Donald Horowitz on the Formation of Democratic Constitutions

Donald Horowitz
Donald Horowitz

Donald Horowitz joins the podcast to discuss the formation of constitutions in democracies. Donald is the James B. Duke Professor of Law and Political Science Emeritus at Duke University. His most recent book is Constitutional Processes and Democratic Commitment.

Listen on SpotifyListen on AppleListen on Google Listen on Stitcher

The most beautiful thing that happened in Indonesia, by the way, which was a polarized society along religious lines more than anything else, was that by the end of the proceedings, everybody knew what everybody else’s problems were, what everyone else’s constituencies wanted. They knew if X noticed that Y was making a demand, before long X figured out what was behind the demand and why Y had to make it and whether it was a real demand or whether it was made just for the sake of being on record.

Donald Horowitz

Key Highlights Include

  • Accounts of constitutional formation in Tunisia, Indonesia, and Malaysia
  • The role of consensus
  • The challenges of negotiated constitutions
  • The need for an inclusive process
  • Why citizen participation is not always beneficial

Podcast Transcript

Ever since I was a kid, I have had a special reverence for the American constitution. After over 200 years it has almost a mythic quality attached to it. But most constitutions do not last long. Tom Ginsburg and Aziz Huq note, “Of the 964 constitutions promulgated during the last two centuries, only 406 remained in force for more than a decade.” 

Donald Horowitz has studied the formation of constitutions in a variety of contexts and situations. In our conversation he references Tunisia, Indonesia, Sri Lanka, Nepal, and India and he has plenty more in his latest book Constitutional Processes and Democratic Commitment. Donald is a Professor of Law and Political Science Emeritus at Duke University. He has studied the creation of constitutions particularly in societies divided by ethnic conflict for decades. 

If you’re like me, you want to understand why some constitutions succeed and others don’t, you want to learn about the formation of constitutions in other countries, and you want to know how to make the process more democratic. This conversation touches on all of that and more. 

And like always I have provided a full transcript at because this conversation references people, places, and ideas you might be learning about for the first time. You can also send me your questions or thoughts about the podcast to my email at But for now… This is my conversation with Donald Horowitz… 


Donald Horowitz, welcome to the Democracy Paradox.

Donald Horowitz

Thank you very much, Justin.


Well, Don. It’s a real honor to have you on the podcast. You’re somebody who I’ve read before, who I’ve always kept an eye on. So, I was real excited to see that your book had just come out and it was absolutely a pleasure to read. So, thank you so much for joining me.

Donald Horowitz

Well, it’s very good of you to invite me and very good of you to say that. I do hope that all of my academic opponents share your view about the book.


Well, Don, I want to start with Tunisia because it feels like a paradigmatic case in a lot of ways. Unfortunately, it’s faced some recent challenges. But let’s set those aside for the moment because I want to focus on its earlier constitutional process that many scholars, including yourself, have praised. Can you walk us through the process where they designed their constitution and explain why it really did offer some very early signs for optimism?

Donald Horowitz

Yes, and I still stand by that despite recent events. You have to start back a little bit before the actual constitution and even before the Arab spring. There was and is a political party in Tunisia, Ennahda, which was described as an Islamist party. I think that description is no longer accurate. I think it’s fair to call it now an Islamic party and I’m not alone in thinking that. Some real Tunisia experts have come to that conclusion. Because the Ben Ali regime was a very harsh regime, a lot of people went into exile including a lot of the leadership of Ennahda.

Many of the functionaries, however were either left in Tunisia in a kind of isolation, because they couldn’t meet easily, the Ben Ali regime was kind of totalitarian regime really, it certainly was a police state, and others of them, of course, were imprisoned in Tunisia. So, the communication between the people in exile who were disproportionately interestingly enough, not in France, which would have been logical given Tunisia’s colonial background, but in London, had one experience and the people left in Tunisia had another experience or a couple of other experiences. And they were dissonant. They weren’t in touch with each other until they got back after the Arab Spring, until Ben Ali was chased out. And even then, they began with a certain number of disagreements when the constitutional process was convened.

It was done by an elected legislature and an expert committee at the beck and call of the legislature. It was pretty much agreed, and I say pretty much with some emphasis, because there wasn’t perfect agreement on the Islamic side, rather than the secular side. The secular side wanted a secular constitution. That’s not a surprise. But I should say the secular sides because it was a very fragmented opposition, extremely fragmented. In the elections, Ennahda, the Islamic Party, got 44% of the seats. They were proportional representation elections, but the fragmented opposition got fragmentary seats. They got the remainder of the seats, but they were very badly fragmented, the secular side was.

