After the Palace Falls

Public Accounting and the Unfinished Politics of Reform in Bangladesh and Sri Lanka

Bangladesh and Sri Lanka have now passed through two different versions of the same post-uprising problem. Sri Lanka removed Gotabaya Rajapaksa through mass protest in July 2022, then waited more than two years before elections gave Anura Kumara Dissanayake and the National People’s Power a decisive mandate. Bangladesh removed Sheikh Hasina in August 2024, installed an interim government under Muhammad Yunus, held a parliamentary election and constitutional referendum in February 2026, and returned to elected rule under the Bangladesh Nationalist Party.

The two countries did not follow the same sequence. Sri Lanka moved from revolt to elite continuity, then to electoral correction. Bangladesh moved from revolt to interim constitutional engineering, then to elections. Their shared problem is that leadership turnover does not by itself dismantle state capture. A ruler can leave office while the civil service, police, courts, procurement system, debt office, public employment rules, media controls, land records, and party finance networks remain available to the next governing coalition.

Public accounting is the discipline that turns state power into visible records, reviewable reasons, and enforceable remedies. It is not only financial reporting. It asks who received the contract, who approved the variation, who holds the debt, who obtained the public job, who delayed the case, who refused the complaint, who ordered force, who paid, who benefited, and what correction followed. Without that discipline, reform remains vulnerable to personality politics. The old ruler leaves, but the administrative method survives.

This is why many uprisings disappoint their own authors. They reveal a captured order, but they do not automatically produce a replacement operating system. Protest can break fear, expose illegitimacy, and force executive removal. It cannot, by itself, rewrite procurement rules, publish debt, depoliticize hiring, protect judges, discipline police, repeal anti-dissent laws, or make party finance transparent. Those tasks require institutional sequencing. They also require a clear distinction between necessary administrative continuity and elite restoration.

Political science gives part of the vocabulary for this problem. State capture describes the distortion of laws, policies, regulations, and administrative decisions by organized political or private interests. Competitive authoritarianism describes systems in which elections, courts, parties, and media exist, but incumbents abuse state resources, coercive power, legal tools, and information control so heavily that competition becomes structurally unfair. Bangladesh under Hasina and Sri Lanka under the Rajapaksas were not identical cases, but both show how formal democratic institutions can be used to preserve unequal access to the state.

The practical question is not whether a post-uprising government uses the language of democracy. Most do. The question is whether it changes the record systems through which capture is reproduced. If the state job is still allocated through opaque advantage, the public contract still hides its real beneficiary, the debt book still conceals risk, the court still delays without consequence, the police file still disappears, and the speech law still turns dissent into security risk, the uprising has changed the holder of power more than the structure of power.

Two Paths After Revolt

Sri Lanka’s path was delayed electoral correction. The Aragalaya removed Gotabaya Rajapaksa, but Parliament, still dominated by the Sri Lanka Podujana Peramuna, selected Ranil Wickremesinghe as president. Wickremesinghe restored parts of macroeconomic management and negotiated through the IMF track, but his government also moved against protest leaders and delayed local government elections that would have tested public sentiment after the uprising. The old order did not return unchanged, but it retained enough institutional position to manage the transition on its own terms.

The 2024 elections later changed the political balance. Dissanayake won the presidency in September 2024, and the National People’s Power won 159 of 225 parliamentary seats in November. The SLPP, which had held 145 seats after the 2020 election, collapsed to three. That result was a real electoral rupture. It also confirmed that the street had opened the political order in 2022, but had not completed the transition. The ballot later converted public anger into formal power.

Bangladesh’s path was an interim cushion. Hasina’s removal was followed by an unelected interim government under Yunus, which used the transition period to set up reform commissions and develop the July National Charter. The February 2026 election and referendum then returned the country to elected rule. Reports differ on the referendum margin, with some sources placing approval around 61 percent and others around 68 percent. The broader point is not the exact percentage. Voters endorsed a reform package, while the BNP won a decisive two-thirds parliamentary majority.

That creates an ironic tension. The interim process was designed to decentralize power and prevent another authoritarian concentration. The election then handed a concentrated legal majority to a legacy political party. The BNP now has enough power to implement the July Charter, dilute it, or selectively accept only those provisions that do not bind the new government. Bangladesh is therefore not simply a failed uprising or a completed transition. It is an active test of whether interim constitutional engineering can survive contact with elected political power.

Sri Lanka and Bangladesh therefore pose different versions of the same sequencing dilemma. Rapid elections provide democratic legitimacy, but they may return established parties before administrative capture has been addressed. Delayed elections create space for reform commissions, legal redesign, and accountability work, but they create legitimacy risks because unelected interim authorities cannot govern indefinitely. The question is not whether elections should be fast or delayed in every case. The question is which records, safeguards, and public duties must be locked in before the next governing coalition controls the machinery again.

