*Rachana Prakash & Uday Dabas  

Professor Tarunabh Khaitan’s academic work has had a major impact on the way we think about constitutional structure, rights, and democratic functioning — most notably through his influential theory of constitutional morality, and his seminal work on anti-discrimination law, proportionality, and the role of the Indian Supreme Court. In this episode, we turn to one of his newest contributions — his chapter titled “Political Parties in a Democracy” in the book The Entrenchment of Democracy. In this piece, Professor Khaitan offers a deeply thought-provoking account of political parties as “bidirectional intermediaries” and invested with “plenary power” in a democratic system, and proposes a principled framework for regulating them constitutionally — one that aims to balance their public and private character, while preserving their vital role in democratic governance.

LISTEN TO THE PODCAST

Rachana: It’s an absolute privilege to welcome Professor Tarunabh Khaitan to the podcast today. He needs no introduction, but I’m going to try and give one anyway. He’s the professor chair of public law at LSE Law School and an honorary professorial fellow at Melbourne Law School. Previously, he’s been the head of research at the Bonavero Institute of Han Rights at Oxford, the Professor of public law and legal theory at the same place, and he’s also been a visiting professor of law at Harvard as well as NYU law schools. His academic work has had a major impact on the way we think about constitutional structure, rights, and democratic functioning; most notably through his influential theory of constitutional morality and his seminal work on anti-discrimination law, proportionality, and the role of the Indian Supreme Court. I’m sure it’s well-known to all of us that he’s been cited very widely by multiple Supreme Courts across the world. Today we turn to one of his newest contributions: his chapter titled ‘Political Parties in a Democracy’ in the book ‘The Entrenchment of Democracy’. In this piece, Professor Khaitan offers a deeply thought provoking account of political parties as bidirectional intermediaries invested with plenary power in a democratic system, and he proposes a principled framework for regulating them constitutionally. This is a framework that aims to balance their public as well as private character while preserving their vital role in democratic governance. Professor Khaitan, thank you so much for joining us. We’re thrilled to have you here. 

Professor Khaitan: Thank you guys, it’s a pleasure to be here, and I look forward to this conversation very much.

Rachana: So, just to start off, I want to begin by exploring what the first portion of your paper does. You begin by defending political parties against the widespread cynicism and critique that traditional neglect of parties within anglophone constitutional theory does have, and you’ve called it the ‘Big-C Silence’. What prompted you to make political parties the normative object of constitutional theory specifically, and why do you believe that this historical neglect has become untenable today? 

Professor Khaitan: I think the starting point for me is that in sufficiently large and complicated societies and polities, there are only two ways of making collective decisions: one is violence, the other is politics. There is no third option. In disparaging political parties, we are actually paving the way for the alternative, which is violence, a “might is right”, “survival of the fittest” state of nature. Because there is no third alternative for collective decision-making in large-scale societies, because politics is the only non-violent viable option, we are stuck with politics. Either we can try and make it better, or we can give it up.

I’m not rosy-eyed about politics. I know all its short-comings. Any institution will have its shortcomings, which does not mean that we accept it as it is. We do have a duty to improve the institutions we belong to at all times. But it also requires an acknowledgement of the importance of these institutions. Given that starting point, and given the absolute indispensability of political parties for any democracy, I think it’s a huge tragedy that constitutional scholars have ignored parties for so long. And because the agenda-setting in the field of constitutional law is done by states, and because the American scholars have largely ignored parties, that has had ripple effects on multiple different jurisdictions. So that was my motivation to try and challenge that blind spot.

Rachana: Thank you for that Professor. There’s this particular line in the chapter that I really liked in which you said that just because we cannot, we will always face the argent that parties don’t live up to the normative ideal that we put in the constitution but obviously that doesn’t mean we don’t try to put that in in the first place as background structures that try to regulate their functioning. So moving into the nature of political parties, you describe parties as Janus-faced institutions, simultaneously public and private. Could you unpack how this dual character both enables their function and complicates their constitutional regulation? What I found particularly interesting was how you argued that liberal constitutions and the sort of public-private binary that they operate under particularly complicates this task.

