Total FIRs - governance burden
solid: monthly · dotted: 3-month rolling averageJanuary 2025: 14,572May 2026: 18,149
Constitution Watchdog · Public-interest accountability
Crime Data Accountability Index
60-Second Guide
- What this is
- A constitutional accountability reader that turns official FIR and recovery tables into rights-facing questions for the public.
- What this isn’t
- It is not a crime-prevalence estimate, conviction tracker, or automatic finding that government has failed.
- What changed this month
- Recorded FIRs increased to 15,389; that movement raises a rights question, while current PHQ monthly outcome disclosure remains 0/8 under a three-phase disclosure roadmap.
- What we still don’t know
- What happened after filing, sequenced by feasibility: custody/remand/arrest first, charge-sheet/disposal next, then court-linked bail, conviction/acquittal and tribunal outcomes.
- What you can do
- Go to the accountability questions, find the data trigger, and use the listed RTI or remedy path where disclosure is missing.
Jan 2025–May 2026 series
Trend discipline: rolling averages before conclusions
The Index privileges series movement and 3-month rolling averages over isolated monthly jumps. Unit-level variance may reflect population, reporting access, police capacity, registration practice, public trust in reporting, or actual incidence; the table therefore supports accountability questions, not final prevalence rankings.
Life/security FIRs - safe life pressure
solid: monthly · dotted: 3-month rolling averageJanuary 2025: 641May 2026: 609
Other Cases share - protection-of-law transparency
solid: monthly · dotted: 3-month rolling averageJanuary 2025: 45.3%May 2026: 38.3%
Narcotics FIRs - liberty-safeguard burden
solid: monthly · dotted: 3-month rolling averageJanuary 2025: 4,386May 2026: 7,092
Women & Child Repression FIRs - protection duty
solid: monthly · dotted: 3-month rolling averageJanuary 2025: 1,440May 2026: 1,952
National FIR rate - equal-protection comparison
solid: monthly · dotted: 3-month rolling averageJanuary 2025: 8.83 per 100kMay 2026: 10.99 per 100k
Empirical caveat: FIR data is registration data, not victimization prevalence. Bangladesh-specific reporting concerns include a police-commissioned survey, reported by The Daily Star, that 73% of 2018 crimes went unreported, and legal/human-rights sources noting risks of false or politically motivated case filing, including M.R. Hossain and Human Rights Watch. The Index therefore asks for charge-sheet, final-report, disposal and audit data before treating registration movement as a constitutional-performance conclusion.
| Month | Total FIRs | Rate /100k | Life/security | W&C | Narcotics | Other share | 3-month rolling avg total |
|---|---|---|---|---|---|---|---|
| May 2026 | 18,149 | 10.99 | 609 | 1,952 | 7,092 | 38.3% | 17,162 |
| April 2026 | 17,180 | 10.41 | 585 | 2,011 | 5,706 | 42.4% | 15,037 |
| March 2026 | 16,157 | 9.79 | 600 | 1,485 | 5,062 | 45.6% | 13,997 |
| February 2026 | 11,775 | 7.13 | 490 | 1,181 | 3,359 | 46.7% | 13,022 |
| January 2026 | 14,058 | 8.52 | 589 | 1,281 | 4,922 | 39.8% | 13,918 |
| December 2025 | 13,232 | 8.02 | 560 | 1,248 | 4,148 | 44.5% | 14,622 |
| November 2025 | 14,465 | 8.76 | 556 | 1,744 | 3,943 | 45.4% | 15,355 |
| October 2025 | 16,170 | 9.80 | 665 | 1,985 | 4,585 | 44.2% | 15,752 |
| September 2025 | 15,431 | 9.35 | 612 | 1,928 | 4,585 | 43.0% | 15,492 |
| August 2025 | 15,656 | 9.48 | 649 | 1,904 | 4,860 | 41.4% | 15,404 |
| July 2025 | 15,389 | 9.32 | 696 | 2,097 | 4,658 | 41.2% | 15,534 |
| June 2025 | 15,167 | 9.19 | 624 | 1,933 | 4,098 | 46.5% | 15,860 |
| May 2025 | 16,045 | 9.72 | 667 | 2,087 | 4,443 | 45.4% | 16,218 |
| April 2025 | 16,368 | 9.92 | 619 | 2,089 | 4,392 | 48.0% | 15,203 |
| March 2025 | 16,240 | 9.84 | 630 | 2,054 | 4,075 | 48.3% | 14,605 |
| February 2025 | 13,002 | 7.88 | 605 | 1,430 | 3,632 | 46.6% | 13,787 |
| January 2025 | 14,572 | 8.83 | 641 | 1,440 | 4,386 | 45.3% | 14,572 |
Showing latest 5 of 17 months.
Primary sourceBangladesh Police Headquarters monthly crime statistics
Source archiveOfficial portal, bundled PDFs and monthly JSON preserved
Schema checkJan 2025–May 2026 columns verified consistent
Source caveatOfficial administrative data; not independently audited prevalence data
Headline Finding
Bangladesh Police Headquarters currently discloses 0/8 monthly outcome fields. The Index now treats the remaining disclosure demand as a realistic three-phase roadmap: custody-death/remand/arrest first, charge-sheet and disposal tracking next, and court-linked bail, conviction/acquittal and tribunal outcomes as the hardest multi-year integration phase.
Core theory
The data is not the constitutional conclusion. It is the trigger for accountable explanation.