It was more or less agreed, as I said, that the constitution would be a secular constitution, but there were some people in Ennahda who probably didn’t really believe that at the outset. I think by the end, they were pretty well convinced that that was the task on which they were embarked, even though they had some rough edges at the end in some provisions, especially if you have a look at Article 6. It says that the job of the constitution is to protect the sacred. That doesn’t sound too secular to just about anybody. But in any case, those were sort of bits and pieces they couldn’t quite edit out at the end.

But it was very widely agreed that this would be a secular constitution and as a matter of fact, the leader of Ennahda, Rached Ghannouchi, wanted to have a British parliamentary system. He was very impressed with the way the British system worked. Unfortunately, the secular opposition said, ‘No, that’s not good because actually what’s going to happen then is you’re going to have a plurality or even a majority of the legislature. And you’ll just run things. You’ll have a parliamentary system and the majority will run over us.’ So, there was a very big compromise. Ghannouchi really was quite strong on the parliamentary system and he eventually agreed to a French style mixed system with a president and a prime minister. This is, I’m sure, something that he is living to regret these days, since the president has made a coup against the constitution.

They really did have a very large area of agreement and when they couldn’t agree, they compromised. For example, they especially compromised as I say on the form of government and on a number of other matters. But for the most part, they really had a consensus. They had a breakdown in the middle of the proceedings which was mediated by four civil society organizations that had managed to survive the Ben Ali period and very effectively brought the two sides back together again after a very severe breakdown in the negotiations. And I say negotiation, but this was not really a negotiated constitution with the exception of a few provisions. And at the end, the consensus was validated by the final vote. The final vote was well over 95% in favor of the constitution. There were a few dissenters, but really a very small number.

So, this was a seriously debated, argued, rather than bargained constitution. They took plenty of time. They took two years. In the background, however, they began to be concerned, because they were worried about the coup that had displaced Morsi in Egypt. That was really something that was overhanging the proceedings altogether. And so, that’s one reason that Section 6 got only as far as it got and it didn’t get the sacred out of there. I think protection of the sacred would have been ultimately eliminated if they’d had a whole lot more time, but that’s just, of course, a speculation. In any case, it’s a kind of paradigm case. And, as you know, I’ve got three paradigm cases in the book of constitutions that are made by consensus rather than made by negotiation or by simply voting on them.


Why is consensus so important? Obviously, when we think about democracy, a lot of people think of it as being a form of majority rule. So, your paradigmatic examples you just mentioned, the key attribute to all of them was that the constitutional process was devised, not through negotiation, not through a strict up and down majority rule vote, rather it was developed through consensus with near unanimous votes behind them. Why is that so important?

Donald Horowitz

Because in order to be achieved, consensus requires extensive reasoning, extensive argument before one can reach agreement. It’s a whole lot easier to negotiate in a certain way, because I know what I want, you know what you want. Let’s see if we can make a deal. I’ll come back to negotiation in a minute, because I want to show that it’s a bit more vulnerable to disaffection than consensus arrangements are. But once we start reasoning about things, we know first of all, it’s going to take us longer. In other words, we’re going to get to know each other’s interests and sympathies and preferences a lot better than if we simply sit down and say, ‘Look, I want this and you want that. And I understand it. And you can have some of this, if you give me some of that.’

So, the process takes longer. It requires something distinctive. That is deliberation. And deliberation is very different from negotiation in exactly the senses that I mentioned. It has to do with reasoning and persuasion. And consensus means I’m persuaded of a point that I perhaps wasn’t persuaded of before and likewise, so are you. And it doesn’t just have to do with my understanding that you want something that I don’t want. So, it’s preferable. Furthermore, there is a very large literature on deliberative democracy. And the deliberative democracy people have found in deliberative proceedings, more public-spirited arguments are made. People don’t just argue their own position. They argue what’s best for everybody. And, of course, what you want in a constitution ideally is the way forward for living together in politics.

I think, that framing is very different from the way forward when you got something and I got something else, or when I just simply out voted you or you out voted me. So, not only do they make more public-spirited arguments in deliberative democracy experiments, but they also listen more attentively. And we have in this country, a lot of studies of juries. And juries have to reach quite a substantial consensus. And the studies of jury deliberations show that there’s far more attention paid to the opposing arguments that are made early in the proceedings when the standard of decision is unanimity. And there’s also much more attention to the details of the argument and minorities have a bigger say.

That is, if I can out vote you and we purport to be having a discussion about where we’re going in the future, I have in the back of my mind that if you’re not going to give in to me, we’ll just take a vote and we’re going to win. Because the one thing we know politicians are very good at, it’s counting votes. So, they will know before the decision time what the decision is likely to be, because they can count up the votes. And, of course, vote counting occurs in consensual proceedings as well as it does in others. But the significance of the vote is really very, very different. And we know from the three cases that I’ve mentioned, India, Indonesia, and Tunisia, that at the end of the day, the vote in favor was just overwhelmingly positive in all of those cases.


