Layer 2 · In depth

Who wants what - the incentive map

Every group that matters for Bangladesh's economy acts sensibly on what it is paid for today. A profile of nineteen groups, a power-and-direction map, and what could change each group's payoffs.

Updated 21 Nov 2026 · 34 min read

On this page
  1. The short version
  2. Two frameworks, and their limits
  3. The vector model
  4. Maslow as a checklist, not a ladder
  5. The map: power and direction of each group
  6. An alternative way to score
  7. Group by group
  8. Prime minister and inner circle
  9. MPs and local party leaders
  10. Party youth and student cadres
  11. Opposition parties
  12. The barred Awami League base
  13. Civil service
  14. Police
  15. Military
  16. Judges and lawyers
  17. Large defaulters and protected conglomerates
  18. Exporters
  19. Small firms, traders and transport owners
  20. Workers, farmers and the general public
  21. Educated youth
  22. Women workers
  23. Qawmi ulama, Hefazat and mosque-based leaders
  24. Teachers and universities
  25. Migrants and diaspora
  26. Outside partners
  27. What the compact protects and asks, group by group
  28. From today’s payoff to a better one
  29. Who blocks and who could carry each reform
  30. A caution on “humans only respond to incentives”
  31. What is uncertain here

This page is the detailed version of everyone has a stake. It profiles each group that can enable or block a more productive economy: what it wants, why today’s behaviour pays, what could change its payoffs, and what that would cost the country. The loops these groups sit in are on circular dependencies; the deals that would change their payoffs are on strategy and deals.

The short version

Every group that holds back development in Bangladesh is behaving sensibly given what it is paid for. Party leaders remove checks because losing power can be dangerous. Cadres collect payments because the party pays them in territory, not wages. Officials delay because deciding is risky and some postings are bought. Large borrowers do not repay because not repaying has been close to free. Religious leaders defend their autonomy because it is their livelihood and standing. The same is true of the groups that help: exporters push for ports and power because they need them.

The task is to change what each group is paid for, at a price the country can afford. Two cautions run through this page. First, the power and direction scores are our judgement, not measurements. Second, people are moved by status, belonging and belief as well as by money, so some of the cheapest levers are not financial.

Every group has something to gain

The map is not a list of opponents. In each profile there is a version of the future in which the group is better off than today: safer, better paid, more respected or less exposed to risk. The reforms in this site only work if that is true for enough groups, which is why the deals are written as exchanges (see strategy and deals).

Two frameworks, and their limits

The vector model

Treat each group as a force with a size (its power to enable or block) and a direction (whether its self-interest today pushes toward or against a more productive economy). The country moves in the direction of the sum. In a more formal version, each group’s contribution is its weight times its strength times the cosine of the angle between what it is maximising and a “prosperity axis”: an angle near zero accelerates development, an angle near 90 degrees is neutral, and an angle above 90 drags.1

Policy can do three things:

  1. Rotate a group: change its payoffs so that self-interest points toward production. Even moving a group from active obstruction to passive neutrality is a large gain.
  2. Re-weight: give more power to groups already pointing the right way (exporters, small firms, women workers, migrants, honest officials) and less to those pointing the wrong way (extortion networks, wilful defaulters).
  3. Decouple: let a group keep what it values without needing to block anything (for example, an army with a secure budget and no role in policing).

Trying to destroy a powerful opposing group (weight to zero) usually triggers unrest or institutional paralysis. That is the practical reason for rotating and re-weighting instead.

Limits of the model

It is a thinking tool. It hides the fact that groups are not single actors: “the bureaucracy” contains honest officers and rent-seekers, and the split matters more than the average. Scores also depend on who is scoring; see the comparison of two scorings below.

Maslow as a checklist, not a ladder

The useful part of Maslow’s hierarchy is the list of needs: income and survival, safety, belonging, status, meaning. The ladder part, where higher needs matter only once lower ones are met, has weak support. A large cross-country study found that people pursue respect and belonging even when basic needs are unmet (Tay and Diener 2011, cited from memory and not re-checked).2 That matters here. An imam on about Tk 10,000 a month (the average in one small survey) and a student activist with no job are both poor, yet what they defend most fiercely is standing and belonging. Offers that raise income but cost status tend to be refused.

