Territory and Power in Constitutional Transitions (Center for Constitutional Transitions, Forum of Federations, and International IDEA, 2019).

Full text

Territory and Power in Constitutional
Transitions
International IDEA Policy Paper No. 17

Territory and Power in Constitutional
Transitions
International IDEA Policy Paper No. 17
George Anderson and Sujit Choudhry

© 2019 International Institute for Democracy and Electoral Assistance
© Forum of Federations
International IDEA publications are independent of specific national or political interests. Views expressed in this
publication do not necessarily represent the views of International IDEA, its Board or its Council members.
References to the names of countries and regions in this publication do not represent the official position of
International IDEA with regard to the legal status or policy of the entities mentioned.
The electronic version of this publication is available under a Creative Commons Attribute-NonCommercial-ShareAlike
3.0 (CC BY-NC-SA 3.0) licence. You are free to copy, distribute and transmit the publication as well as to remix and
adapt it, provided it is only for non-commercial purposes, that you appropriately attribute the publication, and that you
distribute it under an identical licence. For more information on this licence visit the Creative Commons website:
<http://creativecommons.org/licenses/by-nc-sa/3.0/>.
International IDEA
Strömsborg
SE–103 34 Stockholm
Sweden
Email: info@idea.int
Forum of Federations
75 Albert Street, Suite 411
Ottawa, ON K1P 5E7
Canada
Email: forum@forumfed.org
Design and layout: International IDEA
DOI: <https://doi.org/10.31752/idea.2019.1>
ISBN: 978-91-7671-228-3
Created with Booktype: <https://www.booktype.pro>
International IDEA | Forum of Federations

Acknowledgements....................................................................................................................... 6
Executive summary....................................................................................................................... 7
Factors shaping constitutional moments......................................................................................7
Constitution-making processes....................................................................................................8
Constitutional design and territorial claims...............................................................................10
Key recommendations................................................................................................................ 12
1. Introduction............................................................................................................................. 14
2. Constitutional dynamics and constitutional moments....................................................... 15
2.1. Factors shaping constitutional moments.............................................................................15
2.2. The dynamics of constitutional moments...........................................................................19
3. Constitution-making processes............................................................................................. 21
3.1. Agenda-setting...................................................................................................................21
3.2. Deliberation.......................................................................................................................26
3.3. Ratification........................................................................................................................35
4. Constitutional design and territorial cleavages................................................................... 38
4.1. Constitutional moments and constitutional design.............................................................38
4.2. Symmetrical federalism or devolution with a majoritarian central government...................40
4.3. Highly devolved federalism with a consociational central government................................44
4.4. Special autonomy for small territories and a majoritarian central government....................46
Contents

5. Conclusions............................................................................................................................. 49
References and further reading................................................................................................. 51
About the authors....................................................................................................................... 54
About the organizations............................................................................................................. 55
Center for Constitutional Transitions.......................................................................................55
Forum of Federations................................................................................................................55
Foundation Manuel Giménez Abad..........................................................................................56
International IDEA...................................................................................................................56

6International IDEA | Forum of Federations
Territory and Power in Constitutional Transitions
Acknowledgements
The authors of this Policy Paper wish to thank Sumit Bisarya for helpful comments
on a draft version. All remaining errors are the authors’ own. This paper is largely
based on the last chapter of an edited volume forthcoming from Oxford University
Press in 2019 that sets out the results of a larger research project. That volume
includes 17 case studies, which are referenced in this paper.

International IDEA | Forum of Federations7
Executive summary
Executive summary
This Policy Paper provides insights into how territorial claims relate to constitution-
making processes and constitutional design during periods of intense political
engagement over constitutional reform or ‘constitutionalmoments’. It considers the
factors that shape such constitutional moments, the dynamics of those constitutional
moments themselves and how these factors and dynamics frame choices regarding
constitution-making processes and constitutional design.
Factors shaping constitutional moments
Three principal factors shape the structure and dynamics of constitutional moments
and set the agenda for constitution-making processes and constitutional design.
1. Political geometry
The relative size, number and character of salient political cleavages (including their
geographic distribution) shape political mobilization. Cleavage patterns reflect
demography and political narratives and fall into four broad types of political
geometry:
1. cases in which there are multiple politically salient territorial cleavages and a
widespread interest in devolved or federal governance;
2. cases in which there is a majority group as well as one or two territorially
based, minority groups of significant size that may demand autonomy and also
some power-sharing at the national level;
3. cases in which there are one or a few relatively small, peripheral, politically
mobilized regions that seek special autonomy but no power-sharing in national
institutions; and
4. cases in which the political salient cleavages are both territorial and non-
territorial, so that demands for territorial accommodation are part of a larger
constitutional agenda, in which demands for devolution may be offset by, and/
or combined with, other claims.

8International IDEA | Forum of Federations
Territory and Power in Constitutional Transitions
2. Peaceful or violent political engagement around territorial claims
The political circumstances leading up to a constitutional moment can strongly
influence the ensuing process of constitution-making. If the lead-up has been
peaceful and within a constitutional order, the process will normally be bound by the
rules governing constitutional processes and amendment set out in the existing
constitution. If the lead-up was a regional insurgency that has led to negotiations
between the government and the insurgents, there will probably be a bilateral
negotiation but any resulting constitutional changes will have to be implemented
through established constitutional procedures. If the lead-up was a state-wide
conflict, the victor (government or rebels) or the parties to a new settlement
(government and rebels) may opt to draft a new constitution according to new rules
or, in the case of a government victory, to follow the process in the existing
constitution.
3. Power positions of the key political actors
Whether the antecedents are peaceful or violent, a critical factor will be whether the
ensuing constitution-making process requires agreement between the key political
actors or whether one dominant party is in a position to impose a new constitutional
settlement. A dominant party may emerge victorious from a violent combat or from
an election; however, even in these cases, such a victor will need to consider whether
or not to accommodate some of the losers’demands as a way to consolidate longer-
term stability.
Constitution-making processes
Constitution-making processes influence the constitutional design that emerges from
a constitutional moment. Although such processes have a high degree of
indeterminacy, they fall into certain patterns when dealing with territorial cleavages.
Such processes can be divided into three stages.
1. Agenda-setting
The initial phase of constitution-making should consider what processes and
decision-rules should be adopted and who the participants should be. Quite often,
there might also be substantive decisions on the principles or some key terms that
must be reflected in the final outcome. During this phase, decisions may be made
either through multiparty negotiations or unilaterally by a dominant party (if one has
emerged). Typically, the agenda-setting phase is less inclusive in terms of
participation than later phases. Major agenda-setting issues are both procedural and
substantive.
Key procedural issues would include whether the existing constitutional procedures
or new rules (perhaps with a break in constitutional continuity) will apply. Under
existing rules, there may be latitude for non-constitutional forums or processes,
including advisory or consultative commissions, national dialogues and consultative
referendums. In stalemates after a state-wide civil war, the parties to the conflict
usually negotiate ceasefire terms and often agree on new rules for making a
constitution. In stalemates after small regional insurgencies, agreements from

International IDEA | Forum of Federations9
Executive summary
negotiations may be ratified by constitutional amendment or by legislation, which
will involve national institutions, but prior to that an indication of the support of the
affected population may be sought, perhaps by referendum.
Given the stakes, especially after a violent conflict, agenda-setting may include
major substantive principles or arrangements that must be respected in any final
outcome. These could include commitments to federalism of some description, to
the protection of minority rights, to the reintegration of former combatants into
civilian or military roles, and to an amnesty for past crimes. There could also be
elaborate interim governance arrangements with power-sharing or local autonomy.
2. Deliberation
This phase should normally be more inclusive and detailed than the agenda-setting
phase. Very often the leading role on major constitutional reviews is given to an
elected body, which may be an established legislature or a newly created one. There
could be dual-purpose bodies, which serve simultaneously as legislatures and
constitutional assemblies, as well as specialist constitutional assemblies that have no
ordinary legislative responsibilities. The argument in favour of specialist assemblies is
that they avoid the conflict of interest that members of a regular assembly might have
in determining the future scope of their own powers, but the experience of specialist
assemblies has been chequered and most countries use the dual-purpose model. In
either case, expert commissions, consultative forums and national dialogues could
supplement deliberation by these assemblies.
The deliberation phase following a negotiated settlement between a government
and a territorial insurgent group may be difficult, given that the negotiations would
typically have dealt with the critical issues in detail, which could leave little room for
substantive input when the proposed deal is sent to the national legislature for
consideration. On occasion, such deals have unravelled once they have come before
the national legislature, which underlines the importance of government negotiators
maintaining a channel to the legislature during negotiations with the insurgents.
When territorial cleavages are a central issue in a constitutional debate, a particular
problem that can arise is the lack of constitutional provision for any particular role
for regional representatives in the process of constitution-making. Although super-
majorities are normally required for constitutional change, such a provision may not
give much leverage to one or more territorial groups that are seeking accommodation.
There may therefore be a need to pay particular attention to the role such territorial
groups could play during the deliberative phase.
3. Ratification
The final stage of constitutional reform is ratification. This may be effected by a
super-majority in the constitutional assembly (although in the United Kingdom a
simple majority suffices) and in many federations some constitutional changes require
a measure of assent from the constituent unit legislatures or populations.
Referendums are required for ratification of certain constitutional changes in some
countries, usually following majority or super-majority approval by the constitutional
assembly. Referendums have also been used on a consultative basis in some countries.
There can be a territorial dimension to approval by referendum, whether this be a

10International IDEA | Forum of Federations
Territory and Power in Constitutional Transitions
requirement that each of two territorial populations vote in favour (as in Cyprus) or
that there be some degree of territorial support beyond a simple, national majority (as
in Australia and Iraq).
Constitutional design and territorial claims
A key test of constitutional designs is how well they function within their context.
Design options may try to respond to claims for accommodation of territorial
cleavages within a country but they may also reject territorial accommodation by
maintaining a centralized, unitary regime. In some cases, the option of maintaining
the integrity of the country could give way to accepting the secession of one or more
territories. For cases of accommodation within a country, there are three broad
options, each being appropriate to specific contexts:
1. Symmetrical federalism or devolution with a majoritarian central
government
This model is common and covers many varieties and degrees of decentralization. It
is most appropriately applied in countries that have a highly territorialized political
geometry or that encompass a mix of territorial and other cleavages. The major
design issues include the number and boundaries of the constituent units, the
protection of minority rights (nationally and within constituent units), territorial
representation in central institutions (notably upper houses), the form of the
legislature and the executive (e.g. parliamentary versus presidential-congressional),
the nature and extent of devolved powers, and special autonomy arrangements. The
political dynamics when there are multiple cleavages and several constituent units can
be quite fluid, with governing coalitions shifting over time, thereby limiting how
often any group or territory is excluded from power. In this model, majoritarian
government at the centre is broadly accepted, although upper houses often give extra
weight to smaller constituent units.
2. Highly devolved federal government with a consociational central
government
When the political geometry is of two (or sometimes three) antagonistic and
territorially separate communities of significant size that must cohabit within a single
state, there may be recourse to this problematic model. Mutual mistrust favours
maximum devolution for each territorial community, but, since there are important
functions that must remain within the central government’scontrol, there should be
recourse to power-sharing at the national level in which the agreement of a majority
of representatives of each community is necessary for certain specified major
decisions. The power-sharing aspects of this model are particularly challenging,
because they can lead to extended periods of blockage and weak government.
However, in some situations, this may be the only model that has any possibility of
being accepted by all communities.

International IDEA | Forum of Federations11
Executive summary
3. Special autonomy for territories in a federal or non-federal state, with
majoritarian central government
This option is more appropriate for cases in which the political geometry includes
very small, peripheral populations with a strong sense of identity distinct from that of
the majority. Although the majority may resist asymmetry for ideological reasons,
this model can work in situations where the population of the relevant territory is
very small relative to the total population and/or where the extent of the special status
is not too extensive. If the relevant population is relatively large and its special powers
are very extensive, the question of how its representatives function in national
institutions, such as voting on issues that do not affect their region, can be difficult.
These three models may be considered ‘idealtypes’and there will be situations when
the political geometry of a country has features that suggest the need for a hybrid
model; these cases can be especially challenging. Moreover, devolution or federalism
is a model that has been very widely adopted even in countries that do not have
highly territorialized political geometries. Each of these models can be adapted to the
specific context of the country in question in many ways. However, it is useful to
bear the major models in mind when considering the likely constitutional design
alternatives for a country.

12International IDEA | Forum of Federations
Territory and Power in Constitutional Transitions
Key recommendations
1. At an early stage in a constitutional transition, the various parties and groups
shouldconsidertheimplicationsoftheircountry’spoliticalgeometryforan
appropriate constitutional architecture. Such a process of deliberation may
help to create a consensus on the most likely general model, as well as an
awareness of particular challenges in adapting the most likely architectural
modeltothecountry’sparticularcircumstances.
2. There are three major constitutional design options:
a. Symmetrical federalism or devolution with a majoritarian central
government is appropriate for highly regionalized countries, but also for
those with a mix of territorial and other cleavages; the extent of
devolution and the number and boundaries of territorial units will be
key issues.
b. A highly devolved federal government with a consociational central
government can be chosen when there are two (or three) antagonistic
and territorially separate communities, each of significant size, that must
cohabit within a single state; this is a difficult system to sustain so there
must be careful attention to the mechanisms to promote political
cooperation.
c. Special autonomy for territories in a federal or unitary state with a
majoritarian central government is most appropriate in countries that
have one or more territories with relatively small populations with a
strong distinct identity; this model is more problematic for larger
territorial populations.
3. Not all countries have a political geometry that neatly relates to one of the
major constitutional design options. This creates a need for more creative
solutions, such as hybrid models. The fundamental objective should be to
design a system that will permit peaceful, competitive politics on a sustainable
basis.

International IDEA | Forum of Federations13
Key recommendations
4. Thecircumstancesleadinguptoaperiodofconstitutionalengagement—
whether these are peaceful and within pluralistic politics or violent, and
whether they are dominated by a victorious government or faction or
characterizedbyabalanceofforces—willinfluencetheprocessthatisadopted,
sotheyrequirecarefulconsideration.
5. In a peaceful, constitutional transition in which territorial cleavages are salient,
the political principals should consider how to achieve territorial
accommodation within the established constitution-making process. Options
might include a political pact among parties to work towards a consensus, the
participation of territorial representatives, the creation of a commission to
examine territorial options, a special ratification procedure or even an interim
constitution and/or power-sharing arrangement during the transition period.
6. In negotiations with regional insurgents, the government side needs to
considerhowtomanagea‘two-levelgame’:reachingasatisfactoryagreement
with the rebels while having an agreement that in substance and legal form is
likely to achieve legal ratification by national institutions. This includes
consideringthecourts’reactiontotheconstitutionalityofproposed
arrangements as well as the politics in the legislature and, potentially, the
population (if a referendum may be required).
7. Although the victors in a major civil conflict or war will naturally seek to
implement a constitutional design that they have fought for, they should
consider using their position of strength to try to build political bridges, with
respect to both process and substance, with those they have defeated, to
promotealonger-term,‘positivepeace’.Limitedconcessionsofa
constitutional type may help reconcile their opponents to the new regime and
helpnormalizethecountry’spolitics.