So, is consensus more important when you’re constructing a constitution than when you’re doing ordinary legislation or should we be finding ways to make the regular governing process more consensus oriented?

Donald Horowitz

Well, that would be a very good thing. And if rules were oriented in that direction, that would also be a good thing. But, of course, in many legislatures when they figured out that they don’t really have a consensus, they’ve shaped the rules to facilitate voting or to facilitate some kind of compromise. I wanted to say one other thing about consensus versus compromise. Compromises are, I think, more vulnerable to reneging or to willful misinterpretation of what was agreed. There are two different kinds of compromises.

One of them is what you might call log rolling. I want X and you want Y where X and Y are just not on the same axis. They’re not commensurables. I want more food. You want more entertainment on television. I want more sports programs and you want more dramas. If you and I are living in the same household we can agree, I think, that you’re going to have two hours of sports and I’m going to have two hours of drama. That’s pretty easy to do and if there’s a later disagreement, we can do an adjustment. But if, on the other hand, we make a deal which concerns delivery in the future, and I get my part, but you don’t get what you anticipated. You are very likely to be discontented and to want to overthrow the deal and to revoke it or to redo it.

And in a constitution, yes you can and should from time to time redo things. There’s no doubt about that. But what you don’t want is people coming up and saying, ‘Well, you promised that this would happen and it didn’t happen. And therefore, we have to alter the constitution for this purpose.’ It’s much less harmful to do that with respect to legislation, because legislation deals not with the generalities of living together and politics, but with particularities of living together. So, that’s the log rolling piece. Also, log rolling doesn’t have any real rationale. You get what you want. I get what I want. And it may very well be dissonant with the rest of the constitution. So, that is to say the scheme of the constitution can be undone by log rolling. I can show you some examples of that.

The other bit though is what you would call splitting the difference. Where we’re talking about the same thing, it’s commensurable. And where you are going to split the difference it’s very important to know where we’re splitting it and what’s going to happen afterwards. There’s a great example of this in the book which is Article 153 of the Malaysian Constitution which guarantees to the Malays, who regard themselves as the indigenous people of the country, that they have a special position and that was underscored by some particular areas in which they would have some preferential treatment. But the legitimate interests of the other communities, which in those days were just about 50%, the Malays were about 49%.

But this 50-50 country made a deal which gave priority to the Malays in certain sectors. But said, ‘But for the rest, the legitimate interest of the other communities will be recognized.’ And citizenship was granted to the members of the other communities, most of whom didn’t have it previously. Later on, when Malay nationalism got going in a very serious and unhealthy way in this case, what happened was a lot of revisionists decided that the citizenship that was granted to the non-Malays was the outer limit of their rights. That is to say the privileges that the Malays enjoyed could be expanded and, indeed, they were expanded drastically in the early 1970s. And expanded even more in the implementation of the expansion that followed.

And the rights of the non-Malays, almost all of them were not Muslims unlike the Malays, were contracted and that extended also to the religious rights which were very much in jeopardy especially over the last 8 or 10 years. So, that’s another way in which compromise is not as good as consensus for constitutional purposes because it’s more subject to reneging. The interpretation can be willfully wrong and it was willfully wrong in Malaysia. Even in the courts the interpretation of the religion clause, which is a slightly different clause, until fairly recently has been willfully wrong and to the disadvantages of non-Muslims in the country.

So, it’s not merely that consensus has some positive attributes. It’s that compromise, while it has positive attributes without a doubt, it gets you to an agreement, can have some negative attributes that are less likely to occur in the case of consensus outcomes.


So, how should constitution makers proceed if consensus is not possible? It’s odd the way that we’re frowning on compromise, because normally when we talk about politics, we’re asking both sides, ‘Why can’t you find a compromise between the two sides?’ But here we’re saying that it’s not compromise we’re looking for, it’s consensus. But it doesn’t feel like consensus is always possible. So, how do we move forward, if consensus can’t happen and a deal needs to get done?

Donald Horowitz

Oh, I don’t mean to denigrate compromise altogether. I’m only comparing it to consensus. The objective is consensus, but where consensus isn’t possible after extended deliberation, argumentation, and persuasion, then of course the next step is compromise. And, I think, that’s fair enough. But if you’re going to compromise, you can take precautions against reneging of the kind I’ve described. For example, a very good history of exactly what was agreed is a useful thing. Sometimes even embedding the history somehow in the clause, if it can be done economically is a good thing. And then, finally, if you can’t even find a compromise, you may have to take some votes. And most constitutional procedures that are specified in the previous constitution have room for votes.

