The Reforms That Cost Nothing
A government swept into power on a two-thirds mandate to implement sweeping constitutional reform. Seven months later, it has spent much of that mandate undoing the reforms that would have constrained it most. In short, that is the story of Bangladesh’s “July National Charter” since the Bangladesh Nationalist Party (BNP) took office in February 2026—a story less about broken promises than about a government quietly discovering which promises it can afford to keep.
The Charter itself was born of crisis. Negotiated during Chief Adviser Muhammad Yunus’s interim government in the wake of the 2024 uprising, it was meant to be the constitutional settlement that made the uprising’s sacrifice mean something—term limits, judicial independence, a rebalanced parliament, safeguards against exactly the kind of one-party dominance that had just been overthrown. Voters were asked to ratify it in a referendum held alongside February’s parliamentary elections, and by most accounts they did, resoundingly: different sources put support at around 61 percent. The BNP, meanwhile, won in a landslide. On paper, mandate and majority pointed in the same direction. In practice, they have pulled apart.
The referendum ballot bundled four broad questions covering only a fraction of the Charter’s numbered provisions—by one count, at least 20 out of the 47 proposals were named explicitly, with the rest left to inference. This ambiguity wasn’t accidental. The Charter text still carries “notes of dissent” from parties that had signed it but disagreed with specific provisions, and the Consensus Commission overseeing the referendum recommended that those dissenting notes not be factored into the outcome at all—meaning a “yes” vote would settle even the provisions parties had explicitly refused to sign onto without reservation. That is no minor ambiguity. It gives the BNP the margin needed to claim fidelity to “the people’s mandate” in its speeches, while treating most of the document as negotiable.
Where the government has yielded, it has done so in a way that cost it nothing. A bicameral parliament, including the Charter’s proposed 100-member upper house, has progressed in some form, and nothing suggests the BNP objects to the institution itself—only to who gets to partake. That distinction has major implications. The Charter wanted the upper house to be filled by proportional representation, a corrective to a first-past-the-post system whose distortions were on full display in February. Jamaat-e-Islami took 32 percent of the national vote (68 of 300 seats). The BNP’s manifesto instead proposes allocating upper-house seats in proportion to what parties already won in the lower house—a formula that would simply reproduce whichever party’s lower-house dominance already exists. Given that the party currently enjoying that dominance is the BNP, the preference is not hard to assume.
The rest of the contested middle ground follows a similar shape. A two-term limit on the prime minister remains exactly where it was left during negotiations: flagged as a BNP dissent item, neither enacted nor renounced. A more radical proposal—separating the premiership from party leadership, so that no single figure holds both simultaneously—never made it into the BNP’s manifesto at all, a detail that reads rather differently once you note it would have applied directly to Prime Minister Tarique Rahman's own chairmanship of the party. And a proposed oversight committee, designed to keep appointments like the Human Rights Commission and the Anti-Corruption Commission at arm’s length from the executive, fared no better. the BNP argued such safeguards belonged in ordinary law rather than the constitution. This seemingly technical preference leaves the whole arrangement one ordinary parliamentary vote away from being repealed.
Underneath all of this sits a question no amount of political manoeuvring can fully address: does the referendum bind anyone to anything? A writ challenging the constitutionality of the Implementation Order behind the vote has been filed and admitted by the courts, alongside a challenge to the procedure for swearing in MPs as members of a new “Constitution Reform Assembly.” The BNP itself refused that second oath, insisting reform must proceed through the ordinary constitutional-amendment process requiring a two-thirds majority—a majority the party happens to hold outright, which is to say the objection doubles as a claim to unilateral control over the reform timeline.
However, not every reform was left to stall—some were actively reversed. The interim government had passed two ordinances giving the Supreme Court an independent secretariat and taking the executive out of judicial appointments. The BNP has repealed both, notwithstanding years of BNP complaints about a judiciary captured by the previous regime. Something similar happened at the central bank, where governor Ahsan Mansur—a former senior IMF official credited with stabilizing the economy through the transition—was removed without explanation and replaced by a party loyalist. Neither action appears as a numbered line item in the Charter, but both point in the same direction as everything else: toward a government more willing to dismantle independence it inherited than to build independence it never wanted in the first place.
Beyond stalling and reversal, there's a third category: the things nobody even pretended to promise. Nowhere in the Charter or the referendum is there a mechanism for transitional justice regarding the mass killing of protesters in 2024, a death toll estimated at anywhere from 800 to 1,400. Nowhere are the anti-discrimination reforms that protesters had chanted for in the streets. A reform promise can only be as ambitious as its drafters allow, and regarding the two demands most directly tied to why people rallied in the first place, the drafters chose silence—one the new government has shown no interest in breaking.
It would be too easy to label all of this as simple bad faith: a party cynically abandoning commitments the moment it gains power. The more nuanced interpretation (and one this pattern would support) is that almost every provision that has stalled or reversed would have limited the concentration of power in the hands of whoever governs next. A party negotiating a constitution it expects to inherit has every incentive to weaken these exact clauses, regardless of who that party is. The Awami League’s ongoing ban completes the picture. Where the Charter’s provisions would have constrained BMP’s power, the ban protects it: leaving the party's only plausible electoral rival excluded requires no action at all. Taken together, the two cases show a consistent rule: the BNP accepts reforms that cost it nothing and resists anything that would. Some observers warn that permanently excluding Bangladesh's oldest party trades short-term convenience for long-term instability, but that is a cost that falls on the system, not the party.
Bangladesh is not the only place in the region where a popular uprising has recently toppled a long-entrenched government, and it will not be the last place where the harder test comes after the celebration—when the coalition that tore something down has to decide what, if anything, it will allow to be built in its place. What makes the Bangladeshi case worth watching closely is how transparently the mechanism of retreat has worked, through a string of small choices—placing it in ordinary law, keeping the old seat formula, and declaring nothing about repeal. Yet together, their behavior resembles that of the institutions the mandate targeted, rewriting its terms in their own favor. Whether the courts, the opposition, or the next election can arrest that process remains the open question.