14International IDEA | Forum of Federations
Territory and Power in Constitutional Transitions
1. Introduction
Collective demands for the constitutional accommodation of territorial cleavages are
pervasive across very diverse contexts. Territorial claims are being made in the Global
North in Spain and the United Kingdom, as well as in the Global South in Myanmar
and Yemen. Territorial claims are a central issue in the negotiations over the
reunification of Cyprus, and were central to the civil war settlements in Bosnia and
Herzegovina and in Nepal. Territorial claims have also arisen in peripheral regions
that have had insurgencies seeking special autonomy, such as for a Bangsamoro
Autonomous Region in Muslim Mindanao in the Philippines and for an autonomous
Aceh in Indonesia.
In many countries, political identification on the basis of territory is a central basis
of political mobilization, around which political claims are framed, political parties
formed, elections contested, governments composed, and constitutional claims made
and resisted. Constitutional transitions dealing with significant territorial cleavages
face political dynamics and challenges that are quite distinct from such transitions
where such cleavages are absent. They can involve different constitution-making
processes, sometimes with representatives of territorial groups having a formal or
informal role, and certainly involve different political dynamics, whatever the defined
roles of territorial groups are in the process. Such transitions also involve very
different constitutional design options from those in countries in which territorial
cleavages have little political salience; some form of federal, quasi-federal or
asymmetric devolution is usually at issue and sometimes special power-sharing or
minority rights protection arrangements are as well.
This Policy Paper seeks to provide insights into how territorial claims relate to
constitution-making processes and constitutional design, and to offer advice that may
be useful to principals and advisors engaged in constitutional moments—that is,
periods in which there has been intense political engagement over how to respond
constitutionally to significant demands for territorial accommodation from one or
more regions. The following chapter considers the factors that shape the dynamics of
constitutional moments, and sets the agenda for design and process.

International IDEA | Forum of Federations15
2. Constitutional dynamics and constitutional moments
2. Constitutional dynamics and
constitutional moments
2.1. Factors shaping constitutional moments
A constitutional moment may be triggered by any number of events, such as the end
of or a halt to a conflict, the arrival in office of a new political party, or a decision by
power-holders to address a major issue of institutional design. Moreover, they may be
sudden and unexpected. Some constitutional moments are relatively brief, whereas
others are spread out over years; some bring significant, enduring change, whereas
others change little or nothing in formal constitutional arrangements; some settle
constitutional disputes, whereas others do not, leading to further rounds of
engagement over future changes. As one scholar has noted, democratization is a
‘complex, dynamic, long-term and open-ended process’ that requires coordinated
collective action over generations (Whitehead 2003: 201). However, while the origin
and trajectory of each constitutional moment is unique, there are patterns across cases
that fall into different types, which can help policymakers in understanding the
relevant issues, the prospects and the possible trade-offs in a particular context.
Three crucial contextual factors shape the structure and dynamics of the
constitutional moment and set the agenda for the choice of constitution-making
process and the menu of potential options for constitutional design. These are (a) the
political geometry of territorial and other salient political cleavages in the country; (b)
the antecedent circumstances of peaceful, legal and institutionalized means or violent,
extra-legal means used to advance demands for territorial accommodation; and (c)
the relative power of the key actors, as shaped by tests of strength both peaceful and
violent.

16International IDEA | Forum of Federations
Territory and Power in Constitutional Transitions
Political geometry
Political geometry refers to the relative size, number and character of political
cleavages, including their geographic distribution. Understanding variations in
political geometry is crucial in coming to grips with important variations in
constitutional design and constitution-making processes. Territorial cleavages are a
product of the interaction of demography and the narratives deployed by regional
political entrepreneurs. There are four main patterns of political geometry:
1. States characterized by multiple politically salient territorial cleavages. Other
salient cleavages may be present but the territorial issue is central to
engagement on the issue of constitutional change, which typically involves the
general architecture of the regime (e.g. federal or devolved). Ethiopia, India
(with respect to its linguistic groups), Nigeria and Spain are examples.
2. States characterized by a majority group and one or two territorially based,
minority groups of significant size. The minority group(s) typically raise(s)
demands for territorial autonomy and, if large enough, may advocate that
central institutions are structured around a pact or partnership with special
power-sharing provisions among these groups. Examples include Cyprus,
which has a Greek-Cypriot majority and a Turkish-Cypriot minority; and
Bosnia and Herzegovina, which has a Bosniak majority and Serbian and
Croatian minorities. Yemen, with its (divided) northern majority and southern
minority, also has some of these characteristics.
3. States that have peripheral, politically mobilized regions. Such regional groups
make claims for territorial autonomy but do not make demands for power-
sharing in the central institutions, given their small size. Examples include the
tribes in north-east India; the Acehnese in Indonesia; the Moro in the
Philippines; Scotland in the UK; the northern Tamils in Sri Lanka; and the
RussophonesinUkraine’sDonbas.
4. States characterized by politically salient cleavages that are both territorial and
non-territorial. Territorial accommodation is part of a larger constitutional
agenda in which the strength of demands for devolution may be offset by other
claims. In Nepal, ethnicity, indigenous identities and caste identities can be
territorially concentrated but are often dispersed. In South Africa, racial
distinctions are largely non-territorial, but linguistic and tribal differences
within the black population often have a territorial character. In Kenya, there
has been extensive intermingling of ethnic communities that historically were
more territorially based. In Bolivia, major differences between the lowlands
and highlands are mitigated by divisions within each group and by other non-
territorial dimensions. In Iraq, the Kurds are distinct from the Arabs and
largely in one region of the country, but both Kurds and Arabs have significant
internal divisions, which are partially territorial in character.
These four patterns of political geometry are ideal types, and not every country fits
neatly into one pattern.

International IDEA | Forum of Federations17
2. Constitutional dynamics and constitutional moments
Peaceful or violent political engagement around territorial claims
Political entrepreneurs must choose the means to pursue claims for territorial
accommodation. They may seek change through peaceful means, working through
existing institutional channels and under legal—including constitutional—
constraints. Alternatively, they may reject the existing constitutional order and decide
to proceed extra-legally through violence.
The choice of means is influenced by the nature of the political regime in which
political entrepreneurs are operating, by their positions on constitutional options,
and by past history or experience in pursuing their goals. A society may have many
underlying tensions that suggest latent territorial cleavages, but in an autocracy, or an
intolerant and assimilationist majoritarian regime, territorial minorities may find no
effective peaceful outlet for debate and mobilization. When such regimes liberalize or
break down, this may open a major constitutional moment, when territorial and
other groups mobilize to promote their objectives. The choice of means will also be
influenced by political leaders’ assessments of the substantive distance between
themselves and the government or other influential actors and of their political
leverage.
Regional political parties are often the key instrument for non-violent, mass
mobilization around territorial claims through institutionalized channels. On other
occasions, the leaders of a territorial group may choose to work within a non-regional
party, making the contingent judgement that this will maximize their political
leverage. Peaceful pursuit of constitutional change occurs in a context of legal and
constitutional continuity, channelling debates over territorial accommodation
through existing institutions and rules. These rules strongly influence the political
leverage of different groups. In addition, high hurdles for amending a constitution
through existing rules may have a conservative bias and therefore make incremental
changes to the existing constitution more feasible than the adoption of a new one.
However, in certain cases, as in Nepal and South Africa, a radically new political
configuration can result in an agreement on new rules that reflect the current forces
at play while finding a way to maintain formal legal continuity by adopting, through
the established procedures, the new rules for developing a new constitution.
Depending on a group’sparticular grievances and circumstances, regional political
elites may reject normal politics, perhaps because they have failed, and turn to
violence. Violence brings a shift in the character of politics, notably in giving those
with guns (both government and rebels) a central role while marginalizing others.
Territorial groups that turn to violence typically do so with a deep sense of grievance
and frustration with peaceful means. Such a choice is easily understandable in a
context of repressive autocracies but it can happen in democracies, especially fragile
democracies or democracies that strongly favour nation-building over the
accommodation of differences. This is particularly true in former colonies, where the
post-independence political environment was largely one of nation-building and
hostility to territorial claims for constitutional accommodation, which were perceived
as an existential threat. For small, peripheral regional populations, the low returns
from such political mobilization may discourage them from forming a political party
and may drive them to violence if they are deeply aggrieved. The outcome of violence
may be victory for one side—either the government or the rebels—or a stalemate;

18International IDEA | Forum of Federations
Territory and Power in Constitutional Transitions
these different outcomes will influence the nature of the constitutional process that
follows. A government victory would normally involve a process within the existing
constitution (although in Nigeria the military, which took power through a coup,
defeated the secessionist rebels in Biafra and then ruled by decree), whereas a victory
by the rebels would lead to a legal rupture and the drafting of a new constitution.
Power positions of the key political actors
The third factor is the relative power positions of the different political actors, including
governments and the leaders of territorially mobilized groups. The issue of territorial
accommodation may have been around for a long time before the key actors decide
to actively engage with it. Government leaders facing territorial demands must
consider how to react. Do they see the claims for territorial accommodation as
extreme and constitutionally unthinkable (e.g. secession in many countries) or are
they open to discussion and possible political resolution? Where subnational political
entrepreneurs have mobilized peacefully, formed regional parties and contested
elections, they have done so in the space offered by constitutional democracy.
However, most constitutions explicitly or implicitly prohibit the unilateral secession
of a region and from this prohibition some states have inferred a series of restrictions
on political activity. The use of violence by the central state is nearly always a
certainty when subnational political entrepreneurs have opted for violence
themselves.
When political actors actively engage with the question of constitutional change,
they may have experienced tests of their relative strength, whether through
competitive politics or through violent clashes. The outcome of such tests of strength
can determine when and how a constitutional reform process is initiated and the
relative power of the different actors to influence the outcome. In democratic politics,
the most important tests of strength are electoral. These can permit a group or
coalition to capture governmental power, either nationally or (if territorial
governments exist) regionally. The power positions of territorial leaders are
influenced by the size of the territory’s population, its political cohesion and the
leverage these assets provide in the light of the power positions of other actors (e.g.
whether or not they can influence the formation of a government).
In violent conflicts, the critical issue is whether the two sides have reached a
stalemate (with neither side being able to defeat the other so they turn to
negotiations) or whether a clear victor has emerged. Victors will be in a strong
position to determine the substance of constitutional change, including how to deal
with the territorial claims that gave rise to the conflict in the first place. Moreover,
they are likely to have great control over the process. Our main interest regarding
post-conflict situations centres on those arising from stalemates. Stalemates typically
lead to negotiations between governments and regional groups, although there is no
guarantee that they will agree and the possibility of violence resuming can affect how
negotiations unfold. Outside actors can also be important, most obviously when they
have engaged militarily themselves (either as invaders as in Iraq or as ‘peacemakers’as
in Bosnia and Herzegovina), but also when they support one side in a conflict.

International IDEA | Forum of Federations19
2. Constitutional dynamics and constitutional moments
2.2. The dynamics of constitutional moments
Different countries’ constitutional moments can be identified in terms of how they
relate to the three dimensions of the contextual factors: four categories of political
geometry; two regarding the means for the antecedent contest; and two with respect
to the power distribution, with a dominant actor versus a divided or diffuse power
arrangement. The structures of these political contexts each present an inherent logic
regarding how the leading political actors set and pursue their constitutional goals.
These variations in contextual factors shape the agenda, process and design of
constitutional moments, which are illustrated with examples in Table 2.1.
Table 2.1. The contexts of constitutional moments in various political
transitions
Antecedent
context
Power
distribution
Political geometry
Multidimensional
territorial cleavages
Majority–minority
territorial
configuration
Small, distinct
peripheral
regions
Mixed territorial
and non-
territorial
cleavages
Violent Victory Ethiopia
Nigeria
Croatia
Sri Lanka (2009)
Stalemate Bosnia and
Herzegovina
Cyprus
NE India
Aceh/Indonesia
Moro/
Philippines
Sri Lanka (2002)
Nepal
South Africa
Iraq
Non-violent Dominant India (linguistic states) Pakistan (1956) Corsica Russia
Divided Spain Belgium Scotland
Ukraine
Bolivia
Kenya
Yemen
Not every case fits neatly into a single box. The situation in Sri Lanka changed
between 2002, when negotiations took place during a military ceasefire, and 2009,
when the government had prevailed militarily. Yemen is hard to place: its North–
South cleavage suggests that it be assigned to the majority–minority column, but
both the North and South are themselves very divided, territorially and in other
ways, which led to it being placed in the mixed territorial and non-territorial column.
Not all cases have led to a consensual accommodation of a territorial claim within a
state: Croatia was a case of unilateral secession; Sri Lanka in 2009 saw a military
defeat of the Liberation Tigers of Tamil Eelam (LTTE); Yemen failed to achieve a
new constitution and descended into civil war; Cyprus is still in a state of frozen
conflict; Corsica has had minimal success in satisfying its demands for autonomy. In
other cases, territorial populations that are seeking accommodation of their demands
have achieved less than they desired.
The main focus in this Policy Paper is on constitutional moments where the
relative power of the key actors is divided, with no party able to dictate the outcome,
so that there needs to be some kind of agreement if the issue is to be resolved. The
political geometry and power positions of the main actors will be the key to

20International IDEA | Forum of Federations
Territory and Power in Constitutional Transitions
determining the nature and character of the constitutional process. The process will
also be shaped by whether the constitutional moment arises out of antecedents of
peaceful, legal, institutionalized politics or extra-legal, violent conflict.
Where power is divided because of electoral results, political actors address the
substantive agenda items in a context of legal and institutional continuity within a
constitutional democracy. The existing constitution provides a procedural framework
to debate, adopt and implement constitutional change, including possible territorial
accommodation. However, there may be political reasons to seek a broader
consensus, beyond what is required by the constitution; such a consensus may bring
greater political legitimacy to the result than would have happened with a decision
that met the minimal constitutional requirements for approval.
Where there is a military stalemate, the agenda combines peace negotiations and
constitution-making. These can work at cross-purposes because a peace agreement is
backward-looking and focused on ending violence whereas constitution-making looks
forward to creating effective institutions that operate under constitutional democracy
and peaceful, legal politics. The question is how to design a peace process among
armed parties that can broaden into a constitutional process among many parties,
including the public, with a much wider agenda. The process needs to provide the
armed parties with ‘on-ramps’ from armed conflict and peace negotiations to
peaceful, democratic politics.
A cross-cutting issue for constitutional moments that involve divided power—
regardless of whether the antecedents are violent or peaceful—is the risk of a lack of
credible commitment to the process of transition. In the short term, where the
constitutional system continues to exist in all or part of the state, this risk can
potentially empower ‘institutional spoilers’(e.g. courts, opposition parties, the
government backbench and/or elements of the political executive) to prevent the
government from implementing any territorial accommodation. Credible
commitment problems can also occur in the medium term across multiple electoral
cycles because opposition parties can campaign against agreements and proposed
constitutional arrangements and may dismantle them if they win power. Finally, just
as spoilers within the system may undermine a constitutional commitment, so too
can political actors who have been excluded from the process. We explore various
solutions to credible commitment problems in section 3.2.
There is a close correlation between political geometry and the main design
options that should be considered during a constitutional transition. In cases of
multidimensional territorial cleavages, the design option should, in general, be
significantly decentralized symmetrical devolution or federalism combined with a
majoritarian central government. For majority–minority territorial configurations, in
which the minority is relatively large, experience indicates that the option of a highly
devolved federalism, with a consociational central government, should be considered.
However, for small, distinct peripheral regions, experience points to the option of
special, asymmetric autonomy and a majoritarian central government as being the
most appropriate approach. Finally, countries that have a mix of territorial and non-
territorial cleavages should consider devolution or federalism, which could be quite
centralized and perhaps have a large number of units with a majoritarian central
government.