Not only do they have room, but they require that votes be taken at the conclusion of the negotiation, if that’s what it is, and that the voting be done by super majority, usually by two-thirds, sometimes three quarters, occasionally by majority vote. But if majority vote, sometimes followed by a referendum with all citizens voting. So, I’m not denigrating the default. If you have to default to compromise, then you do it. And as I said, in the case of Tunisia, where they agreed on most things, they did have a very big compromise with respect to what the form of government was going to be. Whether it was going to be semi-presidential or whether it was going to be parliamentary.


Now getting back to Tunisia, Elizabeth Nugent made the case that one of the reasons why the Tunisian constitutional process was able to find such strong consensus in the process was that it did not occur in an environment of severe polarization. Whereas she compares it to Egypt, which did have a very severely polarized political environment. Is it possible to create new constitutions in an environment of severe polarization?

Donald Horowitz

Well, let me say first, the beginnings of that process were characterized by severe polarization. The secularists thoroughly distrusted the Islamic party. As a matter of fact, even in Washington, before the constitutional proceedings kept going, I went to a seminar, in which a number of a secular Tunisians spoke. And they said that Ennahda, the Islamic Party, was speaking in code. That when they spoke of women’s rights, yeah, they meant women’s rights once they wear the Hijab and once, they stay home without a male guardian. And there was nothing that the Ennahda people could do to dissuade them except by their actual behavior during the proceedings. So, the proceeding started off very badly that way.

And on the Ennahda side, as I said earlier, there was not complete unanimity on a secular constitution at the outset. The people in London, the exiles in general, had to persuade the other members of the party who had stayed behind in Tunisia during all of those terrible years of the Ben Ali regime that the constitution was going to be a secular constitution. And, I think, in the end, that’s exactly what happened. I’m sure there were still some holdouts on the Ennahda side without a doubt and probably people now saying, ‘See what you got us into.’ Now that there’s been a coup. But in spite of that, I think, it did start out that way.

What really was the difference, I think, between Egypt and Tunisia, and I’m not a middle east expert, but the difference was that, in Tunisia, the military and the police were thoroughly detested. So, that one could have the idea that one was writing on a blank slate, but in Egypt the atmosphere was very definitely more polarized, because the military was not detested. Remember these were the heroes of the 1973 war with Israel. And as a result, the military still enjoyed some respect. Now it may be that the person heading the military and civilian government did not have great respect, but among the population the military still had some respect. So, that was a very severely divided situation.

There was nobody, I think, in the proceedings in Tunisia who was willing to say a good word about Ben Ali or about the previous regime or about the police and the atmosphere was really quite different. And there was a real fear in Tunisia, as I said, that they might go down the way Morsi went down, if they stayed a little too long. So, they did wrap things up pretty quickly toward the end and, I think, largely for that reason.


So, you’ve mentioned a few times that Tunisia has had a recent coup, but at the same time, it’s lasted a lot longer than some other constitutions. It lasted much longer than Egypt, for example. But should we be blaming Tunisia’s recent problems on its constitution or can we look solely at political actors who took advantage of the moment and maybe took advantages of opportunities within their political system?

Donald Horowitz

Well, it’s more the latter than the former, that’s for sure. The president who made this coup has specified in his list of powers, powers that he doesn’t actually enjoy under the constitution with respect to declaring emergencies and the conditions for emergencies. He by the way, was a member of the drafting committee, rather than an elected member of the legislature that sat as a constitutional assembly. There are also allegations of foreign involvement that are not to be taken lightly. There’re some allegations that some Gulf countries allegations by experts, and not by me, allegations that there was involvement by some Gulf states that were fearful of Islamic democracy in the Middle East. And therefore, didn’t want this precedential case to survive.

Now, if that’s true, of course, that may strengthen the hand of the president, because perhaps he’s got some resources that he might not otherwise have and maybe that tempted him into what he’s done. I haven’t studied the coup, I can’t claim to be an expert on the coup, but I think it doesn’t have very much to do with the constitution.

Certainly, it doesn’t have anything to do with the constitutional process, because, I think, secularists and the Islamic Party both are alarmed by these developments and unhappy about them. And you know that the Islamic Party took its people off the streets so as to prevent bloodshed as a result, because bloodshed was certainly threatened. And Ghannouchi said, ‘Let’s negotiate our way out of this.’ Well, so far the negotiation has improved fruitful, and maybe, it won’t prove fruitful for a very long time.


I always go back to Sheri Berman’s work. Her book Democracy and Dictatorship in Europe really emphasized that countries don’t just have authoritarian legacies, but they’re also democratic legacies. It oftentimes takes multiple attempts before a country gets it right. And this is really Tunisia’s first attempt. So, I don’t think that people should say that it’s impossible for them to democratize or look down on their effort because it may take 2, 3, 4 times. It took France at least four times before they got democracy right. They’re on the Fifth Republic now.

