What Drives Opposition to Economic Reforms? The Role of Ex Ante Uncertainty

Abstract

The theoretical literature on support for pro-market economic reforms identifies uncertainty about the future consequences of reform as an important determinant of opposition to pro-market policies. However, this literature does not rigorously test the role of uncertainty on support for pro-market reforms. This paper addresses this gap by testing the role of uncertainty using a new geocoded dataset of state-owned enterprises in India and a difference-in-differences strategy. I find that voters who are threatened with privatization but do not experience the policy vote against the privatizing party whereas voters that actually experience privatization vote in favor of the privatizing party. These results provide strong empirical support for the role of uncertainty to explain opposition to economic reform programs.

Keywords

economic reform privatization India economy

Introduction

The literature on economic reforms and public opinion revolves around a very basic question: if economic reforms are beneficial for an economy, why do they generate widespread resistance?¹ In countries across the world, major reform programs are slow to be implemented and frequently rolled back after implementation due to public opposition from large segments of society. However, the reason why economists frequently recommend and politicians attempt to implement economic reforms is generally because they believe they will provide long-term economic benefits. So, why is political opposition to reforms so widespread?

One explanation that appears in the literature is that individuals, including both the general public and vested interests, have high levels of uncertainty about the future consequences of economic reforms and so support the status quo. This uncertainty may lead to opposition to reforms even if the expected utility of reforms is higher than under the status quo (Bates and Krueger 1993; Fernandez and Rodrik 1991; Przeworski 1991).

While intuitively appealing, an uncertainty-based explanation for opposition to economic reforms has not been rigorously tested. In this paper, I fill this gap by analyzing the privatization of major state-owned enterprises (SOE) in India in the 1990s. I use satellite imagery to geocode the locations of SOEs in India so as to isolate the set of voters (i.e., employees and communities surrounding SOEs) who have an interest in the privatization process. Many scholars have noted that privatization² should be among the most electorally salient economic reforms because it is easy for the broad public to understand, and directly affects a large and organized vested interest in the form of labor at state-owned enterprises (Haggard and Webb 1994; Varshney 1998).

Privatization in India is an ideal case to test the proposed theory because privatization was only implemented at a subset of firms based on their respective finances. Prior to implementation of the policy, it was unclear which firms would be privatized, and, therefore, opposition groups mobilized against privatization across firms. Ultimately, however, the privatizations were only minority sales and had a minimal effect on employment and wages in the period under analysis. My primary empirical tests use a difference-in-differences strategy, in which I compare constituencies with an SOE that was not privatized to constituencies with an SOE that was privatized before and after the initiation of privatization policy in 1991. I find that constituencies with an SOE that was not privatized show a significant decline in vote-share for the incumbent after the initiation of privatization policy in other locations, whereas constituencies with an SOE that was privatized show an increase in vote-share for the incumbent. These results provide strong evidence that uncertainty explains ex ante opposition to economic reforms even in cases when affected voters ultimately support the policy.

The article proceeds in four sections. In the first section, I discuss the literature on why voters oppose economic reforms, with a focus on uncertainty-based explanations. In the second section, I outline the case of Indian privatization, and why it is an ideal case to study the role of uncertainty in voter support/opposition to economic reforms and present the data I use in my analysis. In the third section, I present the empirical test of the uncertainty-based theory for opposition to economic reforms. In the last section, I discuss my results and conclude.

What Causes Opposition to Economic Reforms?

There are three common answers to the question of why economic reforms are politically difficult to implement despite the fact that they are intended to improve overall economic performance: (1) retrospective voting makes the general public myopic and unable or unwilling to give economic reforms sufficient time to improve the economy; (2) the general public does not block reforms, and instead it is special interests that are the main obstacle; and (3) uncertainty about the impact of economic reforms causes opposition to reforms even in cases where the expected net impact of reforms on society is positive. None of these explanations is mutually exclusive. In fact, there is no doubt that the first two explanations play an important role in explaining opposition to economic reforms in some cases. The focus of this paper is on the third explanation, voter uncertainty, which has yet to be rigorously tested in the scholarly literature.

A retrospective voting model as applied to economic reforms implies that voters only consider past or current economic conditions in their voting decisions, and given that economic reforms generally hurt the economy in the short term (Przeworski 1991), voters oppose them. However, a number of works have found this explanation to be lacking. Case studies of economic reforms from Stokes (1996), Przeworski (1996), and Haggard and Webb (1994) show that in some cases voters are able to support economic reforms that hurt present economic performance based on a promise that they are necessary for better outcomes in the future. Similarly, Malhotra and Margalit (2013) find experimental evidence that voters condition their economic voting in the present period on expectations about economic conditions in the future.

A second explanation is that the general public’s opinion on reforms is not what makes reform difficult to implement, and, instead, it is vested interests that benefit under the status quo that block reforms. In most cases of pro-market economic reform, the benefits of reform are diffuse whereas the costs are concentrated for a set of vested interests (Haggard and Kaufman 1992; Olson 1996). Because of this, a number of scholars have noted that if pro-market reforms improve the overall performance of the economy, then it should be possible to compensate the vested interests who lose from reforms and make everyone at least as well off as they were prior to reforms, and, therefore, remove opposition from vested interests. However, in many cases, it will not be possible to credibly commit to compensation prior to reforms. North, Wallis, and Weingast (2009) and Acemoglu and Robinson (2000) argue that the process of economic reform can permanently alter who is in power, and make ex ante promises of compensation by existing elites not credible. Krueger (1974) and Tullock (1975) create models in which the rent-seeking costs to be among the winners pre-reform are already priced into the benefits of the winners. Furthermore, these works point out that the benefits to winners ex ante reform are often illegal, as are the costs needed to acquire benefits (e.g., in the form of bribes). This also makes a credible commitment to vested interests who lose from reforms difficult.

An alternative explanation for opposition to economic reforms, and the one tested in this paper, is that voters are highly uncertain about the future effects of reform policies and so oppose them even if they believe the reforms will benefit the economy as a whole. In uncertainty-based theories, voters support the status quo because they are uncertain about how reforms will affect them at an individual level. Note that this explanation may play a role in both the general public theory and vested interest theory discussed above. For example, in the context of vested interests, uncertainty about whether vested interests will be compensated for losses is critical to the arguments of Krueger (1974) and Tullock (1975).

