12 Governance and institutions

Chapter 12 of 60 in the Bangladesh 2036 research base. Contents of the series.

The 2024 transition corrected the accountability deficit at a stroke and left the capacity deficit intact, and the decade is decided by which deficit the state works on first

For fifteen years Bangladesh's governance record moved in one direction on every index this research base can verify, and in July and August 2024 it broke. The thesis of this chapter is that the uprising corrected the half of the governance problem that indices measure as voice and accountability, and did almost nothing by itself to the other half, the state's capacity to collect revenue, execute a budget, deliver a verdict or publish a statistic, and that the decade to FY36, the horizon the chapter 15 scenarios assume, is decided by the sequence. On the accountability side, Polity5 fell from plus 5 in 2010 to minus 6 in 2018 and ends there [Polity5 2018]; the V-Dem electoral democracy index fell from 0.420 in 2011 to 0.250 in 2019 [VDem 2022]; Freedom House fell from 56 of 100 in 2013 to 40 in its 2024 edition [Freedom House 2024]. On the capacity side, the World Bank's government effectiveness estimate has been stuck between the 35th and 41st percentile for fifteen years [WB WGI 2024], the CPIA efficiency of revenue mobilization rating fell from 3.0 in 2010 to 2.0 in 2024 [WB CPIA 2024], and the budget transparency score fell from 48 in 2010 to 31 in 2023 [IBP Open Budget 2023]. The transition then delivered, in one year, a new Election Commission output (a parliament), a restructured revenue administration, a reset statistics office, a rebuilt procurement rulebook and an enforced bank classification regime, the last three documented in chapters 04 to 06. Whether the elected state that exists from 2026 institutionalises the accountability correction, or governs through the same machine the correction exposed, is the governance question of the decade.

Every index that measures checks on power fell for a decade while every index that measures state capability stood still

The Worldwide Governance Indicators place Bangladesh, for 2024, at the 36.5th percentile on voice and accountability, the 40.9th on political stability, the 39.3rd on government effectiveness, the 39.9th on regulatory quality, the 40.4th on rule of law and the 25.5th on control of corruption [WB WGI 2024]. Voice and accountability sat at the 51.3rd percentile in 2010 and has fallen in almost every update since; political stability fell from the 48.2nd percentile for 2023 to the 40.9th for 2024, the transition registering in the index within one year; control of corruption fell from the 27.2nd percentile in 2010 to 25.5 in 2024, while government effectiveness moved from 35.6 to 39.3, under 4 points in fifteen years of 6 percent growth [WB WGI 2024]. The Global Peace Index, which samples violence rather than electoral order, moved the other way, from 2.128, rank 101 of 163, in 2019 to 1.954, rank 84, in 2024 [GPI 2024]. A state that grew at 6 percent while its effectiveness percentile stood still grew on factors its institutions did not supply.

The electoral record is where the decline concentrated. Polity5 scored the country plus 5 from 2010 to 2012, plus 4 in 2013, plus 1 from 2014 to 2017, and minus 6 in 2018, the year of an election the series scores as a breakdown, and the series ends there [Polity5 2018]. The V-Dem electoral democracy index tells the same story continuously, 0.420 in 2011, 0.293 in 2014, 0.250 in 2019, 0.274 in 2022, the last year held here [VDem 2022]. The two parliamentary elections of 2014 and 2018 are the inflection points, and the January 2024 election under the old rules is the event the uprising turned on. Freedom House records the same slide, a total score of 56 of 100 in 2013, 45 in 2018 and 40 in the 2024 edition, with political rights at 15 of 40 and civil liberties at 25 of 60 [Freedom House 2024]. The Bertelsmann Transformation Index places the 2024 status at 5.10 of 10 with a governance index of 4.55, ranking 94th of the countries assessed [BTI 2024].

Corruption perception moved within a narrow band while the enforcement machine stayed small. The Corruption Perceptions Index score fell from 26 of 100 in 2012 to 23 in 2024, its lowest reading in the series held here, after a best of 28 in 2017 [TI CPI 2024, owid corruption-perception-index]. The Commission's own reports show the machine behind that score: cases filed per year ran 363, 348, 347, 406 and 404 across 2019 to 2023 [ACC 2024, annual report 2023]. The 2024 report, covering the first transition year, shows the tempo holding: 1,894 complaints approved for enquiry produced 451 cases filed and 403 charge sheets, while the Special Judge's Courts disposed of 295 cases, of which 138 ended in conviction; on the 274 Commission cases disposed, 132 convictions give a 48.17 percent conviction rate, down from 67.13 percent in 2023, against a docket of 3,410 cases of which 2,988 were under trial [ACC 2025]. An institution that files about 400 cases a year through a boom, a pandemic and a change of government is calibrated to a steady state, not to the scale of the economy's corruption; the lower conviction rate is what absorbing cases it did not choose looks like.

