The institutional improvisations of Indian democracy confound predictions and principles. The decision to make voting compulsory in local body elections in Gujarat is a deeply intriguing move, normatively and empirically. J P Agrawal of the Congress also has a private member's bill to this effect. There are reasons to be sceptical about this move, but it is important to be clear about its pros and cons.

Despite the Supreme Court's dismissal of an earlier petition, it is not prima facie clear that a law requiring compulsory voting is unconstitutional. In fact, that petition was dismissed more on practical and procedural grounds than on a due consideration of the constitutional issues. Even if not desirable, compulsory voting is not outside the bounds of constitutional permissibility.

The debate over compulsory voting appeals to a common democratic language on four dimensions: choice, legitimacy, equality, and participation. The distinguished political scientist Arend Lipjhart wrote an article in the late '90s suggesting, in desperation, that compulsory voting was the only way to counter voting apathy and declining legitimacy in democracies in the developed world. The Indian framing is curious because it comes from a different angle. India is exceptional in that voting turnouts have remained decent. But most importantly turnout amongst the poor has been no worse, and usually much higher, than amongst the privileged. Compulsory voting is being framed as a device not to empower the poor, but to save the middle class, "the drawing room wallahs" from self-destruction and apathy, another reminder of how easily the discourse of democracy can be mobilised on behalf of the privileged.

But a democratic sensibility should concentrate less on the framing and more on the substance of the issue itself. The first value at stake is choice. Proponents would argue that the compulsion involved in voting is at best minor and no more onerous than the compulsions the state poses on us in so many areas, from taxation to regulation. Indeed, this move requires not that you vote but that you turn out. It requires you to expressly register your act of non-choice, and by aligning it with an option to reject everyone, it may enhance choice.

On the other hand, not turning out is also a significant act of choice; and it may not necessarily register your dissatisfaction. You may simply not care who wins, and there is no reason why this disposition should not have a legitimate place in democratic politics.

The legitimacy argument is less compelling. Here the worry is that candidates often get elected with a small plurality of votes, and increased turnouts would enhance the legitimacy of elected representatives. This argument is flawed for a number of reasons. First, how many votes someone may be elected with is a function of a number of things, including number of candidates. There is no evidence that in democracies where candidates win with small pluralities, the system as a whole is less legitimate than in democracies with compulsory voting. Legitimacy is mostly endogenous to electoral rules; by changing rules you do not necessarily enhance legitimacy. You simply replace one decision-making rule by another.

Finally legitimacy is a complex idea and cannot be reduced to a single arithmetical point. You could argue that a plurality system with low winning margins that allows rapid turnover of different groups in power is more legitimate because it assures different groups that they have a shot at power. It polarises the distinctions between winners and losers much less.

Third, there is the equality argument, used more in the West than in India. The argument was that low turnouts disenfranchise the poor more. This argument empirically does not apply to India. There is also the paradox that large participation of the poor is nowhere associated with a more social democratic empowerment or greater equality. To be fair to the proponents, it is very difficult to predict the ideological consequences of greater turnouts if they get routinised. But it is not clear that they are a tool for enhancing equality.

The fourth argument concerns participation. Requiring citizens to participate improves their civic engagement. The interesting thing about Gujarat is that the experiment has started at the level of local government, paradoxically the one level of government we are the least engaged with and care least about. Engagement with local government is a challenge in our system. But it is not clear whether this is a function of the administrative structures of government, or voter apathy. Whether high turnouts can transform civic and local bodies is a potentially interesting experiment.

But compulsory voting does not necessarily enhance the quality of civic engagement. Lipjhart assumed that forcing people induces them to have greater interest in the political process. But several papers have demonstrated the opposite effect: it gives disproportionate voice to those who are apathetic and therefore likely to vote more randomly. There is merit in privileging those who want to vote, not those who are forced to.

One side argument in the compulsory voting literature concerns election spending. On this view much spending is devoted to either getting voters out to vote, or from deterring them from showing up by increasing their disenchantment with certain candidates. Compulsory voting induces less spending and less negative campaigning. But this argument has also turned out to be very fragile.

But the most powerful objection to compulsory voting is that it criminalises non-voting. Whatever one's views on the duty and right to vote, there is something deeply problematic about punishing non-voting. On merely practical grounds it is likely to give the state immense powers of harassment. It is a deeply democratic value to worry about state intrusion in our lives. One is not reassured by the fact that Gujarat government officials have been quoted as saying that one of the penalties under discussion is denying BPL cards to those who do not vote. Not having prescribed punishment as part of the legislation but making that part of the rules is itself an odd interpretation of democracy.

It is not an accident that most of the countries that have compulsory voting have to desist from enforcing penalties. It will also be interesting to see how lists for local government elections are created and used. One unintended consequence of this drive may be to reduce the incentives to have your name included in a voting list.

Democratic participation is a laudable value. But a flourishing democracy requires a diversity of dispositions, including the option of disengagement. Voting is an important duty. But giving the state coercive power ostensibly in the name of saving the people from themselves is undemocratic paternalism. Voting must remain an act of choice, not propelled by coercion or inducement.