Professor Khaitan: Let me first make a comment on the line you quoted. Giving up on norms because we can never realise them fully is just silly. There are many things in life, in fact most of the valuable things in life are things we cannot realise fully. And maybe we should not seek to realise them fully, because we don’t care about just one thing, we care about many things. Just because you cannot have a completely free society does not make freedom undesirable. Just because you cannot create a completely equal society, a completely democratic and kind society, does not make these norms not worth having. Norms are valuable because we try to get closer to them and that is the point of having norms. I’ve already mentioned the geopolitical reason why parties have been ignored by this field, because of the dominance, the gravitational pull of the United States’ practice and scholarship. But there is a conceptual reason as well, deeply embedded in liberal practice, and the conceptual reason is that liberal constitutionalism makes a fundamental distinction between public actors and private actors. The working default prescription for each of these actor types is diametrically opposite. So public actors, which is primarily the state, but certain state-like agencies may be included (and this is what article 12 does in the Indian constitution), primarily have duties not rights. Whereas private actors paradigmatically individual citizens but in liberal constitutionalism also non-state corporations, paradigmatically have rights not duties. And some states experiment with this, and they can make exceptions, and complicate this – it’s not as tidy in practice, but that’s your starting point in liberal constitutional theory.

There are many different entities, not just parties which are complicated to regulate because in reality the public-private divide is not as black and white. I don’t reject the divide, I don’t think everything is public, and I think privateness is important to recognize. What I do reject is the idea that the public-private divide is a binary divide, that there is one box of public actors and another box of private actors. All agents, actors, individuals, corporations, the state, various agencies of the state, sit on a spectrum of publicness. On one end of the spectrum you have the completely public actor, by which I mean actors who normatively can only act for public policy and rights-regarding reasons. That is the essence of publicness, that you your raison d’être, your point is other-regarding. You exist for the sake of others. Whereas on the other end, the private end of the spectrum, there may be a private individual sitting alone in a bedroom immersed in thoughts, which is perhaps the paradigmatic example of the private end of the spectrum, which is permitted to be entirely self-regarding. Everything else sits somewhere on the spectrum. The legal effort to force into a binary division things that are somewhere in between creates complications, and political parties are the perfect example of this. These are entities that are paradigmatically public in that the point of political parties is to acquire state power and to exercise it for others. You could not get more public than that. On the other hand, however, because they are a bridge between the state and its citizens, their ability to interact  regularly and informally without the paraphernalia of state offices is what permits them to forge public opinion and transmit it to the state.

If we saddle political parties with the same  requirements of procedural fairness and formality that apply to state institutions, they will not be able to perform their functions. The district magistrate cannot do the mobilizational function of a private party that political parties can do precisely because they are not permitted the informality and privateness of parties. Parties have this Janus-faced character but liberal constitutionalism tries to put actors into one box or the other. You’ve seen these debates: “Should parties be subject to the right to information act?”, “Should parties be right holders under article 19 or duty bearers under article 19?” These are complicated questions with no easy answers because of the dual character of parties. 

Rachana: You’re right. Could this be connected to the Big-C constitutional silence in that because the role of political parties goes unacknowledged, it’s easier to ignore these nuances and create the liberal binary and go ahead with the constitution in a fairly less complicated manner?

Professor Khaitan: So this would be one conceptual reason for ignoring parties in the Big-C constitution. Now I say that with two caveats. First, not all Big-C constitutions ignore parties. After the second world war, several continental European constitutions were very explicit about the duties and functions of parties, tending to  focus on their public character, which I think has imposed certain costs on parties. Even the Indian constitution, through the 10th schedule, which was added through an amendment in the 1980s, does concern itself with the party. Now the party is many things. The party is anything from your local block level organization, to your state level party, to the national party, to the polit bureau or the national committee, to the legislative party, to the party worker, to the party leader. These are all different things and the 10th schedule of the Indian constitution only deals with the legislative party. It does engage with the party, but only a very small dimension thereof. Small-C constitutions – constitutional statutes like elections law, like the Representation of People Act, things like that do engage with parties in almost all democratic jurisdictions, but they often come as an afterthought. You design these big institutions of the State, or what Roberto Gargarella calls the engine room of the state first, and then you retrofit parties through Small-C constitutional statutes to work around this engine room that has been designed without a clear sight on parties. That is what I think the big mistake is. 