The accountability institutions around the budget show the same calibration. The Open Budget Survey transparency score for Bangladesh rose from 48 in 2010 to 58 in 2012, then fell through 56 in 2015, 41 in 2017 and 36 in 2019 to 30 in 2021 and 31 in 2023, with public participation scored 9 of 100 and legislative oversight 30 of 100, while the supreme audit institution scored 67 [IBP Open Budget 2023]. The audit numbers show why the score matters: in FY22 the Office of the Comptroller and Auditor General reports 10,537.28 crore BDT recovered or adjusted through audit activity against its own revised budget of 274.86 crore BDT, a ratio it states as 1:38; the same 10,537.28 crore BDT is the amount involved in the 357 observations the Public Accounts Committee examined, of which 642.36 crore BDT is listed as settled through committee action [CAG 2022]. The oversight chain works to the point of detection and stalls at the point of sanction.

The judiciary and the media carry separate indices with one message. The World Justice Project Rule of Law Index scored the country 0.42 in 2015, rank 92 of 102, and 0.39 in 2024, rank 127 of 142 [WJP 2024]. The V-Dem judicial constraints on the executive index, the one judicial independence series in the data lake, fell from 0.211 in 2010 to 0.176 in 2014 and 0.125 in 2022 [VDem 2022]. Reporters Without Borders ranked the country 144 of 179 in 2013 and 165 of 180 in 2024; the score series changes methodology in 2022, so this research base reads the ranks [RSF 2024]. The case backlog the judicial record is known for has no series in the data lake, and the Law and Justice Division annual reports held there stop at FY15, so any recent backlog figure could not be confirmed; the division's current reports would settle it.

The civil service appears in the record as a survey, not a payroll. A motivation survey of 1,081 Bangladesh Civil Service officers in foundation training at the Bangladesh Public Administration Training Centre, fielded from August 2023 to January 2025 across 24 named cadres and held in the data lake as a Figshare research deposit, records a female share of 26.55 percent, a mean age of 33.56 years, and an administration cadre subsample of 450 officers, 27.56 percent female, mean age 36.65 years [BPATC 2024]. Asked what drew them to the service: on a 10 point scale, an entry process free of corruption scored 9.38 and job security 9.08, while public interest commitment on a 7 point public service motivation scale scored 6.56 [BPATC 2024].

The electoral state at drafting time is the newest fact here. By 8 April 2026 the Election Commission was administering the business of a newly elected thirteenth Jatiya Sangsad: its notification of that date announces the schedule for the parliament's 50 reserved women's seats, nominations closing 21 April 2026 and polling on 12 May 2026 [EC 2026]. The general election that produced the thirteenth parliament therefore took place before April 2026, but its polling date, turnout and result are not available in any source this chapter draws on and could not be confirmed here; the Election Commission would settle it. The institutional inheritance, not the composition, is this chapter's subject.

Four mechanisms produced the record: an electoral dominance machine, oversight calibrated to survive, a capacity ceiling and the 2024 correction

The first mechanism is the electoral dominance machine. After 2014 the cost of losing elections fell out of the system: a parliament elected without opposition could appoint the Election Commission, amend the caretaker arrangement, and legislate the digital security framework chapter 13 references from the digital side. V-Dem's free and fair elections index measures the consequence directly, 0.444 in 2010, 0.161 in 2014, 0.137 in 2018 and 0.160 in 2022 [VDem 2022], almost two thirds of that score gone before the uprising. Dominance then funded itself: control of the enforcement apparatus made business finance flow to the incumbent coalition, which is the related party lending machine chapter 06 documents, and control of permits, land and procurement made the coalitions that fund politics dependent on continued incumbency.

The second mechanism is oversight calibrated to survive rather than to bite. The ACC filing record above is one half; the audit chain is the other: 10,537.28 crore BDT involved in recoverable findings against 642.36 crore BDT settled through the parliamentary committee in FY22 [CAG 2022], a sanction rate at the committee level of about 6 percent of the money involved, in a parliament the index record above scores as non competitive. Budget transparency fell to 31 of 100 [IBP Open Budget 2023] because a dominant executive does not need to publish; participation scored 9 of 100 [IBP Open Budget 2023] because consultation without competition is decoration. The statistics office sits in the same column: the CPIA transparency rating has been 2.5 of 6 in every year since 2011 [WB CPIA 2024], and the chapters that use BBS macro data, chapter 01 on growth and chapter 05 on prices, both carry the rebasing caveat that follows from it.