International IDEA | Forum of Federations21
3. Constitution-making processes
3. Constitution-making processes
Constitution-making processes are critical in that they strongly influence the ultimate
constitutional design that emerges from a constitutional moment. They also play a
significant role in determining how that design is received by the different forces at
play over time (Arato 2016; Choudhry and Ginsburg 2016). Although these
processes, especially when they take place during transitions from authoritarian rule
to democratic rule, can have a high degree of indeterminacy (O’Donnell and
Schmitter 1986), we will show how they fall into certain broad patterns when dealing
with territorial cleavages. It is helpful to divide constitution-making processes into
three stages: agenda-setting, deliberation, and decision or ratification.
3.1. Agenda-setting
The initial phase of constitutional agenda-setting reflects the scope of the substantive
issues to be addressed—i.e. either a broad constitutional review or one that is targeted
on a specific territorial question—as well as the process, participants and decision-rules
for constitution-making. Decisions on these matters will be influenced by the
political geometry and the power distribution among political actors and they may be
constrained by the existing constitution. In situations of deep mistrust after a
conflict, agenda-setting may address and resolve some principles or issues of substance
regarding the ultimate outcome. There is no one formula for this phase, which may
be done quickly or over a very extended period. It can overlap, in varying degrees,
with the ensuing phases of deliberation and ratification. The choices made may be
recorded in a formal or informal document. Agenda-setting is typically led by elites,
although the representativeness of the elites and their influence at this stage will
reflect their relative power positions.
Procedural agenda
The crucial factor in the procedural agenda is whether the antecedents to the
constitutional moment are legal, peaceful and institutionalized or violent and extra-
legal.
When the antecedents to a constitutional moment are legal, peaceful and
institutionalized, constitutional procedures govern the process, and may include the
creation of a constituent assembly and recourse to a constitutional amending

22International IDEA | Forum of Federations
Territory and Power in Constitutional Transitions
formula. If constitutionalized, these procedures may serve as constraints. These
procedures also presumptively designate the constituent actors in the process, which
may include national and regional governments, opposition parties, civil society and
(in the case of referendums) the electorate. For example, in the UK, since the
adoption of the Scotland Act 1998, the agenda for discussions around further
devolution and the holding of a referendum has been set by bilateral negotiations at
the cabinet level between the governments in Westminster and Holyrood.
Ensuing agreements are political, such as a pact, joint statement or accord, which
may then require legislative implementation. However, constitutions often give no
direct role to regional governments or legislatures in constitutional change. For
example, in Bolivia from 2006 to 2008, the newly elected departmental prefects had
to act outside the official process, while the elected members of regional parties in
Congress—and later, in the Constituent Assembly—were the main actors within the
formal process. In Spain, in 1978, the elected heads of the regional governments had
no formal role in initiating and determining the process of constitutional change, but
came to play a role because the governments of autonomous communities negotiated
and subsequently ratified the quasi-constitutional statutes of autonomy once the
1978 post-Franco Constitution had been approved. In practical terms, depending on
circumstances and the political strength of the political parties, this phase may be
largely determined by the government alone (e.g. in India) but more frequently there
is interparty discussion and debate (e.g. in Bolivia, Kenya, Spain and Yemen).
In a situation of both legal and institutional continuity, there can be latitude about
various non-constitutional forums, such as special commissions and public
consultations. Sometimes a national conference or dialogue outside the normal
institutional framework can be important. Membership and rules of procedure of
these forums can be loose and fluid and depart significantly from normal
parliamentary or intergovernmental procedures. Participants may include political
parties and governments, as well as a much broader range of non-elected actors,
including labour movements, businesses, organized religions and other civil society
organizations.
In Scotland during the 1990s, a large proportion of the political elite joined an
informal, cross-party Constitutional Convention that, in the absence of any such
action from elected regional representatives, developed a blueprint for devolution.
When the Labour Party came to power in 1997, it endorsed this plan and submitted
it to a Scotland-wide referendum, before Parliament enacted the Scotland Act of
1998.
In Spain, debates over territorial accommodation occurred even before the
transition to democracy formally started. In 1975, before Franco’s death, several
national political parties came together to agree on principles of a democratic
transition; they were joined, following Franco’sdeath, by the main regional parties. A
first step was to legalize all parties and adopt a general political amnesty. The
outgoing Francoist parliament approved a law on political reform, which set the
framework for a peaceful transition to democracy and a new constitution. This was
approved in a popular referendum. This led to free elections, after which the
democratic parties agreed to resolve their differences so that the army or other anti-
democratic forces would have no pretext to usurp the transition to democracy.

International IDEA | Forum of Federations23
3. Constitution-making processes
Whatever the exact mix of forums, under conditions of legal and institutional
continuity, the participation of political elites (including elected representatives
where possible) in agenda-setting is integral to ensuring that the constituencies they
represent will accept the legitimacy of the rest of the constitution-making process,
participate in it, and accept it and implement the resulting constitutional settlement.
In stalemates after violence, the parties to the conflict normally negotiate to set the
agenda. This can be an elaborate and time-consuming process (e.g. in South Africa
from 1982 to 1992). External parties or mediators can assist the process, as in
Cyprus, Nepal, Indonesia (regarding Aceh) and the Philippines (regarding the
Bangsamoro). Unusually, for Bosnia and Herzegovina in 1995 and Iraq in 2005,
outside powers largely set the agenda and the process for negotiations. Very
occasionally, as in Cyprus, external powers are direct parties in the conflict and are
participants in agenda-setting.
The participants in agenda-setting are, by default, the main armed combatants, as
happened in Nepal in 2006. In many long-running territorial conflicts, armed
combatants have evolved into de facto regional governments, as was the case in Sri
Lanka with the LTTE. Even as the civil war continued prior to the 2002 ceasefire,
President Chandrika Kumaratunga decided to lead a constitutional reform process
that, on paper, could have gone some way to meeting the demands of Tamil
nationalists; one reason for its failure was that the LTTE was excluded from the
process. In Kurdistan and in North Cyprus, elected regional institutions and leaders
have created state-like entities that have had to be central to any constitutional
negotiations. The resolution of peripheral rebellions, even without de facto regional
governments—such as in the Philippines, Indonesia and north-east India—has
required that armed rebels be at the table, with the central government representing
the state.
In situations of violence, both the constitutional process and the peace process
begin with a ceasefire. A ceasefire agreement, in principle, can honour the substantive
and procedural constitutional agenda. However, the capacity of a ceasefire agreement
to do so must reckon with the legal and institutional context. In some cases, the
constitutional order and its institutions have effectively collapsed. The ceasefire and
agenda-setting negotiations determine the constitution-making process and also set
substantive constitutional principles or parameters that must be respected by the
ultimate constitution. This presents both an opportunity—because there is no
constitutional constraint on the creative crafting of the process—and a challenge—
because there is no default script for the process, which may therefore be contested.
Moreover, in cases of state collapse, parties will often create provisional governing
arrangements, including constituent assemblies that wield legislative functions as
well. For example, in Nepal the civil war came to an end in 2005 with a 12-point
agreement between seven political parties and the Maoists; that agreement included
key commitments to an interim government that included the Maoists and to an
elected Constituent Assembly. The parties and the Maoists then created a small,
representative commission to prepare an interim constitution, the drafting of which
took over a year. In 2006, the parties reconvened the 2002 parliament, which had
been dissolved in 2005; this parliament then adopted the interim constitution, which
designated the reconvened parliament as the interim parliament; that body then

24International IDEA | Forum of Federations
Territory and Power in Constitutional Transitions
admitted Maoist members and passed a law for the election of a new constituent
assembly.
In other cases of violence prior to a constitutional moment, there may be legal and
institutional continuity across most of the state, even if the central government’s
presence in the regions under conflict may be greatly diminished or barely existent.
The government views its constitutional order as extending to areas under armed
conflict and as governing the constitution-making process. The challenge is to
coordinate existing procedures for legislative and constitutional amendment, in
which armed rebels have no formal role, with bilateral negotiations between armed
parties and central governments. Institutional spoilers could raise constitutional
objections at the agenda-setting stage (e.g. in Sri Lanka, where the mere fact of the
ceasefire agreement with the LTTE was attacked as being unconstitutional because it
was built around a line of control that de facto accepted the limited reach of the Sri
Lankan state). Precisely because of this risk, armed rebels may seek to negotiate
assurances regarding downstream processes during agenda-setting. In Sri Lanka,
although the 2002 ceasefire agreement did not address constitutional issues, the
bilateral dialogue launched by the ceasefire yielded the Oslo Declaration of 5
December 2002, which set down some key markers for the shape of a future
constitutional process (an agreement would need to be ‘acceptable to all
communities’) as well as for the substance of a constitutional agreement (a federal
structure within a united Sri Lanka). This led Sinhalese hardliners to attack the
ceasefire process for arising from an extra-constitutional process and for pursuing
unconstitutional ends; their opposition was one reason why the constitutional process
never got off the ground.
This conundrum has been addressed in some cases through hybrid processes that
formally comply with the existing constitution but work around it to determine both
the process and substantive principles of a constitutional transition. A striking
example of this took place in South Africa, where there was a stalemate between the
African National Congress (ANC), which rejected the existing constitutional system
as racist and illegitimate, and the South African Government, which insisted on legal
and institutional continuity. This process began with years of ‘talks about talks’
before the National Party government repealed the ban on the ANC and informal
negotiations began. This bilateral process culminated in the National Peace Accord,
which set the stage for the Convention for a Democratic South Africa (CODESA)
among 19 political parties. After CODESA collapsed it was succeeded by another
multiparty forum that was underpinned by a consensus between the ANC and the
National Party.Because these talks took place outside formal parliamentary
institutions, they could include the ANC, which had no elected representatives, as
well as other entities—such as the leadership of Bantustans, especially KwaZulu, as
well as the Afrikaner-nationalist VF (Vryheidsfront—Freedom Front), which also
lacked parliamentary representation.
These observations suggest a challenge in Ukraine at present (2018). Ukraine has
launched a constitutional reform process that involves a presidentially mandated
constitutional commission and other activities that are centred around the executive
body (the Rada). However, following the Russian invasion of the Donbas region of
the country, citizens in the non-government-controlled areas have had no effective

International IDEA | Forum of Federations25
3. Constitution-making processes
representation in Kyiv and the Ukrainian Government has refused to engage with the
leaders of the self-proclaimed republics in Donbas. Negotiations on the reintegration
of the Donbas region into Ukraine took place in 2018 within the Minsk process, an
international forum where negotiations took place with Russia, an external actor, as
well as with Western powers. The challenge is therefore how to give voice to those in
the non-government-controlled areas as the constitutional and Minsk processes
proceed.
Substantive agenda
Agenda-setting agreements can include significant substantive provisions that must
be respected in the content of an eventual constitutional settlement. The substantive
scope of the constitutional exercise can influence the nature of the process—for
example, a process that aims to fundamentally rewrite a whole constitution will be
quite different from one that is aimed at a far more targeted issue, such as the
accommodation of a particular region’sterritorial claims.
When territorial claims for autonomy are a potential driver of constitutional
reform, an important decision is whether or not to broaden the scope of the agenda
to include other issues. There are good arguments for doing so. Regional autonomy
may be seen as the sole means to diffuse political power and accommodate the
country’svarious cleavages, when other mechanisms could be considered. Including
certain rights in the constitution—such as language rights, especially for education
and local government—can reduce the pressure for a proliferation of federal units.
Electoral system design at the national level can make minorities more likely to be
represented in national institutions. When the central issue is relations between a
majority and a large minority, the prospect of some power-sharing in central
institutions may be potentially valuable and reduce the degree of regional autonomy
that is ultimately needed. The idea of power-sharing in central institutions was
entirely absent from the 2002–2008 Sri Lankan peace process, when federalism was
briefly considered. The absence of that issue from the agenda created legitimate
concerns about the long-term viability of any settlement. Finally, having many
possible instruments for addressing territorial and minority concerns on the
constitutional agenda also permits trade-offs and bargaining, which facilitate
agreement.
The scope for broadening the constitutional agenda also depends on the relative
importance of the territorial issue. In India and the UK, the pressures from large,
peaceful regional movements made addressing the territorial issue a central concern,
whereas in Spain and Bolivia, although the territorial issue was fundamental, the
constitutional agenda was much broader and encompassed the whole constitution. In
Kenya, Nepal, South Africa and Ukraine, the territorial issue was one among several
issues associated with a comprehensive constitutional agenda. Agendas after a civil
war usually involve a host of constitutional and non-constitutional issues beyond any
territorial questions. They include transitional justice, the resettlement of displaced
populations, property restitution, security arrangements and the reform of the rule of
law. However, the agenda to resolve peripheral rebellions is normally focused
narrowly on the territorial issue.