Sri Lanka’s Captured Fiscal State

Sri Lanka’s crisis was not only a shortage of foreign exchange. It was the result of fiscal and institutional weaknesses that accumulated before the public experienced them as queues. The IMF’s 2022 Article IV consultation recorded that income tax and VAT cuts in late 2019 produced estimated revenue losses exceeding 2 percent of GDP. It also noted that the automatic fuel pricing mechanism was discontinued, increasing fiscal risk from state-owned enterprise losses, and that planned reforms on central bank independence and fiscal rules were suspended.

The World Bank later recorded the social cost. Real GDP contracted by 7.8 percent in 2022. National poverty was projected to double to 25 percent. Public and publicly guaranteed debt reached an estimated 122.5 percent of GDP. Shortages of fuel, cooking gas, food, medicines, and inputs turned fiscal failure into daily punishment. Citizens did not experience public debt as a spreadsheet. They experienced it as waiting, scarcity, inflation, lost income, and humiliation.

The fertilizer ban exposed another form of institutional weakness. The 2021 decision to ban chemical fertilizer imports was presented as a shift toward organic agriculture, but it was implemented abruptly and without the transition capacity required for such a change. Subsequent analysis and reporting showed serious harm to agricultural production, including tea and rice. The policy failure was not merely agronomic. It showed what can happen when executive power outruns institutional correction.

Sri Lanka’s constitutional structure contributed to this risk. The Nineteenth Amendment had sought to reduce presidential discretion and strengthen Parliament, the cabinet, and independent institutions. The Twentieth Amendment reversed key parts of that arrangement, restoring a more centralized presidency and weakening oversight around appointments and commissions. This did not cause every later policy failure. It did help create an environment in which concentrated executive authority could marginalize institutional caution.

The Aragalaya answered the collapse of public life with public refusal. It forced out Mahinda Rajapaksa as prime minister and Gotabaya Rajapaksa as president. Yet the institutional structure was able to route succession through Parliament and preserve elite continuity for another two years. This was not a failure of protest. It was the limit of protest. The street can remove the visible author of failure. It cannot automatically replace the fiscal, constitutional, and administrative systems that made failure possible.

Bangladesh’s Captured Hiring State

Bangladesh’s uprising began with a public employment dispute because public employment was not a narrow administrative issue. After the High Court reinstated a quota arrangement that reserved a large share of government jobs for descendants of 1971 freedom fighters and other categories, students mobilized against what they saw as patronage inside the state’s hiring system. The Supreme Court later reduced the freedom fighter descendants’ quota to 5 percent and directed that 93 percent of jobs be allocated on merit, with 2 percent reserved for minority and disadvantaged categories.

That legal adjustment did not end the crisis because the quota issue had already become a symbol of unequal citizenship. In much of South Asia, a government job is not only a salary. It is security, pension rights, status, family mobility, and proximity to the state. A captured hiring register therefore becomes a governing fact. It tells citizens whether the state recognizes merit, lineage, loyalty, identity, or access.

The Hasina government had already built a record of coercive and legal control. Rights organizations documented repression of opposition activity, pressure on journalists, enforced disappearances, and legal action against speech and dissent. The Digital Security Act was replaced by the Cyber Security Act in 2023, but Amnesty International concluded that the new law retained much of the earlier law’s repressive architecture. In its analysis, 58 of the 62 provisions of the Digital Security Act were retained in some form.

During the 2024 protests, state violence moved the crisis beyond administrative grievance. The UN human rights office later estimated that as many as 1,400 people may have been killed between July 1 and August 15, 2024. It found reasonable grounds to believe that serious violations included extrajudicial killings, arbitrary arrests, detention, torture, and efforts to conceal evidence. The interim government’s cooperation with the UN inquiry was important. Accountability, however, depends on whether the records of command, deployment, detention, weapons use, hospital obstruction, and evidence preservation survive political transition.

The July Charter process was Bangladesh’s attempt to use the morning after differently. The interim government created reform commissions on the constitution, judiciary, electoral system, police, public administration, and anti-corruption. The Charter gathered more than 80 reform proposals, including constitutional and administrative changes. Its referendum gave the reform agenda public force, but not automatic clarity. ConstitutionNet warned that the bundled referendum question made it unclear which proposals voters had approved and what binding force each carried.

The risk is now specific. If the BNP treats the Charter as a symbolic inheritance from the uprising while avoiding constraints on executive power, judicial independence, police accountability, election administration, and public service neutrality, the transition will become electoral restoration rather than institutional reform. If it gives those reforms legal force and accepts limits on its own majority, Bangladesh could become a serious case of post-uprising constitutional repair.