Uday: At the same time, when we’re discussing the public-private distinction, you bring in the idea that rather than understanding them as two discrete boxes, they’re better understood as a spectrum of action-ness and actor sensitivity. So to identify this principle you bring in the purposive autonomy principle. I had two questions. First, how did you arrive at such a principle? And second, how do parties effectively practically apply these principles into their daily political functioning?

Professor Khaitan: The autonomy principle is not a principle that parties have to apply for their functioning. It is a principle that the state and its constitution ought to adopt in relation to parties. It’s a principle about how the state should treat parties.The purpose of autonomy principle is in some ways based on classical liberal principles. It’s an oxymoron, a contradiction in terms, because as I explained earlier, in liberal thought the purpose of private persons is self-regarding (or at least there’s a permission to be self-regarding) and therefore the starting normative assumption is that of autonomy. As a private individual, you can do whatever you want, unless there are justified restrictions on it. Whereas for public entities, because public power is other-regarding, it is directed only towards specified purposes. As a public entity, you may not do anything unless it is justified by prescribed purposes. The purposive autonomy principle is simply a way of recognizing the duality of parties, to say that their privateness demands a presumption of autonomy. The party should be free to do anything they like so long as it does not complicate their public character. So regulation on parties is justified only to the extent that it is in service of their purpose.

What is the purpose of parties? To  organize and mobilize public opinion, for taking turns, and for sharing in political power. And if that is the public purpose of parties what is the practical implication of this? Let me give you  two examples. A regulation requiring parties to  transparently declare their sources of funding regularly to the people, for example, is entirely in keeping with the purposive autonomy principle because the regulation is purposive. It is in sync with the purpose of parties, which is to mobilize and organize public opinion in order to acquire a share in state power. A regulation on parties to compulsorily have internal party elections, in my view, is not justified by the autonomy principle because a party can easily and legitimately perform its public functions without being internally democratic. In fact, there is good evidence to think that inner party democracy may actually be bad for democracy in the state. We have seen in the US and in the UK that the extreme party base that shows up for inner party elections often ends up choosing leaders who seek to undermine democracy in the state. Trump is one glaring example of this. So that would be too much of an interference with the party’s autonomy. Parties should be free to  organize their internal affairs subject to regulations that may be justified by their purposiveness. That is the point of the purposive autonomy principle: to start with the default of autonomy but welcome regulation if it enhances the competitive political role of parties.

Rachana: I understand the purposive autonomy principle, which is to locate that regulatory middle ground which recognizes the importance of privateness and also helps guide how we have to regulate their public presence. I wanted to get a little bit of insight on this: you talk about how this principle can be mobilized and you say that second order regulation is perhaps more conducive to upholding party autonomy, which I thought was interesting given that we started from a place of saying the Big C has this sort of gap in this regulation, so why are we particularly looking at a Small C constitution solution for this?

Professor Khaitan: Great question. Second order regulation is not necessarily Small C regulation; both Big C and Small C constitutions can be either second order or first order. What is the difference between first order and second order regulation? First order regulation is command and control style regulation – your 10th commandment regulation, don’t steal, don’t lie, etc. You have a norm saying don’t do X: most of your criminal laws, for example, or even do X (it can be a positive or negative norm). If you breach that norm, a responsive body can punish you for it, or can investigate the breach, usually courts. So that’s your straightforward first order norms. This is what lawyers are very used to, this is what we think about all the time. Second order norms, or second order enforcement of norms does not work like that. It’s not interested in ex-post remedial action once the norm has been breached. It organizes the institutional environment in a manner that makes the ex ante breach of the norm less likely. Right? So the focus of second order norms is preventative rather than remedial. It tries to reduce the likelihood of breach in the first place.

