The logic of peacemaking is often at odds with the logic of governance. The outcomes of Tajikistan's June 1997 General Agreement seems to offer yet another demonstration of this contradiction. Before the war, the system of governance in Tajikistan was based on centralisation of political power in disregard of regional allegiances, monopolisation of state authority by executive organs to the detriment of Parliament and the courts, exclusion of opposition elements from meaningful participation in political life, and flawed electoral procedures. These elements rendered that system inadequate to contain conflict and may have played a direct role in generating it. The General Agreement was intended, among other things, to provide a mechanism – through the work of the Commission on National Reconciliation (CNR) – to recommend a set of reforms to the political system and institutions. The recommendations would then be taken up through the existing constitutionally mandated legal procedures.

The General Agreement therefore provided Tajikistan with an extraordinary opportunity, unique among post-Soviet states, to refashion its institutional arrangements for governance. The CNR could have convened a constitutional convention or a similar far-reaching process to make its recommendations for reform. The peace process could have been the catalyst for Tajikistanis to undertake deep-rooted evaluation of the systems underpinning their country's governing structures and of the legal system needed to address inherited problems. Yet the General Agreement did not address the serious democratic deficits of the political system derived from Soviet rule. Furthermore, in the transitional period, reforms were made that created a new layer of bureaucrats beholden to the central authorities and, ultimately, to the President. This was complemented by an increased concentration of power at the centre and the gradual incorporation of the opposition into government. It now seems that the reformed political institutions cannot adequately facilitate participatory political life, pluralism, and the protection of rights. Although the end of fighting and a resumption of a degree of political stability have been welcomed by most Tajikistanis, there are concerns that the reformed system of governance may be unable institutionally to manage problems that could generate future political conflict.

The limited extent of post-Agreement institutional reforms is probably due to the relatively narrow political vision of the negotiators, rather than to technical non-compliance with the Agreement's provisions. The international sponsors of the negotiations are arguably complicit in the failure of the opposing parties to design a peace treaty that would create systems to promote democratic consolidation. The international community sponsored the negotiations; foreign states were guarantors of the General Agreement; the UN and organisation for Security and Cooperation in Europe (OSCE) were expressly empowered to provide assistance to the CNR. They used their leverage to encourage the settlement of other political disputes during the CNR's work. Yet it appears that they did not exert influence on the parties to adopt constitutional and legal measures that would strengthen the balance of power or other tools for enhancing democratic governance and the rule of law. This raises important questions about the commitment of the international sponsors to promoting these principles while carrying out their peacemaking work.

To explore these issues, it is necessary to analyse the laws and institutions proposed in the General Agreement and the history of institutional change in the three years since it was concluded. The focus is on: 1) the constitutional amendments drafted by the CNR that were adopted and approved by popular referendum; 2) the parliamentary elections of 2000; and 3) the fate of the power-sharing arrangements contemplated in the agreement and the structure of Tajikistan's governance now that the CNR has completed its work and been disbanded.

The texts of the Protocols that constitute the General Agreement did not explicitly address institutional reform. Their primary focus is expressed in the all-encompassing but only partly elaborated term: 'national reconciliation'. In addition to providing amnesty, the agreement appeared to contemplate a set of structural reforms, based on changes to laws, in order to achieve and institutionalise 'national reconciliation'. For example, the General Agreement provided for new electoral laws and constitutional amendments. Yet the text is as significant for its omissions as for its provisions. The virtue of the Agreement lay in its flexibility and the potentially comprehensive role it carved out for the CNR. Its vice lay in its failure to identify clear directions or guidelines for the CNR in its work of constitutional reform and in the structural bias in favour of the pre-existing system.