The map: power and direction of each group

Scores are our judgement from the evidence on this site. Power runs from 1 (little) to 5 (decisive). Direction runs from -2 (strongly against a productive economy) to +2 (strongly for). “After” is where the deals might plausibly move a group in five years, not a forecast.

These scores are judgement

Power, direction and the “after” column are our assessments. Someone else could score the military as negative (business interests, a history of coups), religious leaders as more strongly negative, or exporters as rent-seekers. The method is what matters: change a score and see whether the order of the deals changes. For most plausible re-scorings, banks and the insurance pact still come first because they act on the largest forces.

Group Power Direction now After Mainly wants Main lever
Prime minister and inner circle 5 0 +1 Stay in power, safety afterwards, legacy Insurance for losers; visible economic wins
MPs and local party leaders 4 -1 0 Recover election costs, feed the local machine Legal party finance; conflict-of-interest rules
Party youth and student cadres 3 -2 -1 Income, protection, status Cadre-to-career package
Opposition parties (Jamaat, NCP) 3 0 +1 Local base, clean image, win in 2031 Real stakes in parliament and local government
Barred Awami League base 2 (latent) -1 0 Safety, a way back Lawful status review; supporters’ rights; individual fair proceedings
Civil service 4 -1 0 Security, promotion, good postings Pay tied to accountability; less discretion
Police 3 -1 0 Recoup cost of post, political cover End purchase of posts; operating budgets
Military 4 +1 +1 Budget, autonomy, reputation, peacekeeping Barracks-for-budget; transparent businesses
Judges and lawyers 3 -1 0 Promotion; fees that rise with delay Stage-based fees; commercial courts; secretariat
Large defaulters and protected conglomerates 4 -2 -1 Cheap capital, protection, safe wealth Close the free option; open a paid exit
Exporters 4 +1 +2 Orders, energy, fast ports, low costs 2029 compact
Small firms, traders, transport owners 2 +1 +2 To be left alone; credit Simple tax and credit record; protection from extortion for all
Workers, farmers and the public 2 (far more when they strike or vote) +1 +1 Prices, wages, steady work, fairness Predictable wage-setting; fewer bribe points; visible fairness at the top
Educated youth 3 (proved in 2024) 0 +1 A job with status; or a way out Skilled private jobs; skilled migration
Women workers 1 +2 +2 Safe, respectable work near home Raise their weight: jobs in reach, safety, care
Qawmi ulama, Hefazat, mosque leaders 3 on some issues -1 on gender, 0 otherwise 0 Autonomy, respect, livelihoods for graduates Livelihoods compact; clear red lines on violence
Teachers and universities 2 -1 0 Income from tutoring; posts by loyalty Pay and career ladder tied to certification
Migrants and diaspora 2 +1 +2 Cheaper, safer migration; a say Lower costs, skills, investment channels
Outside partners (IMF, World Bank, EU, buyers, Japan) 3 +1 +1 Repayment, stability, standards Deadlines and conditions as anchors
Power and direction today (our judgement)
Re-aim these
Lead with these
Lower priority
Give more weight

Chart with Direction of self-interest today on the horizontal axis (from Pulls against production to Pulls toward production) and Power to enable or block on the vertical axis (from Less power to More power). Placement of each item:

