Hasina's data fudging & the white paper

The draft White Paper on state of the Bangladesh economy mentions “data” more often than either “development” or “economy”. It simply had to. If the murder of Bangladesh’s economy was the crime, data manipulation was the cover up. As the White Paper puts it, data fudging by Hasina’s kleptocratic regime probably started as an illusion to establish a development narrative but perhaps ended up being a delusion even to the wire-puller. The regime distorted, delayed or altogether withheld official data to support its development myth.

The good: The white paper dedicates a chapter (XXII) to this issue and describes how official data became increasingly unreliable and opaque over the years under the fascist regime. It explores data quality issues across multiple domains including gross domestic product or GDP (national accounts), trade, balance of payments, inflation, labour, poverty, food security, etc. It explains how lack of availability of and access to raw data, poor coordination among data producing agency, use of outdated methodology, methodological cherry-picking and outright political meddling affected official figures.

The draft paper makes some good recommendations for the reform of Bangladesh’ statistical system. These include independence and increased autonomy for the Bangladesh Bureau of Statistics (BBS) through improved governance and statistical legislation, creation of an Independent Expert Committee for Review and Re-estimation of National Data Series, preparing a Data Quality Assurance Framework, improved coordination and access, capacity building, improved IT infrastructure, etc.

These reforms will take time and political will. The interim government should initiate them and the next elected government should continue the effort. For effective policymaking, the importance of relevant, accurate and timely data cannot be overstated. For a government to destroy its own statistical capacity is the same as blindfolding itself with own hands. We saw where it led the Hasina regime to. The political parties should learn a lesson here and vow to rebuild Bangladesh’s statistical system, even if they do not like the published figures at times.

The Achilles’ heel: But ironically, data, or the poor quality of it, is also the Achilles’ heel of the draft White Paper. In many places, it argues how official data do not align with model estimates and conventional theories. Undoubtedly, there is a strong case against data fudging. But the problem is, model estimates and theories cannot disprove official figures. Rather, the effectiveness of models and theories are evaluated against official data. If one were to play the devil’s (read Hasina’s) advocate, she/he could reject the conclusions in the report out of hand.

In fact, discrediting the report and covertly exonerating the Hasina regime’s falsely claimed achievements have already started (e.g., Prothom Alo, 24 Dec, 2024).

The draft paper repeatedly emphasised Hasina’s data fudging. However, a few important aspects are missing. For instance, it does not provide any plausible adjustment factor for GDP figures. This has important implications, e.g., for having a realistic picture of our external debt sustainability – the actual debt-GDP ratio (with downward adjusted GDP) is likely to be higher than what was reported by the kleptocratic regime. It will be far worse with under-reported debt and inflated GDP data.

Perhaps because of the failure to provide realistic estimates of adjustment factors for kleptocratic regime’s various socio-economic data, the report struggled to explain how Bangladesh met all the criteria for graduation from LDC status. It referred to a UNDESA study ignoring the fact that the UNDESA database is still the same fudged one. Thus, by opting against the postponement of LDC graduation, the White Paper contradicted itself – while claiming data fudging, yet acknowledging that “unprecedented” development did take place.

So, what do we do: Well, we need to produce the smoking gun. We need to move on from models and theories and start digging into the raw data. That Independent Expert Committee for Review and Re-estimation of National Data Series could be set up immediately or the interim government can direct BBS to revise some key official data series. It is understood from the White Paper that GDP estimation suffers from poor quality of raw data on a number of sectors. But there are some low hanging fruits which can conclusively prove that data manipulation of epic proportion did happen and can create strong justification for the implementation of the recommendations.

For example, revising trade data is both possible and very important. It is possible because the customs administration should forever have the customs declarations which trade statistics are based on. So, the raw data are there. It is important because it is believed that most illicit financial flows happen through trade mis-invoicing. Accurate and detailed trade data are our primary source of information for estimating how much money has been siphoned off from Bangladesh by the past regime. The year 2015 comes up often in the White Paper as the beginning of a free fall in multiple sectors. For official data, it adds that “…checks and balances BBS had through technical committees completely broke down from 2015 onwards”.

It is perhaps no surprise that Bangladesh stopped sharing trade data with the United Nations since 2015 and siphoning of money out of the country gained momentum.

International trade data system has the special characteristic that it is a two-sided account. Bangladesh’s export of cotton T-shirts to United States (US) is also US’ import of cotton T-shirts from Bangladesh. In practice, there are some other factors at play but overall, this is how it is. Users can easily compare international trade data between partner countries and any glaring disparities become immediately apparent.

BBS, Exports Promotion Bureau (EPB) and Bangladesh Bank (BB) all publish external trade data. But their data lack the necessary details to be comparable. They are at an aggregated level and not disseminated in a comparable manner. EPB publish exports data only.

The gulf between reality and official trade data came to a head when,weeks before Sheikh Hasina’s ouster, BB revised exports data stating that EPB’s figure was US$10 billion higher than actual exports. The Chief Adviser Muhammad Yunus in his second address to the public promised to publish accurate trade data. It is a very necessary and welcome step.

However, accurate is not sufficient. We also need the necessary details in the data to allow for comparison with our trading partner countries’ data. In particular, we need calendar year (Jan-Dec) monthly trade flow (exports, imports, re-exports, etc.)data (value, net weight and quantity) with breakdown by partner and commodity codes, i.e., using the Harmonised System 6-digits level. (There are around 6,000 HS 6-digits codes available from the World Customs Organization).

The same should, in theory, be possible for inflation (CPI) and labour statistics. BBS should have the raw data. Inflation data could have been manipulated by either tempering with the prices of commodities and services in the basket or the composition of the basket itself or both. Depending on the nature of the shenanigan (i.e., dishonest activity or manoeuvring), revisions may or may not be possible but a close look at the raw data should reveal the manipulation.

The White Paper mentioned that the absurdly low unemployment rate was due to use of the 13thInternational Conference of Labour Statisticians (ICLS) methodology and not the updated 19th ICLS which is stricter and more relevant in its definition of unemployment. Is it possible to apply the 19th ICLS methodology on the existing survey data to revise at least some of the labour statistics? It demands serious consideration.

Even if GDP calculation is a lot more challenging and involves estimation for a number of sectors, it is nevertheless important to empirically establish an adjustment factor for correcting the inflated figures. This can be done for the most recent years for which raw data might be available and applied to previous years. The corrected GDP figures should be published with detailed metadata for full transparency. In general, going forward, BBS should adhere to the quality standards detailed in the 2019 United Nations National Quality Assurance Frameworks Manual for Official Statistics (UN NQAF).

Data to policymaking is what salt is to cooking. You never consider salt to be the main ingredient in your dish, but it is the only ingredient you cannot cook without, the one input which can make or break it. A new Bangladesh will need sound policymaking which in turn requires an independent and strong statistical system.

Khalid Saifullah, Fellow, Save Bangladesh USA Inc. Statistician by profession with experience in working in international organizations. [email protected]

Anis Chowdhury, Emeritus Professor, Western Sydney University, Australia. Held senior positions at the United Nations in New York & Bangkok. [email protected]