The institutional development of the ECI occurred in two phases: a pre-1991 period, in which the ECI played an important-but-circumscribed role; and a post-1991 period, in which, the ECI, under a weakened executive and led by an entrepreneurial CEC, began to expansively interpret its constitutional mandate and play an increasingly powerful role. Below we substantiate the ECI’s expanded role by examining two key indicators: Model Code Implementation and Election Duration.19 We use these two indicators because: (1) they represent the most significant manifestations of the ECI’s expansively interpreted mandate and (2) they can be tracked longitudinally.
Model Code Implementation
The Model Code of Conduct began in the South Indian state of Kerala, in 1960, as a consensus between political parties regarding their electoral conduct. It delineates the types of appeals that may be made during the run-up to an election (e.g. no ethnic or religious appeals, no criticism of candidates’ private lives), outlines the procedures that must be followed for meetings and processions, describes what members of the ruling party cannot do while acting in official capacity, describes permissible election manifesto material and lists polling day rules (e.g. distance parties must maintain from polling booths, how parties must cooperate with authorities, how party workers must identify themselves, etc.). While infractions occur on a regular basis, parties largely adjust their conduct once officially notified, suggesting that the Model Code does represent the rules of the game.20
Until the late 1980s, the ECI merely watched how the Model Code was updated and gradually adopted by additional states. It wasn’t until 1990, however, that the ECI enforced it. In December 1990, T. N. Seshan became CEC and quickly began pursuing ECI independence and mandate expansion. He did so, in part, by formalizing the Model Code. These efforts drew the ire of Narasimha Rao’s Congress government and the executive pushed back, expanding the ECI to three members to check Seshan’s power. These two additional election commissioners failed, however, to curtail both the ECI’s power expansion and Model Code institutionalization.
As Singh (2012: 153) explains, ‘since 1991, the Model Code has come to be seen as an integral part of elections, making the electoral contest democratic by ensuring that the party in power and those who staked claims to power would abide by certain rules, and by pruning the powers of the ruling party to reduce the advantage that it may have in the electoral arena’. One former CEC pointed out, ‘[we] are interested in catching violations of the Model Code. It requires substantial manpower. But in such a competitive environment rivals (parties and candidates) monitor each other. A large number of complaints and Model Code violation reports come from political parties. We take these complaints very seriously and investigate them immediately’. ECI officials knew that a proliferation of television channels and media platforms ensure that such incidents are well-publicized, so the ECI has to respond swiftly.
Examples of Model Code enforcement abound. In January 2017, Arvind Kejriwal was censured for remarks at a rally in Goa.21 Kejriwal, the Aam Admi Party leader, suggested that voters, when parties offer Rs. 5000 for their vote, should ask for Rs. 10,000. In 2015, in the lead-up to Bihar’s state elections, BJP President Amit Shah was censured for stating that, if the BJP loses in Bihar, ‘firecrackers will go off in Pakistan’, a violation of the provision prohibiting aggravation of existing differences or creation of mutual hatred. In the same election Rahul Gandhi, Congress Vice President, was cautioned for suggesting that the BJP makes Hindus and Muslims hate each other; these were unverified allegations used to criticize other candidates or their workers. A more subtle Model Code violation occurred in the lead-up to the 2009 general elections. The ECI notified the Ministry of Youth Affairs and Sports, the Cabinet Secretary and the Chief Secretary of Delhi, for taking out a full-page advertisement on the 2010 Commonwealth Games in major Delhi newspapers listing the infrastructure built for the event and the thousands of job opportunities created. The ECI found this list of the government’s achievements to be a clear Model Code violation.
Since it is not based on legislation passed in parliament, the Code is not judicially enforceable. The action against a violator usually takes the form of an advice, warning or censure. No punitive action can be taken. But this does not make the Code toothless. Its moral authority outweighs its legal sanctity. Its impact is instant. Political leaders are scared of inviting a notice for a violation, as it creates negative public perception about them and their party just before elections. Importantly, while individuals sometimes contest whether their behaviour truly violated Model Code regulations, both candidates and parties almost never argue that the Model Code is illegitimate or should be abandoned.
In addition to Model Code institutionalization, the ECI’s expanded mandate manifested itself in the duration of both national (parliamentary) and state-level elections. With the exception of the 1952 and 1957 national elections, early Indian elections, as evidenced by Fig. 2.1, were brief. Most national elections, from independence to 1996, took only a few days. This has since changed dramatically. The first three parliamentary elections were held over four months, 17 days and ten days, respectively.22 The three elections conducted between 1967 and 1977 were each completed in less than seven days. The same trend continued through the 1980s. The 1991 election was supposed to be completed in a week; however, Rajiv Gandhi’s assassination in the middle of the election extended its duration.
Since 1991, however, the general trend has been towards substantially longer elections. In 1996, election duration began to increase sharply. The 1996 election was completed in three phases over the course of a month; the 1998 election in four phases over three weeks; the 1999 election in five phases over four weeks; the 2004 election in four phases over three weeks; the 2009 election in five phases over four weeks; and the 2014 and 2019 elections in nine phases over five weeks.
This pattern holds at the state level as well, where the ECI also conducts elections. While longer election duration is particularly pronounced in larger states, it also exists in smaller states. For example, in Bihar, the average pre-1991 election duration was 1.4 days, and 17 days since then. In Maharashtra, the average pre-1991 election duration was 1.86 days, and 8.75 days since then. Meanwhile, in Odisha, the pre-1991 number is 3, and 6 since then. In Andhra Pradesh, pre-1991 election duration averaged 1 day, while post-1991 duration averaged 8 days. Finally, in Assam, a much smaller state, in terms of both geographic size and population, average pre-1991 election duration was 1.75, while post-1991 duration was 3.8.
Entrepreneurial CECs justified the increase in election duration by pointing to demands for cleaner elections. The ECI has to move security force personnel and other parts of the administrative apparatus over long distances. These logistical constraints help justify longer election duration because they are directly related to the process of ensuring free and fair polling.
The ECI’s expanding mandate in the post-1991 period is also visible across other indicators. Since 1991, the ECI has mandated voter ID cards, curtailed campaign periods, engaged in large-scale voter mobilization, tried to regulate political parties and entry of candidates into the electoral process, and assumed both executive and quasi-judicial control during elections (Quraishi 2014).