Bates and Krueger (1993) point out that professional economists are uncertain about the effect of proposed reforms, so it is not surprising that individuals are uncertain whether they are themselves future winners or future losers from proposed economic reforms. Przeworski (1991) argues that even when voters believe that reforms may benefit the economy in the long term, they may oppose them because they are risk averse. Specifically, Przeworski (1991) argues that voters who are risk averse will oppose pro-market reforms because the private sector will not provide the stability that the public sector can. A similar argument comes from Bai, Lu, and Tao (2006), in which the authors create a model in which state-owned enterprises can be used to maintain “social stability” in countries that have not developed a social safety net. The authors argue that the fact that SOEs do not have a hard budget constraint allows them to expand hiring in economic downturns and, therefore, reduce economic risk for individuals within the economy. Dewatripont and Roland (1995) develop a model of economic reform implementation that accounts for the enormous levels of uncertainty in how the reforms will affect the the economy as a whole and individuals. As the authors point out, “Large-scale reforms involve great aggregate and individual uncertainty . . . it is not clear whether the outcome will be closer to the West German miracle or the Weimar republic, not to speak of former Yugoslavia.”

The most explicit theoretical work on the role of uncertainty and opposition to economic reforms is Fernandez and Rodrik (1991). The authors create a model in which reforms that would benefit the majority of the public when complete are not supported by a majority ex ante because of individual-specific uncertainty even when individuals are not risk averse. The insight is that even if ex post, a majority is guaranteed to benefit under reforms, ex ante, the expected benefit of reforms can be negative for the majority of voters. There are a number of conditions under which this scenario is possible, for example, when the potential losses of losers are very high relative to potential benefits or when uncertainty is very high.

This literature presents uncertainty as a plausible explanation for widespread opposition to economic reforms that are intended to improve the economy. However, there has been little to no empirical work that has tested the role of uncertainty in opposition to economic reforms. Empirically testing the uncertainty explanations requires specific predictions of an uncertainty explanation that differ from competing explanations. Specifically, there are two competing explanations that need to be accounted for: (1) voters are myopic and only consider short-term costs and not long-term gains, and (2) it is only vested interests who are certain that they are losers from a proposed reform program that opposes the policy. Again, these explanations are not mutually exclusive, but an empirical test of uncertainty must be able to differentiate from these two explanations.

An uncertainty-based explanation for opposition to economic reforms predicts that most voters will oppose reforms ex ante and support reforms ex post. This would be true of a myopic voter story as well; for example, voters oppose the reforms because of the short-term costs but support them once the benefits emerge. However, there are two predictions of an uncertainty-based explanation that are unique: (1) voters will oppose the economic reform even before observing any negative effects from the policy, and (2) voters will support the policy ex post implementation even if the diffuse economic benefits have not emerged as long as they observe that they are not losers from the policy.

In regard to vested interest explanations of opposition to reform, uncertainty on the part of vested interests may drive their opposition to reform. In fact, this is the scenario that I argue occurred in the case of privatization in India where the vested interest was labor in SOEs and surrounding communities. In this case, uncertainty and vested interest opposition are connected, as the uncertainty is experienced by vested interests. Nonetheless, uncertainty can still be separately identified because without uncertainty, we would not expect vested interest opposition to shift to support. The only case where we may expect vested interest opposition to shift to support without uncertainty is if compensation were larger than the losses of vested interests from reform. However, even in cases where this were true, uncertainty about whether the compensation would be delivered must play a role or else vested interests would have always supported the reforms as they knew they would eventually benefit from the delivery of compensation greater than their losses.

Thus, what is needed to test an uncertainty-based explanation for economic reforms is a policy that caused visible uncertainty about its future impact among a segment of the population ex ante implementation, but ultimately had no real effect on that same subset of voters. If voters oppose the reforms ex ante, but support them ex post after observing they are not among the losers from the reforms, then this indicates that uncertainty was driving opposition ex ante. This is because the only variable that would change is uncertainty—when this is removed, voters support a policy that should benefit them in the long term.

In the next section, I outline how the case of privatization in India in the 1990s is a uniquely good case to test an uncertainty-based explanation for voter opposition to reforms because it fulfills these requirements.

Case Selection and Data

Privatization in India

The initiation of privatization in India provides a unique case for testing whether uncertainty underlies opposition to economic reforms. While anti-reform groups mobilized labor and local communities against privatization in the early 1990s, privatization had a negligible impact on labor due to the extremely strict employee protection laws in India (Teitelbaum 2006; Varshney 1998) and because the sales were only minority offerings. Furthermore, because only a subset of SOEs actually experienced privatization in this time period, I can compare how labor responded to the threat of privatization versus how labor responded to actual privatization without negative consequences. Effectively, one group has continuing uncertainty about the effects of privatization (those at firms that were not privatized but are still under threat of privatization), whereas for the other group (those at firms that experienced privatization without negative consequences), uncertainty has been reduced by experiencing the policy.

In 1991, just prior to the initiation of privatizations in 1991, SOEs employed 2,180,000 formal employees and approximately that number again in temporary employees. Public-sector manufacturing employed 1,871,100 permanent and temporary workers (30% of total manufacturing employment) and public-sector mining (permanent and temporary) employed another 959,140 persons (91% of total mining employment; Joshi and Little 1994). SOEs have accounted for approximately 20 percent of total Indian gross domestic product (GDP) since the 1990s (Government of India [GOI] 2010). As a point of comparison with large private firms in India, the fifty SOEs that are listed on stock exchanges account for more than 20 percent of total market capitalization on Indian exchanges (GOI 2013).³

Leading up to the 1991 parliamentary elections, the Indian National Congress (INC) billed itself as the centrist party between the right-wing Bharatiya Janata Party (BJP) and the left-wing parties of the former National Front coalition, and the only party that could solve India’s numerous state succession struggles (Varshney 1998). Congress won the most seats in the 1991 elections and Prime Minister P. Narisimha Rao formed the government. By the time of the election, the Indian economy was entering a major balance of payments crisis, and Rao used the crisis to undertake major reforms. The New Economic Policy (NEP) was instituted within months of the election, and International Monetary Fund (IMF) loans were secured to cover the balance of payments crisis.