So, I’d like to move on to a different country. Chile’s undergoing a constitutional convention. And it’s very different than Tunisia, because they were in a transition from authoritarianism to democracy. Chile has been a democracy now for quite some time, for decades, but their constitution is an inheritance from the Pinochet regime. So, there’s a lot of distaste with the constitution that they have, at least by some people, particularly on the left. What are your thoughts on the Chilean constitutional convention that is in process, more or less, right now?

Donald Horowitz

I haven’t been following it day by day at all, only in very general terms. They made, I think, one terrific mistake actually. If you think that the real threats to democracy come more from politicians these days than they do from the military, and I think the real threats come from defecting politicians, people who renege on their commitments. And you can see that military coups are much less frequent now than they had been in earlier decades. What the Chileans did at the outset was to prevent sitting politicians from being part of the constitutional convention.

I think that’s a mistake, because politicians are the ones who will mess it up. Politicians are the ones who have to decide on how we’re going to live together in politics, because in the end, this constitution will have to give rise to elections for a legislature and maybe elections for other offices, depending on what they decide on. And for that reason, politicians ought to be the ones who are making it, so they understand each other’s positions.

The most beautiful thing that happened in Indonesia, by the way, which was a polarized society along religious lines more than anything else, was that by the end of the proceedings, everybody knew what everybody else’s problems were, what everyone else’s constituencies wanted. They knew if X noticed that Y was making a demand, before long X figured out what was behind the demand and why Y had to make it and whether it was a real demand or whether it was made just for the sake of being on record. This is invaluable information for legislatures. And so, when the elections occurred afterwards, those politicians understood what was going on. I don’t mean to be too dramatic and say that there’s a cathartic element to constitution making.

But it is the case that you can learn a lot about your opponents and you can learn that they’re not really your enemies too, if you’re involved in this process. Once again, you may not learn that, but you have the capacity to figure out what’s possible. What’s not possible. What’s reasonable. What’s not reasonable. Why your opponents think differently and so on. So, that’s one mistake that I think the Chileans made. For the rest. I, don’t know where they’re going to come out. And so, it’s very difficult to say from the outside without having a very close look at the proceedings.


So, you hinted at some of the successes within the Indonesian constitutional process. It’s a country that most of us don’t know as much about. It’s an incredibly large country. It’s one of the largest democracies in the world. Can you fill in some of the gaps? Tell us the story of how the constitutional process came to be in Indonesia.

Donald Horowitz

Well, the Indonesians were lucky in one way and unlucky in another. They were unlucky because they had to start out, as many constitution makers do, in the middle of a very bad crisis, the Asian financial crisis. The Indonesian constitution is an indirect product of the Asian financial crisis of 1997, 98 and Suharto had been urged to accept recommendations of the International Monetary Fund. He proved to be very stubborn about it. The Indonesian economy was falling apart and eventually people from his own party came to him and also from the military, which had a significant role because he was originally a general who came to power in a coup after the coup that had undone Sukarno in 1966.

Suharto was confronted by politicians and by the military and told his time was up and he finally was induced to leave office, but there were crowds on the streets. The army had been shooting into crowds and it had killed some people. In another case, Suharto’s then son-in-law who was a Colonel commanding a renegade military operation killed a lot of people apparently by setting a fire in a department store which was filled with people. So, this was a very serious crisis, not just of governance, but of civil order. And there were students who were outside, and even made their way inside parliament at the invitation of some parliamentarians, who were demanding Suharto be hanged, that revolutionary committees be established.

So, that’s the background in which Suharto‘s successors took power. And they had an authoritarian constitution, but the constitution had one really big saving grace and that was it had a very clear amendment clause. So, that the body to do the amending was specified and so, with the standards for amendment, which were a super majority vote and that would produce the amendments.

And I should say by the way, Suharto’s vice-president, BJ Habibie, who had never expressed a democratic thought in his life, took power as the successor, the legitimate successor to Suharto. He was a very unpopular man with the military. He was a very unpopular man in his own party. He was just a very unpopular man. He was educated in Germany and he thought of himself as a kind of builder of Volkswagen type enterprises in Indonesia, the kind of big enterprises that are terrifically wasteful when they’re constructed by states and seeking prestige projects. So, Habibie had very good advice as it happened, and he declared immediately, well almost immediately, that there would be elections within a year and that democracy was the destination.

Whereupon his opponents from other political parties got together and said, ‘Well, despite our differences, we’re okay with that. We’ll go for the democratic elections.’ And slowly the crisis was calmed down and this legislature, which had not been really legitimately elected, proceeded to take some early measures to calm things down and to do some reforms. It was obviously a reforming inclination that took over and eventually the crowds disappeared from the streets. The elections were duly held and they were more or less fair elections for the first time since 1955. That’s the close of the last elections that Indonesia had that had any credibility at all. They had plenty of elections, but not with credibility. And the result was that there was a fragmented legislature that reflected the various streams of thought present among both Muslims and among non-Muslims and secular nationalist people.