26International IDEA | Forum of Federations
Territory and Power in Constitutional Transitions
Placing territorial accommodation on the agenda of a constitutional process does
not guarantee a desired outcome if there is no indication of the expected substantive
resolution. Groups advancing territorial demands may therefore seek guarantees in
the agreed agenda that bind the process to certain substantive outcomes on key
points. In South Africa, for example, the Constitution of the Republic of South
Africa Act 200 of 1993 created an Interim Constitution that contained 34
constitutional principles that legally bound the Constitutional Assembly, including
one that required that the final Constitution create national, provincial and local
levels of government, and further stipulated that provincial powers ‘shall not be
substantially less than or substantially inferior’ to those under the Interim
Constitution. The Interim Constitution also provided that the Constitutional Court
would have to certify that the final Constitution complied with these principles for
the Constitution to come into force. The Court rejected the first version of the final
Constitution on the basis that it did not comply with this stipulation, necessitating
that the text be amended. Nepal’sinterim constitution also bound the constitution-
making process to require territorial accommodation.
Similarly, in Bolivia those favouring departmental autonomy insisted that the issue
must not be left to future negotiations. In negotiating the terms of reference of the
Constituent Assembly (CA), the parties adopted a law for a referendum on autonomy
that obliged the CA to create a regime of departmental autonomy and automatically
opted-in those departments where the ‘yes’ side in the referendum prevailed. This
arrangement constrained the Constituent Assembly, where there would not have
been a super-majority in favour of autonomy arrangements. Such an option was
possible because institutionally departments already existed, so they could conduct
referendums.
3.2. Deliberation
The deliberation stage is usually distinct from the agenda-setting stage in two
respects: participation is more inclusive and the consideration of substantive issues is
more detailed and concrete. Agenda-setting will normally have established some key
terms regarding process (e.g. participation, forums, decision-rules and timing), but
these often require significant elaboration and even modification. A major question is
how to provide for political pluralism on all sides of the constitutional discussions in
order to build a broad political consensus in support of any resulting framework, one
that supports credible commitment in both the short and medium terms.
Participants
It is usual that the deliberation stage broadens participation beyond the actors who
set the constitutional agenda.
When the antecedents of the constitutional moment are violent and extra-legal, at
least one party will be an armed group. In these instances, the lack of political
pluralism can be an acute problem because leaders of such a group are not elected
and may in fact oppose an electoral test of their strength. An extreme example was
the LTTE in Sri Lanka, which was opposed to political pluralism and ruthlessly
attacked moderates within the Tamil community. Armed groups that command

International IDEA | Forum of Federations27
3. Constitution-making processes
significant support and possess sufficient capacity but are excluded may turn into
spoilers that can undermine a settlement, as happened with early agreements with the
Moro population in the Philippines. When one group has won a civil war, it will be
able to determine both the process and the substance of constitution-making. In
Ethiopia, for example, there was a constitutional convention for the deliberative
phase, but its members were decided by the victors of the civil war, who chose to
exclude some former allies. A cautionary tale comes from Bosnia and Herzegovina,
where the local combatants were completely marginalized. Representatives from the
United States designated the Serbian and Croatian presidents as the representatives of
those communities—with the leadership of the Bosnian Croats and Bosnian Serbs
reduced to consulting with the two presidents outside the official debate. The three
ethnic communities had no buy-in to the constitution agreed to under the Dayton
Accords, although, paradoxically, they were institutionally empowered by it to spoil it
once it was implemented.
‘On-ramps’ for armed groups to transition from violent conflict to peaceful,
multiparty politics provide a way to broaden a constitutional process to promote
political pluralism, by enmeshing such groups in a system of political competition. In
Nepal, for example, the armed Maoist rebels agreed to take seats in an Interim
Parliament which then quickly held elections for a Constituent Assembly, which the
Maoists contested with considerable success, winning the greatest number of seats
and heading the first post-civil war government. This kind of sequenced approach,
from peace negotiations to a democratically elected constituent assembly, was
acceptable to the Maoists because they anticipated sufficient representation in the
Constituent Assembly to drive their agenda forward. The Constituent Assembly
elections also permitted a broad cross-section of political parties to field candidates
and win seats, which had the effect of promoting political pluralism at the
deliberation phase. However, these disparate groups were unable to reach the
required degree of consensus after five years of deliberation, so the assembly was
dissolved and new elections were held, its successor completing the job.
In India’snorth-east, this process of an ‘on-ramp’from violent conflict to electoral
contestation for rebel groups occurred in a number of stages, and with a focus on
regional politics. It converted armed militants into stakeholders in political life as
they achieved greater autonomy. The Nagaland Peace Accord of 1947 carved the
Naga Hills out of the state of Assam and made it a Union Territory under central
control with some local autonomy. In 1963, the government and Naga
representatives negotiated a 16-point agreement that led to the creation of the
Nagaland state in which former rebels competed in elections as a political party. The
government acceded to rebel demands for special autonomy on such subjects as
religious and social practices, customary law, and ownership of land and resources,
which were uncontroversial at the national level because of the small population and
remoteness of these regions. In parallel fashion, Tripura, Manipur, Mizoram and
Arunachal Pradesh acquired status first as Union territories, with little or no
autonomy, as a stage before acquiring statehood. Meghalaya achieved autonomy
within Assam, and then was granted statehood. ‘On-ramps’ may require locally
specific accommodations for political competition. In Indonesia, in Aceh, for
example, the peace agreement provided for the right to organize local political parties,

28International IDEA | Forum of Federations
Territory and Power in Constitutional Transitions
which was an exception to the requirement that all political parties have a national
outlook; this enabled the guerrillas to become a political party.
The deliberative phase can also be broadened to promote political pluralism and
the legitimacy of the process when there are elected governments—whether the
antecedents of the constitutional moment are peaceful or violent—but this too can
pose issues and can play out differently depending on the institutional arrangements
as well as the political context. In parliamentary systems that operate on the basis of
simple majorities, opposition parties can be excluded or marginalized if the
government controls the legislature. To counteract this risk, parliamentary
constitutions usually require a super-majority for constitutional amendments,
potentially giving opposition parties leverage. Semi-presidential regimes may have
cohabitation, in which the president and the government come from different
political parties. While, in principle, cohabitation provides an opportunity for
political pluralism, its potential to do so depends on the prevailing political culture.
In Sri Lanka during the 2002–2008 process, notwithstanding cohabitation, the
hyper-partisan nature of Sri Lanka’spolitical party system acted as an impediment to
building a cross-party consensus on the peace negotiations. The exclusion of
President Chandrika Kumaratunga of the People’s Alliance (PA) from the
negotiations conducted by Prime Minister Ranil Wickremesinghe of the United
National Party (UNP) led her to dismiss three cabinet ministers and trigger early
elections, which in turn led to the collapse of the talks. Presidential systems present a
genuinely hard case. In the Philippines, the control of the presidency and the Senate
by different political parties has blocked the implementation of commitments made
at the negotiating table, as occurred in 2015. There is no easy solution to this
difficulty.
Political leadership can be a major driver of inclusion and pluralism in deliberative
processes when there is competitive, pluralistic politics. In Spain, given the objective
of a successful transition to democracy, the outgoing Francoist parliament approved a
democratic law on political reform, while a broad cross-section of parties, including
those that had been excluded from politics, developed a broad consensus that kept
the lid on partisanship during the critical transitional phase. By contrast, in Yemen,
there was an absence of leadership in the wake of the fall of Ali Abdullah Saleh. The
national dialogue, which conducted extensive consultations to address the
constitution, did not engage the real power-holders in serious negotiations: Saleh was
excluded in order to mark a sharp break from the past, and southern separatists
boycotted the meetings. Political tests of strength during a constitutional process,
whether through elections or referendums, can provide an occasion for greater
inclusion in the debate and potentially shift the power positions of the players. In
Bolivia, for example, the population was mobilized by elections, referendums and
street demonstrations in repeated tests of strength, which in turn shaped negotiating
dynamics to the favour of Evo Morales.
Forums and decision-rules
The options considered for the forum for constitutional deliberations are influenced
by political geometry, and whether the antecedents have been peaceful and legal or
violent and extra-legal.

International IDEA | Forum of Federations29
3. Constitution-making processes
Under legal and institutional continuity, the consideration of territorial claims
occurs within existing constitutional provisions regarding constitutional amendment
and the forum for deliberation. Most constitutions leave the power of constitutional
amendment with the legislature voting with a super-majority, by implication making
it the deliberative body, although some require ratification by referendum (and
several federations require some level of approval by the legislatures or populations of
the federal units). However, legal rules rarely foreclose the creation of supplementary
deliberative bodies, such as the constitutional commissions in Kenya or the state
reorganization commission in India. In Bolivia, while Congress possessed the formal
authority over constitutional amendment, it passed a law creating the Constituent
Assembly (although in the end the process reverted to Congress).
Although super-majorities are the norm for constitutional change, the UK is the
outlier in that it requires no more than a simple majority for constitutional change as
part of the ordinary legislative process. In the UK, regional devolution occurred
through statute, although political decisions were made to incorporate referendums
into the process prior to the creation of the Scottish Parliament, and later about the
question of Scottish independence. In India, changes to state boundaries and the
creation of new states can be enacted through the very low threshold of a simple
majority in parliament (a state whose borders are changed must be consulted but has
no veto), although most other changes require a super-majority. Kenya’s
Constitution required a double simple majority (legislative and from a referendum).
Spain’s required a super-majority in parliament and a simple majority in a
referendum. In Bolivia, Congress had the power to amend the constitution, but there
was a political agreement to hold a referendum to ratify any amendment.
When the antecedent of the constitutional moment is a stalemate in a regional
insurgency or rebellion, the government will assert legal continuity for constitutional
change, but this may channel the deliberations into forums, such as the national
parliament, where elected representatives representing the rebels would be too few to
have any significant influence. In Indonesia, for example, the central government
initially had the parliament (where representatives from Aceh and Papua had a
minimal presence) enact autonomy laws unilaterally. These laws were unsuccessful in
stemming armed struggle even though, for Aceh, the laws had favourable terms
regarding resource revenues. By contrast, the 2006 Aceh autonomy law has been
largely successful because its terms were spelled out in the negotiated peace agreement
of 2005. In the Philippines, non-consensual, unilateral change was also unsuccessful.
The adoption of the 1987 Constitution authorized Congress to create the
autonomous regions, which it did in 1989 through a law that set up the Autonomous
Region in Muslim Mindanao (ARMM). It did this over the objections of the Moro
National Liberation Front (MNLF). The legislation required a referendum to
determine the geographic scope of the ARMM, which resulted in only one city,
Marawi, voting to be included; even Cotabato City, the regional centre for
Mindanao, voted to stay out. Although having a smaller region with limited powers
won the support of some Moro politicians, the referendum result failed to satisfy
Moro insurgents and did not succeed.
The initial and central forum for resolving a territorial insurgency is a bilateral
table (between government and rebels), in which peace and a constitutional

30International IDEA | Forum of Federations
Territory and Power in Constitutional Transitions
settlement are negotiated. However, a single focus on bilateral peace negotiations
risks failure downstream in the process, when the normal procedures for
constitutional review and deliberation and ratification must apply. There is a need,
especially for governments, to consider how they will address all stages of the process.
In the Philippines, the 2008 Memorandum of Agreement on Ancestral Domain that
was entered into by the government and the Moro Islamic Liberation Front (MILF)
was ruled unconstitutional by the Supreme Court because its terms promised to
adopt any constitutional amendments that were necessary to implement it, whereas
the power of constitutional amendment rested exclusively with Congress. In due
course, the MILF entered into new negotiations and reluctantly opted to proceed by
ordinary statute to avoid the difficulty in getting a constitutional amendment
approved, given the very high threshold required. A hybrid body, with 50:50
government/MILF membership, drafted the Bangsamoro Basic Law (also known as
the Bangsamoro Organic Law), but legislative approval, which had seemed highly
likely, was derailed in the Senate after a botched and very bloody government anti-
terrorist raid. Another example of the problems posed by institutional spoilers comes
from Sri Lanka, where government backbenchers and opposition legislators
challenged the constitutionality of an interim governing arrangement for LTTE-
controlled areas (the Post-Tsunami Operations Management Structure, P-TOMS),
which was struck down by the Sri Lankan Supreme Court on the basis that the
proposed entity’s spending powers infringed on parliament’s control of public
finances. This decision ended any prospect of further cooperation between the
parties.
These examples show the potential challenges that can be posed by institutional
spoilers during constitutional moments. Spoilers have legitimate interests in the
results of negotiations but can upset delicate negotiations because of their late
involvement and negative opinions. The question is how to mitigate this risk.
Legislators who must eventually implement an agreement could be brought into the
negotiating process in some ways at the agenda-setting or deliberative stages, but this
is a highly tactical matter that requires subtle political management. In the case of
court systems, this is much harder because they cannot be parties to negotiations.
The South African transition, which occurred under legal continuity, developed a
creative option whereby the Interim Constitution created a new Constitutional
Court, and an appointments procedure that ensured the creation of a court with
judges who were acceptable to the major political parties. The Constitutional Court
was assigned the task of assessing the final constitutional text for compliance with the
34 Constitutional Principles agreed to by a multiparty consensus. The Constitutional
Court was therefore an external judicial check on the constitutional process; at the
same time, however, the Court conceived of its role as being an institution of the
transition to a new constitutional order.
In contexts of complete legal and institutional rupture, as is often the case after
civil wars, there may be no legal constraint on the choice of forums and decision-
rules. The Dayton Peace Accords’constitutional premise was one of a complete break
with the previous constitution, which opened the way for the Accords to be produced
by a unique constitutional process, in which armed parties negotiated with direct
international participation and yielded an agreement that simultaneously ended

International IDEA | Forum of Federations31
3. Constitution-making processes
armed conflict and adopted a new constitution within three weeks. The case of
Cyprus, where fighting stopped long ago in its ‘frozen conflict’, resembles that of
Bosnia and Herzegovina in that negotiations over a formal end to the civil war are
tied to negotiations for a new constitutional settlement, although they have had to
deal creatively with the Greek-Cypriot view that there is continuity of the state (not
of the constitution) and the Turkish-Cypriot view that two sovereignties are being
brought together.
In most civil wars, the transition from civil war to ceasefire to peace agreement
and, finally, to constitution takes place over an extended period. If the previous state
has collapsed or if the government lacks any effective presence and/or legitimacy,
there is often a need to create provisional governing arrangements during the period
in which a longer-term constitution-making process can occur. These provisional
arrangements, whose negotiation will have been largely led by elites, create
transitional governing institutions, determine their membership, define decision-
making procedures and establish a constitution-making process. Interim
constitutions may set out these provisional arrangements in situations where the form
of legal continuity is maintained but where the existing regime is so tainted and
weakened that it cannot continue, even during an interim period. In South Africa,
the transition from the old order to an interim constitution preserved legal continuity
but fundamentally changed the regime by providing for universal suffrage elections to
a Constitutional Assembly and establishing a power-sharing executive that would
have responsibility for the constitutional process.
Interim constitutions can create substantive and procedural path dependencies for
the shape of a more permanent constitution. Substantively, they can provide
baselines for negotiations over the final constitution. Iraq’s Transitional
Administrative Law recognized Kurdish autonomy and created a procedure for
ratifying a new constitution by referendum that effectively gave the Kurdish region a
veto, and therefore set a baseline for Kurdish autonomy under the permanent
constitution. Procedurally, inclusive interim government arrangements can generate
buy-in to the constitution-making process and build trust among political opponents.
The South African experience illustrates the value of a power-sharing interim
government for the success of a constitution-making process. Both the National Party
and the Inkatha Freedom Party (representing Zulus) were represented in the cabinet,
participated in the constitutional process and supported the final constitutional text
even though they were not able to achieve all of their goals.
Constitutional and/or constituent assemblies
Elected constitutional and/or constituent assemblies (this paper uses the terms
interchangeably) may be proposed for major constitution-making processes. There
are two kinds of constitutional assemblies. First, there are specialist constitutional
assemblies that have no ordinary legislative responsibilities and are elected as part of a
specific constitutional reform process, at the end of which they are dissolved
permanently. The appeal of specialist constitutional assemblies lies largely in the
theory, which is questionable, that they may be put above ordinary politics and
beyond the control of political parties (e.g. if political parties are expressly excluded
from fielding candidates). Second, there are dual-purpose elected constitutional