So why a preference for second order norms rather than first order norms? First, courts are astonishingly bad at regulating parties, and we have a huge  amount of experience  and data on this. I would recommend your attention to a paper by Tom Daly and Brian Christopher Jones, which is in the same vole that the chapter we are discussing appeared. They look at the militant democracy provisions that arise in the German constitution which allow the constitutional court to ban parties that are a threat to the democratic democratic order. They argue quite convincingly that such  ex-post first order regulations are either used by courts effectively in cases where there is no need: when the party threatening democracy is so small and so politically weak anyway that the court can get away by banning it. Or the courts are completely helpless because the party is too strong and there will be a political backlash if the court acted against it. The recent NPD case in Germany is a case in point, where the court shrunk from banning the NPD because it wasn’t politically a non-entity.  There are cases, for example in Turkey, where  the right-wing Islamist party was banned multiple times by Turkish courts and it always came back in a new avatar. Even if the courts do act, it’s usually pointless because it’s an unspent political force and a court order will not kill it. What is even worse is when courts are captured. These types of first order norms are weaponized by the ruling party against the political opposition. We have so much data from Singapore where  the judiciary is used routinely to bankrupt the political opposition using defamation laws. The point of the opposition is to criticize the ruling party. The ruling party gets compliant courts to deem criticism of the ruling party as defamation and orders exorbitant costs which bankrupts political opposition. Recently in Pakistan an inner party democracy rule was used by the election commission captured by the ruling party to deny election symbols to Imran Khan’s party. The courts did eventually intervene to reverse that but that judgment was never implemented.

What I’m saying is that ex-post judicial intervention is either useless, or unnecessary, or counterproductive, or weaponized, and it is an extremely risky business to control parties through courts except in the most limited cases. Procedural regulation like for declaration of assets and things like that, or discrimination against some members, etc.: those things are fine. Courts can do those little things. But the existence or non-existence of a party is too much burden for the weakest branch to bear.

So how can second order regulation help? Second order regulation can make certain types of parties less likely to emerge. If we take anti-system parties, for example, parties that are against the democratic and egalitarian political order. Right now, we know different electoral systems have different implications for polarizing, bigoted, anti-minority parties. We know that in a multi-party first-past-the-post system where the prize is 35% of the vote, it is frequently possible for a party to win by polarizing a population. We also know that in the proportional representation system, smaller parties are incentivized to polarize and be hateful because if you get 10% or 20% of the votes, you will get 10% or 20% of seats, and once in a while you will come to control the balance of power in a coalition government, which is the norm in PR systems. We’ve seen this in Israel and in Italy, where extreme far-right parties are controlling the levers of power and the existence of the government depends on them. We also know that in a ranked-choice vote system where citizens have to vote not by just choosing one of the parties or their candidates but ranking at least their top two or top three candidates or parties is terrible for extreme parties because winning parties are either your first choice or your second choice. In a rank choice voting system, parties have to appeal not just to their base. They also need to tell voters of other parties, “You may not love us, but can you tolerate us? Can you put us down as your number two?” And because these far-right parties don’t get anybody’s second choice vote, they are either your first vote or your last vote. A ranked-choice system keeps parties in the middle, and forces them to build broad follower bases and a large coalition. That is a good example of second order regulation. There’s no court telling a party don’t act like a faction, or not to act against the interests of a subset of the population. But the system is designed in a way that  big tent parties, parties that build broad coalitions that are tolerant, that are pluralistic are politically rewarded and extremist narrowly focused polarizing parties are politically penalized. So that is an example of second order regulation and there are many others as well.

Rachana: So we’re talking about broad systemic choices that we make which then have an  impact on how politics plays out?

Professor Khaitan: It has a path-dependent impact. By making certain choices, we reward certain types of political behavior and not others. Federalism is an example of this. Federalism has huge consequences on the party system. We know that presidentialism leads to weaker parties (weakness here means the party vis-à-vis the leader). In presidential systems, the party has very little control over the leader. In parliamentary systems, in general, parties have much greater control over their leader. So, the institutional design, at least some of the design features, has strong correlations with the kinds of parties and party systems we will have. This relationship is not perfect, because institutions don’t work like the law of gravity. Chance, leadership, and lots of other things intervene. But in general, where we know that certain relationships have held for decades in multiple places, we can deploy that knowledge to incentivize certain types of desirable political behavior and disincentivize bad behavior without the involvement of courts.

Uday: So when we discuss how these second-order, third-order regulations come in, I feel this relates closely to your work on what democracy-related costs are for such parties. I remember you mentioned four of these costs. Could you slowly walk us through what all of these costs are, and what you think is the most important factor that most liberal democracies and their parties are underestimating in the modern age?