The CNR was composed of 26 members, designated by the Tajik government and the United Tajik Opposition (UTO) 'on the basis of the principle of parity' and headed by the leader of the UTO with a government representative as deputy. Members could not be removed unless they failed to fulfil their duties. The Commission was to be financed under a special line item of the state budget. The CNR's Charter made it exclusively responsible for:

• developing a mechanism to oversee the fulfilment by the parties to the Agreement of their obligations under it;
• implementing measures for the safe return of refugees and their active inclusion in the social-political and economic life of the country, and for rehabilitation of war-damaged housing, industrial and agricultural objects;
• developing proposals for reform of laws on the functioning of political parties, movements, and the mass media.
• It conferred on the CNR and the President the following joint responsibilities 'for the transitional period':
• submission to national referendum of the proposed constitutional amendments;
• preparation and submission for approval by Parliament, and if necessary to nation-wide referendum, of a new law on parliamentary and local government elections;
• creation of a Central Election Commission;
• government reform: recruitment of UTO representatives to the structures of executive power including ministries, agencies, local administrative bodies, judicial and law-enforcement organs, in the light of the principle of regional parity;
• management and control of demobilisation, disarmament, and reintegration of armed opposition forces, reform of security forces and the procuracy;
• supervision of the exchange of all prisoners-of-war and liberation of forcibly detained persons;
• adoption of an Act of mutual forgiveness and the drafting of a law on amnesty, to be adopted by Parliament and the CNR;
• submission for parliamentary review of proposals for the date of elections to a new professional parliament under the supervision of the UN and the OSCE with the participation of those with observer status in the inter-Tajik negotiations.

For a body with broad administrative and law-drafting powers, the CNR was ill-defined and dependent on the government. Its charter lacked procedural safeguards to ensure its independent functioning, apart from the provision for 'close coordination with the UN and OSCE missions'. Potentially the most significant role of the CNR – to draft constitutional amendments and new legislation – was shared with the President, who already had the power to appoint half the Commission members. This put the UTO, an alliance representing diverse political groupings, at a structural disadvantage that could hamper its capacity to shape policy and decision-making in the transitional period. It can thus be argued that the CNR was from its inception vulnerable to pressure and to being ultimately co-opted by government leaders.

The focus of the CNR's work was to facilitate cooperation between former opponents so that they could reach consensus on how to implement the General Agreement and shape the constitutional and legal arrangements that would guide the country's future. Yet during the two years of the Commission's activity, the UTO appeared to shed much of its oppositional character. Many members increasingly adopted government-friendly positions and in some cases even switched party affiliation. Although the capacity for opposing parties to cooperate on matters of overarching concern is important in all democracies, the experience of Tajikistan's politics during the transitional period may have undermined the development of a strong 'loyal' opposition that could maintain accountability and political pluralism.

On 30 June 1999 Tajikistan's parliament, the Majlisi Oli, debated and adopted the amendments to the constitution prepared by the CNR and formally proposed by the President. Twenty-eight out of 100 articles of the constitution were substantially revised, twenty-one of them completely. The main structural changes introduced were the creation of an upper house of Parliament representing the regions and the extension of the presidential term from five to seven years, with a one-term limit. Amended Article 28 provided the constitutional framework for parties based on religion for the first time. The logic, essence and contents of the proposed amendments are troubling from the perspective of democratic consolidation.

The creation of an upper house, the Majlisi Melli, was intended to address the perennial problem of regionalism in Tajikistan's politics by ensuring regional representation and reducing inter-regional tensions. Bicameralism – when a country's legislature consists of two 'chambers' – is typically but not exclusively a feature of federal polities, where the upper house affords representation to the federal units, as in the US Senate or the Russian Federal Council. The CNR did not recommend a federal system for Tajikistan. Yet its bicameralism might be seen as a quasi-federal compromise. It provides some mechanisms to give a voice in decision-making at the centre to the regions without granting them formal regional autonomy (except in the special case of Badakhshan, which already had a degree of constitutionally mandated autonomy). Although members of the Majlisi Melli are chosen on the basis of regional representation, the primary powers conferred on the upper house are over the justice system: to elect and recall judges and to approve the appointment and dismissal of leaders of the prosecutor's office. Therefore they lack the legislative powers to respond to concerns raised by their regional constituencies.