  • PM and inner circle: high Power to enable or block, middling Direction of self-interest today; in the quadrant 'Lead with these'.
  • Civil service: high Power to enable or block, low Direction of self-interest today; in the quadrant 'Re-aim these'.
  • Large defaulters: high Power to enable or block, low Direction of self-interest today; in the quadrant 'Re-aim these'.
  • MPs and local leaders: high Power to enable or block, low Direction of self-interest today; in the quadrant 'Re-aim these'.
  • Military: high Power to enable or block, high Direction of self-interest today; in the quadrant 'Lead with these'.
  • Exporters: high Power to enable or block, high Direction of self-interest today; in the quadrant 'Lead with these'.
  • Outside partners: middling Power to enable or block, high Direction of self-interest today; in the quadrant 'Lead with these'.
  • Opposition parties: middling Power to enable or block, middling Direction of self-interest today; in the quadrant 'Re-aim these'.
  • Police: middling Power to enable or block, low Direction of self-interest today; in the quadrant 'Re-aim these'.
  • Party cadres: middling Power to enable or block, low Direction of self-interest today; in the quadrant 'Re-aim these'.
  • Ulama and mosque leaders: middling Power to enable or block, middling Direction of self-interest today; in the quadrant 'Re-aim these'.
  • Educated youth: middling Power to enable or block, middling Direction of self-interest today; in the quadrant 'Give more weight'.
  • Judges and lawyers: middling Power to enable or block, low Direction of self-interest today; in the quadrant 'Lower priority'.
  • Workers, farmers, public: middling Power to enable or block, high Direction of self-interest today; in the quadrant 'Give more weight'.
  • Awami League base (latent): low Power to enable or block, low Direction of self-interest today; in the quadrant 'Lower priority'.
  • Teachers: low Power to enable or block, low Direction of self-interest today; in the quadrant 'Lower priority'.
  • Small firms: low Power to enable or block, high Direction of self-interest today; in the quadrant 'Give more weight'.
  • Migrants and diaspora: low Power to enable or block, high Direction of self-interest today; in the quadrant 'Give more weight'.
  • Women workers: low Power to enable or block, high Direction of self-interest today; in the quadrant 'Give more weight'.
  • Politics and party machines
  • State institutions
  • Business, finance and partners
  • People and communities
  • Education and faith
Positions follow the table above, nudged slightly so that labels do not overlap. The barred Awami League base is placed from the table (the original chart left it out).

Three things stand out.

  • The top-left corner (powerful, pulling the wrong way) holds the groups that run the state day to day.
  • The groups pulling hardest the right way (women, migrants, small firms) have the least power.
  • The single most powerful actor sits in the middle and can tip either way.

A plan that only “fights” the top-left loses. A plan has to re-aim some of them, raise the weight of the bottom-right, and give the centre reasons to tip.

An alternative way to score

The same groups can be scored on a coarser basis: a 1-10 power weight and an alignment angle (above 90 degrees is a net drag, below 90 a net help) for eleven broader groupings. Reading it next to the map shows where reasonable assessments differ.

Group (broader grouping) Weight (1-10) Angle today Target angle Closest rows in the map above
Ruling party and leadership 9.5 130 35 PM and inner circle; MPs and local leaders
Opposition parties (including the barred party’s base) 7.5 135 40 Opposition; Awami League base
Youth and student cadres 7.0 160 20 Party cadres
Religious institutions and movements 8.0 125 80 Ulama and mosque leaders
Civil bureaucracy 9.0 140 25 Civil service
Police 8.0 145 30 Police
Judiciary and lawyers 8.0 120 20 Judges and lawyers
Military 9.0 75 35 Military
Business: protected firms (165) and exporters, tech founders, SMEs (20) 9.0 165 and 20 15 Large defaulters; exporters; small firms
General public, informal workers, farmers, migrants 8.5 85 10 Workers, farmers, public; migrants; women
Education sector, teachers, non-political students 8.0 110 10 Teachers; educated youth

All values are judgements. Group names are neutral descriptions.

Where reasonable scorings differ

  • Ruling party and opposition. The main map treats the leadership as neutral and able to tip (0) and the opposition as neutral (0), because the opposition gains by waiting and building rather than by forcing a street confrontation.3 A reasonable alternative scores both as net drags (about 130 and 135 degrees), on the assumption that an opposition with no formal power resorts to agitation. A check of 2026 violence found much of it is mob action and clashes inside and between governing and allied camps, which supports a refinement more than either extreme.4 We keep the main map because it rests on more recent, better-argued evidence.
  • Public power. An alternative gives the public and migrants a very large collective weight (8.5); the main map scores workers, farmers and the public at 2, rising sharply when they strike or vote. The difference is between latent collective power and ordinary-day power. The garment strike of 1 November 2026 supports the view that latent power is real.
  • Religious institutions. An alternative gives high weight and a clear drag (125 degrees); the main map scores them 3 and mildly negative only on gender issues. We keep the narrower reading because the evidence on religion as a cause is mixed (see culture and religion), but evidence on trends in religiosity and on madrasa graduates’ earnings is thin, so this is the score held with least confidence.
  • Military. Both scorings place it as the most stabilising large actor (about +1, or 75 degrees).