32International IDEA | Forum of Federations
Territory and Power in Constitutional Transitions
assemblies that simultaneously serve as legislatures. These bodies are also created as
part of a specific constitutional reform process and are almost always elected in a
public campaign in which the constitutional agenda is the main or a central issue.
Political parties openly field candidates and operate according to the conventions of
government and opposition (although there may be a broad-based government).
After the constitution comes into force, this body may become an ordinary legislature
that exercises power under the constitution. A specialist constitutional assembly was
elected in Bolivia (alongside the existing legislature, and named the Constituent
Assembly) and dual-purpose constitutional assemblies were elected in India (as part
of the lead-up to independence) and also in Iraq, Nepal, South Africa and Spain.
There can also be alternatives to elected or fully elected constitutional assemblies.
The Bomas process in Kenya used a hybrid body that brought together
parliamentarians and non-politicians.
As these examples illustrate, constitutional assemblies may be used in cases of
multiple politically salient territorial cleavages (e.g. in India and Spain), and
politically salient cleavages that are both territorial and non-territorial (e.g. in Bolivia,
Iraq, Kenya, Nepal, South Africa and Yemen). In addition, they can be used as part
of peaceful democratic politics (e.g. in Bolivia, Kenya and Spain) or in civil war
stalemates (e.g. in Iraq, Nepal and South Africa). In Bosnia, peace negotiations and
constitutional negotiations occurred simultaneously, driven by elite and international
actors, which precluded a constitutional assembly, whereas in Cyprus the
negotiations have been led by the heads of the two elected governments, with
ratification subject to approval in parallel referendums. These latter cases suggest that
constitutional assemblies may face particular difficulties when dealing with contexts
of a majority alongside a large, territorially based minority.
The main argument made by proponents of specialist constitutional assemblies is
that they guard against the risks of institutional self-interest and self-replication by a
body dominated by political parties (Elster 1995). However, there are good reasons
to be sceptical of this argument. As a practical matter it is difficult to keep parties out
of politics, and their buy-in will be necessary for implementation. Additionally,
constitutional assemblies can fall prey to abuse. In Latin America and the former
Soviet Union, presidents have convened constitutional assemblies for partisan ends,
in order to wrest control of the constitution-making process from legislatures
controlled by opposition parties. Urged on by the executive, these constitutional
assemblies have also inferred from their unlimited constitution-making powers the
authority to exercise plenary power unrestrained by any constitutional limitations.
They have seized legislative powers and reconstituted other institutions at will and
have expanded executive powers and shrunk legislative powers. Sometimes, to
provide democratic legitimacy, presidents have used plebiscites to ratify a
constitution produced by a constituent assembly.
These experiences suggest that there is a need for caution regarding the role of
specialist constitutional assemblies, especially in the context of territorial cleavages,
which are in large part claims for the dispersal of centralized public power. Bolivia
provides a cautionary tale. The Constituent Assembly (CA) was created by legislation
that was the result of a compromise between the Movimiento al Socialismo (MAS)
that favoured a centralized regime and Poder Democrático Social (PODEMOS) that

International IDEA | Forum of Federations33
3. Constitution-making processes
advocated regional autonomy. Each party controlled a chamber of Congress, making
any agreements on regional autonomy and on broader constitutional reform difficult.
They therefore opted for a CA that could approve a new constitution only by a two-
thirds majority. The CA was obliged to create a framework for regional autonomy
that departments had already opted into by referendum. It was far from non-partisan
and although MAS won a majority in the CA it fell short of securing the super-
majority. The CA became deadlocked, which led MAS to argue that the CA had
unlimited constituent authority to make decisions by a simple majority, and therefore
to overturn the two-thirds voting requirement. This led PODEMOS to walk out of
the CA. MAS then used its control of the CA to draft a constitution and to attempt
to take it to a referendum, which was blocked by the Electoral Court, which held that
any referendum required congressional assent. Given the risk of civil war, Bolivia’s
neighbours intervened with an offer of mediation. The parties went back to the
negotiating table and Congress (not the CA) ultimately approved a new constitution,
which passed by 60 per cent in a national referendum.
The purported defect of a dual-purpose constitutional assembly—that it vests a
primary constitution-making role in the same political parties that will wield power
under the constitution—may in fact be a strength. A constitution-making process
that excludes elites may be counterproductive because it is those elites who will
operate the gears of the constitutional democracy created and can undermine it from
within. Allowing political parties to bargain around self-interest favours the stability
of the resulting constitution, whose framework they will see as being more
advantageous than the alternatives of force and fraud. Even when a constitution has
been adopted in a referendum, if political leaders do not buy into it, they may not
implement it. A vivid example comes from Iraq, where the hastily drafted and
incomplete constitution was approved by a referendum in October 2005. However,
its provisions on federalism are unimplemented because most Arab political leaders,
who dominate the central government, felt no ownership of the constitution-making
process, resented extensive autonomy for the Kurdish region, and did not support
federalism in the Arab areas.
The Bomas process in Kenya provides another example of the potential risks of
excluding political elites from the process of constitution-making. Public
consultation and the initial preparation of a constitutional draft were overseen by the
arm’s-length and largely non-partisan Constitution of Kenya Review Commission.
This draft was then referred to the National Constitutional Conference (NCC), a
hybrid body with members drawn from the legislature, the districts, political parties
and civil society, which was to review, potentially amend and adopt the commission’s
draft and transmit it to Parliament. Parliament was then to vote on it without
amendment, with a two-thirds majority required for approval. This arrangement
proved unsustainable because the earlier phases had not satisfactorily accommodated
the views of the political parties, which used their control of the final, parliamentary
step to make substantial revisions to the draft. This draft was rejected in a
referendum that was, in many ways, a plebiscite on the president. The subsequent
Kenyan constitutional process set up a Committee of Experts to work directly with
Parliament via a Parliamentary Select Committee; this draft passed in a referendum.
Yemen provides another example of the same problem because the national dialogue

34International IDEA | Forum of Federations
Territory and Power in Constitutional Transitions
failed to engage, at least directly, with the most important actors in the country. The
members of this large body were in many cases non-partisan, but they were loosely
organized and not truly capable of negotiating on difficult subjects. Those who were
affiliated with political parties or political interests had leaders who were pulling
strings from outside the process and who often tried to obstruct agreement. In the
absence of elections, it was difficult to determine who could really speak for different
parts of the population, but at no point did the major leaders come together with a
view to negotiating.
A crucial issue for regional and political minorities advancing claims in a dual-
purpose assembly is to avoid majoritarian decision-making that disregards the wishes
of regional minorities. For example, in the context of a civil war stalemate in Iraq, the
Constitution Drafting Committee of the National Assembly acted on the basis of a
simple majority, which meant that the views of the Kurds were consistently outvoted.
However, senior representatives of the US Government intervened and took control
of the process because of their concern to meet established deadlines. This gave the
Kurds, who already had a well-established regional government, considerable weight
in the final drafting of the Constitution (even beyond the advantages they enjoyed
under the decision-rules for the ratifying referendum: see section 3.3). The most
obvious and powerful tool for protecting minority interests in a constitutional
assembly is to require super-majority agreement—such as through a two-thirds
threshold—as was used in Bolivia, Nepal, Spain and South Africa. However, super-
majority requirements increase the risk of deadlock, as happened for a time in both
Bolivia and Nepal, which can be highly destabilizing in a major transitional context.
Minorities may still be outvoted even when there is a super-majority requirement, as
happened in Nepal, where the Madhesi and Tharu minorities were marginalized.
Mechanisms can be set down in a constitution or in an interim agreement to
overcome deliberative deadlock. South Africa’sInterim Constitution provided that,
in the event that the Constituent Assembly could not reach the super-majority of
two-thirds, a simple majority in the assembly could submit a draft to carry out a
referendum with a 60 per cent approval threshold; if that referendum failed, there
would be elections for a new assembly. In cases where deadlock-breaking mechanisms
have not been spelled out in advance, it has fallen to the courts to craft a solution,
which may be to turn matters over to the electorate. In Nepal, the Supreme Court
dissolved the deadlocked Constituent Assembly after a constitutionally imposed
deadline had been extended four times; the new elections changed the composition of
the Constituent Assembly, which finally succeeded in meeting the two-thirds
threshold. In Bolivia, the Electoral Court ruled some proposed referendums
unconstitutional, but held that Congress could enact referendum legislation, which it
eventually did as part of a constitutional settlement. In effect, the court remanded the
issue to Congress, which would have faced the electorate on a fixed timetable had it
not resolved the issue.
Finally, although there are challenges in achieving agreements that include support
from territorial and other minorities, occasionally there can be risks in the other
direction, when somehow a majority has been marginalized in a constitutional
settlement. Arguably this happened in Iraq, as discussed above.

International IDEA | Forum of Federations35
3. Constitution-making processes
3.3. Ratification
The final stage of a constitutional process is ratification. When there is constitutional
continuity, the ratification of constitutional amendment by a legislature typically
requires a super-majority as opposed to the simple majority needed for ordinary
legislation. Some ratification procedures end when the national legislature or
constitutional assembly has voted in the required numbers (if necessary, by recourse
to the deadlock-breaking mechanisms described above). Other procedures give the
final say to the electorate in a referendum. In these cases, the legislature may be able
to approve a draft by a simple majority, with final ratification then requiring a
majority in the referendum. Several federations also require the consent of some
proportion of regional legislatures or populations (by referendum). In South Africa,
for example, constitutional amendments that affect the powers of the upper house
(the National Council of Provinces), provincial boundaries, functions, powers or
institutions, or provincial competences require the approval of two-thirds of the
provincial delegates in the upper house as well as two-thirds of the members of the
lower house. In India, most constitutional amendments require a two-thirds majority
in both houses of parliament; given that the upper house is indirectly elected by the
state legislature, the states therefore have an indirect role in the process. India is
unique among federations in providing that the creation of new states or changes to
state boundaries can be approved by simple majority in the two houses of parliament,
without requiring the consent of the affected states (though they must be consulted).
In Nepal in 2005, the seven-party alliance of Nepali political parties that entered
into negotiations with the Maoists wished to maintain legal continuity, but a
constitutional process based on the old constitution was unacceptable to the Maoists,
and the compromise was an interim constitution. The interim constitution provided
that the Constituent Assembly could ratify the constitution by a two-thirds vote, but
there was also a constitutional commitment to making decisions by ‘consensus’. In
the final stages of the Nepali process, there were major tensions over whether
ratification could proceed based on a two-thirds vote or whether there should be
consensus in addition. Legally, the former was applicable, but the constitutional
reference to consensus weakened the legitimacy of the two-thirds rule and the process
was opposed in major demonstrations by the Madhesis and Tharus, who did win
some minor concessions.
Ratification is a different matter when there has been legal rupture, in that those
who have the power can decide the rules. An extreme case was the Dayton
Agreement, where the USA basically drove the process, and the principals to the
negotiation, the presidents of Croatia and Serbia, did not hold office in Bosnia and
Herzegovina. The constitution was an annex to the peace agreement with no further
process for ratification. However, even in such cases of arguably illegitimate
constitutions, the politicians who are meant to operate within the new regime must
decide whether to respect the new constitution or risk system breakdown. In Bosnia
and Herzegovina, local politicians have certainly tested the limits of the regime. The
procedure for ratification in Iraq was also essentially imposed by the occupying forces
of the USA and consisted of approval by the transitional National Assembly and then
a national referendum vote. Ratification in the referendum required a national

36International IDEA | Forum of Federations
Territory and Power in Constitutional Transitions
majority, but would fail if more than two-thirds of registered voters in more than two
governorates voted no. This latter rule was originally designed to protect the Kurdish
minority, which in the end voted for the constitution (partially because they had so
much influence over it). Ultimately, Sunni governorates almost torpedoed it (in two
governorates more than two-thirds voted no, and in one more governorate the two-
thirds threshold was almost met).
A referendum is the most inclusive approach to ratification in that it provides the
electorate with a chance to pronounce on a draft constitution (or amendment)
prepared by its representatives. A positive result in a referendum can add to the
legitimacy of constitutional change; this was the result in Bolivia and Spain.
However, referendums are risky, especially in polarized situations where the
constitution contains difficult and controversial compromises. In Cyprus, the Annan
agreement was subject to two parallel referendums, one in each community, and was
rejected by Greek Cypriots. Of course, a negative vote in a referendum may prove
salutary, as in Kenya, where voters in the constitutional referendum in 2005 rejected
the draft, partly because of its content, but equally because of opposition to the then
president. A new constitution was approved in Kenya by referendum in 2010.
Referendums are sometimes used not for formal ratification but as tests of public
opinion. This was true in Scotland, where voters rejected independence in a
referendum that had no legal status but was accepted as politically decisive. In
Bolivia, national and local referendums were used as tests of strength (along with
elections) in the contest over the country’sconstitutional reform; these eventually set
the context for the final negotiations and the ratification of the new constitution by
referendum. In Ukraine, President Kuchma’s threats to hold consultative, non-
binding referendums (which public opinion polls suggested he would win) gave him
the leverage to force the Rada to adopt his constitutional proposals. Elections,
whether to constitutional assemblies or to ordinary parliaments, are the usual field of
political battle in democracies and they can critically influence the course of a
constitutional process. This happened in Nepal, when the configuration of the new
Constitutional Assembly was able to reach the two-thirds majority necessary to ratify
a constitution after the previous Constitutional Assembly had failed, over many years,
to do so.
Important post-ratification steps can also be required for the full implementation
of a constitutional settlement. In Spain, for example, the new autonomous
communities (ACs) had to be created. In cases of mergers of provinces this was done
by votes of municipal councillors, or in a few cases by local referendum. Moreover,
each AC needed to negotiate its autonomy statute, which then was ratified by
legislatures both national and regionally. In the Philippines, in the settlement on
Bangsamoro a tough negotiating issue related to the use of referendums at the
implementation stage to determine the exact boundaries of the new territory.
Very few recent constitutions (in contrast to many established federations) give a
formal role to territorial representatives or populations in ratifying constitutional
change. The Iraqi referendum rule stands out as an exception, as does that of Cyprus,
which requires both the Greek-Cypriot and Turkish-Cypriot communities to vote in
favour of it for ratification. The British Government’sacceptance of the legitimacy of
a Scottish referendum on independence as a binding expression of opinion falls short

International IDEA | Forum of Federations37
3. Constitution-making processes
of formal ratification, but certainly gives a decisive voice to the regional population.
Similarly, the negotiation of settlements with insurgents in peripheral territories in
north-east India, Aceh and Mindanao gave representatives from these areas a central
role in the substance of the agreements, but no role in their formal ratification.
Super-majority rules can sometimes give significant political leverage to territorial
minorities around the ratification of a constitution or amendment, as happened in
South Africa and in Spain. However, these rules are not normally equivalent to a
power of veto by a regional minority. In Nepal, the aim to achieve consensus
prolonged the process but eventually gave way to a two-thirds majority that overrode
strong objections, notably from the Madhesis, who vociferously objected to the new
provincial map. In Yemen, the national dialogue could not resolve the number and
boundaries of the new regions, so the president created a mechanism to approve his
preferred option; both the Houthis and the Southern Movement rejected the
proposed regional structure, which was a key factor in the outbreak of civil war.
In summary, ratification rules are fundamental for shaping the events preceding
ratification and for the legitimacy or acceptance of the eventual result. The principal
parties may be constrained by the ratification rules, but in cases of state collapse, or
even in cases of total regime change within a context of constitutional continuity, the
lead actors may be able to determine or shape the rules of ratification.