In the context of Tajikistan's highly centralised executive branch, which has responsibility for the administrative functions of government, the methods of electing members of the Majlisi Melli arguably strengthen the hand of the executive at the expense of the legislature and the judiciary. The new system may, in fact, diminish the legislature's independence and authority in comparison to those of its unicameral predecessor. Election to the Majlisi Melli is indirect. Three-quarters of the membership are elected by deputies of the aggregate local assemblies of each province, of the capital, Dushanbe, and of the Mountainous Badakhshan Autonomous Province. The system provides equal representation for the five regional administrative units or provinces. The remaining quarter is appointed directly by the President. The parliamentary election law of 1999 permits local executive officials – who are ultimately accountable to the President, as employees of the executive branch of government – to stand for election to the Majlisi Melli. Over 80 per cent of the members elected or appointed in the 2000 elections were such officials. Furthermore, the Majlisi Melli convenes only when called by the President (the lower house, the Majlisi Namoyandagon, convenes regularly).

The extension of the presidential term from five to seven years makes it the longest constitutionally mandated term of office for a chief executive within the former Soviet Union. The restriction to one term in office appears to shorten the term any one individual might serve by three years (from two consecutive five-year terms to one seven-year term). Yet the incumbent President Rakhmonov had already been in office for almost a decade before his election to a newly extended term in November 1999.

Another amendment authorised the formation of political parties 'among others, parties of a democratic, religious, or atheistic character'. This article was drafted primarily to legalise the activity of the Islamic Renaissance Party. Yet the disjunctive legal formula is unfortunate because it suggests that parties can be either democratic or religious, but not both. It further suggests that democratic parties are only part of a much larger universe of parties.

In addition to the substance of the amendments, some people have expressed concerns about the lack of transparency and participation in drafting and adopting the amendments. Owing to the circumstances in which it was created, the CNR was not a democratically representative body. It might have thought to seek legitimacy by a consultative process. Its drafting work was instead conducted in virtual secrecy, without consulting the other political parties, civic organisations, the scholarly community, or the general public. The government enjoyed a virtual monopoly of trained legal drafting skills and therefore had a considerable advantage in the actual drafting of amendments. Moreover, the process of referring amendments to a referendum did not comply with the constitutionally mandated procedures: the draft text was given to parliamentarians two days before the vote and the final text only one hour before it.

Perhaps the most significant aspect of the constitutional amendments is what they failed to do. They did not introduce real structural reform or alter the highly centralised form of government characteristic of all post-Soviet states (with the arguable exception of the Russian Federation). They did not contemplate any substantive regional autonomy or devolution, much less federalism. For example, the President has the power to appoint the heads of regional authorities, although members of local assemblies are elected directly. Nor did the amendments provide safeguards for greater transparency and accountability in the exercise of administrative authority or seek to curb undue administrative discretion. They did not enhance the system for the protection of rights or the freedom from administrative interference of the political process and the activity of political parties.

They also did not seek to redress pervasive gender inequalities in the public and private sectors. Tajik institutions have yet to incorporate a contemporary approach to gender or to introduce norms of non-discrimination and differential treatment, as appropriate or fair in given circumstances. The constitution addresses gender inequality as a separate matter, although only in the context of grounds for equal treatment. Moreover, the language used in the text relies exclusively on the masculine grammatical forms. Separate anti-discrimination or rights-protection legislation on the basis of gender does not exist. Legislation requiring differential treatment on the basis of gender (such as labour legislation) employs the vocabulary of assistance, thus perpetuating an ideology of female passivity.

Significant opposition to the proposed constitutional amendments arose from many quarters, as reflected in the independent press. Twenty-nine per cent of votes cast in the referendum rejected the amendments, a significant number in a region where ruling parties regularly poll 98 per cent or more in national elections. Most Tajikistanis nevertheless accepted the referendum's outcome, as indeed did the UN and the OSCE.

The February-March 2000 elections were a watershed in the history of independent Tajik politics, with national elections contested on a multi-party basis. Despite the shortcomings of the CNR and the constitutional reform process, and despite the extraordinary political circumstances in which they unfolded, these elections represented a step forward in the search for a new and authentic form for expression of the popular will.

The conduct and results of the elections nonetheless revealed a number of serious substantive and procedural flaws in the election law drafted by the CNR. As in the case of the constitutional amendments, the law-creating process itself was problematic. Drafting was for the most part in the hands of the government. Opposition members of the CNR were limited to the role of observers. The OSCE made available some external technical drafting expertise, but with uncertain effect. The law was adopted swiftly and with minimal debate. It was passed on 10 December 1999 and published the following day, when the President set the elections for 27 February 2000. The constitutionally mandated periods of notification were not observed.