Official FIRFirst Information Report: the police record that starts a criminal case. It is an allegation record, not proof of guilt. and recovery data are converted into public questions about constitutional governance. Constitution Watchdog uses Police Headquarters data as a monthly evidence signal and as a public record of the State’s own account of crime registration. The Index asks whether the State is taking effective, lawful and transparent measures to protect safe life, protection of law, lawful enforcement, women and children protection, equal protection across places, and visible justice follow-through after cases enter the system. It does not say that more FIRs automatically prove State failure, or that fewer FIRs automatically prove success. It asks a narrower and more disciplined question: when the State records serious crime, liberty-impacting enforcement, women and children cases, or unexplained residual categories, what public explanation and follow-through should a constitutional government provide? Article 31 is the primary protection-of-law anchor; Article 32 supports the life and personal-liberty frame where the facts justify it. The wider positive-duty reading for ordinary third-party crime is expressly marked as an extension supported by comparative authority, not as a settled Bangladeshi appellate rule. Article 21 is used only as interpretive support unless paired with an enforceable Part III right. The result is a burden-of-explanation model, not a verdict model.
Tier One
Direct disclosure and phased advocacy roadmap
The 0/8 publication gap remains visible, but the demand is no longer presented as an all-at-once institutional ask. The Index sequences the eight fields by constitutional urgency, likely internal availability and implementation difficulty.
Current publication status0/8PHQ monthly outcome fields currently published
Custody-death data, remand and arrest baseline
Highest-stakes and most feasible first disclosures: custody-death data, remand and arrest baseline.
Charge-sheet and disposal tracking
Investigation follow-through disclosures: charge-sheet and disposal tracking with common case-status definitions.
Bail, conviction/acquittal and tribunal outcomes
Hardest court-linked outcome disclosures: bail, conviction/acquittal and tribunal outcomes through multi-institution integration.
Show full roadmap
Custody-death data, remand and arrest baseline
Fields: Custody-death data, Remand, Arrest.
This phase starts with the highest constitutional stakes: death or injury in State custody, coercive remand, and the arrest denominator needed to interpret both. These records are also the most likely to exist somewhere inside police, magistracy, prison or internal incident-reporting systems, even if they are not yet published as a clean monthly table.
Capacity required: A PHQ monthly template, unit-wise reporting, incident reference numbers, arrest/remand totals and manual reconciliation with prison or magistracy records where needed.
Timeline: Achievable within 12 months of sustained RTI, parliamentary and civil-society pressure if PHQ standardises the reporting format.
Charge-sheet and disposal tracking
Fields: Charge-sheet status, Disposal time.
Once liberty and custody-risk fields are visible, the next question is whether registered FIRs move through investigation. Charge-sheet and disposal tracking is harder than Phase 1 because it requires case-status continuity across months, but it remains largely within the police/prosecution workflow before final adjudication.
Capacity required: Common case-status definitions, unique FIR or case identifiers, monthly pending/charge-sheet/final-report/disposal fields, correction logs and unit-level reconciliation.
Timeline: Realistically achievable within 18–24 months if PHQ standardises case-status reporting and pressure continues after Phase 1.
Bail, conviction/acquittal and tribunal outcomes
Fields: Bail, Conviction or acquittal, Tribunal outcomes.
This phase is the hardest because PHQ cannot publish reliable final outcome figures from police records alone. Bail status, conviction/acquittal and tribunal outcomes require case-linkage across police, prosecution, court and tribunal systems, plus clear counting rules by FIR, accused person, offence or judgment.
Capacity required: PHQ–Judiciary–prosecution data integration, shared identifiers, digital court-status linkage, tribunal registry coordination and published definitions for how outcomes are counted.
Timeline: Requires PHQ-Judiciary data integration and formal inter-institutional coordination; this is a multi-year reform rather than a PHQ-only publication fix.
| Outcome field | Roadmap phase | PHQ monthly disclosure status |
|---|---|---|
| Arrest | Phase 1 - Custody-death data, remand and arrest baseline | Not disclosed |
| Bail | Phase 3 - Bail, conviction/acquittal and tribunal outcomes | Not disclosed |
| Remand | Phase 1 - Custody-death data, remand and arrest baseline | Not disclosed |
| Charge-sheet status | Phase 2 - Charge-sheet and disposal tracking | Not disclosed |
| Disposal time | Phase 2 - Charge-sheet and disposal tracking | Not disclosed |
| Conviction or acquittal | Phase 3 - Bail, conviction/acquittal and tribunal outcomes | Not disclosed |
| Tribunal outcomes | Phase 3 - Bail, conviction/acquittal and tribunal outcomes | Not disclosed |
| Custody-death data | Phase 1 - Custody-death data, remand and arrest baseline | Not disclosed |
Tier Two
Movement signal
Movement signals, thresholds, rolling averages and doctrinal classification are presented as interpretive readings, separate from the direct-disclosure count above.
Current movement readingBurden of explanation
July 2025 movement signal
This tier uses adjacent 3-month rolling-window movement and per-series volatility thresholds to identify questions that need follow-through evidence.
Movement signalWatchTrigger: Life/Security +4.0% (Watch threshold 2.8%; n=696; High confidence). Calibration: preliminary calibration, based on 16 observed monthly changes.
Governance burden: current 3-month rolling average15,534current month: 15,389
High confidence
Underlying n = 15,389; confidence tier: High confidence.
Equal protection: current 3-month rolling average rate9.41per 100,000 · current: 9.32
High confidence
Underlying n = 15,389; confidence tier: High confidence.
Life/security right: current 3-month rolling average662current month: 696
High confidence
Underlying n = 696; confidence tier: High confidence.