The first round of equity sales in SOEs occurred in December 1991, just five months after the INC took office. In total, between 1991 and 1996, the INC government sold an average of 19.2 percent of forty SOEs (out of the more than 200 SOEs; Makhija 2006). The government sold equity shares above 40 percent in only 10 percent of the companies offered, privatization revenue was only half of the amount targeted in budgets, and there were no SOEs that were fully privatized during the INC government of 1991–1996.⁴

The main impetus for privatization in India was to raise funds to cover the deficit and for more popular subsidies (Mani and Bhaskar 1998; Panagariya 2011; Varshney 1998). This explains why privatization went from being entirely off the table prior to the fiscal crisis, to a reality following the crisis. Because the main motivation for privatization in India has been to raise funds, the firms that were privatized initially were the largest and most valuable (Arun and Nixson 2000; Mani and Bhaskar 1998). The manufacturing, oil and gas, and metal production sectors account for almost 75 percent of privatizations in this period because these made up the most attractive investment opportunities (these sectors account for less than half of the total number of SOEs).⁵

The online appendix shows the number of employees working at privatized firms for the years 1991–1996. Although close to a third of SOE employees worked at firms that were privatized in this period, only forty out of the approximately 220 SOEs were privatized. Again, this reflects the fact that the largest SOEs were the ones that were subject to privatization in this period, and also indicates that employee numbers were not part of the selection process.⁶

Ultimately, though, labor and local communities were not negatively affected by privatization after enactment. Baijal (2002) states that, “A fear is often expressed that the employees would lose if companies were privatized. Recent privatizations have shown that these fears are totally unfounded. Privatized companies have not retrenched a single person.” The government also provided compensation to employees to dampen labor opposition to privatization. For example, voluntary retirement packages were offered at SOEs to reduce the pains of reforms while also reducing the workforce. Average compensation for voluntary retirements was US$17,108 per employee, which is one of the largest in the world.⁷ A number of authors have argued that this dramatically reduced opposition to privatizations in the long run (Kapur and Ramamurti 2002; Makhija 2006; Uba 2008), despite the fact that union leaders of opposition-affiliated unions almost always opposed voluntary retirement packages (Srinivasan 1999; Uba 2008). Workers at privatized firms have also been offered equity at highly reduced prices (Kapur and Ramamurti 2002), and in some cases offered retraining programs (Roychowdhury 2004).

There are a number of cases where workers at SOEs initially opposed privatization, only to then support it after the privatization occurred. For example, after a sixty-seven-day strike against privatization and the eventual privatization of an aluminum-producing SOE, the leader of a non-affiliated union at the main plant stated, “It [the strike] was a terrible mistake. There is virtually no pressure. Even idle employees are being paid and the company pay package is the best in the last three decades.” The union even offered to work overtime free of charge for the new private management until the plant achieved 100 percent output to make up for the strike following privatization. A similar outcome appears in a study of utility privatization at the state level by Kale (2007). The author argues that,

In both cases of utility privatization (Orissa and Delhi), although the labor unions opposed the policy initially, their compliance was ensured through tripartite agreements signed by the government, the unions, and the new private owners. These agreements secured the wages and benefits of current employees.

The above analysis indicates that the high ex ante negative expectations did not occur in the SOEs that were actually privatized. If ex ante uncertainty about the future effects of privatization is what is driving public opposition to the policy, then we should observe ex ante opposition change to ex post support in cases where the policy was implemented. In contrast, in locations where the policy was not implemented (but that have an SOE), we should observe the opposition to the policy to continue because the uncertainty has not been removed. Thus, there are two testable hypotheses that an uncertainty-based explanation predicts in this context:

Hypothesis 1: In constituencies with an SOE that was threatened with privatization but where it did not occur, there will be a decline in support for the party that proposed the policy (the INC).

Hypothesis 2: In constituencies with an SOE that was privatized, there will be an increase in support for the party that proposed the policy (the INC).

The first hypothesis is the case where uncertainty is not removed. The second hypothesis is the case where uncertainty is removed. In the empirical analysis to follow, I test these hypotheses.

Data and Method

I collected information on the location of all central-level SOEs in India. To do this, I used a master list from the GOI’s Department of Public Enterprises. I then found the location of all units of each SOE. I relied primarily on the relevant SOE websites to find locations. If I was not able to find the information there, I then turned to other sources such as local newspaper reports. In general, I was only able to find the general area in which the SOE is located, but using satellite imagery, I was able to find the exact locations of the vast majority of SOE units.⁸

The dataset does not include SOEs that do not have specific locations or have a multitude of small locations. For example, banks, insurance companies, and telecommunication service providers were excluded from the dataset because the units of these SOEs are too small to expect any noticeable effect on voting, and they are so ubiquitous that there is almost no variation across constituencies.

I then located these SOEs in State Assembly constituencies using geographic information system (GIS) data from ML InfoMap (2008). There were 4,120 State Assembly constituencies throughout India in this period. Assembly election dates are staggered by state, that is, only some states have elections in every year. I use State Assembly constituencies as opposed to Parliamentary (Lok Sabha) constituencies because State Assembly constituencies are a fraction of the size (approximately six per Parliamentary constituency), and so we can expect an SOE to affect voting at this level. While privatization of centrally owned SOEs was very much a national-level policy and not a state-level one, I am assuming that within State Assembly constituencies with a large SOE and a large percentage of voters employed in an SOE, privatization by the national-level, INC-led government affected support for the INC in state-level elections. This assumption is justified in the next subsection.

The dependent variable of interest is the change in support for the INC before and after the initiation of privatization policies. In my basic models, I use the change in INC vote-share, however, I also use other measures of vote outcomes to check for robustness in the online appendix. I compare the election closest to but before the 1991 initiation of privatization policy (in no case earlier than 1989) with the first election following the completion of the INC privatizations (i.e., elections between 1996 and 1998).

The independent variables of interest in the study are (1) a dummy variable for whether there is an SOE in a State Assembly constituency, and (2) a dummy variable for whether the SOE in a constituency was privatized. Thus, there are two treatment and control sets in the analysis: constituencies with an SOE compared with constituencies without an SOE, and constituencies with an SOE that was privatized by the INC between 1991 and 1995 compared with constituencies with an SOE that was not privatized.

Vote-Share of Incumbent as Proxy for Public Opinion

Ideally, the dependent variable in this study would be survey data that directly measures support and opposition to privatization during this time period. Unfortunately, there is no survey data on this from this time period that are representative at a small enough level to compare opinions at SOEs that were privatized with SOEs that were not privatized. Therefore, I rely on changes in the vote-share of the incumbent party that implemented privatization in state-level elections as a proxy for support for privatization policy.

This involves a number of assumptions. First, I am making the assumption that constituencies with SOEs were affected more by privatization policy and were more likely to vote on this issue than constituencies in other parts of the country. Second, I am making the assumption that privatization, which was a policy implemented at the national level, impacted the vote-share of the INC in state-level elections. In this section, I provide justification for these assumptions.