38International IDEA | Forum of Federations
Territory and Power in Constitutional Transitions
4. Constitutional design and
territorial cleavages
4.1. Constitutional moments and constitutional design
A key test of constitutional designs or innovations is how well they function, which
means how appropriate the design is to the context. Even designs that were contested
initially may come to be accepted if they function well. And although design does
matter, the history of a constitution-making process can also have lasting effects in
terms of how different groups view the legitimacy of the outcome. There are often
powerful political imperatives to develop ‘permanent’ constitutions, not least in
conflict-affected countries. Although the processes of making such constitutions can
take more or less time and be done in different ways, there comes a point when a new
constitution is deemed to be operational, and that becomes a fact with its own
consequences (Choudhry 2008).
Nonetheless, there are three major constitutional design options to respond to
claims for the accommodation of territorial cleavages. The applicability of each
option has a presumptive logic, most closely related to the underlying political
geometry, and therefore should be considered closely in the appropriate political
context. These logics are summarized briefly in the following subsections.
Symmetrical federalism or devolution with a majoritarian central
government
Symmetrical federalism or devolution is common and covers many varieties and
degrees of decentralization. It should be considered in countries whose political
geometry is highly territorialized (e.g. Ethiopia, India, Nigeria and Spain) but also in
those countries that have a mix of territorial and other cleavages (e.g. Bolivia, Kenya,
Nepal and South Africa). Moreover, symmetrical federation may be combined with
asymmetric, special autonomy arrangements for smaller units (see section 4.2). The
leading example is India. The major design issues with symmetrical federalism
include the number of the constituent units and their boundaries, the protection of
minority rights (nationally and within constituent units), territorial representation
within central institutions (notably upper houses), the form of the legislature and the
executive, and the extent of the devolution of powers. When there are multiple

International IDEA | Forum of Federations39
4. Constitutional design and territorial cleavages
cleavages the political dynamics can be quite fluid, with governing coalitions varying
over time, which potentially limits how often any group or territory is excluded from
power. Majoritarian government at the centre is broadly accepted, but upper houses
often give extra weight to representatives of small territorial units. The examples of
Iraq and Yemen are hard to place. In principle, Iraq opted for symmetrical
federalism, but as the design was incoherent and was never implemented the Kurdish
region effectively has special autonomy. A constitution could not be agreed to in
Yemen, but the design that was considered provided symmetrical devolution to 6
regions and 21 governorates, while having special power-sharing provisions within
the legislature between northern and southern representatives.
Highly devolved federal government with a consociational central
government
This difficult institutional arrangement is sometimes chosen when the political
geometry is of two (or sometimes three, as in Bosnia and Herzegovina) antagonistic
and territorially separate communities that must cohabit within a single state.
Typically, in this system of government there is a majority and a minority
community and the latter must be of significant size for this option to be considered.
Mutual mistrust suggests that the communities should carefully consider maximum
devolution or self-government for each community. However, the inevitable fact that
some important functions remain with the central government should draw these
communities to consider consociational government, a particular form of power-
sharing in which the agreement of a majority of representatives of each community is
necessary for certain specified major decisions. Belgium has a highly developed form
of consociational federalism. The example of Cyprus illustrates the difficulties of
negotiating such an arrangement, even when it appears to be necessary.
Special autonomy for territories in a federal or non-federal state with a
majoritarian central government
This option is most appropriate in cases where the political geometry includes very
small, peripheral territorial populations that have a strong sense of distinct identity.
The national majority may resist claims by such groups for special autonomy for
ideological reasons, but as a practical matter such arrangements can work well if the
population is very small relative to the total population (as in Aceh in Indonesia or in
north-eastern India) or if the extent of special status is not too extensive. Special
autonomy can be more problematic, as in Scotland, when the population is relatively
large and the devolution is very extensive. In Ukraine, the Minsk agreement envisages
special autonomy for the districts in Donbas that are not government controlled, but
this idea has attracted no support in the Rada. In Sri Lanka, special autonomy might
have made sense for the Northern Province, but it was never the focus of negotiations
and this option has been foreclosed by the military defeat of the LTTE.
Of course, there are alternatives to these three options, even when a country is
highly territorialized or has strong claims for autonomy in some regions. One is a
centralized, unitary regime in which there is little or no accommodation of territorial
cleavages, and there is possibly an attempt to assimilate minorities. Another is
secession, where the break-up of the state is accepted as preferable to a difficult

40International IDEA | Forum of Federations
Territory and Power in Constitutional Transitions
cohabitation. Secession was the avowed aim of rebels in India, in Indonesia, in the
Philippines and in Sri Lanka, but in all of these cases the rebel factions were forced to
settle for less. Majorities can often favour a strong, indivisible ‘nation’ and resist
options for devolving or sharing powers. It is therefore not inevitable that the
apparently most appropriate model for territorial accommodation will be adopted:
that will depend on the details of political geometry and the relative strength and
ideologies of the key players. Moreover, when a country is riven by particularly deep
territorially based conflicts, it may be that there is no constitutional design option
that will produce relatively functional and sustainable democratic politics.
Although a constitutional moment that results in change is usually thought to end
with an agreement on a new or revised constitution, in reality it is one step in a
longer process, which includes the critical issue of implementation. A territorial
accommodation, whether in the form of a constitutional amendment or a statute,
launches further processes whereby that accommodation is operationalized in
practice, in the course of which it will be interpreted, reviewed and adapted over time
to changing circumstances. Moreover, on many occasions, the very same
disagreements that gave rise to the demand for the territorial accommodation in the
first place will resurface in debates over how to translate that accommodation into
practice. Constitutional territorial accommodations do not necessarily resolve issues;
rather, they may manage them by establishing some points of consensus and
channelling remaining issues into institutional mechanisms, where they are dealt with
through continuing political processes.
4.2. Symmetrical federalism or devolution with a majoritarian central
government
Most federal and devolved systems are symmetrical in that the territorial subunits
have the same powers and responsibilities and their national institutions are
majoritarian in design. This system seems suitable for countries that are highly
territorialized with several significant regions, such as India and Spain—which also
have asymmetric, special autonomy arrangements for some subunits—and to those
where there is a mix of territorial and non-territorial cleavages, such as Kenya or
South Africa. Key design issues within this model are the number and boundaries of
the constituent units, the type of central legislature and executive, the upper house,
the electoral system, and the allocation of powers and fiscal resources between levels
of government.
Of these design issues, the determination of the number and boundaries of the
constituent units was a major issue in several countries (e.g. Bosnia and Herzegovina,
Ethiopia, India, Kenya, Nepal, Nigeria, South Africa, Spain and Yemen). The criteria
for forming subunits can include historic boundaries, sociocultural identities,
economics, administrative efficiency, geographic features such as rivers and
mountains, the desired number or size of units, and, inevitably, political pressure
(Anderson 2014). In Spain and in South Africa, the approach to creating units was
largely successful. The redesign of the political maps of India and Nigeria has also
helped stability and political dynamics but there are still pressures for more states. In
Nepal, the strong dissatisfaction among the Madhesi over the failure to get a single

International IDEA | Forum of Federations41
4. Constitutional design and territorial cleavages
province poses a significant political challenge, but a single Madhesi province with 40
per cent of the population could prove even more destabilizing. In Yemen, the
proposed regional boundaries in the North were a casus belli for the Houthis, while
the Southern Movement was hostile to the proposal to establish two regions in the
south.
Most highly territorialized countries have more than one level of territorial cleavage
(e.g. the territories of big confederated tribes versus smaller subtribes) that can
provide an element of choice in drawing boundaries, especially in relation to dealing
with sociocultural identities, such as ethnicity, language and religion, which are all
potentially very symbolic and politically volatile. Ethiopia, India, Nigeria and Spain
are sometimes viewed as ‘ethno-federations’, because territorially based ethnic
political mobilization has been at the root of demands for constitutional
accommodation and has been important in determining unit boundaries. These
countries, however, have taken different approaches. In some cases, ethnic boundaries
largely align with subunit boundaries. In India, the leaders of the independence
movement put off addressing the call for linguistically based states, as they feared the
potential impact of linguistic states on national unity. But within a few years of
independence India acceded to the call for linguistically defined states. The earlier
fears about linguistically based states proved largely unfounded; their creation helped
bring government closer to the population, and increased the competitiveness and
fluidity of national politics because of the rise of new regional political parties that
could enter into coalition governments at the central level. In Spain, virtually
everyone can speak the national language, but the three ‘historic nationalities’(the
Basque Country, Catalonia and Galicia), who have their own languages and represent
over a quarter of the population, harboured strong grievances about their treatment
under the Franco regime. These regions led the push for major devolution, but the
view was that the 50 provinces were too numerous to be the principal units of a
quasi-federal regime based on ACs. The three historic nationalities and the island
provinces were therefore guaranteed status as ACs, whereas the primarily Spanish-
speaking provinces of the mainland were required to amalgamate to reduce the
number of units.
By contrast, in Nigeria, subunit boundaries have been designed to break up the
major cleavages. Nigeria’s original three federal states each embraced a major tribe
and language—Hausa, Yoruba and Igbo—as well as many smaller tribes and
subtribes. The political dynamics were dysfunctional, especially because the northern
state held a majority of the country’s population. After the army coup and war in
Biafra in 1967, the generals moved to restructure the federation into an eventual 36
states where the subunits were designed to break up the cleavages between the three
major populations that were so destabilizing for the country’s politics. The chosen
design of Kenya’s new devolution is somewhat similar to the example of Nigeria
above. Kenya has 40 to 60 ethnic groups and three or four major tribes. Its politics
have been heavily tribal, especially around the few large tribes, but there is no neat
territorial division, especially between its largest tribes. Given the history of tribal
rivalry and conflict, politicians feared drawing up a political map that would reinforce
tribal tensions. The centralizers’ preferred option was a presidential regime with
many, weak districts; others wanted larger, more powerful units. Ultimately, the

42International IDEA | Forum of Federations
Territory and Power in Constitutional Transitions
former option largely prevailed. Many of the 47 counties in Kenya have a majority
tribe, which has posed issues for minorities, but devolution has enabled some groups
in opposition nationally to form regional governments, thereby diffusing power.
The risk of discrimination against minorities within constituent units is a common
challenge of devolution. In India, for example, linguistic minorities within states are
disadvantaged with respect to public employment, political representation and public
services. In Nigeria, such discrimination has even given rise to the constitutional
claim that those classified as ‘indigenes’have more rights (e.g. to state employment
and some educational programmes) than ‘settlers’whose historic roots lie outside the
state, even if the families of these ‘settlers’ have been in the state for generations.
Some federal constitutions protect the rights of internal minorities in constituent
units—notably rights to education and public services in a minority language—but
in practice such protections are often weak because central governments can find that
intervening on behalf of a minority can alienate the subunit’smajority. Marginalized
minorities within federal states may therefore turn to civil unrest and, where a
minority is regionally concentrated, may demand a new constituent unit, as has
happened in India—although there is a limit to such a logic. Sometimes minorities
that have suffered discrimination within constituent units align with those in power
nationally, which can provide some protection for them.
Not all devolved or federal regimes are in countries with a highly territorialized
political geometry. Even in countries whose political geometry is a mix of non-
territorial and territorial cleavages, careful attention should be given to the territorial
dimension when considering the design of its devolved arrangements. This has
certainly been the case in Nepal. It is a country of astonishing diversity, with 125
recognized caste and ethnic groups, and a history of political dominance by the
upper-caste Chetri and Bahuns, who form fewer than 30 per cent of the population.
Although the country’s deep cleavages became evident during the People’s War
(1996–2006), Nepal is not highly territorialized; groups are territorially intermingled,
with only 15 of 75 districts having a single group as the majority. A key demand of
the Maoists, which became the largest party in the first constitutional assembly, was
for ‘ethnic federalism’. This demand was highly contentious, both because of the
initial resistance of the old-line parties to federalism and because there were almost no
natural ethnic units. A consensus was ultimately reached on a form of federalism in
which provinces typically have at least two significant population groups. However,
the Madhesi and Tharu groups argue that the federal map was designed to
disempower them, and this issue continues to fester.
South Africa is a highly diverse country, with 11 official languages. The ANC
deeply resented the Bantustans (the territories set aside for black-only populations) of
the apartheid regime, and it, along with the white minority National Party, was
determined to erase these manufactured territorial cleavages. The interim
constitution created 9 new provinces to replace the 4 previous provinces and 10
Bantustans. These provinces were based largely on past regions of economic
development, not ethnic or linguistic criteria. South Africa’s new devolved structure
has brought some pluralism into its politics, including pluralism within the ANC,
but ‘territorialcleavages’are relatively minor, although the prospect of abolishing the
provinces, which once was mooted, now seems remote.