The law and its implementation deserve careful study to guide further refinement and reform of the electoral process, perhaps including the development of a new Electoral Code. Although the norms established by the new election law appeared to open the floodgates to genuine political contest, they were for the most part declaratory and lacked procedural means for implementation. The concrete mechanisms of the law conferred on the central authorities and the Dushanbe elite several strategic and tactical advantages that, in practice, limited the participation and influence of opposition forces. Under the election law, twenty-two of the sixty-three members of the Majlisi Namoyandagon would be elected from republic-wide party lists on the basis of proportional representation, with a 5 per cent minimum threshold. (In initial versions of the law, only ten seats would have been filled by proportional representation.) The remaining forty-one members would be elected from regional constituencies on a simple first-past-the-post system. These seats would be contested by both party-affiliated candidates and self-nominated (independent) candidates able to collect 500 signatures.

Contesting the seats chosen on the basis of proportional representation meant, on a practical level, an ability to mount a country-wide electoral campaign. This in effect required a national party base and finances for mass media advertising. A consequence of this system is that those parties with the social base and political resources to contest a country-wide campaign were at an advantage. In the first round of elections, President Rakhmonov's People's Democratic Party did especially well, while parties with a primarily regional base did less well. Eligibility to contest parliamentary seats required prior party registration with the Ministry of Justice. Several parties were excluded on this basis, among them the 'New Opposition' – including the Agrarian Party, Junbesh National Movement of Tajikistan, and the Party of Justice and Development Tajikistan (renamed the Social Democratic Party in October 1999), and some others. They were variously denied registration, banned, or removed from the register.

In the elections, seventeen of the twenty-two deputies elected to the republic-wide seats were from Dushanbe and fifteen from the People's Democratic Party. The centralised character of the elections was reinforced because party-nominated candidates were able to stand simultaneously for the national constituency and the single-mandate regional constituencies. This rule placed independent candidates at a further disadvantage. Interestingly, aggregate voter support for the party-list candidates of the national constituency was little more than half the support for candidates in the single-mandate constituencies.

The provision for government reform in the CNR Charter was amended by protocol to give the UTO a 30 per cent quota of positions throughout the government and 25 per cent of members in the Central Election Commission. This quota was filled at most senior levels but not at the level of line officials. Yet it is possible to argue that a power-sharing arrangement on the basis of a quota, even if fully implemented, is an unpromising instrument of reform. It leaves the structure intact, does not provide for meaningful integration of outsiders, fails to stipulate participatory procedures for policy- and decision-making, and creates incentives to preserve an unreformed system. Most UTO representatives brought into government experienced pressure to adopt the government's views if they wished to retain their appointments.

Tajik political institutions after the CNR era are, if anything, more centralised than they were before. Of particular concern is the trend towards concentration of state power in the hands of the executive, which is ultimately accountable to the president, at the expense of the legislature and the judiciary. The structural bias of the new Parliament has been described. The Majlisi Melli is dominated by members who are simultaneously agents of executive power in their role as officials of the local hukumat administrations. The legislative staff of the Majlisi Oli (but not the parliamentarians themselves) and the judicial staff of the courts (with the exception of judges themselves) are all employees of the executive. The executive also retains control over local administrations. The president is both head of state and head of government and appoints or dismisses the prime minister and cabinet. In addition, the presidential administration duplicates many of the positions and functions of the government, exhibiting a parallelism of ruling structures (known traditionally as dvoevlastie or double rule) familiar from Soviet times.

The parliamentary elections marked the ending of the transitional period envisioned in the General Agreement. In the minds of most people, this signalled a welcome end to the Tajik conflict of 1992-97. Yet a legal-institutional analysis reveals that the underlying foundations of the country's new governance structures appear to replicate significantly the features of the old system that failed to mitigate political conflict and may, in fact, have contributed to it. This reveals the tension between the need to reach agreements that will restore stability after a war and the risk of exacerbating problems in governance that may give rise to future conflict.