Public opinion studies have shown that knowledge of economic reforms in India in the 1990s was extremely low (Kumar 2004; Yadav 2004). Furthermore, the number of voters directly impacted by SOEs (discussed below) was small at the national or state level (average turnout in state elections in this period was 64%; Yadav 1996). Thus, privatization was certainly not a major voting issue in constituencies without an SOE. The bigger assumption is that privatization was a major voting issue in state-level elections in constituencies with at least one large SOE.

The average number of SOE employees in a constituency with an SOE is approximately 6,500. The average size of the electorate in an assembly constituency is 162,000. Therefore, SOE employees represent approximately 4 percent of the total electorate in an average assembly constituency with an SOE. In terms of the average number of voters in an assembly election, the percentage accounted for by labor is closer to 7 percent. Note that most employees at SOEs are from areas local to the SOE location as government policy is to hire from the local population to avoid conflict with local communities (Subramanian 2007). This should limit the possibility that any observed relationship between SOE location and voting patterns is a function of migrant populations.

In addition to workers being affected by privatization, communities surrounding SOEs were also affected. There are news reports from this period of spouses and family members participating in anti-privatization strikes (OneIndia News 2006) as well as community members not directly employed by the SOE (Gurtoo 2006). This is not surprising as SOEs play a significant role in communities beyond employment. Large SOEs in India are required by law to use a percentage of their profits for Corporate Social Responsibility projects, and these projects are generally located in the immediate vicinity of an SOE location and bear the name of the SOE. For example, the state-owned National Aluminum Company employed approximately 7,500 individuals in three separate locations. Each of these locations has a township named after the company, and across the townships, the company funded a public hospital, a public community center, a club for workers and their families, a human resource development center, a city park, and numerous public grade schools and high schools; and all of these are named after the company. Attaching the name of an SOE to what is effectively government-provided social infrastructure is common for all Indian SOEs; especially in the case of hospitals, schools, and sport stadiums. In the online appendix, I provide maps and a discussion of other community infrastructure projects in areas with SOEs.

Furthermore, extensive strikes against privatization policy prior to its implementation increased the visibility of the privatization issue in constituencies with an SOE, and these strikes were clearly associated with parties other than the incumbent party (the INC). Because all national political parties in India have affiliated unions, partisan opposition and support of privatization was highly apparent at the shop-floor level of SOEs. The INC, BJP, and Communist Party (CPI[M])-affiliated unions had the largest membership at the time (Saxena 1993), and were particularly dominant in SOEs (Agarwala 2006). Party-affiliated unions tend to tow the party line on all political issues. For example, labor union strikes were a contributing factor in the lead-up to the 1975 State of Emergency in India, however, the INC-affiliated union (the largest in the country) did not participate in strikes against the INC government in this period (Rudolph and Rudolph 1987).

Between 1991 and 1996, the INC-affiliated union did not participate in national-level strikes against privatization, whereas all other major unions did (Uba 2008).⁹ For the major unions, leadership is frequently chosen on political grounds. Union leaders are also frequently aspiring politicians and use union leadership as a political launching pad (Teitelbaum 2006). Thus, at the SOE level, party and union are close to synonymous.

In total, there were forty-five national-level protests (between 1991 and 2003) against privatization involving an average of two million participants each (with the largest ones approaching ten million participants; Uba 2008). In addition, there were ninety-one strikes, fifty-four demonstrations, seventeen sit-ins, and thirty-two other protests including blocking roads and occupying buildings near SOEs. And these were large events, for example, the ninety-one strikes had an average of 161,000 participants (Uba 2008). All of these strikes were called by opposition-affiliated unions, and government-affiliated unions abstained from the strikes. From 1991 to 1996, the INC-affiliated union encouraged labor restraint, whereas the opposition-affiliated unions denounced the potential ill effects of privatization on labor.

While privatization of central SOEs was enacted at the national level in India, privatization was still an important voting issue in State Assembly elections. In many cases, SOEs sit on land that is owned by the states and leased to the central government, giving State Assemblies the ability to veto privatizations (Kale 2007; Kapur and Ramamurti 2002).¹⁰ In addition, there are a host of cases where privatizations have become a major issue in State Assembly elections. The Telugu Desam Party campaigned on a strongly anti-reform policy in Andhra Pradesh State Assembly elections in 1995 and won a majority in the state (Suri 2004; Uba 2008). In the 1995 elections in Maharashtra, the allowance of a foreign private firm to enter the power generation sector, which was previously reserved for SOEs, became a central issue in the campaign (Dugger 2001; Smith 1999). In the privatization of Aluminum India Limited, the State Assembly supported union strikes, offered to purchase 51 percent of the firm to keep it from being purchased by a private buyer, and threatened to cancel the state lease of bauxite mines if the privatization occurred (Gurtoo 2006). Almost identical situations occurred in privatizations of other SOEs such as the National Aluminum Company, Bharat Aluminum Company Ltd. (BALCO), and the hotels owned by the India Tourism Development Corporation (Kapur and Ramamurti 2002). In the case of BALCO, the Chief Minister of the State Assembly “instigated labor in the plant to strike, claiming that their future would be jeopardized if the sale went through” (Kapur and Ramamurti 2002).

It should also be noted that support for reforms was certainly not unanimous among the INC leadership, largely because reforms were not popular among the general public. As noted in Bardhan (2005), “Even politicians in any ruling party over the last decade who support reforms play them down during election time; a party that initiates some reforms is quick to oppose them when out of power.” Privatization in India was even more politically risky than broader deregulation and tariff reforms.

However, at the state level, I was not able to find any example of an INC Chief Minister that actively opposed economic reforms. This in no way means INC leaders actively supported reforms or campaigned on the issue. In fact, very few INC leaders actively supported economic reforms. Nonetheless, opposition parties sought to make the election about the INC proposing and supporting economic reforms in constituencies with SOEs. In this way, the INC was associated with privatization at all levels of government because the opposition used it as a method of seeking electoral advantage. Thus, the assumption that privatization was a major voting issue in state-level elections in constituencies with an SOE appears to be justified. Furthermore, while there was limited united support for reforms among the INC, the opposition painted the INC as pro-reform and “anti-people” (Bardhan 2005) so as to make electoral gains.