International IDEA | Forum of Federations43
4. Constitutional design and territorial cleavages
Although countries that have multidimensional territorial cleavages (or a mix of
territorial and other cleavages) generally do not adopt power-sharing within central
institutions, they can nonetheless use other means to balance territorial interests and
reduce political confrontation among regions. The Nigerian Constitution requires
that (a) parties that compete in national politics have a national character, which
means a certain level of membership and number of candidates across the country;
(b) a president must win not just a majority in the country, but also at least one-
quarter of the vote in two-thirds of the states; (c) the cabinet has a member from each
state; and (d) the composition of the civil service must reflect the country’s ‘federal
character’.These arrangements have helped mitigate the most dangerous cleavages.
Nigeria’sconstitutional prohibition on regional parties is unusual, but it points to
the potential importance of electoral laws for translating territorial and other
cleavages into electoral representation and the party system. In a highly territorialized
system, first-past-the-post systems, which favour the largest parties within electoral
districts, can give a distorted representation of local sentiment (Scotland provides an
extreme example, in which the Scottish National Party won 4 per cent of seats with
22.5 per cent of the vote in the 1992 UK election and then 95 per cent of seats with
50 per cent of the vote in the 2015 election). Proportional representation systems
limit these distortions and give a better sense of support for national and regional
parties and so can limit the potentially destabilizing over-representation of regional
parties. Constitutions may be silent on the choice of electoral system or, as in South
Africa, indicate what kind of law should apply but leave the details to legislation.
Most federations have upper houses in their parliaments that serve as the ‘regional
chamber’. Countries that have opted for new constitutions with symmetrical
federalism or devolution—Bolivia, Ethiopia, Kenya, Nepal, Nigeria and South Africa
—all opted for such bodies. Most have equal representation of constituent units in
the upper house, although their methods of selection (e.g. direct versus indirect
election) and terms of office vary. The powers of upper houses in congressional
systems are usually greater than those of upper houses in parliamentary systems,
where the lower house is the confidence chamber and also controls finances. The
upper houses in Ethiopia, Kenya and South Africa are largely empowered to deal
with issues that relate to the territorial units, while the upper house in Nepal is
essentially an advisory body (as is the one in South Africa on national issues). It is
perhaps telling that Spain put off the reform of its Senate, and Iraq’s constitution
provides for an upper house to be determined in due course. The experience with
upper houses is that their members vote more along party than regional lines, thereby
limiting their regional role. To the extent that they rebalance political weight
compared with popularly elected lower houses, this favours smaller constituent units,
so their role in accommodating major territorial cleavages is highly contextual and
usually incidental to design. Another issue that can be very contentious in designing
federal constitutions is the choice of a presidential-congressional regime, a
parliamentary regime or a semi-presidential regime. In practice, the choice made
tends to reflect past institutional history and other factors more than a conception of
how to manage territorial issues, even though these institutions do have implications
for the functioning of the federal or devolved system.

44International IDEA | Forum of Federations
Territory and Power in Constitutional Transitions
Finally, a critical issue is the allocation of powers and finances to the central
government versus the constituent unit governments. South Africa opted for a
centralized form of federalism, based on the German model in which there is
extensive concurrent jurisdiction whereby federal law prevails in the event of conflict
between federal and provincial legislation. In Bolivia, a good part of the conflict was
over revenue allocation; the central government emerged strengthened, yet the gas-
producing departments still retained significant advantages. In Iraq, the hurried effort
to draft a constitution led to a draft that was significantly ambiguous regarding the
allocation of powers, nowhere more glaringly than on the critical issue of powers over
petroleum resources. Of the countries that have written new constitutions in recent
times, Spain probably created the most devolved regime, which reflects the depth of
its territorial cleavages.
4.3. Highly devolved federalism with a consociational central
government
Countries whose political geometry is characterized by two or three major and highly
territorialized communities that are antagonistic to one another should consider the
option of a highly devolved federal structure combined with formal power-sharing or
consociational arrangements at the centre. However, although such arrangements can
have a strong underlying logic, it is important to note that they are difficult both to
design and to manage, and they risk reifying political divisions in the country. They
may be especially difficult if the regional populations are significantly different in
size.
A number of federations that had only two or three units broke up after a relatively
brief existence (Pakistan in 1971, Czechoslovakia in 1992, Sudan in 2011). Nigeria
was fundamentally restructured to escape its largely bipolar north–south dynamic—
although initially with three states—which had contributed to the outbreak of civil
war. None of these bipolar federations had effective power-sharing in the central
government, although Sudan was meant to. Belgium, which was initially a unitary
country, has become the one example of a democracy that has developed a federal
system with a constitutionalized, consociational government at the centre.
As difficult as such federal-cum-consociational arrangements can be, their logic
arises from two (or three) mutually antagonistic communities finding that they are
forced to cohabit within one country, often because of international pressure. The
populations are territorially largely separated—perhaps after regional homogenization
following conflict, as in Bosnia and in Cyprus—so it is possible to create territorially
defined subnational governments for the respective communities that can assume
important powers. At the same time, it is inevitable that some decisions must be
made by the central government. As the population of the smaller community or
communities fears majority decision-making would disempower it, an accord requires
arrangements for power-sharing with each community having a collective veto within
the central government. In Bosnia and Herzegovina, this logic led to the imposed
settlement in the Dayton Accords, and the same logic still underlies negotiations
regarding a possible settlement in Cyprus. There has been some very creative
thinking about how to improve the design of such consociational arrangements

International IDEA | Forum of Federations45
4. Constitutional design and territorial cleavages
(Loizides 2016) but, given its difficulties, this model should be considered only in
exceptional circumstances.
Although the consociational power-sharing arrangement in Bosnia and
Herzegovina has brought peace, it is widely recognized to have created a deeply
problematic government. The arrangement has hardened ethnic identities, so that,
despite international support for free and fair elections and incentives for multi-
ethnic and moderate parties and constraints on ethno-nationalist parties, multi-ethnic
parties have fared poorly and moderate ethnic parties have been drawn to more
extreme positions. The reason is that the constitution organizes politics around
ethnicity, which gives rise to outbidding within ethnic communities. Moreover,
because ethnic appeals occur simultaneously across multiple offices, there is a coattail
effect of campaigns that appeal to hard-line ethnic positions, even at the level of the
constituent entities, where there should be space for greater political pluralism. The
system of ethnic parties has become the basis of a political economy of patronage.
The mechanism for corruption is the requirement that government appointments be
ethnically representative, coupled with government ownership of a large percentage
of the economy. This has led ethnic parties to reward their supporters and themselves
via concessions, procurement and financing. There has been great difficulty in
achieving effective decision-making at the centre because of mutual vetoes. For a long
time, the system relied on the internationally appointed High Representative to break
deadlocks.
The Annan Plan for Cyprus was based on the idea of a bizonal, bicommunal
federation, with consociational arrangements in the central government. The plan
was rejected by the Greek population by referendum in 2004, but this model
remained the basis for the negotiations that failed to reach an agreement in 2017 and
is the likely basis for future negotiations. Although there are many difficult issues in
the negotiations, including security, rights of return, property and territorial
adjustment, the possible arrangements for governance have proven very difficult. As
the Turkish population is only 15 to 20 per cent of the population on the island
(depending on how it is counted), the Greek community has understandable
concerns about equal power-sharing, while the Turkish community feels the need for
it to avoid Greek dominance. The Senate that was proposed in the Annan Plan
would have had an equal number of representatives from both communities, while
the Chamber of Deputies would have had no fewer than 25 per cent of its members
from the Turkish-Cypriot community. Special voting arrangements would have
meant that a significant minority of members from both communities had to approve
the appointment of the executive and make other decisions of vital interest. As in
Bosnia and Herzegovina, the Supreme Court would have had an equal number of
judges from each community plus foreign judges to avoid deadlocks.
Two of the most successful consociational arrangements have been in Belgium and
in Northern Ireland, but in these cases the ratio of majority to minority is close to 60
to 40 or 55 to 45, which makes power-sharing difficult but potentially manageable.
Similarly, in Bosnia and Herzegovina, the minority Serbs and Croats represent 31 per
cent and 16 per cent of the population compared with 51 per cent for the Bosniaks.
In Cyprus, the Turkish community constitutes less than a fifth of the total
population. The constitutional deliberations in Yemen’s national dialogue were

46International IDEA | Forum of Federations
Territory and Power in Constitutional Transitions
driven towards a federal system, but the South’spreferred option of two regions was
rejected; the accommodation of the South was meant to come through significant
power-sharing in the central legislature (but not in the presidency or the courts). In
this case the focus was on giving southern representatives in the parliament vetoes on
matters that touched their interests, yet, with southerners standing at only about 20
per cent of the population, it was unclear how viable this arrangement would be.
In summary, the already difficult challenges of consociational arrangements are
that much greater when the minority is relatively small in that the majority is less
prepared to share power. For the same reason, small peripheral regions that obtain
special autonomy lack sufficient power to plausibly make claims for power-sharing in
central institutions. Indeed, for that reason, consociationalism has never been
seriously considered in Iraq, where the Kurds make up around 15 to 20 per cent of
the national population, or in Sri Lanka, where the Tamil-majority Northern
Province represents around 5 per cent of the national population.
Although consociationalism is rare and difficult to achieve as a longer-term
constitutional design, this type of power-sharing has been used quite effectively on an
interim basis for a period of transitional government during a constitutional
transition in which longer-term arrangements are to be worked out. This strategy
worked well in South Africa and reasonably well in Kenya, but it was clear in both
cases that the longer-term arrangements would be majoritarian—although with
important constitutionalized protections for the formerly dominant minority in
South Africa.
An apparent alternative to a complex federal-cum-consociational structure for two
territorially distinct and mutually hostile populations would be partition and the
creation of new independent countries. Circumstances matter greatly when it comes
to break-up. The international community has been prepared to sanction mutually
agreed break-ups, as it did in several post-communist countries, and has even set the
stage for potential break-up, as it did in mediating the peace agreement in Sudan in
2005. However, the international community is extremely reluctant to reward
military intervention, as in Cyprus, or ethnic cleansing, as in Bosnia and
Herzegovina; moreover, the majority populations in these cases are still not prepared
to let the minority Turks and Serbs secede. The international community is even
reluctant to support more democratic but unilateral attempts at promoting secession,
as can be seen in the isolation of the Catalan and regional Kurdish governments
following their attempts to force independence through referendums. Most
constitutions explicitly or implicitly exclude secession, but there are several significant
exceptions. Moreover, there is some quantitative evidence that partition as a solution
to nationalist wars is more likely to promote post-conflict peace and democracy than
the alternatives of unitary government or regional autonomy, though this may
depend on context (Chapman and Roeder 2007).
4.4. Special autonomy for small territories and a majoritarian central
government
Countries with one or more peripheral regions that are highly territorialized relative
to the rest of the country should consider special autonomy arrangements for these

International IDEA | Forum of Federations47
4. Constitutional design and territorial cleavages
territories, especially if the population of the territory is small relative to the rest of
the country. Such arrangements may be considered following a stalemated
secessionist insurrection and/or in a liberal democratic country that is not doctrinaire
about centralized sovereignty. Given their small size, peripheral territories with
special autonomy are not accorded power-sharing arrangements at the centre.
This option is one of asymmetric autonomy. Such demands for asymmetric
treatment arose in Indonesia, for Aceh and Papua; in the Philippines, for the
Muslims of Mindanao; in India, notably in the north-eastern tribal areas; and in the
UK, for Scotland. (There are also special fiscal arrangements in Spain’s Basque
Country and Navarre, but these represent a historical arrangement.) In most of these
cases in which special autonomy has been granted, the relevant territorial population
is very small relative to the national population: Aceh, 2 per cent; Bangsamoro, 5 per
cent; and in India’s north-east, six new states carved out of Assam together account
for 1 per cent of India’spopulation. This practice is consistent with other comparable
cases, such as the Åland Islands, 0.5 per cent; Jammu and Kashmir, 1 per cent; and
Zanzibar, 2 per cent. There is a strong logic to special autonomy being largely
restricted to territories with relatively small populations. Small peripheral populations
do not expect or demand power-sharing in the central government. They can
maintain their normal representation in central institutions, and their representatives
can even vote on matters that do not affect them because they are such marginal
players. In federations, the larger constituent units are more likely to accept special
autonomy for a very small, very distinct territory than they would for a large
constituent unit, which they would see as their equivalent.
The small relative size of most regions that have special autonomy helps to explain
the constitutional dynamics that surround devolution for Scotland. Devolution in
the UK is asymmetric, both because there is no devolution to England (and so
nothing that resembles federalism) and because the three devolved administrations
each have different arrangements. The devolved administrations cover 15 per cent of
the UK’spopulation, with Scotland accounting for 8 per cent. In comparative terms,
Scotland has a very large relative population for a unit with special autonomy and it
also has exceptionally extensive devolved powers. This arrangement has created the
still unresolved structural issue of the role of MPs from Scotland and the other
devolved regions voting on legislation that applies only to England. This is an
especially acute question in a parliamentary regime in which the House of Commons
is the confidence chamber for the government. There is no neat solution. The UK
example shows the risks of special autonomy if it applies to a significant part of the
country and to a wide range of powers. The Kurdish region in Iraq, with about 15
per cent of the country’spopulation, effectively has very extensive special autonomy
so this may pose similar issues about the role of its representatives in the national
government in the longer term.
Although special autonomy can function well in countries that have small
minorities, majorities can object strongly to special autonomy, even for very small
and peripheral regions, often for symbolic reasons or out of fear of promoting
eventual secession. There has been frequent opposition to special autonomy for
Mindanao from the Philippine majority, so the deal done by the Aquino
administration has not been ratified and it may yet be replaced by the alternative of

48International IDEA | Forum of Federations
Territory and Power in Constitutional Transitions
‘statestatus within a federal Philippines’proposed by President Duterte. The idea of
special autonomy for the Tamils in northern Sri Lanka never found its way onto the
agenda, perhaps because it would have been unacceptable to the Sinhalese majority,
but in principle the population in question (especially if the Eastern Province were
excluded) was small enough that such an arrangement might not have been
destabilizing. Structural indicators favourable to special autonomy are therefore no
guarantee it will be adopted. It may be that, on occasion, symmetric devolution
across the whole country prepares the way for special autonomy in some region or
regions. In Indonesia, for example, decentralization laws in 1999 and 2004 for all
292 regencies preceded the introduction of the special autonomy law in Aceh. The
general decentralization law changed the political baseline and made Aceh’s special
autonomy more acceptable across Indonesia.