Recorded Continuity, Not Indiscriminate Purge

A post-uprising state cannot govern by pretending that every inherited official, contract, office, and process can be removed overnight. That path leads toward administrative collapse. The alternative to capture is not a De-Ba’athification-style purge that destroys state capacity in the name of cleansing the old order. Courts must sit, ports must function, salaries must be paid, hospitals must open, imports must arrive, and police must maintain basic order.

Continuity is therefore necessary. Unrecorded continuity is dangerous. Officials may remain, but their decisions should become reviewable. Contracts may continue, but beneficial ownership, variation orders, delivery status, and payments should be disclosed. Courts may proceed, but delay data should be published. Police may operate, but complaints, arrests, detention registers, and use-of-force incidents must be preserved. Interim authorities may propose reforms, but their powers should be time-limited, publicly debated, and transferred into a constitutional process with legal clarity.

Recorded continuity is the middle path between elite restoration and reckless institutional rupture. It recognizes that the state must keep operating, but it refuses to let inherited machinery operate in the dark. It gives honest officials protection, citizens evidence, legislatures oversight, courts a record, and future investigators a file that has not already disappeared.

The Seven Pillars of Public Accounting

1. Fiscal and Sovereign Debt Control

Sri Lanka’s crisis showed that public debt, tax concessions, foreign reserves, state-owned enterprise losses, and contingent liabilities are not technical matters reserved for ministries and lenders. They determine whether citizens later face shortages, austerity, inflation, and loss of income. A post-uprising government should publish a debt and fiscal risk register covering central government debt, guarantees, state-owned enterprise obligations, arrears, collateralized borrowing, creditor exposure, restructuring terms, and major fiscal risks. Citizens cannot share sacrifice rationally if the balance sheet remains incomplete.

2. Open Procurement Architecture

Public contracts should be traceable from planning to tender, award, implementation, variation, payment, completion, and audit objection. The Open Contracting Data Standard already provides a model for publishing data and documents across the contracting cycle. A serious post-uprising government should make contract files searchable by agency, contractor, beneficial owner, contract value, procurement method, variation, payment status, delivery status, and audit finding. Procurement transparency should not be treated as a slogan. It should be an operating system.

3. Depoliticized Public Employment

Bangladesh’s uprising began with public employment because hiring is one of the clearest ways citizens encounter the state. A public employment ledger should publish vacancies, eligibility rules, examination weightings, interview scores, quota use, waiver decisions, appointment lists, appeal outcomes, and disciplinary decisions where privacy rules allow. Representation policies may be legitimate when openly designed and publicly justified. Patronage begins when allocation rules become unclear, selectively applied, or insulated from review.

4. Judicial and Administrative Delay Control

Delay redistributes power toward those who can afford time, lawyers, possession, and political access. Courts and administrative agencies should publish case age, adjournment patterns, reserved judgment periods, enforcement delays, and agency compliance with court orders. Routine services should carry receipt numbers, official fees, timelines, written reasons for refusal, appeal routes, and remedies for unlawful delay. A right that cannot be used on time is not fully available to the citizen who needs it.

5. Coercive Command Accountability

Transitional justice cannot depend only on headline prosecutions. It requires deployment orders, command responsibility records, ammunition logs, arrest and detention registers, hospital access records, casualty lists, CCTV or body-camera preservation where available, and independent forensic procedures. Bangladesh’s UN report makes this especially urgent because the alleged violations involved patterns of force, detention, torture, and concealment. If the record is not preserved early, accountability becomes slower, weaker, and easier to politicize.

6. Decriminalization of Digital Dissent

Bangladesh’s experience under the Digital Security Act and Cyber Security Act shows how legal regulation of speech can become part of competitive authoritarian control. A post-uprising legal order should separate legitimate cyber law from anti-dissent law. Fraud, hacking, child exploitation, data protection, and digital security require legal tools. Peaceful criticism, journalism, satire, protest speech, and political dissent should not be treated as security offenses through vague provisions, warrantless search powers, or criminal penalties that chill public debate.

7. Transparent Political Finance

Uprisings often weaken incumbents but leave party finance opaque. If parties are funded through invisible business interests, public contracts, state advertising, patronage appointments, or illicit rents, elections can rotate office without changing the source of influence. Disclosure of party income, campaign spending, donor categories, public advertising allocation, conflict-of-interest rules, and beneficial ownership links should be treated as part of post-uprising reconstruction, not as a later democratic luxury.