Empirical Strategy

My empirical strategy is to compare election results in constituencies with an SOE with constituencies without an SOE, as well as constituencies with an SOE that was privatized. My primary empirical tests consist of difference-in-differences tests, in which I compare the same constituencies before and after the initiation of privatization policies in 1991. Constituencies with an SOE that was not privatized in this period experienced the threat of privatization without implementation, and constituencies with an SOE that was privatized experienced the threat of privatization and with implementation.

For the difference-in-differences strategy to identify the effect of privatization policy on locations with SOEs, there needs to be no other change that occurred in the period when privatization policy was initiated by the INC (post-1991) that would affect voting in constituencies with an SOE differently than constituencies without an SOE.

It is also central to my empirical strategy that SOEs were not selected for privatization based on some feature of the area that is related to changes in voting behavior between 1991 and 1996. As discussed in the prior section, SOEs were selected for privatization in this period based on size and value. The fact that the largest and most valuable SOEs were chosen for privatization in this period means that each SOE privatized had multiple large locations. This should make it difficult if not impossible to have selected SOEs with all locations in assembly constituencies with specific traits.

Unfortunately, census data do not match with electoral data in India (a problem that all researchers of Indian politics face). To analyze the similarity between areas with an SOE and an area without an SOE, I located SOEs into census subdistricts (a unit of similar size to a State Assembly constituency but with different boundaries). The online appendix presents balance statistics for subdistricts with an SOE and without an SOE location. The results indicate that subdistricts with and without SOEs are similar across observable variables with the exception that areas with an SOE have a higher urban population. However, we see no difference between subdistricts with an SOE that was privatized relative to subdistricts with an SOE that was not privatized. This provides support for the difference-in-differences strategy.

The base specification of the difference-in-differences at the State Assembly level is,

\begin{array}{l} Y_{i},_{j} = β_{0} + β_{1} S O E_{i} + β_{2} Y e a r_{j} + β_{3} (S O E_{i} \times Y e a r_{j}) + \\ β_{4} P R V_{i} + β_{5} (P R V_{i} \times Y e a r_{j}) + f_{i} + e_{i}, \end{array}

where $S O E$ is a dummy variable for whether a constituency has an SOE, $Y e a r_{j}$ is a dummy for pre- or post-initiation of privatization by the INC in 1991, $S O E_{i} \times Y e a r_{j}$ is an interaction between the SOE dummy and a dummy for pre- and post-1991, $P R V_{i}$ is a dummy for whether a constituency has an SOE that was privatized, $P R V_{i} \times Y e a r_{j}$ is the interaction between the privatization dummy and the dummy for pre- and post-1991, $f_{i}$ is a vector of state-level fixed effects and state-level fixed effects interacted with the $Y e a r_{j}$ dummy, and $e_{i}$ is the error term. The dependent variable ( $Y_{i},_{j}$ ) is the vote-share of the INC. The interactions $S O E_{i} \times Y e a r_{j}$ and $P R V_{i} \times Y e a r_{j}$ are the variables of interest in this specification, and show how election outcomes change from election to election in constituencies with an SOE (and an SOE that was privatized) relative to constituencies without an SOE.

In Table 1, I use all State Assembly units in the sample (N = 3,043).¹¹ I include individual state dummy variables and individual state dummy variables interacted with the $Y e a r$ variable to control for time variant trends that are specific to individual states. Thus, my results are robust to scenarios such as if all constituencies in a limited number of states show a large increase in vote-share for the INC during this period and these same states have a large proportion of total SOEs.

Table 1.

INC Vote-Share Pre- and Post-privatization with Entire Sample.

	Model 1	Model 2	Model 3	Model 4
SOE	0.069		0.948
SOE	(1.421)		(1.700)
Year	−1.533	−1.650	−1.437	−0.268
Year	(3.963)	(3.926)	(3.942)	(2.441)
SOE × Year	−1.191		−3.938*
SOE × Year	(1.503)		(1.884)
Privatized		−2.240	−3.153	−2.966
Privatized		(2.038)	(2.529)	(3.708)
Privatized × Year		5.864**	9.665**	12.281**
Privatized × Year		(1.678)	(1.900)	(2.733)
Constant	15.926**	15.926**	15.926**	38.65*
Constant	(3.205)	(3.204)	(3.203)	(16.645)
State FE × Year	Yes	Yes	Yes	No
N	6,086	6,086	6,086	360
r ²	.337	.338	.337	.455

INC = Indian National Congress; SOE = state-owned enterprises; FE = fixed effects.

†

p < .10. *p < .05. **p < .01.

In model 1, I look at the overall comparison of constituencies with an SOE relative to constituencies without an SOE before and after the initiation of privatization by the INC. We see that the interaction between SOE and the time variable is negative but not significant, meaning that constituencies with an SOE showed no significant change in support for the INC following the initiation of privatization relative to constituencies without an SOE when we look at the entire sample. In model 2, I look at the comparison of constituencies with an SOE that was privatized relative to all other constituencies before and after the initiation of privatization by the INC. Here, we see evidence (at the 99% confidence level) that having an SOE that was privatized during this period is associated with an increase in the support for the party that initiated privatization (the INC). In model 3, I look at SOEs that were not privatized and SOEs that were privatized relative to constituencies without an SOE. In this model, the estimated coefficients on both interactions are significant: privatized SOEs show an increase in support for the INC, non-privatized SOEs show a decrease in support for the INC. In model 4, I limit my sample to only the constituencies that have an SOE (N = 180) and look at the change in INC vote-share in constituencies with an SOE that was privatized relative to a constituency with an SOE that was not privatized. The results indicate that the vote-share for the INC increased in constituencies with an SOE that was privatized by more than 12 percent relative to constituencies with an SOE that was not privatized.

These models indicate that SOEs have a heterogeneous relationship with INC vote-share determined by whether they were privatized in the period under consideration (1990–1996). Without privatization, constituencies with an SOE show a decline in support for the INC following the initiation of privatization policy relative to constituencies without an SOE. This result is intuitive in that we would expect labor and communities that rely on SOEs for patronage from SOEs to oppose privatization as it may reduce the benefits they receive from SOEs. The second result is much less intuitive. We see that constituencies with an SOE that was privatized show an increase in support for the INC in this period.

There are a number of assumptions that we need to make for these results to be seen as causal. While it is not necessary to assume that there is nothing that differentiates constituencies with an SOE from constituencies without an SOE other than the presence of the SOE, we do have to assume that any systematic differences did not cause the change in INC vote-share following the 1991–1996 period. In the balance statistics, we saw that constituencies with an SOE, constituencies with an SOE that was privatized, and constituencies without an SOE were relatively similar across a number of variables, but not in regard to urban concentration. Constituencies with an SOE are systematically more urban than constituencies without an SOE.

