About this site
Why this site exists, who it is for, the principles it follows, how the research was done, what its limits are, and how to read its boxes and layers.
Purpose
This site exists to help Bangladeshis, and everyone who wishes the country well, see its economic situation clearly and think together about what could make it prosper. It is not written to praise or to condemn any party, leader or group. It is written out of care for the country and its people.
Its core message is that whatever our differences, we can all work together for Bangladesh’s prosperity. That is a conclusion of the research as well as a hope: when we map what each group wants and what it is paid for, almost every group turns out to have something to gain from a more productive economy.
Who it is for
- Bangladeshis at home and abroad who want to understand the economy and take part in improving it.
- Non-Bangladeshis who are interested in the country’s economy or who want to help, as investors, buyers, development partners, scholars, students or friends.
- Officials, journalists, teachers and anyone who needs a clear, sourced account.
You do not need an economics background. Technical words are explained when they first appear and in the glossary.
Editorial principles
These rules apply to every page. They are written out in full for the people who maintain the site; here they are in plain words.
- Written with care. The voice is warm but professional and neutral.
- No blame. The site does not blame parties, leaders, religions, regions or groups. It describes incentives and systems, not villains. But it also does not hide facts. Where the evidence is uncomfortable for any group, it is stated plainly and without drama. Neutrality never means leaving out evidence.
- Facts are separate from judgement. Every figure has a source, with a link where one is available, and none is exempt as common knowledge. Where sources disagree, both figures are shown. Our own assessments are labelled as ours. Figures that cannot be sourced are marked unverified or left out. We never invent figures, sources or links.
- Three layers. Short overviews for anyone, detailed analysis for those who want the reasoning, and data and sources for those who want to check.
- No duplication, and no omission. Each fact or argument lives in one place; other pages summarise it in a line and link to it. Nothing important from the research is dropped: a coverage record notes where every point landed or why it was cut.
- One voice. Pages are written as a single argument. Source documents are credited as sources, not quoted as competing voices.
- Plain language. Short sentences, everyday words.
- Only reviewed work is final. Pages are marked as drafts until an independent review has been assessed and its accepted findings applied.
Method
The research behind the site was done by several frontier and open-weight AI models working from published sources: government and central-bank data, international institutions, think tanks, academic papers and news reporting. Separate analysises covered the economic diagnosis, the loops among problems, the incentives of each group, strategy and sequencing, and the critique of the whole. Their results were cross-checked against each other. Where they disagreed, the more recent and better-reasoned view was preferred, and the disagreement is recorded on the relevant page and in the disputed figures list. A separate AI review, run slightly adversarially, then examined each page before it moves from draft to final, and an editor assessed each finding.
The synthesis was written, cross-checked and reviewed by AI systems, with the project owner setting the questions and the editorial rules. We say this plainly so that readers can judge it. Every claim carries its source so it can be checked by a person.
Limits
This is a first iteration, and it has real limits.
- Most figures are taken from reports and press coverage of the underlying data, not from the databases themselves.
- Several important things are unmeasured, such as the total size of extortion, party finance and what madrasa graduates earn.
- Some sources could not be read in full or could not be traced, and their figures are marked unverified.
- Power and direction scores, growth ranges and confidence levels are our judgement.
- The research was a snapshot as of 21 November 2026, and the facts will change.
The strongest objections and the gaps are on risks and open questions and, in full, on objections and critique.
How to read the boxes
The site uses three kinds of callout so that fact, dispute and opinion are never mixed up.
Key point
A key box states a main message, the point to take away if you read nothing else on the page.
Sources disagree
A disputed box shows a place where sources disagree and gives both figures or views, with their sources. It is not a verdict.
Our judgement
A judgement box marks our own assessment, such as a power score, a growth range or a confidence level. It is reasoned, but it is not a sourced fact, and others may reasonably differ.
Footnotes carry the source name and, where the research gave one, the link. Every figure needs a source; if a figure has no footnote, it is marked unverified or should not be there, and we would like to hear about it.
What comes next
This site is read-only for now. Future iterations will add ways to take part: to suggest corrections, to add sources, to challenge scores and judgements, and to contribute local knowledge that the research could not reach. The structure is built to be updated, with dated figures and a change log, so that it can stay current and become better with the help of the people it is for.
If you spot an error or have a better source, that is exactly what the next iteration is for.