Professor Khaitan: In any sufficiently large society, in any sufficiently big polity (by “sufficiently big,” I mean anything bigger than a modest-sized Indian village), you cannot have representative democracy without parties. And this is because a partyless democracy would incur four big costs that parties alone can reduce. The first of these costs I’ve characterized as the voter participation cost. Suppose you live in a democracy with, I don’t know, a million, two million, or a hundred million people, and you want to participate in politics, you want to make a difference on certain issues. Say you really, really care about climate change, and you want to make a difference in climate politics and climate laws, but there are no parties in your democracy. Elections happen, individuals stand for elections, but there’s no party. In such a democracy, your ability to make a difference is absolutely minuscule. It’s vanishingly small. Because even if you become a full-time activist on climate and devote all your time and energy to seeking climate reform, the participation cost in politics for you is huge. Parties by aggregating interests and preference business give us economies of scale. Now of course in the above scenario you could start an NGO, you can gather like-minded people similarly interested in climate change action and then realize that acquiring some state power will be  more effective in changing policy. You may then start putting up candidates for elections, and voila, you have created a political party. So parties reduce the participation costs of voters and also allow voters to participate to different degrees. You can be an otherwise disinterested voter who only shows up for elections. You can be a party activist or a loose affiliate. You can be a donor, you can be a party member, you can be in party leadership, there are different levels of involvement which  is also not something  that’s possible in a partyless democracy.

The idea of sortition and citizen assemblies, which is based on ancient Athenian democracy, doesn’t help with this either. The chances of your being selected by lot to sit on a citizens assembly in any reasonable sized democracy is tiny. Among several tens of millions of people, the chances that you will be selected for a 100 member assembly is small, and if these are issue based citizens assemblies the chance that you will be selected on an assembly that is tasked with the issue that you care about, say climate change, is next to negligible. This cost is not at all reduced by direct sortation methods. The second cost is voter information cost. In a partyless democracy, if there are 10 candidates to vote for, you have to do individual research on all of those 10 candidates to find out what their program is, what their policy is. Parties give us proxy information, rule of thumb information, not perfect information, but we know broadly where these people stand, what their ideology is, within a spectrum. We have some sense because of party affiliation, at least in healthy functional democracies and party systems, and we get proxy information without doing a huge degree of research. Of the other two costs, one is the policy packaging cost.

The one thing we don’t appreciate sufficiently is that no democracy can fulfill or realize the interests and preferences of all the people all the time. There has to be some give and take, there have to be compromises. There has to be some sort of packaging of values and preferences and interests to create a common platform. The best packages are those which leave everybody only slightly unhappy, because nobody gets everything they want, but nobody gets nothing either. Parties are very good at doing that because parties, at least good, functional parties, build big coalitions and they can convince their members that while they may not get everything, this package comes closest to your dream scenario. Even if you vote for us holding your nose it’s better for you than any others. Today, compromise has become a bad word, but we cannot, in large societies, do politics without compromise and compromise is impossible without parties aggregating our interests and values and preferences. The final cost that parties reduce is the ally prediction cost. Once elections have happened for MPs and for MLAs, to know who else is likely to be an ally, who may be persuadable but will be difficult, and who is absolutely repugnant to you, to the point of even refusing to shake their hand. These are not unimportant decisions. Even in a small assembly of 200 people, if you had to do this without parties capable of making simultaneous 150-200 negotiations, things would become extremely dysfunctional. Sooner or later, these people will start evolving into blocs coming together. Parties facilitate and reduce that ally prediction cost. All of these things are essential for a democracy to run smoothly, and if parties did not exist, we would have invented them and we did invent them. There’s a reason why no representative democracy of any reasonable size today can function without parties.

Uday: And do you think that for allowing such costs we require a certain democratic environment in which such political parties work? Would you say that there are prerequisites where parties would require a certain political atmosphere, a certain democratic environment in which they can realize these costs and work towards their efficiency?

Professor Khaitan: This idea that democracy can only work  given certain preconditions used to be extremely common in the middle to late 20th century, and it was actually the basis on which many scholars and politicians predicted the infeasibility of Indian democracy. There were assumptions, like that democracy can only thrive if you have a certain degree of affluence, or that democracy can only thrive under certain institutional conditions, or that democracy can only thrive if you have a certain political culture. India is one of the key examples  which proves these assumptions wrong. The one thing you do need for democracy to thrive is the ruling party’s willingness, when it loses an election, to step down. That is the key cause of the internal fragility of democracy. Democracies are internally fragile because the most powerful political actor is expected, on the happening of certain events, to voluntarily give up power. No political party in India has ever refused to do that so far. Not even the Congress party under Indira Gandhi – when she lost elections there was a peaceful transfer of power. That is the first prerequisite for democracy, and you can’t legislate that into existence. That is a matter of raw power but also ideological power, if democracy at least has to be accepted to that degree.