In an effort to reduce this potential bias, I drop all constituencies from the analysis that either do not contain an SOE location, or do not border a constituency that has an SOE location. Instead of using state fixed effects as in the previous analysis, I use fixed effects for clusters of adjacent constituencies.¹² I use fixed effects for each cluster of constituencies in the figure. This reduces the sample to 819 constituencies and sixty adjacent constituency clusters.

Therefore, I am now making the less demanding assumption that there is no difference across adjacent State Assembly constituencies that could cause an increase in INC vote-share in this period other than the SOE. This allows me to compare urban constituencies with an SOE to urban constituencies without an SOE in the same urban center, and rural constituencies with an SOE with rural constituencies without an SOE in the same rural area.¹³

In Table 2, I present models 1 to 3 as in the prior results, but with the reduced set of constituencies and adjacent cluster fixed effects. The results are substantively the same as in the first set of models with all constituencies included and no adjacency fixed effects, although the magnitude of the estimates on a privatized SOE are reduced from 12 percent to 10 percent. Constituencies with an SOE that was not privatized show a reduced level of support for the INC following the initiation of privatization policies, and constituencies with an SOE that was privatized show an increase in support for the INC following the initiation of privatization policy.¹⁴

Table 2.

INC Vote-Share Pre- and Post-privatization with Reduced Set.

	Model 1	Model 2	Model 3
SOE	−0.347	−2.914^†	−0.413
SOE	(1.488)	(1.548)	(1.805)
Year	−23.917**	−25.120**	−22.580**
Year	(0.989)	(0.000)	(0.958)
SOE × Year	−2.406		−5.081**
SOE × Year	(1.978)		(1.915)
Privatized		2.481	0.235
Privatized		(2.931)	(2.993)
Privatized × Year		4.631	9.212*
Privatized × Year		(3.882)	(4.006)
Constant	25.294**	26.577**	25.327**
Constant	(0.744)	(0.774)	(0.903)
Adjacency × Year FE	Yes	Yes	Yes
N	1,638	1,638	1,638
r ²	.395	.397	.399

INC = Indian National Congress; SOE = state-owned enterprises; FE = fixed effects.

†

p < .10. *p < .05. **p < .01.

To further ensure that the results are not driven by the urban location of many Indian SOEs, I used ESRI’s World Urban Spatial Dataset (which spatially denotes the densest urban population areas in a country¹⁵) to limit my sample to SOEs outside of urban areas. In the case of India, there are 279 of these areas, and constituencies with an SOE in one of these areas was dropped from the dataset. Because I am using fixed effects at the cluster level, any cluster in which the only SOE was located in an urban area was dropped from the analysis completely. This reduces my sample to only 140 constituencies with an SOE, of which thirty-six were privatized. In Table 3, I run the same models as above but without the urban SOEs. The results are insignificant in the first two models but in the predicted direction. In the fully specified models (models 3 and 4), the coefficients on the interaction between privatization and year are significant and in the predicted direction, while the interaction between SOE and year in model 3 is significant at the 90 percent confidence interval level with the reduced sample.

Table 3.

Base Regressions without SOEs in Urban Constituencies and Reduced Sample.

	Model 1	Model 2	Model 3	Model 4
SOE	−0.738	−2.155	−0.121
SOE	(1.677)	(1.480)	(2.022)
Year	−24.736**	−0.165**	−0.165**	0.643
Year	(1.150)	(0.000)	(0.000)	(3.421)
SOE × Year	−0.769		−4.146^†
SOE × Year	(2.301)		(2.301)
Privatized		−0.435	−2.295	−4.808
Privatized		(3.311)	(3.569)	(3.863)
Privatized × Year		8.480*	12.291**	11.747*
Privatized × Year		(4.126)	(4.555)	(4.751)
Constant	41.667**	42.010**	41.509**	43.146**
Constant	(8.734)	(8.720)	(8.724)	(11.739)
Adjacency × Year FE	Yes	Yes	Yes	No
N	1,530	1,530	1,530	280
r ²	.408	.411	.412	.453

SOE = state-owned enterprises; FE = fixed effects.

†

p < .10. *p < .05. **p < .01.

As an additional robustness check, I use estimates of the number of employees at each SOE to show that the impact of privatization on voting was stronger at firms with more employees, as we would expect if the results are due to the SOEs and not an unobserved variable associated with SOEs. This analysis is presented in the online appendix, and the results are substantively unchanged.

Testing the Parallel Trends Assumption

A major assumption of the difference-in-differences strategy for identification of a causal relationship is that the treatment and control groups follow parallel trends in the outcome variable prior to the time period under study and would continue on these trends if not for the shock (the initiation of privatization policy). While this assumption is not directly testable, there are a number of ways of suggesting that it holds.

As suggested in Angrist and Pischke (2008), one way to examine this assumption is to graph the coefficients from regressions of the treatment on the lagged dependent variable. I run the model:

Y_{i} = β_{0} + β_{1} S O E_{i} + β_{2} P R V_{i} + f_{i} + e_{i}

on all constituency elections in each yearly cross-section with elections from 1985 to 2008 (nineteen years with elections), where $f_{i}$ is the adjacent constituencies fixed effects. In Figure 1, I graph the regression estimates for $β_{1}$ and $β_{2}$ over time. The solid line maps the estimates for the coefficient on $S O E$ , and the dotted line maps the estimates of the coefficients on $P R V$ . The estimated coefficients follow a similar trend over time until 1994 (and are not statistically different from zero), at which point we see a divergence in the estimates. The estimate for $S O E$ becomes negative and significant, whereas the estimate on $P R V$ becomes positive and significant. By the 2000 elections, the estimates return to a similar trend and to no longer being statistically different from zero.

Figure 1.

Regression estimates for election years from 1985 to 2008.

Another means of testing the parallel trend assumption is to run the same models on alternative outcome periods and to use alternative sets of independent variables. If the assumption is correct, then we should not see the treatment have any effect on these other outcomes. In Table 4, I present placebo regressions of my main model (model 3 from the previous tables) with alternative national-level parties in India (the BJP and Left Parties¹⁶) in the same elections, and for the elections surrounding the 1999 parliamentary elections. None of the models are significant in any period or for any non-INC party, thus lending support to the necessary assumptions of the difference-in-differences strategy (i.e., parallel trends assumption).¹⁷

Table 4.