Protection of law: current 3-month rolling average Other share44.4%current month: 41.2%
High confidence
Underlying n = 6,333; confidence tier: High confidence.
Liberty safeguards: current 3-month rolling average Narcotics FIRs4,400current month: 4,658
High confidence
Underlying n = 4,658; confidence tier: High confidence.
Women/children protection: current 3-month rolling average2,039current month: 2,097
High confidence
Underlying n = 2,097; confidence tier: High confidence.
Accountability movement board
Movement is a rights question, not a verdict
Every movement card begins with the right being tested and then ties the accountability question to the month’s official FIR signal.
Life/security pressure
696+72 / +11.5%High confidence
Underlying n = 696; confidence tier: High confidence.
Data link: murder, kidnapping, robbery and dacoity total 696 in July 2025; change vs previous month: +72 / +11.5%.
Movement in serious-offence registrations asks for prevention, charge-sheet conversion, victim-support and disposal data before any breach conclusion.
Accountability question: are serious recorded harms moving in a way that demands a prevention and follow-up answer from the State?
Coercive enforcement pressure
5,079+538 / +11.8%High confidence
Underlying n = 5,079; confidence tier: High confidence.
Data link: coercive/recovery categories total 5,079; narcotics is 93.4% of recovery cases.
Special/recovery categories require arrest, remandCourt-authorized police custody or continued detention during investigation before trial., bail, detention and disposal data before liberty impact can be assessed.
Accountability question: are enforcement cases being handled with visible safeguards after people enter the criminal-process system?
Legal transparency gap
41.2%-5.3 percentage pointsHigh confidence
Underlying n = 6,333; confidence tier: High confidence.
Data link: “Other Cases” contains 6,333 FIRs, or 41.2% of all registered FIRs.
The remaining unclassified share requires statutory category disclosure and public-information follow-through.
Accountability question: can people understand what kinds of cases sit inside the official “Other Cases” bucket?
Women/children protection pressure
2,097+164 / +8.5%High confidence
Underlying n = 2,097; confidence tier: High confidence.
Data link: Women & Child Repression FIRs total 2,097; change vs previous month: +164 / +8.5%.
A decline cannot be called progress without reporting-access, tribunal disposal, compensation, shelter/referral and case-outcome data.
Accountability question: are women and children cases followed by protection, tribunal progress, compensation or referral support?
Doctrinal basis
Not every constitutional link carries the same confidence level
The Index separates textual/legal propositions from extended constitutional arguments. This preserves the burden-of-explanation frame without overstating doctrine.
Article 31 (protection of law), Article 33 (arrest and detention safeguards) and Article 35 (trial and punishment safeguards)
Article 31 (protection of law) is used for legality and protection-of-law questions; Article 33 (arrest and detention safeguards) for arrest/detention safeguards; Article 35 (trial and punishment safeguards) for fair-trial and punishment-process safeguards.
Women & Child Repression
The Prevention of Women and Children Repression Act, 2000 supplies the offence framework; Articles 31–32 (protection of law; life and personal liberty) supply the protection-of-law and life/security frame. Extended reading: comparative support only; no settled Bangladeshi appellate holding creates a general third-party-crime protective duty. No affirmative-action exception clause is used as the basis for this protection claim.
Articles 31–32 (protection of law; life and personal liberty) positive protective function
The Index treats serious life/security FIRs as creating a public-law burden of explanation primarily from Article 31’s protection-of-law guarantee, with Article 32’s due-process safeguard read alongside it by analogy. Extended reading: comparative support only; no settled Bangladeshi appellate holding creates a general third-party-crime protective duty.
Article 102 (constitutional writA court order requiring a public authority to do, stop doing, or justify something. jurisdiction)
Article 102 (constitutional writ jurisdiction) is used with attention to Bangladesh’s liberalized public-interest standing, especially where a public wrong or systemic rights question affects an indeterminate class of people.
Part III rights vs Article 21 (duties of public servants) - persuasive/interpretive only; not independently justiciable per Article 8(2)
Part III rights such as Articles 27, 31, 32 and 35 are directly justiciable through Article 102 where the legal requirements are met. Article 21 is a Part II principle: persuasive/interpretive only - not independently justiciable per Article 8(2).
Case-law anchors used for extended readings
- BLAST and others v Bangladesh, 55 DLRDhaka Law Reports: a widely used Bangladesh law-report citation series. (HCD) 363Arrest/remand, custodial death, life/liberty, torture and breach of public duty are treated as judicial anchors for custody and official-action questions; third-party-crime extension remains analogical.
- Md. Saifuzzaman v State, 56 DLR (HCD) 324Supports the arrest, detention, investigation and treatment-of-suspects safeguard line used for Article 33 (arrest and detention safeguards)/remand questions.
- Dr. Mohiuddin Farooque v Bangladesh, 49 DLR (AD) 1This case is used as analogical support for a broad life-and-security frame, not as a settled Bangladeshi holding that ordinary third-party crime automatically creates a constitutional breach. The Index treats FIR movement as a trigger for public-law explanation. No Bangladeshi appellate court is presented here as having recognized a general positive protective duty for third-party crime.
- Comparative (persuasive, not domestic)Osman v United Kingdom (ECtHR) and Velásquez-Rodríguez v Honduras (IACtHR) are cited as comparative support for reading state positive obligations to extend to foreseeable third-party harm. These are not Bangladeshi authority and are not treated as binding.