They proceeded then slowly to get into constitution making and they then deliberated over everything. And it just took forever to get it done. They produce four enormous amendments which dealt with just about all of the important features of democracy and of civil rights and civil liberties. And the result of those four amendments over the course of almost five years, four and a half years, something like that. The sum and substance of those amendments was to transform the constitution into a democratic constitution in which something like 87% of the words were changed, but it’s still the 1945 constitution of Indonesia except that 87% of the words are different.

And so, those who relish continuity and there were some secular nationalists who were attached to the constitution because they were diehard Sukarnoists, which is, I think, an unfortunate state to be in. But there they were and those people proceeded to want merely to amend. And in some cases, they didn’t really want amendment. And so, slowly they were persuaded to opt for amendment of the constitution rather than to have a wholly new constitution, but essentially, it’s a new constitution. It’s unfortunately not always counted as a new constitution in quantitative studies.


That’s not uncommon though. A lot of countries, a lot of the post-communist countries, Hungary is one that immediately comes to mind that held onto their old constitution, but made dramatic amendments that transformed it from authoritarian into democratic constitutions. And many countries, especially in the third wave, inherited constitutions that oftentimes were even designed by the authoritarian dictators or a single party that transitioned the country into a democracy. Chile’s a perfect example that inherited a constitution from the Pinochet regime as it transitioned into a democracy. Should newly formed democracies though, should they redesign new constitutions for themselves or are there advantages to just working with an existing constitution that they inherit?

Donald Horowitz

Well, if you mean for the long haul, I would say that there are great disadvantages to working with an authoritarian constitution. They’ll have to do a lot of work arounds, which are really going to be unconstitutional. It’s not a good idea to encourage people to think that the provisions of the constitution are not legitimate. And they are not legitimate in the authoritarian cases. But Indonesia is a perfect case to emphasize this. The Indonesian constitution had so many things wrong with it, it would take us hours to describe what was wrong with it. It was actually written very much under the auspices of the Japanese occupation. So, it had a kind of Japanese overlay of integralist constitution making. Never mind what that means, but it’s not nice.

t had, however a very serviceable element. It had a body that was used to making constitutional amendments and had the authority to make constitutional amendments. Once you reelect that body in a democratic election, that authority can be used. And the old Indonesian constitution was used to transform itself. And if you have a serviceable constitution, that’s the threshold inquiry I would always want to make. Is this serviceable and can it be transformed or do we really need to start from scratch? And if so, do we have an idea of what really needs to go? And they won’t really know everything that needs to go until they’re well into the process too.

So sometimes that’s a tentative judgment they’ll have to make. But serviceable constitutions that can then be utilized to produce something else and whether you call a new constitution, a new constitution, or whether you call it an amended constitution depends very much on whether there are still some attachments to the old constitution or to the old regime that need to be mollified, even as we’re making these changes. And that’s the Indonesian story. Those people are now virtually gone, but they certainly did need some mollification. You can do this using that constitution, but the threshold test is always serviceability I would say.


So, constitutions are not always democratizing events. David Landau and Rosalynn Dickson have written, “All too frequently, constitutional processes are used to promote distinctly anti-democratic ends or to advance the cause of would be autocrats by removing democratic checks and balances on the exercise of political power.” And we’ve seen that in countries from around the world. We could point to Hungary and we could point to Venezuela when they had their constitutional convention. We could point to a lot of cases within Latin America that we look at the constitutional process as really being a form of backsliding of democracy. How can we identify whether the constitutional process is legitimately a democratizing event rather than something that is a grab for power by political leaders?

Donald Horowitz

Only with difficulty. I’m not sure I have a general rule for this, but I do think that you’ll be able to detect this along the way if legitimate processes are perverted. For one thing, I think it’s very important if we’re going to have consensual or even negotiated outcomes, rather than simply arbitrary votes, it’s very important that the constitution makers be elected and the literature says that inclusion of elected representatives is very important and is positively associated with democratic outcomes. So, one way to tell is to have these people who’ve been elected by free and fair elections. If they have been elected by free and fair elections, that’s a very good start. I don’t think that’s enough. I think you need inclusion, deliberation and consensus and all of the rest that we’ve discussed.

But some other aspects that are said to be associated with democratic outcomes like public participation have no evidence in support of the proposition that they are indeed correlated with democratic outcomes. So, you can look, first of all, to see whether the election was good and if the election was free and fair. The odds are pretty good, unless the authoritarian people somehow managed to make a coup inside the body making the decisions. Now, that’s not perfect I want to tell you.

