Bangladesh GDP audit finds discrepancies as government promises independent statistics

GDP review exposes gaps in Bangladesh's economic figures

A forensic audit has identified discrepancies in Bangladesh's past gross domestic product calculations, ministers said on 8 November 2026, following a review that examined national income statistics from the previous decade. The findings concern the figures used to describe the size and growth of the economy, with earlier scrutiny identifying unusual manufacturing trends and missing records explaining how some estimates were produced.

Finance and Planning Minister Amir Khasru Mahmud Chowdhury and State Minister for Planning Md Zonayed Abdur Rahim Saki announced the findings at the Bangladesh Bureau of Statistics auditorium in Agargaon, Dhaka, during National Statistics Day. They said the audit report would be published after expert review. No publication date, revised GDP total or numerical adjustment to past growth rates was announced.

The government also outlined changes to how statistics are collected, checked and released. These include giving the Bangladesh Bureau of Statistics, or BBS, greater independence, removing the need for a minister's signature before publication, checking survey visits through location data and connecting government population registers.

GDP measures the value of goods and services produced within a country over a particular period. Errors in that calculation can change the picture policymakers see when deciding how much the economy can support, which sectors need help and whether development policies are working.

What investigators have been checking

A committee led by Statistics and Informatics Division Secretary Md Firoz Sarker has been tracing BBS estimates back to their underlying records. Its remit includes GDP, the Consumer Price Index and progress on the Open Government Data initiative. The Consumer Price Index, or CPI, tracks changes in the prices households pay and is used to calculate inflation.

The review compares sector estimates with Bangladesh Bank records and examines how calculations evolved across different statistical reference periods. Committee members described incomplete documentation, including cases where records explaining the origin of particular figures could not be found.

The investigation followed an expert taskforce established by the interim government to assess the quality, transparency and accessibility of official statistics. The subsequent BNP led government pursued a deeper examination of figures produced during the former Awami League administration.

The finance minister said the purpose was to understand past irregularities and correct the process, rather than target individuals. In July, committee members had described their central task as closing weaknesses in estimation methods rather than recalculating historical growth figures. By November, ministers were publicly stressing that the review had found evidence of discrepancies.

From a missed deadline to a promised report

The audit has developed over several months, with an initial submission deadline passing before technical findings were discussed publicly. The sequence helps explain why ministers are announcing findings while the final report still awaits publication.

• March 2025: Research and Policy Integration for Development published an independent assessment of Bangladesh's GDP using energy consumption.
• June 2026: Saki said past statistics were being reexamined for incorrect or distorted information.
• 15 July 2026: The audit committee's submission deadline passed without completion.
• 22 July 2026: Completion was expected within that month as investigators continued checking underlying records.
• 14 August 2026: A paper on weaknesses in national income measurement was presented at a Bangladesh Institute of Development Studies roundtable.
• 8 November 2026: Ministers said discrepancies had been identified and the report would be released after expert review.

The August paper was prepared by Bangladesh Institute of Development Studies Director General AK Enamul Haque and Research Director Atonu Rabbani. It examined national income statistics from the past decade and identified substantial room for inconsistencies in industrial data.

Why manufacturing and reference years matter

The technical review compared GDP trends calculated using the older 2005 reference framework and the later 2015 framework. The July description traced the older series to the 2005-06 financial year. Investigators found abnormal patterns, particularly in manufacturing, and examined the methods that might explain them.

A base year is a reference point used to compare economic activity over time. Updating it helps national accounts reflect changes in prices and the structure of production. Comparing results under different reference years can help investigators locate where a change in assumptions or coverage altered the measured growth path.

The August paper also highlighted reliance on figures from the National Board of Revenue, the Export Promotion Bureau and state owned enterprises. It identified a risk of overreporting when these secondary records feed into national accounts.

Another area requiring scrutiny was the Quantum Index of Industrial Production, or QIIP. This index tracks changes in the volume of industrial output. Its use in economic estimates involves assumptions about how the production being measured represents the wider industrial sector. The review flagged technical, procedural and methodological assumptions for closer examination.

Wahiduddin Mahmud, a former adviser to the interim government who chaired the August discussion, recommended a dedicated digital framework for national income accounting and a similar template for inflation measurement. Such a framework would give statisticians a consistent structure for recording inputs and calculations.

An independent estimate puts a number on the debate

Research and Policy Integration for Development, known as RAPID, approached the issue from a different direction. Its March 2025 energy consumption study, by Mohammad Abdur Razzaque and Syful Islam, used data covering 196 countries from 1990 to 2023 to examine the relationship between energy use and economic output.

The researchers estimated Bangladesh's GDP in 2023 at $314 billion, with a 95% confidence interval of $238 billion to $416 billion. That interval expresses the statistical uncertainty around the model's estimate. The brief described the official figure as 40% above its central estimate and about 5% above the upper end of the interval.

Bangladesh's exports required an adjustment to the model. The authors found merchandise exports were 20.5% higher than the global average for equivalent energy consumption. After accounting for the domestic value generated by export production, they raised their GDP prediction by an average of 13.6%.