Group by group

Each profile gives what the group wants, why today’s behaviour pays, what could re-aim it, and the price.

Prime minister and inner circle

  • Wants: to stay in power; to be safe after power; a legacy.
  • Why today’s behaviour pays: the governing party holds 209 of 300 elected seats on about half the vote. Control of judges, police and the anti-corruption body is worth more to it today than constraints on them, which is why several interim-era reforms on those bodies were repealed or allowed to lapse.5 The fate of the Awami League teaches every ruler that losing can mean destruction. Keeping party cadres fed with local contracts and postings is also a way of keeping loyalty before the next election in 2031.1
  • What could re-aim it: make losing survivable. Research on courts finds that politicians build independent referees when they expect to lose power some day; the referee is insurance.6 The same logic covers a caretaker system, term limits and opposition-chaired oversight. Second, give wins that muscle cannot deliver: in one poll 89.4% said electricity was handled badly while 78.6% backed the Family Card.7 Third, legacy: finishing a term and handing over peacefully would be an achievement without precedent in the leader’s own family.3
  • Price: the fair-process guarantees that protect losers also apply to today’s rulers if they lose power; they are not immunity from properly brought proceedings.

MPs and local party leaders

  • Wants: to recover the cost of winning, keep the local machine fed, win again.
  • Why today’s behaviour pays: about 60% of MPs are businesspeople (about 70% of the cabinet) and half carry loans or liabilities.8 A seat pays back through contracts, licences, loan rescheduling and recommendations for jobs and postings. These are rational responses to a system in which regulated legal funding exists but disclosure and enforcement are weak and informal funding is common (a hypothesis; the size of the shortfall is not established).44
  • What could re-aim them: audited public funding per vote won and disclosed donations; rules (conflict-of-interest and recusal) that keep MPs with overdue loans out of decisions on finance and banking; credit for local projects delivered through open tenders. Parties funded this way would also be asked to dissolve armed or extortionist student and labour wings as a condition (a proposal in one source). MPs gave up plots and duty-free cars this year, which shows privileges can be surrendered when the public is watching.9
  • Price: taxpayers fund parties, which is unpopular; MPs who are also businesspeople lose an advantage they paid for.

Party youth and student cadres

  • Wants: income, protection, status, belonging and a route up. A poor rural student gets a hall bed, daily income from collections and a shortcut to contracts.
  • Why today’s behaviour pays: see the party-machine loop in circular dependencies. The party pays in territory, and violence is how a young activist proves his worth.10 About 31% of young people are not in work, education or training (one source gives 30.9%; see the disputed-figures page), so the alternatives are thin.
  • What could re-aim them: a package, because no single piece works alone: a legal income and status route (apprenticeships, elected student unions), the removal of the territory they are paid in, and a higher cost for extortion. The design, and why the wording is “separate party machinery from campuses” rather than “ban student politics”, is in Deal 4.
  • Price: some money reaches people who behaved badly, and formalising a “fee” that began as extortion makes it legitimate. Parties lose cheap street muscle and must rely on paid staff and popularity.

Opposition parties

  • Wants: a local base, a clean image, victory in 2031.
  • Why today’s behaviour pays: Jamaat’s first-time MPs and the NCP both gain more by waiting and building than by forcing a street confrontation.3 Jamaat is acting like a government-in-waiting: a shadow budget smaller than the government’s, interest-free loan pledges, an anti-corruption pitch.12 Its favourability slipped from 53% to 48%.13 Opposition parties took no seats in the committee drafting the constitutional amendment.14
  • What keeps them inside the system: real stakes: opposition chairs for the public accounts and oversight committees, local governments they can win, an upper house by vote share, and the legal pathway to contest through independent courts.
  • Price: a stronger Jamaat has more say over gender and family policy; the governing party loses its monopoly on legislation and municipal budgets, so lawmaking becomes slower and more negotiated.