If you look carefully at the Nepal constitution, it absolutely had all the preconditions for democratic outcome, all of them. There were two constituent assemblies that sat in Nepal. Each one sat for years. The first one actually ran so far out of time that the third go round for an extension of time the court in charge of granting the extension decided, ‘No, you can’t have any more time.’ So, they eventually had to have further elections and yet another attempt to get a consensus. And they were very far along when events, namely, a very severe earthquake took over the process and resulted in a lot of public demands that the process be finished, so they could get on with the work of repair and reconstruction.

And at that point with some people in the convention, whose preferred outcome wasn’t a hundred percent democratic, did was to engineer a vote that would produce a federal state that was not in accordance with the interests of almost a third of the population. The people living along the plane that abuts India, the so-called Madheshis who are stigmatized by other Nepalese as Indians. They’re not Indians, they’re actually Nepalese.

But the constitution that was produced created a terrible violent backlash with hundreds of people killed. And it went on for months and months and months, because Nepal is a caste ridden and ethnicity ridden society. And the high caste and the privileged ethnic groups united at the end to deprive the Madheshis, who are as I intimated not a preferred people in the country, and to deprive also another indigenous group also in that same area of the Madhesh region to deprive them of things that, I think, it’s fair to say that they would legitimately have had in a decent dispensation.

So, things can go wrong. There’s no doubt about it. To some degree, some of those things have been righted by further legislation. But they’re never going to be completely right until people in the Madesh region have a great deal more satisfaction with the outcome than they do have.


So, Don you hinted at a point that I thought was really central in the book. You said that participation is not as important as inclusion and in the book, you actually go even farther where it’s not just inclusion through elections, but it’s inclusion of different groups, making sure that they have representation during the process. Why is inclusion so much more important than citizen participation?

Donald Horowitz

Well, let’s take them separately. Many countries, especially in the developing world, have important ethnic divisions. And in countries of that sort, which happened to be the ones I’ve spent most of my life studying, in countries of that sort, you want to be sure that you haven’t excluded a significant ethnic group. By significant, I don’t mean merely large in size, but I mean a group, for example, that’s small in size, but has some serious grievances against the way politics been carried on before.

So, you’d want to have an electoral system that makes sure to embrace them. If they’re widely dispersed, you would almost surely want to have a proportional electoral system for that. Although I’m not a very big fan of proportional systems for legislatures after the fact of constitution making in such societies, because you get a lot of fragmentation or even polarization and you make things very difficult to get done. Where the groups are territorially concentrated, you can do this with first past the post elections. That is to say plurality elections or majoritarian elections or two-round elections on the French model. There are many different ways of doing it, but it’s very important to have inclusions of groups.

Now with respect to citizen participation, as I said before, there’s many studies that test the relationship between citizen participation and democratic outcomes. And there is now literally no evidence that citizen participation generates or is associated with democratic outcomes. Citizen participation has very, very high costs. That is to say you spend a lot of time running around the country at an early stage of the proceedings when you may actually have momentum. And sometimes that momentum is squandered as you travel around the country trying to get an idea of what the public would want.

And lo and behold, when this question has been unpacked, what does the public want? It turns out that even when an educated public is asked, do you want a constitutional court or do you want the court to test the constitutionality of statutes? Do you know what they say? They say, ‘I don’t have the slightest idea because I don’t have any expertise in this field.’ If you ask them any of the other questions about components of the constitution, they say, ‘Sorry, that’s for experts to decide.’ And that’s, if you manage to get to that. But it turns out also that in many countries where there’s been a lot of publicity about this, it turns out that a lot of people don’t even know that this process is taking place.

Now affected interest groups do know and they submit memoranda and they show up at meetings, and as it happens, the memoranda sometimes run to the thousands. I don’t mean thousands of pages. I mean, thousands of memoranda multiplied by however long the average memorandum is. And there’s typically very little expert staff. In fact, very little staff for constitutional proceeding to sort through this mass of material. So, shortcuts are taken. Sometimes there are disputes over how to process it. Everything gets slowed down and the momentum gets lost.

In Sri Lanka, this was actually a crucial, but barely noticed factor in the failure to produce a constitution after the very good 2015 elections that were highly favorable to producing a new and much fairer constitution. Because the previous two constitutions had been made by two parties who happened to be in office sequentially and one produced a constitution to its taste and another produced a constitution to its taste and not since the initial post-colonial constitution in which a British commission sat has a really public constitution been produced. And the chance was missed partly because there was so much delay at the outset in ensuring public participation.

I want to tell you that foreign advisers are almost unanimous in recommending one thing for a constitutional process and that’s public participation. They will compromise on many other things because they don’t have firm views on those other things. But the one firm view that they have, which is insufficiently supported by the literature is that public participation, extensive public participation, produces a more legitimate and democratic outcome. It’s a sad story that so much of external advice goes in this direction rather than in the direction of trying to figure out what institutions might serve it well but that’s the state of affairs at the moment.