Questions for public accountability
Data-backed questions the State should answer
Each card starts from the right being tested, then shows the exact monthly data trigger, then asks the formal accountability question. This keeps the constitutional purpose attached to the evidence instead of floating as a generic slogan.
Life and security
Data trigger: July 2025: 696 life/security FIRs were registered across murder, kidnapping, robbery and dacoity; 3-month average: 662; change vs previous month: +72 / +11.5%.
Which unit-wise charge-sheet conversion rate, final-report rate and average disposal time are available for life and security FIRs in this reporting period?
In short: were these cases ever followed up, or just filed?
Enforcement pathRTIRight to Information: a legal route to request information from public authorities under the Right to Information Act, 2009. request for prevention/outcome data; Article 102 (constitutional writ jurisdiction) writA court order requiring a public authority to do, stop doing, or justify something. if legal duties or disclosure failures become justiciable.Liberty under coercive enforcement
Data trigger: July 2025: 5,079 coercive/recovery-category FIRs are shown, with narcotics at 93.4% of recovery cases; the monthly table still gives no arrest, bail, remand or detention follow-through.
Which arrest, remandCourt-authorized police custody or continued detention during investigation before trial., bail, detention, charge-sheet and disposal data are published for narcotics, arms, explosives, smuggling and speedy-trial categories?
In short: what happened to people after enforcement cases were opened?
Enforcement pathRTI request to Home/Police/Courts; Article 102 (constitutional writ jurisdiction) habeas corpusA court remedy requiring authorities to produce a detained person and justify the detention./writ pathway where detention legality or systemic non-disclosure is challenged.Legal transparency and public information
Data trigger: July 2025: “Other Cases” contains 6,333 FIRs, equal to 41.2% of all registered FIRs; change vs previous month: -5.3 percentage points.
Which constituent offences, statutes and case counts make up the full “Other Cases” category in this reporting period?
In short: can the public see what is hidden inside “Other Cases”?
Enforcement pathRTIRight to Information: a legal route to request information from public authorities under the Right to Information Act, 2009. request first; Information Commission complaint/appeal if refused; Article 102 (constitutional writ jurisdiction) if refusal or opacity raises public-law legality concerns.Women and children protection
Data trigger: July 2025: 2,097 Women & Child Repression FIRs are shown; 3-month average: 2,039; change vs previous month: +164 / +8.5%.
Which W&C Tribunal disposal rate, average disposal time, compensation disbursed and shelter or referral counts are published for this period?
In short: did these women and children cases lead to protection, support or tribunal action?
Enforcement pathRTI request for tribunal/outcome data; Parliamentary question for ministry-level accountability; NHRCNational Human Rights Commission: Bangladesh’s statutory human-rights body that can receive and examine rights-related complaints. complaint where protection failure is alleged; Article 102 (constitutional writ jurisdiction) writ where systemic protection failures are evidenced.Public-servant duty and police-conduct accountability
Data trigger: July 2025: police-assault/public-order related FIR count is 39, but the PHQ monthly table does not publish custody deaths, custodial injury, complaints against police or disciplinary outcomes beside it.
Which unit-wise data are published on police-as-victim/public-order cases, custody deaths, custodial injury, complaints against police and disciplinary outcomes?
In short: can the public see whether police-related cases also track custody, complaints and discipline?
Enforcement pathRTI request to police/prison authorities; NHRC complaint where rights abuse is alleged; Article 102 (constitutional writ jurisdiction) only where an enforceable Part III right, unlawful custody or official inaction is independently in issue.Fair trial and remedy follow-through
Data trigger: July 2025: 15,389 total FIRs are published, but current PHQ monthly outcome disclosure is only 0/8; the missing fields are now sequenced through a three-phase roadmap rather than demanded as one undifferentiated block.
Which Phase 2 charge-sheet, final-report and disposal-time fields are published by FIR category, and what Phase 1 custody/remand/arrest fields are already available internally?
In short: did the cases move toward a real legal outcome, or stop at filing?
Enforcement pathRTI request and Parliamentary question first; Article 102 (constitutional writ jurisdiction) writ where an enforceable rights/public-law failure is evidenced; NHRC complaint for rights-linked enforcement failures.Territorial equality
Data trigger: July 2025: national FIR rate is 9.32 per 100,000, while the unit table shows raw police-unit totals beside proxy rates; fair comparison requires denominator, staffing and registration-audit data.
Which population-normalised FIR rates by police unit are published, and what denominator, staffing and registration-audit data support those rates?
In short: are different areas being compared fairly, or just by raw case counts?
Enforcement pathRTI request for denominator, staffing, registration and audit data; Parliamentary question for unit-level variance; Article 102 (constitutional writ jurisdiction) only if unequal protection or systemic official failure becomes evidenced.Codebook
Why each Police Headquarters category maps to a constitutional question
This section shows which rights can go to court directly, and which ones only help explain the issue.