The barred Awami League base

  • Wants: safety and a way back.
  • Why it matters: a large group with no legal channel has little to lose, and in Bangladesh’s history excluded parties have often pressed their case through the street (our reading). Past political shutdowns were costly: one widely quoted UNDP estimate put them at 3-4% of GDP in the 1990s, though researchers question the method.15
  • Three separate questions, kept separate: (1) the legal status of the party as an organisation, which is a matter for lawful procedure and review; (2) the rights of ordinary supporters and voters, who keep their civic rights and protection from arbitrary action whatever their affiliation; and (3) individual proceedings, where accusation is not guilt, and responsibility is established person by person in fair courts with victims and witnesses protected. Affiliation is never a test of guilt, and an accusation is never a test of who may take part in politics.
  • A counter-view: one economist writes that the party “has long ceased to matter” as a political force, and another argues it will not return for about two decades.316 If they are right this group’s weight is lower than scored.
  • Inclusion and accountability are compatible duties: they are not a trade of justice for stability. Fair, timely proceedings for those credibly accused go ahead, and the political rights of everyone else are protected.

Civil service

  • Wants: security, promotion, a good posting, a safe retirement.
  • Why today’s behaviour pays: doing nothing is safe and deciding is risky. A fast approval earns no bonus if it succeeds but risks charges if anything goes wrong, so delay is personally safe. Some postings cannot be had without payment, and the payment is recovered later.17 Pay has just risen 100-142% for 2.4 million employees with no conditions attached.18
  • What could re-aim it:
  • The pay rise is phased to July 2027 with allowances from 2028, so later phases can still be tied to asset declarations, as TIB has asked,19 and to published service standards.
  • “Deemed approval”: if a standard licence or clearance is not refused with written reasons within 7-14 days, the system approves it (a proposal; see the roadmap).
  • Team-based bonuses for service speed or digital tax collection, a competitive senior-services pool with lateral entry, statutory protection against being parked as an “officer on special duty” for refusing illegal orders, and a limited retirement package for senior staff who cannot adapt.
  • Evidence from elsewhere: announcing certain audits cut missing road funds in Indonesia by about 8 percentage points;20 biometric payments in Andhra Pradesh saved about $40m a year against $6.3m of cost;21 performance pay for tax collectors in Pakistan raised revenue (and in places raised bribes where oversight was weak).22
  • Price: strikes, as in the 2025 tax-service dispute; a higher senior pay bill; and results from other countries may not carry over.

Police

  • Wants: to recover what the post cost; cover from political masters; promotion.
  • Why today’s behaviour pays: the Police Reform Commission reported bribes of up to Tk 1 million for a constable’s post and said money paid for postings is recovered through bribes.23 A government survey found about 62% of people seeking police services paid a bribe.24 Stations are underfunded for basic operations, which pushes costs onto the public.25 Morale also suffered after August 2024 amid fears of reprisal (unverified; no primary source located), and the police are the only institution with more public disapproval than approval.26
  • What could re-aim them: postings by published rule; real operating budgets; a police commission with a binding three-name shortlist for the top job; shift caps, housing and body cameras; separate investigation and law-and-order wings; and one visible island, such as rebuilding highway and traffic policing on two corridors with new recruits and higher pay. Georgia dismissed about 16,000 traffic police and raised pay sharply, and bribery fell.27
  • Price: Georgia’s reform needed a dominant executive, and democratic control of the police stayed weak.28 MPs lose a tool for influence over local rivals, and the budget must fund better pay up front.

Military

  • Wants: a secure budget, autonomy, reputation, peacekeeping income, its welfare businesses.
  • Where it stands: the army has 78% public approval.26 Bangladesh had 5,689 peacekeepers in January 2025 and the UN pays $1,428 a month for each.29 The forces run welfare trusts with hotels, real estate and other firms (an estimate puts the commercial footprint at about $700 million or more; unverified).30 In August 2024 the army leadership did not use force against protesters, according to one input (not independently sourced here).
  • What keeps it pointing the right way: a predictable budget and continued peacekeeping, in return for staying out of policing and politics. Peacekeeping depends on the forces’ human-rights record, which is an outside anchor. On the businesses, publish the accounts, apply ordinary tax and audit rules, and stop expansion into new sectors; do not attempt a forced sell-off. Indonesia passed a law in 2004 to take over all military businesses within five years, and at the end of 2007 (the last year for which the 2010 report had data) the military still held 23 foundations, more than 1,000 cooperatives and 55 companies.31 There may be room for disciplined project execution in infrastructure and logistics.
  • Price: accepting a military business base that competes with private firms.