Well, it’s unfortunate too, because these are high stakes moments. There’s a lot riding on getting things done right. In the book you write, “Premature constitutional processes can do worse than fail. The results can be so disappointing as to escalate conflict and perhaps discourage future efforts at accommodation.” So, Don, I do want to know here, if the moment is not right, if it doesn’t feel like you can get the process into the right circumstances, should citizens really continue to suffer under a poorly designed constitution or are there really any exceptions to this rule?

Donald Horowitz

Oh, there may be some exceptions, but I think on the whole the rule is probably right. But mind you. When I say premature, I usually mean because there’s violence occurring in the country. Take the case of Somalia or Yemen for that matter. Both of those, I would say, embarked on constitutions prematurely. Somalia has to be a federation. Somalia is a society badly divided along what are called in Somalia clan lines, but what are really. essentially ethnic lines. And there’s no way that Somalia can be ruled from the center without dispersing quite a lot of authority into regions.

So, Somalia has a constitution now, which by the way, was written by a pretty illegitimate process. Somalia has a constitution that’s a federal constitution, but they can’t agree on the demarcation of the federal boundaries. They don’t control all of the territory because of Al-Shabab, the Islamist radical insurrectionist group. it would have been much better if Somalia had done something a little bit less ambitious, namely adopt an interim constitution. You can adopt an interim constitution. Iraq, ironically enough had a very good interim constitution, which had been written by three people and would have served very well for a longer period of time than it was called upon to serve or it would have been a good model for the final constitution, which is a very poor model. But Somalia might’ve functioned with a skeletal interim constitution until things calmed down.

And that might take very long time. Yemen did something that I think was very, very good. Yemen had a dialogue process. And the dialogue process had very good rules. It wasn’t trying to draft any constitution. It just wanted to get a few principles out on the table. What can we agree on? And by agree, they meant by consensus. By the way, when I say consensus, I don’t mean 100% unanimity because that gives spoilers too big a role. But in Yemen, they had a process that was designed to make sure that they had cross ethnic cross political agreement on some basic principles.

The problem is that in Yemen, they also went further and they decided to say, how many units there would be in the north and how many in the south. The north has the Houthis. Everybody now knows about the Houthis. In those days, almost no one knew who the Houthis were, but the Houthis are Shia, of a different set from the Iranian Shia, but sufficiently similar so that the Iranians are supplying and funding them. The Houthis then took up arms against this arrangement because they didn’t like what was going down and so, many southerners had wanted to secede from Yemen, which is an amalgamation of two previous territories to begin with. And both of them then became active because neither of them was happy with this quite tentative federal designation.

It’s preferable to keep dialogues to dialoguing rather than to absolute decision-making. And the alternative to a thoroughly autocratic constitution is an interim constitution, which can be proclaimed more or less autocratically as long as it’s intended in good faith to get things rolling at the appropriate time. And one could have relied on it. The Iraqis would have been much better off if they had taken their time rather than finishing early. By the way in Somalia and in Iraq there are provisions that are not yet filled in, constitutional provisions to be filled in later, in Iraq by a certain deadline, which has long since passed. And they’re not filled in yet.

Some of this has to do with the fact that Iraq emerged from a constitutional process as a federation when only some people wanted a federation. The Sunnis certainly didn’t want a federation and other parts of it just have to do with bad drafting because this was very rushed and a very amateurish document The United States had a bad role in Iraq in the sense that it cut short the possibility of a six month extension which had been explicitly provided for in the transitional administrative law. They did that, I think, for what were honest and good motives, but they made a terrible mistake doing it.


Well Don, thanks so much for joining me. I really did enjoy your book, Constitutional Processes and Democratic Commitment. The title comes across as extremely academic and it is a very academic book, but I love the way that you’re able to weave in all of these examples, so that it both integrates the theory, the big picture ideas, with some very concrete examples that really bring out the ideas, put them into context that you can really understand and kind of grasp. I thought it was an excellently written book. Thank you so much.

Donald Horowitz

You very much for having me. I enjoyed this thoroughly. I’m very glad to be on your podcast.

Key Links

Constitutional Processes and Democratic Commitment by Donald Horowitz

Ethnic Power Sharing: Three Big Problems” by Donald Horowitz in the Journal of Democracy

Reconsidering Democratic Transitions Francis Fukuyama, Donald Horowitz, Larry Diamond on YouTube

Democracy Paradox Podcast

Aldo Madariaga on Neoliberalism, Democratic Deficits, and Chile

Hélène Landemore on Democracy without Elections

More Episodes from the Podcast

More Information

Democracy Group

Apes of the State created all Music

Email the show at

Follow on Twitter @DemParadox

Follow on Instagram @democracyparadoxpodcast

100 Books on Democracy

Leave a Reply

Up ↑