Placebo Parties and Placebo Periods.

	BJP91	BJP99	CP91	CP99
SOE	−0.772	0.005	1.714	0.670
SOE	(1.096)	(1.326)	(1.066)	(0.947)
Year	−2.907**	0.293	0.699^†	0.023
Year	(0.734)	(0.661)	(0.375)	(0.575)
Privatized	0.724	1.460	0.756	0.768
Privatized	(2.054)	(3.195)	(2.255)	(1.773)
SOE × Year 91	0.645		−1.398^†
SOE × Year 91	(1.468)		(0.751)
Privatized × Year 91	−0.701		0.924
Privatized × Year 91	(3.348)		(1.539)
SOE × Year 99		−0.802		−0.106
SOE × Year 99		(7.556)		(1.900)
Privatized × Year 99		−1.467		1.462
Privatized × Year 99		(13.042)		(3.271)
Constant	0.132	7.383**	−0.857	−0.335
Constant	(0.274)	(0.331)	(0.533)	(0.473)
Adjacency × Year FE	Yes	Yes	Yes	Yes
N	1,656	1,648	1,656	1,650
r ²	.389	.374	.527	.475

SOE = state-owned enterprises; BJP = Bharatiya Janata Party; CP = Communist Party.

†

p < .10. *p < .05. **p < .01.

Finally, in Table 5, I present my basic models but instead of using the SOEs that were privatized from 1991 to 1996, I use the SOEs that were privatized after this period. If there is something about constituencies with privatized SOEs that makes them more likely to show a strong increase in INC vote-share between 1991 and 1996 and it is not related to the SOE, then we might also expect this to be true of firms privatized in later periods (twenty-six total locations under the BJP and INC from 1998 to 2013), assuming that the selection was based on a similar criteria in later periods. The interaction of $P R V \times Y e a r$ does not approach significance in any model.¹⁸ This again supports the argument that the main findings are not the result of an unobserved feature that privatized SOEs may have been selected upon.

Table 5.

INC Vote-Share with Placebo of Future Privatized Firms.

	Model 1	Model 2	Model 3
SOE	−1.349	0.368
SOE	(1.243)	(1.464)
Year	−25.120**	−23.380**	−3.865
Year	(0.000)	(1.008)	(3.319)
SOE × Year		−3.872^†
SOE × Year		(2.319)
Privatized after 1998	−3.755	−5.165	−1.976
Privatized after 1998	(4.065)	(3.991)	(5.325)
Privatized after 1998 × Year	4.811	7.694	5.270
Privatized after 1998 × Year	(5.049)	(5.202)	(5.216)
Constant	41.817**	41.388**	26.868**
Constant	(0.310)	(0.366)	(2.329)
Adjacency × Year FE	Yes	Yes	No
N	1,638	1,638	356
r ²	.395	.396	.008

SOE = state-owned enterprises; FE = fixed effects.

†

p < .10. *p < .05. **p < .01.

Conclusion

This paper tests the role of uncertainty in driving voter opposition to economic reforms using the case of privatization in India. The empirical results are very much in line with an uncertainty-based explanation for ex ante opposition to economic reforms. Prior to the implementation of privatization in India, affected constituencies show a significant decline in support for the party that proposed the policy relative to constituencies without an SOE. However, after experiencing privatization, affected constituencies show an increase in support for the party that implemented privatization relative to both constituencies that do not have an SOE and to constituencies that have an SOE that was not privatized. In contrast, constituencies with an SOE that did not experience privatization in this time period maintained their opposition to the party that proposed privatization policy.

Thus, the uncertainty surrounding the future effects of privatization appears to drive the decline in support for the INC in this period, whereas the actual policy appears to have led to an increase in support for the INC. Essentially, ex ante opposition turns into ex post support among communities who experience the reform and observe minimal consequences of the policy. Furthermore, the result cannot be driven by ex post economic effects of the policy because the policy did not have observable costs or benefits in the period analyzed, and the result cannot be driven by ex ante support for privatization in general because we observe ex ante opposition in threatened constituencies (i.e., constituencies with an SOE).

These results provide strong support for an uncertainty-based explanation of opposition to economic reforms. Voters were uncertain about how they would be affected by privatization and so opposed the policy ex ante. Voters who experienced the implementation of privatization ultimately supported the policy despite the fact that the policy provided no benefits in the period under analysis. Essentially, voters appear to have supported the economic reform once the uncertainty was removed.

While this work provides evidence of the role of uncertainty in explaining ex ante opposition to economic reforms, the results are limited to a single case of reform in a single country. Furthermore, privatization in India during the time period analyzed was limited in scope to a fraction of total SOEs and minority shares in these firms. Future work should attempt to broaden these results to other contexts. This includes additional countries where reform was broader in scope. An additional avenue worth exploring may be an experimental design that manipulates levels of uncertainty by providing information about the effects of an ongoing real-world reform.

Last, the support for uncertainty as a causal variable in determining opposition to economic reforms found in this paper in no way implies that opposition is not also driven by other factors—such as retrospective voting and vested interest opposition as discussed in section “What Causes Opposition to Economic Reforms?” Despite these limitations, this analysis provides strong evidence that uncertainty about the future effects of reforms plays an important role in opposition to economic reforms.

Footnotes

Supplemental Materials

Supplemental materials for this article, as well as replication data, are available with the manuscript on the Political Research Quarterly (PRQ) website.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

References

Acemoglu

Robinson

J. A.

. 2000. “Political Losers as a Barrier to Economic Development.” The American Economic Review 90 (2): 126–30.

Agarwala

2006. “From Work to Welfare: A New Class Movement in India.” Critical Asian Studies 38 (4): 419–44.

Angrist

J. D.

Pischke

J. S.

. 2008. Mostly Harmless Econometrics: An Empiricist’s Companion. Princeton: Princeton University Press.

Arun

Nixson

. 2000. “The Disinvestment of Public Sector Enterprises: The Indian Experience.” Oxford Development Studies 28 (1): 19–32.

Bai

C.-E.

Tao

. 2006. “The Multitask Theory of State Enterprise Reform: Empirical Evidence from China.” The American Economic Review 96 (2): 353–57.

Baijal

2002. “Privatisation: Compulsions and Options for Economic Reform.” Economic and Political Weekly 37:4189–95.

Banerjee

Munger

M. C.

. 2004. “Move to Markets an Empirical Analysis of Privatization in Developing Countries.” Journal of International Development 16 (2): 213–40.