The second prerequisite is free and fair elections, meaning that you have a system where the ruling party does not rig elections in its favor. With that we have played fast and loose many times in our history. Many elections have been sufficiently free and fair. For others there is a question mark, but it’s only rarely (if ever) been below the threshold of acceptability at least to the opposition. Those two preconditions are cultural conditions, and you have to accept them. Everybody has to accept them. But beyond that, there are many other things that you need for democracy. You need functional institutions, you need rule of law, you need freedom of expression. You need a lot of other things that are extremely important for democracy to flourish. But again, democracy is not a value that you either have or don’t have. It admits to degrees. You can have more or less of it. And of course, the goal should be to strive towards a better democracy, not necessarily a greater democracy. We have started assuming that  democracy is the only legitimizing factor for all institutions, and that’s not true. The system is not more democratic if we start electing our courts, our election commissions, or our police chiefs. Certain institutions have to be appointed if democracy is to function well as a system. A colleague of mine used to tell me that there is a lot of scholarship on the power of unelected judges and how that’s abusive, and he says if you want to see the abuses that elected judges, come to Texas and I’ll show you. So be wary of what you ask for. But anyway, with those qualifications, I think yes, there are certain minimal preconditions, but they’re not as expansive as many people might assume. 

Uday: So, in addition to how political parties have to regulate these issues, there is also discussion about what kind of constitutional design a party should have within its own structure to ensure that it achieves these very objectives. You discuss these objectives over a series of five to ten pages. How would you suggest that democracies achieve these objectives, especially in the case of parties that do not have a particularly strong history of internal democratic practice?

Professor Khaitan: As I said, I do not care whether a party is internally democratic. What matters is whether the party is committed to democracy at the level of the state. Of course, parties are better and healthier when they are based on ideologies and values rather than individuals and personality cults. Institutional design can help in creating these dynamics. It is one thing to require inner party democracy, but I think the purposive autonomy principle would tolerate the idea that the leader of the legislative should be elected by the elected representatives in that legislature. For example, the leader of the Bahujan Samaj Party in Parliament should be elected by BSP MPs, and not appointed by an external authority. I think it is entirely acceptable for the purposive autonomy to encourage collective leadership in parties, so that parties don’t become the personal fiefs of individual leaders. I have already said that a parliamentary or semi-parliamentary system is generally much better at facilitating the collective leadership of parties not controlled by individuals. In a presidential system, parties are much more likely to become personal fiefdoms. But design is not the only thing that determines the kind of political outcomes you get; it is one of the many factors, but certainly the one we have most control over. There are definitely ways to encourage a healthier party democracy. In the Indian context, I think the appointed state offices, like constitutional offices, and offices in guarantor institutions like Election Commissioners, Supreme Court and High Court judges, should be appointed through a Sri Lanka-style Constitutional Council, which has equal representation from the ruling party and the opposition. We cannot afford to continue in a democratic system where the ruling party calls the shots and the opposition only has a tiny voice. The say has to be equal for a democracy to function, because the ruling party has all the incentive to game the system in its favour. There are things we need to change to increase the likelihood that the democracy will survive. The key is not to make one side of the game also the referee of that game. That is just obviously bad design. There are things that need to change to improve the quality of democracy in our country and to arrest the decline of democracy in our country. 

Rachana: I think to connect the concerns we have been raising about constitutional design, I want to bring the focus back to the way you theorize political parties. You theorise them as bidirectional and infused with plenary power, and you describe them as intermediaries. I want to understand what distinguishes this particular characterization of parties from, for instance, a basic expectation of electoral responsiveness. Why is this particular characterisation of parties constitutionally significant? 