International IDEA | Forum of Federations49
5. Conclusions
5. Conclusions
Constitutional transitions can be difficult and even perilous. Many transitions fail to
reach a conclusion or produce a result that does not endure. Although transitions in
which significant territorially based populations assert claims for autonomy can face
an extra level of difficulty, there are well-established models (e.g. federalism,
devolution, special autonomy for small regions, or even highly devolved federalism
with central power-sharing) that have provided many countries with a political
architecture that suits their needs. This paper has indicated how these different
models should be considered in particular contexts of political geometry—that is, of
particular arrangements of territorial and non-territorial political cleavages. That said,
there is no ‘best’ model, even for societies that may have broadly similar political
geometries; each country must work to develop its own arrangements, taking
inspiration and lessons from others as appropriate. Whatever general architecture is
adopted, myriad details must be worked out. Some deeply divided societies have a
political geography that is particularly challenging. They may have little sense of a
shared national identity, and the disaffected territorial minority or minorities may
seek very substantial autonomy (even secession) that would not be acceptable to the
majority as the basis for either symmetrical federalism or asymmetrical special
autonomy. This might be especially the case if the territorial minority’s (or
minorities’) relative population size is not large enough to make consociational
power-sharing at the centre acceptable to the much larger majority, but is large
enough that extensive special autonomy could destabilize the functioning of
representative institutions at the centre.
There are some cases where a potentially applicable model is not considered or is
excluded from consideration because one side or the other insists on one major
option or a very limited range of options. This is often for ideological and
nationalistic reasons; majorities often resist territorial accommodation because they
fear it will eventually lead to secession. Empirically, there is little evidence to support
this fear, and territorial accommodation has often been a critical part of stabilizing
democratic regimes. Unitary regimes themselves are not immune to secessionist
movements; some unitary countries have fallen into civil war or dissolved.
Whatever the political context or design options to be considered, the process of
constitution-making itself deserves careful thought, and different actors may try to
shape it to their advantage. In circumstances where constitutional continuity is being

50International IDEA | Forum of Federations
Territory and Power in Constitutional Transitions
maintained and the political context is peaceful, the forums and formulas for
constitutional change will be established and act as a constraint. Depending on the
circumstances, an amending formula that requires a super-majority may give a
territorial group (or groups) political leverage, but often this will not be the case.
Most unitary constitutions are silent on how to deal with the special issues of
territorial politics in constitution-making, but the many examples of significant
process innovations serve as supplements to constitutionally mandated processes;
these can be important when dealing with claims for territorial accommodation.
Constitution-making in post-conflict situations is distinct from that in peaceful
circumstances of constitutional continuity. When the violence has been an
insurgency in a small, peripheral region in which there has been a stalemate in the
fighting, attempts at resolution will normally be through negotiations between the
government and insurgents. The government will maintain that the existing
constitution and its processes apply; however, constitutions do not establish
procedures for dealing with rebels. The substantive issue in such cases is almost
always special autonomy, but a major question can be whether the new arrangements
will be effected through a constitutional amendment or through ordinary statute.
The former provides greater protection, but often involves a high threshold of
approval that may be difficult to achieve politically. Resort to an ordinary statute may
be easier politically, but it needs to comply with the existing constitution in the view
of the courts. The courts, the political opposition and the government backbench are
potential institutional spoilers in giving effect to a political agreement, which will
require a strategic response by government negotiators.
The situation is quite different after a conflict in which a clear victor has emerged.
If it has been a general civil war, the substantive issue may be the drafting of a new
constitution. Victorious rebels can do this in a rupture from the previous
constitutional order; however, sometimes victorious generals on the government side
operate by decree, unencumbered by the constitution they purportedly defended. In
these cases, it is natural that the victors will move to impose their preferred
constitutional model, but in doing so they may alienate the losers or even some of
their allies. When the combat has been more regional and the government has
prevailed, it may be constrained procedurally by the constitution, but in any case
have little interest in substantive reform that would accommodate the wishes of the
territorial insurgents. In both of these cases, there is a risk of ‘victor’sjustice’in which
the political arrangements that emerge do little or nothing to address underlying
structural problems in the country.
The processes adopted for constitutional transitions will be shaped by many of the
same factors that influence the range of possible constitutional designs. Each of the
three major design options that accommodate territorial cleavages within a state
(federalism-devolution; asymmetric autonomy; federalism with consociational power-
sharing) has particular relevance in the context of a specific type of political
geometry, and they are all subject to many possible variations and refinements. Even
so, there can be political contexts where no one of these seems quite right and where
a hybrid might be considered. Of course, secession and strongly enforced unitary
government are also options, but they do not respond to the political call to
accommodate territorial cleavages within a continuing state.

International IDEA | Forum of Federations51
References and further reading
References and further reading
References
Anderson,G.,CreationofConstituentUnitsinFederalSystems(‘Meetingthe
ChallengesofEmergingConstitutionalDemocracy’),CenterforConstitutional
Transitions Working Paper (2014), <http://www.constitutionaltransitions.org/
publications/creation-of-constituent-units-in-federal-systems>, accessed 19
November 2018
Arato, A., Post Sovereign Constitution Making (Oxford: Oxford University Press,
2016)
Chapman,T.andRoeder,P.,‘Partitionasasolutiontowarsofnationalism:the
importanceofinstitutions’,American Political Science Review, 101/4 (2007), pp.
677–91
S. Choudhry (ed.), Constitutional Design in Divided Societies: Integration or
Accommodation? (Oxford: Oxford University Press, 2008)
S. Choudhry and T. Ginsburg (eds), Constitution-Making (Cheltenham: Edward
Elgar, 2016)
Elster,J.,‘Forcesandmechanismsintheconstitution-makingprocess’,Duke Law
Journal,45(1995),pp.364–96
Loizides, N., Designing Peace: Cyprus and Institutional Innovations in Divided Societies
(Philadelphia, PA: University of Pennsylvania Press, 2016)
O’Donnell,G.,andSchmitter,P.,Transitions from Authoritarian Rule: Tentative
Conclusions about Uncertain Democracies (Baltimore, MD: Johns Hopkins
University Press, 1986)
Whitehead, L., Democratization Theory and Experience (Oxford: Oxford University
Press, 2003)

52International IDEA | Forum of Federations
Territory and Power in Constitutional Transitions
Further reading
Amoretti, U. and Bermeo, N., Federalism and Territorial Cleavages (Baltimore, MD:
Johns Hopkins Press, 2004)
Anderson, B., Imagined Communities: Reflections on the Origin and Spread of
Nationalism (London and New York: Verso, 1998)
Anderson, G. and Choudhry, S., Constitutional Transitions and Territorial Cleavages
(Stockholm: International IDEA, 2015), <https://www.idea.int/publications/
catalogue/constitutional-transitions-and-territorial-cleavages?lang=en>, accessed
13 December 2018
Bell, C. and Zulueta-Fülscher, K., Sequencing Peace Agreements and Constitutions in
the Political Settlement Process, International IDEA Policy Paper No. 13
(Stockholm: International IDEA, 2016), <https://www.idea.int/publications/
catalogue/sequencing-peace-agreements-and-constitutions-political-settlement-
process?lang=en>, accessed 28 November 2018
Brancati, D., Peace by Design: Managing Intrastate Conflict through Decentralization
(Oxford: Oxford University Press, 2009)
Cheibub,J.,Elkins,Z.andGinsburg,T.,‘Beyondpresidentialismand
parliamentarism’,British Journal of Political Science,44/3(2014),pp.1–30
Collier,P.andHoeffler,A.,‘High-valuenaturalresources,developmentandconflict:
channelsofcausation’,inP.LujalaandS.AasRustad(eds),High-Value Natural
Resources and Post-Conflict Peacebuilding (Abingdon: Routledge, 2011), pp.
297–312
Ghai, Y. and Woodman, S. (eds), Practising Self-Government (Cambridge: Cambridge
University Press, 2013)
Gurr, T., Peoples vs. States: Minorities at Risk in the New Century (Washington, DC:
United States Institute for Peace, 2002)
Horowitz, D., Ethnic Groups in Conflict, 2nd edn (Berkeley, CA: University of
California Press, 2000)
Kymlicka, W., Multicultural Odysseys: Navigating the New International Politics of
Diversity (Oxford: Oxford University Press, 2009)
Levitsky, S. and Way, L., Competitive Authoritarianism: Hybrid Regimes after the Cold
War (New York: Cambridge University Press, 2010)
Lijphart, A., Patterns of Democracy: Government Forms and Performance in 36
Countries (New Haven, CT: Yale University Press, 1999)
Norris, P., Driving Democracy: Do Power-Sharing Institutions Work? (New York:
Cambridge University Press, 2009)

International IDEA | Forum of Federations53
References and further reading
Simeon,R.andConway,D.-P.,‘Federalismandthemanagementofconflictin
multinationalsocieties’,inA.GagnonandJ.Tully(eds),Multinational
Democracies(Cambridge:CambridgeUniversityPress,2001),pp.338–65
Watts, R., Federal Second Chambers Compared,Queen’sUniversityInstituteof
Intergovernmental Relations Working Paper 2008-02 (2008)
Wimmer, A., Ethnic Boundary Making: Institutions, Power, Networks (Oxford:
Oxford University Press, 2013)
Wimmer, A., Waves of War: Nationalism, State Formation and Ethnic Exclusion in the
Modern World (Cambridge: Cambridge University Press, 2013)

54International IDEA | Forum of Federations
Territory and Power in Constitutional Transitions
About the authors
George Anderson was a deputy minister (permanent secretary) in the Canadian
Government and subsequently Chief Executive of the Forum of Federations. He has
been a member of the Standby Team of Experts in the UN’sDepartment of Political
Affairs and has consulted extensively around the world. His Federalism: An
Introduction (Oxford University Press, 2009) has been translated into 23 languages,
with over 100,000 copies in print, and his Fiscal Federalism: A Comparative
Introduction (OUP, 2010) is also widely translated. He has edited books on
federalism and oil and gas, internal markets in federations, and rivers in federal
systems as well as having written several chapters and working papers on issues related
to devolved government. He is currently a fellow at the Centre for Democracy and
Diversity at Queen’s University, Kingston, Canada, and has had earlier resident
fellowships at Harvard University and New York University.
Sujit Choudhry is an internationally recognized authority on comparative
constitutional law, and has been an advisor to constitution-building, governance and
rule of law processes for over 20 years, including in Egypt, Jordan, Libya, Nepal,
South Africa, Sri Lanka, Tunisia, Ukraine and Yemen. He founded and directs the
Center for Constitutional Transitions. He has published over 100 articles, book
chapters, policy manuals, reports and working papers. His edited volumes include
The Oxford Handbook of the Indian Constitution (Oxford University Press, 2016),
Constitution-Making (Edward Elgar, 2016), Constitutional Design for Divided Societies
(OUP, 2008) and The Migration of Constitutional Ideas (Cambridge University Press,
2005). His current project is Comparative Constitutional Law: A Global and
Interdisciplinary Approach (with M. Versteeg, D. Landau and T. Ginsburg) (OUP,
forthcoming). He sits on the Board of Editors of the International Journal of
Constitutional Law (ICON) and the Executive Committee of the International
Society of Public Law (ICON-S).

International IDEA | Forum of Federations55
About the organizations
About the organizations
Center for Constitutional Transitions
The Center for Constitutional Transitions (CT) generates and mobilizes knowledge
in support of constitution-building by assembling and leading international networks
of experts to produce evidence-based policy options for decision-makers and agenda-
setting research, in partnership with a global network of multilateral organizations,
think tanks, and NGOs. CT has worked with over 50 experts from more than 25
countries. CT’s projects include Security Sector Reform and Constitutional
Transitions in New Democracies; Territory and Power in Constitutional Transitions;
Security Sector Oversight: Protecting Democratic Consolidation from Authoritarian
Backsliding and Partisan Abuse; and Semi-Presidentialism and Constitutional
Instability in Ukraine.
<http://www.constitutionaltransitions.org>
Forum of Federations
The Forum of Federations, the global network on federalism, supports better
governance through learning among federal experts and practitioners. Active on 6
continents and supported by 10 federal countries, it manages programmes in
established and emerging federations and publishes scholarly and educational
materials.
<http://www.forumfed.org/>

56International IDEA | Forum of Federations
Territory and Power in Constitutional Transitions
Foundation Manuel Giménez Abad
The Foundation Manuel Giménez Abad for Parliamentary Studies and the Spanish
State of Autonomies is a Foundation with a seat at the regional Parliament of Aragon
in Zaragoza. Pluralism is one of the main features of the work of the Foundation. In
fact, all activities are supported by all parliamentary groups with representation at the
Parliament of Aragon. The main objective of the Foundation is to contribute to the
research, knowledge dissemination and better understanding of parliamentary studies
and models of territorial distribution of power. In general terms, the activities of the
Foundation are concentrated in four key areas: political and parliamentary studies;
territorial organization; Latin America; and studies on terrorism.
<http://www.fundacionmgimenezabad.es/>
International IDEA
The International Institute for Democracy and Electoral Assistance (International
IDEA) is an intergovernmental organization with the mission to advance democracy
worldwide, as a universal human aspiration and enabler of sustainable development.
We do this by supporting the building, strengthening and safeguarding of democratic
political institutions and processes at all levels. Our vision is a world in which
democratic processes, actors and institutions are inclusive and accountable and
deliver sustainable development to all.
In our work we focus on three main impact areas: electoral processes; constitution-
building processes; and political participation and representation. The themes of
gender and inclusion, conflict sensitivity and sustainable development are
mainstreamed across all our areas of work. International IDEA provides analyses of
global and regional democratic trends; produces comparative knowledge on good
international democratic practices; offers technical assistance and capacity-building
on democratic reform to actors engaged in democratic processes; and convenes
dialogue on issues relevant to the public debate on democracy and democracy
building.
Our headquarters is located in Stockholm, and we have regional and country
offices in Africa, the Asia-Pacific, Europe, and Latin America and the Caribbean.
International IDEA is a Permanent Observer to the United Nations and is accredited
to European Union institutions.
<http://idea.int>

Collective demands for the constitutional accommodation of
territorial cleavages are pervasive across very diverse contexts.
In many countries, political identification on the basis of
territory is a central basis of political mobilization, around
which political claims are framed, political parties formed,
elections contested, governments composed, and
constitutional claims made and resisted.
Constitutional transitions dealing with significant territorial
cleavages face political dynamics and challenges that are
quite distinct from such transitions where such cleavages are
absent. They can involve different constitution-making
processes, sometimes with representatives of territorial
groups having a formal or informal role, and certainly
involve different political dynamics, whatever the defined
roles of territorial groups are in the process.
This Policy Paper seeks to provide insights into how
territorial claims relate to constitution-making processes and
constitutional design, and to offer advice that may be useful
to principals and advisors engaged in constitutional
moments: periods in which there has been intense political
engagement over how to respond constitutionally to
significant demands for territorial accommodation from one
or more regions.
International IDEA
Strömsborg
SE–103 34 Stockholm
Sweden
Email: info@idea.int
Forum of Federations
75 Albert Street, Suite 411
Ottawa, ON K1P 5E7
Canada
Email: forum@forumfed.org ISBN: 978-91-7671-228-3 (PDF)