Judges and lawyers

  • Wants: judges want promotion and postings, which the executive controls; lawyers’ fees in some arrangements rise with the number of hearings, which can make delay pay (a general point that varies by lawyer and case; fee schedules are not documented in the research).
  • What could re-aim them: fees by stage of case instead of by hearing; caps on adjournments; paid mediation and arbitration as a new fee stream; commercial courts with published clearance rates; a Supreme Court secretariat so judges’ careers do not depend on ministers; and, for loan recovery, an escrow-before-stay rule with a hardship exemption and judicial discretion.
  • Price: bar associations are politically organised and will resist; the benefit of strong support from the Chief Justice and commercial banks is needed.

Large defaulters and protected conglomerates

  • Wants: cheap capital, protection from competition, safety for wealth already moved.
  • Why today’s behaviour pays: default is close to free. Each new rescheduling window confirms that waiting beats paying. Protective tariffs (one source cites a 70.4% effective tariff on footwear, a measure and product scope not specified; unverified) and guaranteed power capacity payments add to the returns of staying protected rather than competing.
  • What could re-aim them: close the free option and open a paid exit: a bankruptcy law, one settlement window with a fixed end date, and consequences for wilful non-payment, with a productive alternative in export finance and serviced zones. Honest firms hit by shocks are treated differently from wilful non-payers. The detailed design is in Deal 1 and Deal 5.
  • Price: a settlement is a partial forgiveness of civil and financial liability, not immunity from independent investigation. Money already abroad is, in one economist’s words, extremely difficult to recover,33 so part of the loss is permanent and the public will pay it. Several protected firms and weak banks will shrink or close.

Exporters

  • Wants: orders, reliable energy, fast ports, duty-free inputs, low wages.
  • Where they stand: aligned with the country on trade, ports and energy; not aligned on wages, unions and their own subsidies. Exporters are the main constituency for the November 2029 graduation date.
  • What could move them further: a 2029 compact. The state delivers tariffs, ports and gas; exporters accept a wage-setting process, the labour standards the EU requires, and targets for women’s employment. Duty-free bonded access could be extended from garments to other exporters.
  • Price: higher labour costs in a weak market.

Small firms, traders and transport owners

  • Wants: to be left alone; credit; predictable costs.
  • Why today’s behaviour pays: paying the local collector buys peace; registering with the tax office buys visits from more collectors. Staying small and informal is rational. About 932 transport associations already collect “donations” from vehicles.34
  • What could re-aim them: make the legal bargain better than the illegal one: a simple tax and a credit record, with protection from extortion that does not depend on registration (Deal 6).
  • Price: lower tax rates than standard VAT at first, and the state has to deliver the protection or the scheme loses all credibility.

Workers, farmers and the general public

  • Wants: affordable food, steady work, safety, dignity, fairness.
  • On “work harder and be less corrupt”: most household bribes are paid to obtain services people are entitled to; one TIB survey found 70.9% of households faced corruption in services.17 Wages have lagged prices (the official wage index trailed consumer prices for 57 months in a row, per one macro summary; unverified).32 People work harder when effort is rewarded and give fewer bribes when there are fewer points at which a bribe is demanded.
  • Their power is larger than it looks: on 1 November 2026 about 987,000 garment workers walked out of 529 factories over the minimum wage.35 When wages trail prices for years and there is no trusted way to settle pay, the street becomes the bargaining table, and exporters and workers both lose.
  • What helps: a predictable rule for setting wages; digital services that remove the official from the transaction; one unified transfer card paid through banks and mobile money;33 targeted cash to the poorest households (a proposal says the poorest 30%; unverified) when blanket fuel and power subsidies are phased out; local taxes that visibly fund ward services; lower recruitment fees for migrants; and visible fairness at the top, because people resent paying tax while large defaulters go untouched. Corruption is the problem named most often in polls (22%), ahead of drugs (15%).36

Educated youth

  • Wants: a job with security and standing, or a way out.
  • Why it matters: this group helped bring down a government in 2024, so its weight is high when it moves together.
  • What could re-aim it: more skilled private jobs (the trade and bank agenda); public recruitment on a fixed, transparent calendar; migration pipelines that teach a language and a skill and so pay back in higher remittances and returning experience; student unions for those who want to lead; and technical diplomas that earn as much on day one as a generalist degree.
  • Price: skilled migration is a loss of talent in the short run.