| Source category | Constitutional provision | Doctrinal basis | Enforceability tier | Signal label | Accountability question |
|---|---|---|---|---|---|
| Dacoity, Robbery, Murder, Kidnapping | Articles 31–32 (protection of law; life and personal liberty) | Extended Articles 31–32 reading - judicially extended / analogical. Extended reading: comparative support only; no settled Bangladeshi appellate holding creates a general third-party-crime protective duty. | Part III enforceable route: Article 102 may apply where a concrete Articles 31–32 rights issue is pleaded and evidenced. | Life/security pressure | Do the official figures correspond to charge-sheet conversion, disposal-time, victim-protection and prevention data? |
| Woman & Child Repression | Articles 31–32 (protection of law; life and personal liberty) + Prevention of Women and Children Repression Act, 2000 | Statutory + constitutional frame. Extended reading: comparative support only; no settled Bangladeshi appellate holding creates a general third-party-crime protective duty. | Part III + statutory route: Article 102 may apply where statutory failure is tied to Articles 31–32 rights. | Women and children protection pressure | Can the State show tribunal disposal, compensation, shelter/referral and enforceable outcomes behind the registered W&C FIRs? |
| Narcotics, Arms Act, Explosive Act, Smuggling, Speedy Trial | Articles 32–35 (life/liberty and criminal-process safeguards) | Textual: for liberty safeguards; outcome impact requires data | Part III enforceable route: Article 102 may apply to arrest, detention, remand, trial or punishment safeguards where evidenced. | Coercive enforcement and liberty safeguards | Which arrest, remand, bail, detention, charge-sheet and disposal data for these coercive-enforcement categories? |
| Other Cases | Article 31 (protection of law) + Article 39 (freedom of thought, conscience and speech) + Right to Information Act, 2009 | Textual/statutory transparency basis | Part III + statutory route: Article 102 may apply if opacity becomes an evidenced public-law legality or rights issue. | Legality, classification and public information | Can the Government disclose which statutes and offence groups sit inside the residual “Other Cases” category? |
| Police Assault / police-conduct follow-up data | Article 21 (duties of public servants) - Persuasive/interpretive only; not independently justiciable per Article 8(2) | Article 21-only public-servant duty; no standalone writ basis without a linked Part III right or statutory duty. | Article 21 alone: monitor/RTI/NHRC route only unless paired with an enforceable Part III right. | Public-servant duty and official-conduct accountability | Can police-as-victim figures be separated from custody, complaint, disciplinary and use-of-force data needed to assess official conduct? |
| Burglary, Theft, Riot | Articles 31–32 (protection of law; life and personal liberty) | Extended Articles 31–32 reading - judicially extended / analogical. Extended reading: comparative support only; no settled Bangladeshi appellate holding creates a general third-party-crime protective duty. | Part III enforceable route: Article 102 may apply only if systemic official failure is independently evidenced. | Protection of law and public order | What public-order prevention and response measures correspond to these registered offence signals? |
| Recovery Total | Articles 32–35 (life/liberty and criminal-process safeguards) | Textual safeguards + empirical follow-through | Part III enforceable route: Article 102 may apply to custody, liberty or fair-process questions where evidenced. | Enforcement output needing outcome data | What happens after recovery: arrests, bail, remand, charge sheets, final reports, convictions, acquittals and case disposal? |
Questions the table cannot yet answer
The watchdog function begins where the official table stops
The Police Headquarters table is a primary source, but it is not a complete constitutional record. Outcome disclosure is therefore pursued through a realistic three-phase advocacy roadmap rather than a single all-or-nothing 0/8 demand.
Statutory breakdown of “Other Cases”
Without this, citizens cannot know which laws are being used inside the largest residual category.
Accountability path: RTI request; Information Commission appeal/complaint; Article 102 (constitutional writ jurisdiction) if opacity becomes a public-law legality issue.Phase 1: custody-death data, remand and arrest baseline
Highest-stakes State-control fields should come first because custody death, remand and arrest are closest to liberty, bodily integrity and immediate official responsibility.
Accountability path: Achievable within 12 months of sustained RTI/parliamentary pressure if PHQ adopts a unit-wise monthly template and reconciles with prison or magistracy records where needed.Phase 2: charge-sheet and disposal tracking
These fields show whether FIRs move beyond registration into investigation and case-status outcomes, but require consistent identifiers and monthly status definitions.
Accountability path: Realistically 18–24 months with PHQ case-status standardisation, charge-sheet/final-report/disposal definitions, correction logs and unit-level reconciliation.Phase 3: bail, conviction/acquittal and tribunal outcomes
These are the hardest outcome fields because they require case-linkage across police, prosecution, courts and tribunal registries rather than PHQ publication alone.
Accountability path: Multi-year; requires PHQ-Judiciary-prosecution data integration, shared identifiers, court-status linkage and clear counting rules by FIR, accused person, offence or judgment.Unit-level evidence
Police-unit table with proxy per-100,000 rates
Current table month: July 2025. Raw volume and proxy rates are shown together. Rates reduce scale distortion but do not create final crime-prevalence rankings.
Comparison caveat: Metro (DMP, CMP, etc.) and Range rows use different population bases - city-corporation vs. division-wide - and are not directly comparable for ranking purposes. Compare within type, not across.