The model predicted GDP above the recorded figure until 2015. After that, the recorded figure moved increasingly above the predicted path. This gives the official review another dated pattern to examine alongside its comparison of statistical reference years.

RAPID also explained the uncertainty in using energy as a guide to economic size. Technology, the mix of industries, urbanisation and informal activity all affect how much output a country produces for a given amount of energy. Formal GDP measurement requires detailed national accounting, while an energy model offers an additional way to check whether the economic picture looks plausible.

Boat counts and welfare records show the practical stakes

Rashed Al Mahmud Titumir, adviser for finance and planning, gave a striking example of weak administrative records: an earlier count had listed 13,000 vessels, while a review found 267,000. A more detailed description of his remarks identified the category as boats powered by engines.

The revised count is about 20.5 times the earlier figure, a difference of 254,000 boats. For agencies planning transport services or measuring economic activity on waterways, that changes the scale of the sector they are trying to understand.

Titumir also said social protection records contained widespread discrepancies and many intended recipients had missed assistance. He attributed an additional eight million people falling below the poverty line to misgovernment and policies based on distorted information. His remarks connected statistical quality to a concrete household concern: whether public assistance reaches people who need it.

One planned response is a proxy means test for selecting beneficiaries. This uses observable household characteristics as indicators of economic need when reliable income records are difficult to obtain. Officials said the aim was to remove false beneficiaries and direct support toward eligible poor people and women.

Three reforms target collection, selection and identity

The government presented three linked reforms at the November event. Together, they address how economic information is gathered, how welfare recipients are chosen and how records about the same person can be connected across public agencies.

• Use the international System of National Accounts for economic measurement and introduce geographic tracking to check whether survey collectors visited the households they were assigned.
• Apply proxy means testing to improve the selection of social protection beneficiaries.
• Build digital public infrastructure connecting national identity records, birth and death registration, and a population register.

The System of National Accounts provides common rules for recording economic production, income and spending. Geographic tracking addresses a different part of the process: whether the fieldwork behind the figures actually took place where it should have.

Connecting identity and population records would allow agencies to check information across registers rather than maintain separate, disconnected lists. Titumir said digital register exchange was being pursued to tackle discrepancies such as those found in vessel counts and welfare records.

The reforms also depend on staffing and practical capacity. Chowdhury said the government was considering BBS staff shortages, while Prime Minister's Political Adviser Zahir Uddin Swapon called for stronger institutions at both BBS and the Statistics and Informatics Division.

Publication without a minister's signature

Saki said BBS now publishes statistics, including the CPI, without waiting for a minister's signature. He described computer assisted processing in which information moves to a central database and is released automatically in real time.

Removing ministerial approval changes a specific point where political authority previously stood between statistical work and publication. Chowdhury said BBS was being granted full independence and argued that making more information public would allow more errors to be identified and corrected.

Researchers are also being offered greater access. Saki said a specialised data laboratory had opened so students and researchers from Bangladesh and abroad could analyse information independently. In August, Titumir had proposed a verification laboratory at the BIDS Graduate School and said future national surveys would be financed domestically rather than through donor agencies.

Swapon said a full Open Data Policy would be developed through discussion with stakeholders. The initiative extends the audit beyond a single report by giving people outside the agencies producing statistics a greater opportunity to examine them.

A scientific process rather than a political script

Swapon, who also chairs the Parliamentary Standing Committee on the Ministry of Planning, said independence must apply throughout the production of statistics, from collecting observations to interpreting them for policy.

Addressing the National Statistics Day gathering, Swapon described the process policymakers should follow:

"Data collection, interpretation and organisation, followed by transforming information into knowledge products and incorporating them into policy formulation, must follow a complete scientific process," he said.

He compared statistical integrity to medical diagnosis: a diagnostic report needs to describe a patient's actual condition, rather than fit a prescription chosen beforehand. Economic figures, in the same way, need to describe conditions before leaders decide what to do about them.

The November meeting was chaired by Firoz Sarker, with a welcome address from BBS Director General Md Shamimul Haque. Speakers also included BBS Director Kabir Uddin Ahmad, Dhaka University professor Syed Shahadat Hossain and BIDS Director General Enamul Haque.

Chowdhury linked reliable data and digital infrastructure to the growing use of artificial intelligence. Titumir described misinformation as part of an ongoing hybrid war, while Swapon argued that accessible, reliable public information would help counter misleading accounts about Bangladesh.

The immediate task is now publication of the reviewed audit, alongside the promised changes to collection and access. Its findings will give policymakers and the public a fuller account of where the calculations went wrong and which procedures need correction.

Key Points

• Ministers said a forensic audit found discrepancies in GDP calculations from the previous government's tenure.
• The final report is to be published after expert review; no release date or revised GDP figure was announced.
• Earlier technical work identified unusual manufacturing trends, risks in secondary data and incomplete documentation.
• RAPID estimated 2023 GDP at $314 billion using an energy model, with a broad uncertainty range of $238 billion to $416 billion.
• Planned reforms include location checks on surveys, new welfare selection methods and connected population records.
• BBS is publishing statistics without ministerial signatures, while officials promise greater research access and institutional independence.