When a government overrides a referendum
In a parliamentary democracy, electorates do not directly choose their leaders. A referendum, by contrast, offers citizens a rare opportunity to speak directly on specific questions. That is why a referendum is supposed to mean more than an opinion poll.
When the question concerns the Constitution, the distinction becomes fundamental: Parliament represents the people, but the people themselves exercise constituent power. Bangladesh now faces a test of that principle.
On 12 February 2026, Bangladesh held a referendum on constitutional reforms under the July National Charter, on the same day as the parliamentary election. The referendum produced a decisive “Yes” vote. The constitutional reform package therefore received a direct popular mandate from almost 70% of voters.
The BNP, under Tarique Rahman's leadership, secured a commanding parliamentary majority on the same day. The political dispute arose when the new government declined to follow the institutional mechanism envisaged for implementing the referendum mandate. Rather than constituting the proposed Constitution Reform Council, BNP lawmakers took their parliamentary oath without taking the additional oath envisaged for members of the Reform Council.
The government has instead pursued constitutional reform through a special parliamentary committee. The BNP maintains that the July Charter Implementation Order and the proposed Reform Council lack a proper constitutional basis, allowing the referendum ordinance to lapse despite having campaigned for a “Yes” vote.
This is where Bolivia's experience becomes relevant.
Bolivia's warning
In February 2016, Bolivians were asked a constitutional question about changing the consecutive presidential term limit so President Evo Morales could run again. Their answer was no. Some 51.3% voted against the proposed constitutional amendment, while 48.7% supported it. The result was unambiguous.
The electorate had rejected the attempt to remove the presidential term limit. One might have expected the matter to end there. It did not.
In November 2017, in a highly controversial ruling criticised by opponents as biased towards the governing party, Bolivia's Constitutional Tribunal held that constitutional and statutory restrictions on re-election could not be applied to Morales because political rights protected by international human rights law entitled him to seek another term.
The decision effectively neutralised the constitutional restriction that voters had refused to remove through a referendum. Morales was consequently permitted to run for a fourth consecutive term in the 2019 election.
The remarkable aspect of the Bolivian case was that the government did not persuade the people to change their answer. An institution of the state effectively changed the constitutional consequences of that answer. Morales's subsequent political trajectory offers a warning about what happens when constitutional legitimacy becomes contested. There is a resemblance here to Bangladesh's apex court nullifying the caretaker government system.
Morales contested the 2019 election. The result became deeply controversial. The Organization of American States' subsequent audit identified serious irregularities and deliberate manipulation that, in its assessment, made it impossible to validate the announced result.
Separately, the organisation's human rights bodies rejected the proposition that unlimited presidential re-election was an autonomous human right under the American Convention. Mass protests followed. On 10 November 2019, Morales resigned as president and left the country.
Bolivia's 2019 crisis had multiple political, institutional and electoral causes. Yet the overriding of the 2016 referendum was central to the legitimacy dispute surrounding Morales's fourth-term candidacy and contributed to his downfall. The lesson extends beyond presidential term limits.
When citizens are told that another institution may overrule their direct constitutional decision, the damage is not confined to one constitutional provision. It can undermine confidence in the entire constitutional order. Once constitutional legitimacy erodes, elections themselves may cease to settle political disputes. That is what makes Bolivia relevant to Bangladesh.
The referendum: mandatory or directory?
A referendum may be legally binding or merely consultative. Its effect depends on the applicable constitutional and legal framework. A government may legitimately argue that a referendum expresses political will without itself amending the constitutional text.
But that is not the same as saying a government can accept a referendum when convenient and disregard it when inconvenient.
Bangladesh must now confront a question: if the February referendum was approved by voters who voted “Yes”, what is the political status of that decision? If Parliament remains entirely free to disregard its institutional consequences, the referendum risks becoming little more than political theatre.
That would create an unusual democratic hierarchy. The February parliamentary election produced a representative mandate for the BNP. The referendum produced a direct mandate from the same electorate on constitutional reform. These were two different votes on two different questions. Can a parliamentary majority automatically be treated as a mandate to overturn the second?
This distinction matters particularly because the BNP's electoral victory was so decisive. A government with such a mandate has substantial democratic legitimacy. But representative legitimacy and constituent legitimacy are not necessarily interchangeable. The government represents the people in Parliament. In a referendum, the people speak directly.
Other countries offer different lessons
History offers several examples of governments addressing this problem differently. In Greece in 2015, voters rejected proposed bailout conditions by more than 61%. Yet the government subsequently negotiated and accepted a new bailout arrangement.
