⁵² Patrick French uncovered a strikingly similar relationship with data he compiled on the members of the 15th Lok Sabha elected in 2009. See Patrick French, India: A Portrait (New York: Vintage, 2012).

⁵³ Ronojoy Sen, “Identifying Criminals and Crorepatis in Indian Politics: An Analysis of Two Supreme Court Rulings,” Election Law Journal 11, no. 2 (2011): 216–25.

⁵⁴ These figures on criminality come from reports compiled by the Association for Democratic Reforms based on the affidavits submitted by candidates to the Election Commission of India. Detailed data is available at http://myneta.info.  

⁵⁵ Trilochan Sastry, “Towards Decriminalisation of Elections and Politics,” Economic and Political Weekly 49, no. 1 (January 2014): 30–41.

⁵⁶ Detailed data are available at http://myneta.info.

⁵⁷ Association for Democratic Reforms, Analysis of Criminal and Financial Background Details of Winners in Navi Mumbai Municipal Election 2015, April 27, 2015, www.adrindia.org/research-and-report/election-watch/local-bodies/maharashtra/2015/analysis-criminal-and-financial-0.

⁵⁸ This debate is summarized in Neelanjan Sircar and Milan Vaishnav, “Ignorant Voters or Credible Representatives? Why Voters Support Criminal Politicians in India,” mimeo, Carnegie Endowment for International Peace, 2015.

⁵⁹ Amartya Sen, Poverty and Famines: An Essay on Entitlement and Deprivation (Oxford: Clarendon Press, 1981); Timothy Besley and Robin Burgess, “The Political Economy of Government Responsiveness: Theory and Evidence From India,” Quarterly Journal of Economics 117, no. 4 (2002): 1415–51.

⁶⁰ Banerjee et al., “Are Poor Voters Indifferent to Whether Elected Leaders Are Criminal or Corrupt?”

⁶¹ See, inter alia, Ward Berenschot, Riot Politics: India’s Communal Violence and the Everyday Mediation of the State (London: Hurst, 2012); Lucia Michelutti, The Vernacularisation of Democracy: Politics, Caste and Religion in India (New Delhi: Routledge, 2009); and Jeffrey Witsoe, Democracy Against Development: Lower-Caste Politics and Political Modernity in Postcolonial India (Chicago: University of Chicago Press, 2013).

⁶² “Rajan Vichare Wins Hands Down in Criminal Cases Race,” Fourth Estate (blog), April 12, 2014, http://fourthestateonline.in/?p=1430.

⁶³ This link is elaborated in greater detail in the author’s forthcoming book on crime, corruption, and democracy in India.

⁶⁴ The Hindu caste system can be divided into three broad categories: Upper Castes, who traditionally have occupied the highest rung of the caste system; Scheduled Castes, also known as Dalits and formerly as “untouchables,” who are located at the very bottom of the caste hierarchy; and Other Backward Classes, a very large, heterogeneous mix of individual castes located between the two poles. Because caste does not apply or applies unevenly to non-Hindu communities, we also included a prompt for a candidate “from your own community.”

⁶⁵ Sircar and Vaishnav, “Ignorant Voters or Credible Representatives?”

⁶⁶ These data are available at http://myneta.info.

⁶⁷ For more on this topic, see Vaishnav and Smogard, “A New Era in Indian Politics?”

⁶⁸ Milan Vaishnav, “Ethnic Identifiability: Evidence From a Survey of Indian Voters,” mimeo, Carnegie Endowment for International Peace, 2015.

⁶⁹ Witsoe, Democracy Against Development: Lower-Caste Politics and Political Modernity in Postcolonial India.

⁷⁰ Simon Chauchard, “Unpacking Ethnic Preferences: Theory and Micro-Level Evidence From North India,” mimeo, Dartmouth College, 2015.

⁷¹ Center for the Study of Developing Societies, Bihar Assembly Election Postpoll Survey 2010 (New Delhi: CSDS, 2010).

⁷² A. K. Verma, Mirza Asmer Beg, and Sudhir Kumar, “A Saffron Sweep in Uttar Pradesh,” Hindu, May 23, 2014, www.thehindu.com/opinion/op-ed/a-saffron-sweep-in-uttar-pradesh/article6037683.ece.

⁷³ K. K. Kailash, “Regional Parties in the 16th Lok Sabha Elections: Who Survived and Why?” Economic and Political Weekly 49, no. 39 (2014): 64–71.

⁷⁴ Vaishnav, “Ethnic Identifiability: Evidence From a Survey of Indian Voters.”

⁷⁵ Mark Schneider, “Does Clientelism Work? A Test of Guessability in India,” Center for the Advanced Study of India Working Paper 14–01, 2014.

⁷⁶ Neelanjan Sircar, “Identity Predicts Vote Choice, Not Electoral Outcomes,” Transitions (blog), March 14, 2014, http://indiaintransition.com/2014/03/14/identity-predicts-vote-choice-not-electoral-outcomes.77

⁷⁷ Ashutosh Varshney, “Hindu Nationalism in Power?” Journal of Democracy 25, no. 4 (2014): 34–45.

⁷⁸ Milan Vaishnav, “Why a ‘Congress-Mukt Bharat’ Would Be Bad for BJP,” Indian Express, April 7, 2015, http://indianexpress.com/article/opinion/columns/why-a-congress-mukt-bharat-would-be-bad-for-bjp/99.