Professor Khaitan: The point of political parties is to act as a bridge between the state and the people. That is why parties come about. All of the costs that I mentioned arise because it is expensive and complex to connect a large number of people with the offices of the state. Given this bridging purpose of parties, I talk about two distinct features of parties. The first is that they act as a bidirectional intermediary. That is, parties do not just collect, mobilize, and organize popular will to convey it to the state. They also, in the case of a ruling party, justify state policies to the people, and if in opposition, criticize those policies. It’s bidirectional in that sense. They convert popular opinion to state policy and vice versa. Responsiveness is integral to and embedded in this bidirectional function. A party that fails to articulate popular will into state policy is unlikely to do well politically, unless it comes up with a system of patronage politics, which is symptomatic of an unhealthy party system.

The second idea, which is often not noticed in existing literature and which distinguishes parties from NGOs, is that parties of governance are plenary in nature. The state ought to have policies addressing almost everything: health, education, defense, sanitation, foreign affairs, and so on. Singular issue parties in governance in the long term are both undesirable and infeasible. These types of parties may act like NGOs or pressure groups, they might win one or two seats but if they enter government, they typically transform into plenary parties. They will start caring about a broad spectrum of issues and will learn the importance of political packaging. Single-issue parties, which focus on maximizing one goal, parties of governance must balance multiple objectives and constituencies. The political system incentivizes this transformation. Over time, single-issue parties are obsessed with one goal, and don’t do packaging. They want to maximise the one thing they care about. Whereas parties of governance have to learn packaging and the system will teach them that. These are examples of second-order regulations – you do not need to legislate that parties must be plenary. The system makes them plenary. Political survival, and the need to flourish politically makes them plenary. The need to function bidirectionally arises from the need to flourish politically. Parties that perform these functions badly will generally die out or won’t do particularly well. These are not things you need to legislate on. In a well-functioning system these are things that will happen automatically. 

Rachana: I have a quick follow-up on that. I was wondering, what is our specific interest in understanding parties as plenary and recognizing that they have this range of power? Why are we theorizing about them when we’re concerned with constitutional law? And how does that become important when we want to regulate parties as a unit at the level of the constitution?

Professor Khaitan: Whether a constitution says anything about parties or not, and here I mean both Big-C and Small-C constitutions, it matters. It matters to the type of party system we are going to have. I’ve already given you multiple examples. The choice between a presidential and a parliamentary system will give you completely different party systems. We have debated presidential and parliamentary systems in India for a very long time, in complete ignorance of what their implications are for the party system. The kind of electoral system you have, for example ranked-choice voting, will have completely different implications for the kind of party system you’ll have. And again, we discuss electoral systems in complete ignorance of the implications it will have for the party system. We’ve debated federalism for a very long time, but we have not realized that India started with a federal state but a unitary party system. What changed in the 1960s and ’70s is that not just the state but the polity itself federalized. We saw the emergence of state parties. So, the party system complicates and scrambles all of your normative assumptions behind institutional design. If you do institutional design without thinking about parties, or in ignorance of parties, you will still have implications for party systems, but they will not be thought through. They will be ignorant, under-educated, and might even surprise you in welcome and unwelcome ways, if you have never thought about what a particular institutional design you’re adopting for your country.

A first-past-the-post system in a unitary party system follows Duverger’s Law – that under first-past-the-post, you’re likely to get two main parties, at most three, with a third small party. Now, India follows Duverger’s Law, but India is not a single polity. Since the Indian polity federalized, Duverger’s Law is followed in each of India’s polities. Uttar Pradesh is an exception, but outside of that every state has two or three main parties of contention. But what that means at the national level is a very fragmented party system, with dozens of parties. In the US, the state system is federal, but the party system is still unitary. The party system is not federalized. It’s the same two parties which remain the key parties of governance at the state level and in every state. So, Duverger’s Law still holds there, it’s still two parties. And under that system, at least the assumption is that when there are two main parties, they will build big tents. But once the party system federalizes, then the implication for democracy, that first-past-the-post will lead to big tent parties, no longer holds. It becomes more complicated, when you can start winning elections with 30%-31% of the vote; you no longer need to build big tent parties. You don’t need to take everyone along. You can win elections by polarizing.