The third mechanism is the capacity ceiling, separate from the accountability collapse. Government effectiveness moved from the 35.6th to the 39.3rd percentile in fifteen years and regulatory quality from the 41.7th to the 39.9th [WB WGI 2024]; the Heritage economic freedom score, a second regulatory composite, fell from 56.5 in 2021 to 51.9 in 2024 [Heritage IEF 2024]. The CPIA ratings that measure the machinery tell the harder story: quality of public administration fell from 3.0 of 6 in 2010 to 2.5 in 2024 after a low of 2.0 in 2020 to 2022, the public sector management and institutions cluster average from 3.0 to 2.6, and efficiency of revenue mobilization from 3.0 to 2.0 [WB CPIA 2024]. The state's capacity declined in absolute rating terms while the economy doubled. The revenue mechanism is chapter 04's: a tax ratio stuck near 7 to 8 percent of GDP [WB WDI 2026] is not a legislative failure, it is an administrative one, and the CPIA revenue rating is its institutional price. The BPATC survey closes the loop on personnel: a service whose officers rank job security at 9.08 of 10 as their entry priority [BPATC 2024] has been managed through transfer, posting and promotion discretion, the everyday currency of the dominance machine, and the discretion that made the bureaucracy pliable also made it slow.

The fourth mechanism is the 2024 correction itself. In July and August 2024 a student led movement against the quota system for government jobs became a mass uprising, the security response killed hundreds, a figure the UN human rights office later put at as many as 1,400, which could not be confirmed, the report not available in any source this chapter draws on and OHCHR not in the registry, the government fell, and an interim administration under Muhammad Yunus took office. Every event claim in this paragraph is a qualitative reading of the 2024 public record and could not be confirmed against a primary file in the data lake; what this research base can measure is what the transition changed in the data, and it changed fast. The bank classification reset produced the 30.6 percent measured classified loan ratio in chapter 06 [BB NPL 2026]. The exchange rate was floated and the lending cap abolished in chapter 05's regime break [BIS 2026]. The revenue board was split into policy and management divisions, chapter 04's administration mechanism. The national accounts were rebased and the CPI basket updated, chapters 01 and 05's measurement caveat. And the procurement rulebook was rebuilt: the Public Procurement (Amendment) Ordinance 2025 and the new Public Procurement Rules 2025 took effect together on 28 September 2025, carrying 154 rules and 21 schedules, removing the plus or minus 10 percent price cap in national works procurement, making the e-procurement platform mandatory for all public procurement, requiring disclosure of the actual beneficial owner of contract awards, and creating a debarment board [BPPA 2025]. Of everything the transition touched, the procurement rewrite is the best documented in the data lake and the reform whose compliance the European Union will monitor directly.

The decade ahead runs on three clocks: the elected settlement, the external compliance calendar and the statistics rebuild

The constitutional and electoral settlement is the first decision point, and its author is the thirteenth Jatiya Sangsad with the Election Commission as counterpart. The interim period produced commission reports on the constitution, the electoral system, the police, public administration, the judiciary and the anti-corruption apparatus, submitted through 2025; the sequence and status of each report could not be confirmed from the sources this chapter draws on; the Cabinet Division would settle it. What converts the reports into institutions is the legislative machinery of the new parliament: ordinances and amendments tabled, and the share that pass within a year. The EC notification of April 2026 [EC 2026] shows the commission functioning on schedule; whether the next two election cycles are competitive is the binary the decade turns on.

The external compliance calendar is the second decision point, and its authors are the Ministry of Commerce and the European Commission. Chapter 02 documents the GSP+ requirement to ratify and effectively implement 27 international conventions on human rights, labour rights, environment and governance before the Everything But Arms window the transition timetable targets closes at the end of 2029 [EU GSP 2023]. For this chapter the point is institutional: GSP+ is a governance programme with a trade payoff and a two year monitoring cycle, the only external mechanism on the table that audits the judiciary, the police and the civil liberties record on a fixed schedule with a price attached. A state that implements the 27 conventions for the tariff obtains the enforcement infrastructure its own commissions have proposed.

The credibility rebuild of the measuring institutions is the third decision point, and its authors are the Statistics and Informatics Division for BBS, the Finance Division for the budget documents, and Bangladesh Bank for the monetary and financial record. The rebasing of the national accounts and the CPI basket during the transition, recorded as caveats in chapters 01 and 05, is credible only if the revisions are published with methodology and back data, because the alternative, a statistics office perceived to move numbers with governments, is the credibility discount chapter 01 names as a macro risk. The budget benchmark is measured: the Open Budget transparency score at 31 [IBP Open Budget 2023] and the CPIA transparency rating at 2.5 since 2011 [WB CPIA 2024] are the numbers a rebuild has to move, and chapter 04's threshold of 50 on the next survey wave is the regime change mark. The IMF programme's structural anchors, revenue mobilisation and ADP execution quality in chapter 04's account and the resolution framework in chapter 06's, are the external audit of the capacity side; the reviews are the resolving source.