| Unit | Total | Rate /100k | Life/security /100k | W&C /100k | Other share | Other | Narcotics | W&C | Murder | Kidnapping | Robbery | Dacoity | Speedy Trial | Police Assault | Theft | Burglary | Arms | Explosive | Smuggling | Recovery Total | Population basis |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| DMP | 1,606 | 15.62 | 1.38 | 1.66 | 34.6% | 555 | 528 | 171 | 75 | 25 | 38 | 4 | 8 | 5 | 101 | 46 | 23 | 13 | 11 | 575 | Dhaka North + Dhaka South City Corporation - city-corporation proxy |
| CMP | 422 | 13.08 | 0.71 | 1.67 | 30.6% | 129 | 156 | 54 | 9 | 6 | 8 | 0 | 4 | 4 | 28 | 16 | 3 | 0 | 5 | 164 | Chattogram City Corporation - city-corporation proxy |
| KMP | 140 | 19.48 | 0.42 | 1.95 | 27.1% | 38 | 61 | 14 | 2 | 1 | 0 | 0 | 1 | 0 | 14 | 7 | 2 | 0 | 0 | 63 | Khulna City Corporation - city-corporation proxy |
| RMP | 203 | 36.72 | 1.63 | 3.26 | 31.5% | 64 | 86 | 18 | 2 | 3 | 4 | 0 | 0 | 2 | 14 | 6 | 4 | 0 | 0 | 90 | Rajshahi City Corporation - city-corporation proxy |
| BMP | 123 | 29.33 | 0.48 | 6.20 | 33.3% | 41 | 40 | 26 | 1 | 0 | 1 | 0 | 1 | 0 | 8 | 5 | 0 | 0 | 0 | 40 | Barishal City Corporation - city-corporation proxy |
| SMP | 107 | 20.10 | 0.94 | 2.07 | 35.5% | 38 | 31 | 11 | 2 | 0 | 3 | 0 | 1 | 0 | 9 | 4 | 1 | 1 | 6 | 39 | Sylhet City Corporation - city-corporation proxy |
| RPMP | 74 | 10.45 | 0.42 | 0.99 | 43.2% | 32 | 20 | 7 | 0 | 1 | 1 | 1 | 0 | 0 | 8 | 0 | 4 | 0 | 0 | 24 | Rangpur City Corporation - city-corporation proxy |
| GMP | 238 | 8.90 | 0.75 | 1.20 | 39.1% | 93 | 70 | 32 | 9 | 6 | 3 | 2 | 0 | 2 | 13 | 5 | 1 | 1 | 1 | 73 | Gazipur City Corporation - city-corporation proxy |
| Dhaka Range | 2,746 | 6.21 | 0.33 | 0.82 | 40.1% | 1,102 | 879 | 364 | 77 | 12 | 36 | 22 | 14 | 6 | 142 | 70 | 20 | 0 | 2 | 901 | Dhaka Division - division proxy, not police-jurisdiction adjusted |
| Mymensingh Range | 934 | 7.64 | 0.29 | 0.93 | 43.7% | 408 | 293 | 114 | 23 | 10 | 3 | 0 | 7 | 1 | 36 | 7 | 1 | 0 | 31 | 325 | Mymensingh Division - division proxy |
| Chittagong Range | 2,421 | 7.30 | 0.30 | 0.72 | 38.6% | 935 | 863 | 239 | 46 | 18 | 25 | 11 | 16 | 7 | 124 | 58 | 50 | 2 | 27 | 942 | Chattogram Division - division proxy |
| Sylhet Range | 746 | 6.76 | 0.25 | 0.93 | 51.9% | 387 | 133 | 103 | 20 | 3 | 3 | 2 | 3 | 3 | 51 | 5 | 2 | 0 | 31 | 166 | Sylhet Division - division proxy |
| Khulna Range | 1,347 | 7.74 | 0.26 | 1.29 | 39.1% | 527 | 389 | 225 | 25 | 7 | 10 | 4 | 14 | 2 | 81 | 23 | 22 | 4 | 14 | 429 | Khulna Division - division proxy |
| Barishal Range | 875 | 9.62 | 0.37 | 2.07 | 47.1% | 412 | 153 | 188 | 15 | 4 | 7 | 8 | 14 | 1 | 43 | 24 | 3 | 2 | 0 | 158 | Barishal Division - division proxy |
| Rajshahi Range | 1,634 | 8.03 | 0.26 | 1.15 | 45.4% | 742 | 468 | 235 | 27 | 7 | 14 | 4 | 6 | 3 | 89 | 9 | 11 | 11 | 8 | 498 | Rajshahi Division - division proxy |
| Rangpur Range | 1,688 | 9.59 | 0.23 | 1.63 | 48.6% | 821 | 446 | 287 | 29 | 6 | 5 | 1 | 0 | 3 | 72 | 7 | 8 | 0 | 3 | 457 | Rangpur Division - division proxy |
| Railway Range | 85 | - | - | - | 10.6% | 9 | 42 | 9 | 0 | 0 | 5 | 0 | 0 | 0 | 16 | 0 | 0 | 0 | 4 | 46 | No territorially comparable census denominator used - not applicable |
Per-100,000 rates are monthly proxy rates based on BBS 2022 denominators. Railway Range is not rate-normalised because no territorially comparable denominator is used.
Methodology
Methodology standard, authorship and caveats
The Index separates source fact, calculated signal, public rights question, doctrinal basis and accountability path. It is built to be readable for the public and still auditable by lawyers, researchers and policy reviewers. Methodology: CW Research Team (lead: Md. Ibrahim Khalilullah, President, Constitution Watchdog). Calculation methodology v1.0, adopted 3 July 2026.
Source discipline
Only official PHQ monthly FIR and recovery tables are used. The State’s own published data is treated as the factual starting point, while the Index clearly warns that FIR data is not conviction data or victimization prevalence.
Signal discipline
Movement is calculated through adjacent 3-month rolling averages, per-series volatility thresholds and count-floor confidence tags. This discipline, treating quantitative movement as a trigger for inquiry rather than a self-executing finding of breach, follows the general methodological posture of frameworks like UN OHCHR's human rights indicators guide and CESR's OPERA framework, adapted here to a domestic criminal-justice dataset neither framework was built around. A signal is a reason to ask for explanation, not proof of breach.
Legal discipline
Article 31 is the primary protection-of-law anchor; Article 32 supports life and personal-liberty questions where the facts justify it. Article 21 is interpretive only unless paired with an enforceable Part III right.