This was not technically a cancellation of the referendum because the subsequent agreement was a new arrangement. But the episode demonstrated how a government can pursue a course that many voters understood as contrary to the political meaning of their referendum vote.
Ireland offers a different approach. Voters rejected the Lisbon Treaty in 2008. Rather than dismissing the result, the government secured additional assurances and returned to the electorate in 2009. Voters then approved the treaty, reversing the earlier decision through another popular vote.
France followed a more controversial route. French voters rejected the proposed European Constitution in 2005. Three years later, France approved the Lisbon Treaty through parliamentary procedures. The two treaties were legally different, although critics argued that the Lisbon Treaty reproduced much of the substance of the rejected document.
These examples show that a referendum can be followed by a different political outcome. The question is who has the authority to produce that outcome. Another Parliament? Another court? A new referendum? Or the same government choosing to disregard the first result? These are constitutionally distinct choices.
Bangladesh's real constitutional dilemma
This controversy must not be reduced to a dispute between the BNP and its adversaries. The deeper issue concerns the relationship between the power that establishes a constitutional order and the institutions established by it. Parliament derives its authority from the Constitution. The people are the ultimate source of constitutional authority.
When citizens exercise that authority directly through a politically accepted referendum, Parliament cannot casually treat their decision as an ordinary election promise.
This does not mean every “Yes” vote automatically inserts new provisions into the Constitution. Constitutional amendments still require legally valid procedures. Nor does it necessarily mean Parliament loses its constitutional amendment powers. The question is whether Parliament is implementing the popular decision through a legally valid procedure or replacing it with its own preferred constitutional arrangement.
If the latter becomes accepted practice, the precedent will extend far beyond the July Charter. Imagine a future government losing a referendum on judicial reform. It could argue that the referendum was politically important but legally inconvenient. Another government could lose a referendum on the electoral system and subsequently amend the Constitution in Parliament to achieve substantially the same objective.
Another could disregard a referendum concerning the balance of powers between the President and Parliament.
Once that principle is established, the referendum ceases to be an exercise of constituent power and becomes a political consultation whose consequences depend on the goodwill of the government in office. That is the danger illustrated by Bolivia.
Morales did not simply lose an argument over presidential tenure. The institutional decision to circumvent the popular rejection helped create a legitimacy crisis surrounding his subsequent candidacy. The 2019 election became disputed, the streets became an arena of political contestation, and the constitutional government itself entered a crisis.
Bangladesh should take that lesson seriously, not because Bangladesh is Bolivia or because the two political situations are identical, but because constitutional legitimacy presents a common democratic challenge. A government that wins an election possesses a mandate to govern.
A government that wins a parliamentary supermajority possesses a mandate to legislate and, where constitutionally permitted, amend the Constitution. But neither proposition automatically answers a third question: what happens when the same electorate has already spoken directly on a constitutional question?
One democratic response would be to return to the electorate if the government considers the referendum unlawful, its implementation mechanism unconstitutional, or the circumstances fundamentally changed. Yet this presents another difficulty. If the government publicly campaigned for a “Yes” vote, it signalled acceptance of the process. How, then, can it return the issue to voters without acknowledging that its earlier position was misleading or evasive?
Another option would be to seek a judicial determination on the referendum's legality or constitutionality. Yet the consequences of judicial intervention in deeply political questions, evident in Bangladesh's caretaker government verdict and Bolivia's Morales decision, make this approach less advisable. Judicial resolution of a political dispute can create further political complications. Not every political question is best settled as a judicial one.
Amid a complete opposition boycott, the government's chosen path is to implement its own version of the July Charter by amending the Constitution through a special parliamentary committee. But this approach risks opening a Pandora's box: establishing a precedent under which a popular mandate can be rejected with little more than a shrug.
What we must fear is a fourth option: the people vote, the government disagrees, and the vote is silently stripped of any consequence.
Concluding remarks
The BNP's overwhelming electoral victory gives the Tarique Rahman government considerable democratic authority. That authority should make it more confident, not less, in respecting a separate mandate given by the same voters on the same day.
The ultimate test of a constitutional democracy is not whether a government can win. It is whether, after winning, it can accept the limits imposed by the people themselves. Bolivia demonstrates what can happen when that distinction is blurred. Bangladesh has an opportunity to learn that lesson before the costs become greater.
For the government's chosen path to succeed in a society as politically divided and confrontational as ours, it must address the resentment and indignation this controversy has generated through careful negotiation, political accommodation and statesmanship.
--
The writer is a researcher on constitutional and legal affairs