International Leverage and Domestic Ownership

External actors should not fund stabilization while treating institutional accountability as secondary. The IMF, World Bank, Asian Development Bank, bilateral donors, and UN agencies cannot rebuild domestic legitimacy for Bangladesh or Sri Lanka. That work belongs to citizens and their institutions. They can, however, avoid reinforcing the same opaque systems that produced crisis.

Sri Lanka’s IMF governance diagnostic was useful because it treated corruption vulnerabilities and governance weaknesses as macro-critical. That principle should travel. Financing should be linked to debt transparency, public financial management, procurement disclosure, state-owned enterprise reporting, anti-corruption capacity, and social protection that can be audited. Human rights support should be linked to evidence preservation, witness protection, independent investigation capacity, and repeal or revision of laws that criminalize dissent.

Conditionality alone will not produce reform. It can create incentives, benchmarks, and documentation requirements. The decisive test remains domestic. A government that wants reform will use external benchmarks to discipline its own bureaucracy and resist captured interests. A government that wants only financing will satisfy the form and evade the substance.

The Operating Tests

The test of post-uprising reform is practical. Can a citizen obtain a routine state service without a political intermediary? Can a contractor see who won and why? Can voters see who funds parties? Can Parliament see the full debt position? Can courts measure delay honestly? Can a student verify hiring criteria? Can a journalist request public information without being criminalized? Can a family of a protest victim trace command responsibility? Can an honest officer refuse unlawful pressure and keep his career?

These are not abstract democratic ideals. They are operating tests of whether the state has changed. A captured state survives through discretion without reasons, delay without consequence, contracts without traceability, debt without full disclosure, appointments without merit, force without command accountability, and speech laws that turn criticism into risk. A recovered republic must make those systems answerable.

Sri Lanka shows that electoral correction may take years after protest opens the political order. Bangladesh shows that interim constitutional reform can create a window before elections, but that the window can close quickly once an established party controls Parliament. Neither path is automatically superior. Both depend on whether reform reaches the records, incentives, and remedies through which public power is used.

Leadership change begins the transition. Public accounting determines whether the transition reaches the state.

Notes

1. Joel S. Hellman, Geraint Jones, and Daniel Kaufmann, “Seize the State, Seize the Day: State Capture, Corruption, and Influence in Transition,” World Bank Policy Research Working Paper No. 2444, 2000.

2. Steven Levitsky and Lucan A. Way, Competitive Authoritarianism: Hybrid Regimes After the Cold War (Cambridge: Cambridge University Press, 2010); Steven Levitsky and Lucan A. Way, “The New Competitive Authoritarianism,” Journal of Democracy 31, no. 1, 2020.

3. International Monetary Fund, “Sri Lanka: 2021 Article IV Consultation,” IMF Country Report No. 22/91, March 2022.

4. World Bank, Sri Lanka Development Update: Time to Reset, April 2023.

5. ConstitutionNet, “The Return of Sri Lanka’s Imperial Presidency: The Twentieth Amendment to the Constitution Bill,” September 2020.

6. Carnegie Endowment for International Peace, “The Aragalaya Protest Movement and the Struggle for Political Change in Sri Lanka,” August 27, 2025.

7. Reuters, “Sri Lanka Postpones Local Elections, Citing Lack of Funds,” February 24, 2023.

8. Inter-Parliamentary Union, “Sri Lanka Parliament: November 2024 Election Results.”

9. Amnesty International, “What Happened at the Quota-Reform Protests in Bangladesh?” July 29, 2024.

10. Al Jazeera, “Bangladesh Top Court Scraps Most Job Quotas That Caused Deadly Unrest,” July 21, 2024.

11. Amnesty International, “Bangladesh: Interim Government Must Restore Freedom of Expression in Bangladesh and Repeal Cyber Security Act,” August 8, 2024.

12. United Nations Office of the High Commissioner for Human Rights, “Bangladesh: UN Report Finds Brutal, Systematic Repression of Protests,” February 12, 2025.

13. International IDEA, “Explainer: A Poll and Referendum to Define Bangladesh’s Next Chapter,” February 10, 2026.

14. ConstitutionNet, “Bangladesh’s Referendum and Reforms: The Need to Return to a Constitutional Process,” March 4, 2026.

15. Congressional Research Service, “Recent Political Developments in Bangladesh,” June 5, 2026.

16. Carnegie Endowment for International Peace, “Bangladesh’s Unfinished Revolution,” May 28, 2026.

17. Open Contracting Partnership, “Open Contracting Data Standard.”

18. International Monetary Fund, “Sri Lanka: Technical Assistance Report, Governance Diagnostic Assessment,” September 2023.

19. Liaquat Ali, The Captured Republic, manuscript chapters on the invisible government of privilege, patronage, and public accounting.