Advocacy discipline
Outcome disclosure is sequenced realistically: custody, remand and arrest first; charge-sheet and disposal tracking next; court-linked outcomes through longer PHQ-Judiciary integration.
Evidence and source disciplineFull notes
FIRs are not convictions
The source records FIR and recovery counts. It does not establish guilt, prevalence, conviction, acquittal or judicial fact.
The State is the source
Police Headquarters is the official publishing authority, and the Index deliberately relies only on PHQ's own published figures rather than an outside estimate or alternative count. Because every number originates from the State's own record, any accountability question raised here is a question about the government's own figures, not about a rival estimate the government could reject. The data is primary administrative data, but it is not independently audited crime-prevalence data; this self-reporting limitation is treated as a caveat, not a reason to ignore the data.
Category classification
Offence categories come from PHQ. CW maps those categories to constitutional provisions, doctrinal confidence levels and enforceability tiers in the codebook; CW does not redefine the source data.
Other Cases definition
Other Cases means the residual category labelled by PHQ. The Index treats the missing statutory breakdown as a legality, public-information and RTI issue.
Extraction and verification
Monthly PHQ PDF tables are treated as immutable source documents. Extracted category sub-totals must sum to the published category total, and category totals must sum to Total FIRs before publication. A minimum 10% random sample of monthly figures should be independently re-extracted with zero tolerance for count-field discrepancies.
Revision policy
If PHQ revises a previously published month, the original extracted value is retained and marked original, while the revised value is added as PHQ-revised with date. Historical classifications are not silently overwritten; a parallel revised-series recalculation may be shown with a dated note.
Calculation and signal disciplineFull notes
Metric construction
Total FIRs are the sum of all category counts for the month. Rate/100,000 is Total FIRs divided by the population denominator, then multiplied by 100,000. Category share is category count divided by Total FIRs, shown as a percentage. Counts are displayed as whole numbers; rates and shares use consistent public decimal precision.
Trend discipline
The Index privileges adjacent 3-month rolling-window movement over isolated single-month jumps. Increase or decrease creates an accountability question, not a verdict.
Movement formula
For month M, the current rolling average is mean(M, M-1, M-2). The prior rolling average is mean(M-1, M-2, M-3). Volume movement is current rolling average minus prior rolling average, divided by prior rolling average, multiplied by 100. Share movement is current rolling-average share minus prior rolling-average share, reported in percentage points, not relative percent.
Volatility calibration
Each series receives its own thresholds from its signed month-over-month rolling-average change history. The standard deviation of that change series is multiplied by permanent constants: Watch = 0.85σ, Concern = 1.7σ, Severe = 3.4σ. The mean is retained as a descriptive calibration statistic but does not shift the classification baseline.
Indicator framework lineage
The Index's separation of direct disclosure (a structural/process fact: does PHQ publish the field at all) from movement readings (an interpreted signal requiring explanation) follows the general methodological posture set out in OHCHR's Human Rights Indicators: A Guide to Measurement and Implementation, which treats administrative statistics as one recognised category of human rights data and cautions that the absence of published data can itself be a signal of a duty bearer's lack of commitment, and the caution against treating quantitative underperformance as a conclusive breach finding set out in CESR's OPERA Framework. Neither framework was built around crime or FIR data specifically; the Index adapts their general structural/process/outcome discipline to a domestic criminal-justice dataset not contemplated by either source. FIR movement is treated as a burden-of-explanation trigger, not as proof of breach.
Classification rule
Only pre-defined adverse-direction movement can trigger Watch, Concern or Severe, and only after the count-floor confidence rule is applied. For the current Index series, increases in Total FIRs, Life/Security, Women & Child Repression, Narcotics and Other Cases share are treated as adverse-direction movement. Favorable movement is reported numerically but does not trigger a risk label.
Calibration confidence
No series receives a calibrated threshold before 12 observed monthly changes are available. Any classification based on fewer than 24 observed monthly changes carries a mandatory preliminary calibration qualifier wherever the classification appears.
Count-floor confidence rule
Small monthly counts can make percentage changes look dramatic even when the actual number of FIRs moved by only a few cases. To avoid false alarms, the Index applies a count floor before displaying any Watch, Concern or Severe signal: if the current monthly n is below 20, the card is tagged Low confidence, insufficient volume for signal, monitor only, and the signal label is suppressed; if n is 20-99, the card is tagged Moderate confidence; and if n is 100 or more, the card is tagged High confidence. The raw number remains visible, but small-count movements are treated as monitor-only rather than alerts.
Confidence badge copy
Signal cards use short public badges: High confidence, Moderate confidence, and Low confidence, monitor only. Each signal card also states: Underlying n = [count]; confidence tier: [tier].
Seasonality rule
Until at least two full years of same-calendar-month data exist, no seasonal adjustment is applied. Once enough same-month history exists, the Index should display both adjacent rolling-window movement and same-window year-over-year movement, defaulting to the lower tier when the two materially disagree until three full years exist.
Population rates
Rate/100,000 is calculated as Total FIRs divided by population denominator, multiplied by 100,000. The denominator uses the most recent BBS population estimate available at publication and is updated only at each annual BBS release, logged as a version change. Current bundled rates use BBS Census 2022 proxy denominators.
