Layer 1 · Overview

Why progress has stalled

Why Bangladesh's growth model is running out of road - in many places taking pays better than making - explained in four layers (institutions, politics, skills, norms), and why culture alone does not explain the outcome.

Updated 21 Nov 2026 · 4 min read

On this page
  1. Making and taking
  2. Four layers of explanation
  3. Why culture alone does not explain it
  4. Why the layers lock together
  5. Everyone has a stake

Bangladesh does not lack hardworking people. The same workers build skyscrapers in Dubai, sew the world’s clothes and send home $35.6bn a year.1 The country’s difficulty is the rules of the game inside it. This page gives the short version of the explanation.

Making and taking

Income per person is the share of people who work, times what each worker produces. What each worker produces depends on tools and capital, on skills, and on whether capital and skilled people end up where they are most productive.

On each of these, the evidence points to ways in which resources are allocated by connection and power as well as by productivity.

Our reading

Our synthesis is that in many places taking pays better than making, and that this is the common thread across the terms above. It is a judgement, not a measured fact.

  • A loan that is never repaid is cheaper than one that is repaid. About a third of outstanding bank lending by value is classified as bad,2 and about half of that sits with 5,775 borrowers (September 2025).3
  • A party post can pay more than a factory job. Posts bought with a payment are recovered later.
  • A certificate leads to a government post, a party link or a visa, so schools reward certificates over skill.
  • Paying a local collector buys peace, while registering with the tax office buys more visits, so small firms stay small and informal.

Each group in the system acts sensibly on what it is paid for. That is why appeals to patriotism, new targets and new plans have not been enough. The test of any proposal is whether it raises the return to making, or lowers the return to taking, for a group strong enough to matter.

Four layers of explanation

  1. A low starting point. In 1972 Bangladesh had about 62% of Pakistan’s income per head,4 after colonial land systems, the split of the jute economy and a war, with a famine to follow in 1974.
  2. A narrow model that worked, and is running out. Garments and remittances grew on informal deals that were reliable when one party dominated. Trade protection, a weak tax state, poor banks and energy costs now limit it.
  3. Institutions and politics. Where winning power gives control of police, courts, banks and contracts, losing is dangerous. Incentives favour short horizons, and independent checks on power can come to look like threats. Reforms that would bind a government usually need that government’s agreement.
  4. Skills, organisations and norms. Schools reward grades; hiring is often by connection; women are largely out of paid work; deference to rank can silence the feedback an organisation needs.

Why culture alone does not explain it

It is tempting to say that culture or religion is the reason. The evidence does not support that as a complete answer:

  • Muslim-majority countries differ greatly. Bangladesh’s income per head was 62% of Pakistan’s in 1972 and about 119% of it in 2025 (constant 2010 dollars).4
  • When garment factories opened near villages, girls stayed in school longer and married later, so norms moved with opportunity.
  • Claims such as lost hours from prayer, or a lower work ethic, have no good evidence behind them; Bangladeshis work about 2,383 hours a year on average.5
  • What does matter is enforced exclusion: rules or pressures that keep women out of paid work, limit peaceful dissent, or impose one view by threat. On the evidence that is a real but second-tier cause, below banks, taxes, trade and the structure of politics.

None of this says anything about the character of any people or faith. The full test of eleven claims is on Culture and religion.

Our ranking

We put the thin tax state, captured finance, trade protection, weak learning and women’s exclusion from work ahead of religion as constraints, and treat religion as a real channel to watch rather than as the root. Confidence is moderate.

Why the layers lock together

Because each layer feeds the others, fixing one in isolation usually fails: the others restore it. Banks fund politics, politics protects the banks; few good jobs push young people toward party work; extortion keeps firms small. These links are traced on The loops. They can be cut, and the best first cuts are where someone strong gains by changing them (The path forward).

Everyone has a stake

Because each group acts sensibly given what it is paid for, the way forward is to change those payoffs, not to blame anyone. A country that is safer to lose an election in, easier to do business in, and better at rewarding skill would be better for the party in power, the opposition, officials, exporters, workers and families (Everyone has a stake).


  1. Prothom Alo on the World Bank Bangladesh Development Update, https://en.prothomalo.com/business/local/qgylemzmn1. ↩

  2. Prothom Alo, NPL ratio 32.78%. ↩

  3. Bonik Barta, about Tk 3 trillion of default sits with 5,775 borrowers. ↩

  4. FRED (World Bank WDI), GDP per capita in constant 2010 US dollars: Bangladesh and Pakistan. ↩↩

  5. FRED (Conference Board), annual hours worked per person engaged, Bangladesh, 2023; comparability limits apply. ↩

Go deeper

Where this comes from

Synthesised from 5 research reports, checked against the sources cited above and reviewed before publication. How we work · All sources.