Bardhan

2005. “Nature of Opposition to Economic Reforms in India.” Economic and Political Weekly 40:4995–98.

Bates

R. H.

Krueger

A. O.

. 1993. Political and Economic Interactions in Economic Policy Reform. Oxford: Blackwell.

10.

Dewatripont

Roland

. 1995. “The Design of Reform Packages Under Uncertainty.” The American Economic Review 85:1207–23.

11.

Dinc

I. S.

Gupta

. 2011. “The Decision to Privatize: Finance and Politics.” The Journal of Finance 66 (1): 241–69.

12.

Dugger

C. W.

2001 “High-Stakes Showdown; Enron’s Fight over Power Plant Reverberates beyond India. New York Times, March 20.

13.

Fernandez

Rodrik

. 1991. “Resistance to Reform: Status Quo Bias in the Presence of Individual-Specific Uncertainty.” The American Economic Review 81:1146–55.

14.

Government of India. 2010. “2010 Public Enterprise Survey.” New Delhi: Department of Public Enterprises, Government of India.

15.

Government of India. 2013. Department of Disinvestment, September. http://www.divest.nic.in/.

16.

Gupta

2005. “Partial Privatization and Firm Performance.” The Journal of Finance 60 (2): 987–1015.

17.

Gurtoo

2006. “Mindset Challenges at Aluminum India Limited: Privatization of a State-Owned Enterprise.” Asian Case Research Journal 10 (2): 261–80.

18.

Haggard

Kaufman

R. R.

. 1992. The Politics of Economic Adjustment: International Constraints, Distributive Conflicts, and the State. Princeton: Princeton University Press.

19.

Haggard

Webb

S. B.

. 1994. Voting for Reform: Democracy, Political Liberalization, and Economic Adjustment. Vol. 94. Washington, DC: The World Bank.

20.

Haltiwanger

Singh

. 1999. “Cross-Country Evidence on Public Sector Retrenchment.” The World Bank Economic Review 13 (1): 23–66.

21.

Joshi

Little

I. M. D

. 1994. India: Macroeconomics and Political Economy, 1964–1991. Washington, DC: The World Bank.

22.

Kale

2007. Power Steering: The Politics of Utility Privatization in India. Austin: The University of Texas at Austin.

23.

Kapur

Ramamurti

. 2002. “Privatization in India: The Imperatives and Consequences of Gradualism.” In India after a Decade of Economic Reforms: Retrospect and Prospects, edited by Srinivasan

T. N.

. Stanford: Stanford University Press.

24.

Krueger

A. O.

1974. “The Political Economy of the Rent-Seeking Society.” The American Economic Review 64 (3): 291–303.

25.

Kumar

2004. “Impact of Economic Reforms on Indian Electorate.” Economic and Political Weekly 39:1621–30.

26.

Makhija

A. K.

2006. “Privatisation in India.” Economic & Political Weekly 41 (20): 1947–51.

27.

Malhotra

Margalit

. 2013. “Do Expectations Condition Retrospective Voting?” Unpublished.

28.

Mani

Bhaskar

M. V.

. 1998. “A Curmudgeon’s Guide to Economic Reforms in India’s Manufacturing Sector.” Economic and Political Weekly:3257–72.

29.

ML Infomap. 2008. India GIS Data in ESRI SHP Format. New Delhi: ML Infomap.

30.

North

D. C.

Wallis

J. J.

Weingast

B. R.

. 2009. Violence and Social Orders: A Conceptual Framework for Interpreting Recorded Human History. Cambridge: Cambridge University Press.

31.

Olson

1996. “Distinguished Lecture on Economics in Government: Big Bills Left on the Sidewalk: Why Some Nations Are Rich, and Others Poor.” The Journal of Economic Perspectives 10 (2): 3–24.

32.

OneIndia News. 2006. “RCF Employee Wives Join Agitation.” March 28. http://news.oneindia.in/2006/03/28/rcf-employees-wives-join-agitation-1143551140.html.

33.

Panagariya

2011. India: The Emerging Giant. New York: Oxford University Press.

34.

Przeworski

1991. Democracy and the Market: Political and Economic Reforms in Eastern Europe and Latin America. Studies in Rationality and Social Change. Cambridge: Cambridge University Press.

35.

Przeworski

1996. “Public Support for Economic Reforms in Poland.” Comparative Political Studies 29 (5): 520–43.

36.

Rodrik

1996. “Understanding Economic Policy Reform.” Journal of Economic Literature 34 (1): 9–41.

37.

Roychowdhury

2004. “Globalisation and Labour.” Economic and Political Weekly 39:105–108.

38.

Rudolph

. 1987. In Pursuit of Lakshmi: The Political Economy of the Indian State. Orient BlackSwan.

39.

Saxena

1993. “The Hindu Trade Union Movement in India: The Bharatiya Mazdoor Sangh.” Asian Survey 33 (7): 685–96.

40.

Smith

1999. Privatized Infrastructure: The Role of Government. London: Thomas Telford.

41.

Srinivasan

1999. “Voluntary Retirement and Workers’ Welfare.” Economic and Political Weekly 34:1873–76.

42.

Stokes

1996. “Public Opinion and Market Reforms: The Limits of Economic Voting.” Comparative Political Studies 29 (5): 499–519.

43.

Subramanian

2007. “A Sociological Profile of a Public Sector Workforce.” Economic and Political Weekly 42 (51): 37–47.

44.

Suri

2004. “Democracy, Economic Reforms and Election Results in India.” Economic and Political Weekly 39:5404–11.

45.

Teitelbaum

2006. “Was the Indian Labor Movement Ever Co-opted? Evaluating Standard Accounts.” Critical Asian Studies 38 (4): 389–417.

46.

Tullock

1975. “The Transitional Gains Trap.” The Bell Journal of Economics 6:671–78.

47.

Uba

2008. “Labor Union Resistance to Economic Liberalization in India: What Can National and State Level Patterns of Protests Against Privatization Tell Us?” Asian Survey 48 (5): 860–84.

48.

Varshney

1998. “Mass Politics or Elite Politics? India’s Economic Reforms in Comparative Perspective.” The Journal of Policy Reform 2 (4): 301–35.

49.

Yadav

1996. “Reconfiguration in Indian Politics: State Assembly Elections, 1993–95.” Economic and Political Weekly 31:95–104.

50.

Yadav

2004. “Economic Reforms in the Mirror of Public Opinion.” The Hindu, June 13.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.23 MB

0.00 MB

10.28 MB