So the main argument for first-past-the-post, and against PR, is that you have more big tent parties. But that need not hold in certain types of systems. Now, if you understood this, then you might opt for ranked-choice voting in order not to have anti-minority parties. So this is why we need to think about parties, because every institutional choice you make has implications for the party system, and the party system you have has implications for institutional design. We know that in a unitary party system, without federal parties, federalism was weak at the institutional level. Indian federalism only flourished after the party system also federalized, after the polity federalized. So if you like federalism, you might worry about a unitary party system, where state chief ministers could be dismissed by the party high command sitting in Delhi. Even today, in national parties, we just saw the BJP chief of Tamil Nadu was removed by the high command from Delhi. These things matter in a democracy. And we better not ignore the key ingredient, a key factor that scrambles all of these choices. We do so at our own peril, and at democracy’s peril. That is why we need to think about parties when we are thinking about institutional design.

Rachana: Thank you so much for that. I think that really puts into perspective why we should be considering political parties as a unit when we make those norms in the first place. I just wanted to conclude with one last question, which I think you may have already covered partially. Because this chapter calls for constitutional imagination to center parties, not as incidental political actors but as primary constitutional subjects, and if we were to take that sort of reconceptualization seriously, how would our reading of constitutional democracy and how we understand democracy itself shift? I just want to ask if you have any particular takeaways that you would like people to leave with once they read this chapter, or anything you’d like to add to that question.

Professor Khaitan: Well, mainly I think the first thing we need to do is to move away from an exclusive focus on courts. Courts are a key player, an important player, and I like courts, and I like what they ought to do when they do it well in a constitutional scheme. But they suck all the oxygen in the room. They get our exclusive attention as constitutional scholars at the expense of other extremely important constitutional actors. When was the last time we saw a course that focuses on election commissions, or on parliamentary committees, or political parties, or the office of governors? And not just on what courts have done in relation to these institutions, but on these institutions as such: how they function, what are their internal modes of functioning, on the bureaucracy, on the civil service, on the Rajya Sabha, on legislative councils that five odd states have, on Union Territories. We just do not pay enough attention, and in a democracy, scholars in part have a role in any system or any scheme of separation of powers. Legal scholars in particular, because they understand and speak the language of power, which is law, have the ability to check power and its exercise by checking its reasons. When was the last time we analyzed the reports or judgments or decisions of election commissioners or the auditor general to see if they are compliant with constitutional logic and morality? I think that is probably the big-picture lesson: that this is not just an intellectual failure, it’s also a political failure and a moral failure to ignore constitutional actors other than courts in our study of how the constitution functions.

Rachana: I think that’s certainly been my biggest takeaway from this conversation – that we can constitutionalize norms outside the courtroom and outside our limited understanding of where we think these norms can be codified, and to think about broader questions of institutional design, and taking all kinds actors into the fold when we think about those choices, rather than just sort of coming up with a basic infrastructure and then retrofitting parties within it. I think it comes back to, like you said, just a matter of shoddy design, and we really need to be thinking about those bigger questions. 

Uday: I just had one question. When you wrote this chapter, which country would you say you had in your mind most when you discussed political parties?

Professor Khaitan: When we do constitutional theory which abstracts out of particular examples, the one thing we cannot, and must not do, is focus on one country. Because a data set of one will result in skewed, bad theory. And we have seen for a very long time American scholars who, in almost complete ignorance of all other systems, present scholarship and theorizing about constitutional institutions not as American constitutional theory but as global, universal constitutional theory. On the other hand, I am an individual with my own intellectual limitations. I absolutely do not know every constitutional system, or even close to it, that functions in the world. But what I’ve been trying to do over the years, through my teaching, through my self-education, through my readings, is to familiarize myself with a wider canvas of systems around the world. This is not a piece in comparative constitutional law that does systematic and rigorous case selection and focuses on a few jurisdictions. It is a work of constitutional theory. But I would like to think it’s based on some understanding of a wide range of jurisdictions rather than a single jurisdiction. Of course, there are some jurisdictions that you know much better than others. Most people in the constitutional theory field need to know the influential exporting jurisdictions, mainly the UK and the US, and a little bit of Germany. Obviously, because of my own biography, I know India and South Asia more generally. But my effort has been to familiarize myself not just with the big hitters and the parochial jurisdictions, but also to educate myself on what’s going on in Africa and Latin America and other such places. Having said that, I of course do not at all insist that everything here is easily applicable everywhere else. These are factors to be considered. This is not a manifesto or a blueprint.

Rachana: Thank you for the discussion, Professor Khaitan.

Professor Khaitan: Thank you.