Legal and constitutional disciplineFull notes
Constitutional enforceability axis
The Index separates constitutional relevance from enforceability. Part III rights, including Articles 27, 31, 32 and 35, are treated as directly justiciable rights: where the facts, standing and remedy requirements are met, a person or public-interest petitioner may invoke Article 102 writ jurisdiction. Article 21 is different. It sits in Part II as a Fundamental Principle of State Policy about duties of citizens and public servants. Because Article 8(2) says Part II principles guide governance, legislation and interpretation but are not judicially enforceable by themselves, Article 21 is not used here as a standalone writ ground. It may strengthen interpretation of a linked Part III right, a statutory duty, custody complaint, unlawful inaction or official-conduct issue, but it cannot alone convert every service failure into a constitutional case.
Article 102 path
Article 102 is used with the public-interest standing line of cases in mind; a monitor or civil-society actor may need to show public wrong, systemic failure or an indeterminate class affected, not merely institutional curiosity.
Positive-duty caveat copy
Inline caveat: Extended reading: comparative support only; no settled Bangladeshi appellate holding creates a general third-party-crime protective duty. Footnote version: The Index’s Articles 31-32 reading treats official FIR movement as a trigger for public-law explanation, not proof of breach. It extends beyond confirmed Bangladeshi appellate holdings, which principally address state action, custody, arrest, detention, or environmental life conditions. The third-party-crime protective-duty theory is supported only by comparative and persuasive authority, including Osman v UK and Velásquez-Rodríguez v Honduras; no Bangladeshi appellate court is cited here as recognizing it generally.
Sub judice caution
The Index does not decide individual cases that are sub judiceA matter currently before a court, where public commentary should avoid prejudging the case.; it asks aggregate public-data questions about prevention, classification, liberty safeguards and justice follow-through.
Disclosure roadmap and remedy realismFull notes
Outcome-disclosure roadmap
The Index still records whether PHQ publishes the eight monthly outcome fields: arrest, bail, remand, charge-sheet, disposal time, conviction/acquittal, tribunal outcomes and custody-death data. Advocacy is sequenced into three phases: Phase 1 seeks custody-death data, remand and arrest baseline fields; Phase 2 seeks charge-sheet and disposal tracking; Phase 3 seeks bail, conviction/acquittal and tribunal outcomes because those require police-court-prosecution linkage. The current publication status remains visible as a score, but it no longer functions as an all-at-once implementation demand.
RTI, litigation and parliamentary route realism
The roadmap does not assume that every missing field can be forced quickly through litigation. RTI requests and Information Commission processes may face delay or backlog, while Article 102 litigation is lawyer-dependent, time-consuming and costly for civic monitors. Parliamentary Questions and standing-committee pressure should therefore run in parallel as a lower-cost channel for ministry-level disclosure, especially for Phase 1 and Phase 2 reporting templates.
Interpretive ambiguity
FIR levels may reflect real incidence, reporting access, police capacity, registration practice, public trust, under-registration, or false or politically motivated filings. The Index therefore asks for outcome and audit data before drawing stronger conclusions.
Live calibration tableComputed from bundled series
| Series | Type | Observed changes | σ | Watch | Concern | Severe | Status |
|---|---|---|---|---|---|---|---|
| Total FIRs | Volume movement (%) | 16 | 5.9% | 5.0% | 10.0% | 20.0% | Preliminary calibration |
| Life/Security | Volume movement (%) | 16 | 3.3% | 2.8% | 5.5% | 11.1% | Preliminary calibration |
| Women & Child Repression | Volume movement (%) | 16 | 10.5% | 8.9% | 17.9% | 35.7% | Preliminary calibration |
| Narcotics | Volume movement (%) | 16 | 7.9% | 6.7% | 13.4% | 26.9% | Preliminary calibration |
| Other Cases share | Share movement (pp) | 16 | 1.3 pp | 1.1 pp | 2.2 pp | 4.3 pp | Preliminary calibration |
Current calibration table: Threshold values are calculated live from the bundled series. They are not hand-entered fixed round numbers.
Population denominator note: These are proxy denominators for monthly per-100,000 rates. Metropolitan units use corresponding city-corporation populations where available. Range units use division populations from Table 2.6. These are not official police-jurisdiction-adjusted denominators and are shown to reduce raw-volume distortion, not to rank actual crime prevalence.
Evidence archive
Primary Police Headquarters documents and local archive copies
Each public accountability signal remains tied to a source document, an official portal reference, and machine-readable evidence. Open a year below to view the months available for that year.
20265 available months
May 2026
PHQ upload date: 09-06-2026
April 2026
PHQ upload date: 06-05-2026
March 2026
PHQ upload date: 08-04-2026
February 2026
PHQ upload date: 08-03-2026
January 2026
PHQ upload date: 09-02-2026
202512 available months
December 2025
PHQ upload date: not available in supplied data
November 2025
PHQ upload date: not available in supplied data
October 2025
PHQ upload date: not available in supplied data
September 2025
PHQ upload date: not available in supplied data
August 2025
PHQ upload date: not available in supplied data
July 2025
PHQ upload date: not available in supplied data
June 2025
PHQ upload date: not available in supplied data
May 2025
PHQ upload date: not available in supplied data
April 2025
PHQ upload date: not available in supplied data
March 2025
PHQ upload date: not available in supplied data
February 2025
PHQ upload date: not available in supplied data
January 2025
PHQ upload date: not available in supplied data
BBS Census denominator
Population and Housing Census 2022 extract used for proxy rates.
Official portal listing: Bangladesh Police Crime Statistics portal. If an official PDF URL changes, the archived copy where bundled preserves the source reviewed for this publication; monthly JSON remains available for each loaded period.