The civil service decision runs underneath all three. The public administration reform commission's proposals on recruitment, promotion and accountability, status not established as above, address the machine described in the mechanism section: a service with 9.08 of 10 job security as its entry priority [BPATC 2024] and a revenue administration rated 2.0 of 6 on collection efficiency [WB CPIA 2024]. The decision that matters is whether posting and promotion discretion becomes rule bound: every other governance reform is executed by the officers the discretion manages. The decade's path, stated so chapter 15 can use it: the reform scenario assumes the constitutional settlement holds through two competitive elections, the GSP+ monitoring cycle is entered, and the CPIA public administration and revenue ratings recover to 3.0 or better by around FY31; the baseline assumes partial execution, competitive elections but shelved administrative reform; the stall scenario assumes the dominance machine rebuilds inside the new parliament, and the index record of 2010 to 2024 resumes its slope.

Reversion, reform fatigue and capacity slide are the risks; the measurement, compliance and personnel dividends are the upside

Risks. First, reversion: the dominance machine rebuilt inside the elected state, with the Election Commission and the enforcement apparatus again calibrated to incumbency, and the revealing indicator is the WGI voice and accountability percentile, at 36.5 for 2024 [WB WGI 2024], falling back toward 30 while the Freedom House score, at 40 [Freedom House 2024], resumes its decline. Second, reform fatigue: the commission reports accumulate, the parliament legislates selectively, and the procurement and revenue reforms survive on paper while the discretionary machinery reasserts itself; the revealing indicator is the share of the 27 GSP+ conventions effectively implemented, which the EU monitoring cycle audits against the 2029 window the chapter 02 transition timetable targets [EU GSP 2023], and the ACC filing tempo, which at about 400 cases a year [ACC 2024] is the signature of the steady state returning. Third, a capacity slide: the administrative and revenue ratings keep falling, the CPIA efficiency of revenue mobilization rating at 2.0 [WB CPIA 2024] dropping further while the tax ratio stays under 8 percent, chapter 04's stall marker, and the state cannot execute the development programme chapter 10 documents even where the money exists; the revealing indicator is the CPIA quality of public administration rating, 2.5 in 2024 [WB CPIA 2024], falling back to its 2.0 low.

Upside. First, the measurement dividend: the transition's honesty about the balance sheet, the 30.6 percent classified ratio in chapter 06 and the rebased accounts in chapters 01 and 05, is the precondition for the credibility premium that the sovereign spread series in chapter 07 prices, and a statistics office that publishes revisions with methodology converts a governance reform into a financing term. Second, the compliance dividend: the 27 conventions, the beneficial ownership disclosure and the debarment board in the procurement rules [BPPA 2025], and the GSP+ monitoring cycle [EU GSP 2023] together form an external audit infrastructure that does not depend on domestic political will, and the revealing indicator is the EU's first monitoring report after the application chapter 02 targets for 2028. Third, the personnel dividend: a rule bound posting and promotion regime aimed at the civil service that the BPATC survey describes would, within one posting cycle, move the officers who execute every other chapter's agenda, and the revealing indicator is the CPIA public administration ratings recovering to 3.0, the level the 2010 state actually had [WB CPIA 2024].

What to watch: five indicators whose thresholds mark the regime

1. WGI voice and accountability percentile. Current value 36.5 for 2024 [WB WGI 2024]. Threshold: a recovery above 45 within two updates confirms the accountability regime change this chapter's reform scenario assumes; a fall below 30 is the reversion mark.
2. Corruption Perceptions Index score. Current value 23 of 100 in the 2024 edition [TI CPI 2024]. Threshold: a score of 30 or higher by the edition the 2030 calendar targets is the corruption regime change; a reading below 20 is the lowest score in the series and the reversion confirmation.
3. Open Budget transparency score. Current value 31 of 100 in the 2023 survey [IBP Open Budget 2023]. Threshold: above 50 in the next wave, the same threshold chapter 04 sets, is the budget credibility regime change; a fourth consecutive wave at or below 31 is the publication ceiling confirmed.
4. Anti-Corruption Commission filing tempo. Current value 451 cases filed in 2024, against a 2019 to 2023 band of 347 to 406 [ACC 2025] [ACC 2024]. Threshold: sustained filings above 800 a year, double the steady state, with senior officials convicted in the Special Judge's Courts, is the active enforcement regime; a return to the 347 to 406 band after the transition prosecutions is the steady state reasserting itself.
5. CPIA ratings for public administration and revenue mobilization. Current values 2.5 and 2.0 of 6 in the 2024 assessment [WB CPIA 2024]. Threshold: both at 3.0 or better in the assessment the early 2030s window targets is the capacity regime change; a further fall is the slide marker chapter 15's stall scenario assumes.

Sources used

[Polity5 2018] Polity5 democracy score series for Bangladesh, ending 2018, via bdpolicy parquet governance/polity5_bd, series: polity2. [VDem 2022] V-Dem Institute electoral democracy index and component indices for Bangladesh through 2022, via bdpolicy parquet governance/vdem_bd, series: v2x_polyarchy, v2xel_frefair, v2x_jucon. [Freedom House 2024] Freedom in the World 2024 edition, Bangladesh scores, via bdpolicy parquet governance/freedomhouse_bd. [WB WGI 2024] World Bank Worldwide Governance Indicators, Bangladesh percentile ranks 1996 to 2024, via bdpolicy parquet governance/wgi_bd. [TI CPI 2024] Transparency International Corruption Perceptions Index 2024, Bangladesh scores 2012 to 2024, via the OWID corruption perception file in the data lake. [BTI 2024] Bertelsmann Transformation Index 2024, Bangladesh status, governance and rank, via bdpolicy parquet governance/bti_bd. [GPI 2024] Institute for Economics and Peace Global Peace Index, Bangladesh score and rank, via bdpolicy parquet governance/global_peace_index_bd. [ACC 2024] Anti-Corruption Commission Annual Report 2023, enquiry, case filing, trial and conviction statistics, OCR text in the data lake. [ACC 2025] Anti-Corruption Commission Annual Report 2024, complaints approved for enquiry, case filing, charge sheets, trial and conviction statistics, OCR text in the data lake. [CAG 2022] Office of the Comptroller and Auditor General, Annual Activity Report FY22, audit recovery, settlement and PAC statistics, OCR text in the data lake. [IBP Open Budget 2023] International Budget Partnership Open Budget Survey 2023, Bangladesh transparency, participation and oversight scores, via bdpolicy parquet governance/ibp_open_budget_bd. [WJP 2024] World Justice Project Rule of Law Index, Bangladesh overall score and rank 2015 to 2024, via bdpolicy parquet governance/wjp_rol_bd. [RSF 2024] Reporters Without Borders World Press Freedom Index, Bangladesh rank 2013 to 2024, via bdpolicy parquet governance/rsf_press_freedom_bd. [Heritage IEF 2024] Heritage Foundation Index of Economic Freedom, Bangladesh overall score, via bdpolicy parquet governance/heritage_ief_bd. [BPATC 2024] BCS officer motivation survey of 1,081 officers in foundation training at the Bangladesh Public Administration Training Centre, fielded August 2023 to January 2025, 24 named cadres, Figshare deposit 31023460, via bdpolicy.db series bcs_survey_* and lake/governance/bcs_cadre_motivation_benchmarks. [EC 2026] Election Commission Bangladesh notification of 8 April 2026, election schedule for the 50 reserved women's seats of the thirteenth Jatiya Sangsad, polling 12 May 2026, OCR text in the data lake. [BPPA 2025] Bangladesh Public Procurement Authority quarterly newsletter, October to December 2025, Public Procurement Rules 2025 effective 28 September 2025, OCR text in the data lake. [WB CPIA 2024] World Bank Country Policy and Institutional Assessment, Bangladesh ratings 2005 to 2024, series IQ.CPA.PADM.XQ, IQ.CPA.PUBS.XQ, IQ.CPA.REVN.XQ, IQ.CPA.TRAN.XQ, via bdpolicy parquet governance/wb_cpia_bd. [BB NPL 2026] Bangladesh Bank classified loan indicator snapshot, referenced for the transition era classification reset, detailed in chapter 06. [BIS 2026] Bank for International Settlements USD/BDT series, referenced for the 2024 float, detailed in chapter 05. [WB WDI 2026] World Bank World Development Indicators, tax revenue to GDP, referenced from chapter 04. [EU GSP 2023] European Commission GSP regulation and EBA transition timetable, referenced from chapter 02.

---

Verified line by line against primary sources: 96 claims checked, 2 corrected.

Previous: 11 Climate and environment

Next: 13 Digital economy and AI

All sixty chapters and the ten narrative chapters.