Women workers

  • Wants: safe, respectable work within reach, and control over earnings.
  • Where they stand: the group most aligned with growth and with the least power.
  • What raises their weight: factories and service jobs outside Dhaka, safe transport, hostels, childcare, transfers paid to women, harassment tribunals. The evidence on payoffs is the strongest on the site: the World Bank has estimated full convergence in participation would add 1.8 points of annual potential growth (an upper bound).37
  • Price: friction with conservative opinion, smaller when the case is made as family income than when it is made as rights.

Qawmi ulama, Hefazat and mosque-based leaders

Religious institutions in Bangladesh have responded to incentives before, and their leaders have interests like any other group. Treating them as a bloc defined by fanaticism is neither accurate nor useful; their landscape is divided (Hefazat, Jamaat and other groups have competed), and engaging pragmatic factions on economic welfare is more productive than a public culture war.1

  • Wants: autonomy from the state; respect; students and donations; livelihoods for graduates; a moral say on family and gender questions.
  • What the institutions do: Qawmi madrasas run on donations and zakat, charge no tuition and often house orphans and children from poor families.38 For a poor rural family they are a free school with meals and dignity. Enrolment estimates range from about 1.4 million to about 5 million; nobody knows (see the disputed-figures page).
  • How the graduates live: mostly, it is believed, as imams, muezzins and madrasa teachers (no national data). One survey of 300 volunteer imams found an average income near Tk 10,000 a month, with 64% on oral contracts.39 If it is representative, a large, respected, poorly paid group that depends on local mosque committees is exactly the kind of group that defends its standing hard.
  • Evidence that incentives work: when the state offered teacher salaries and stipends for girls, many traditional madrasas registered, added secular subjects and admitted girls; perhaps a third of today’s state-recognised Alia madrasas are such converts.40 A previous government kept Hefazat quiet through concessions such as recognising the Qawmi degree.41 This evidence dates from the 1980s-2000s, and there is no recent trend data on religiosity.
  • What could give them a stake in growth: voluntary add-ons (maths, English, digital and vocational certificates) funded per student and independently accredited (the boards are one eligible route); skilled-migration routes that use Arabic as an asset; a transparent welfare fund for imams linked to community roles they already play; a seat at the table on matters in their field; and a shared campaign against corruption, bribery, food adulteration and loan default, all of which religious teaching already condemns (a proposal; unverified).
  • Equal rules, protected individuals: violence, mob action and attacks on shrines or minorities are prosecuted whoever commits them; peaceful disagreement is protected; grants have neutral eligibility; and basic protection never depends on community endorsement or on living inside a particular zone. Individuals within any community keep the right to decline a programme and to dissent.
  • Price: state money brings state entanglement with religion, and may strengthen Islamist parties. Deferring family-law reform leaves the cost with women. Graduates’ actual earnings are still unmeasured. Proposals differ on whether liberal economic or tourism zones should come early or later; all agree that private life should be free from vigilante and arbitrary state interference, and that basic protections apply nationwide and not only inside zones.

Teachers and universities

  • Wants: a living income; posts.
  • Why today’s behaviour pays: private tutoring earns more than teaching well in class, and loyalty earns posts. Pay is low (an assistant teacher in the government-aided system earned Tk 12,500 in 2023, and over 100,000 such teachers were paid only from tuition fees), so “unqualified” is partly the result of a labour market that pays too little to attract good graduates.42
  • What could re-aim them: a career ladder that ties pay to certification; public school-level results; independent search committees for vice-chancellors; industry-linked technical colleges. One proposal would double base teacher pay in phases, conditional on competency exams and an end to compulsory private coaching of one’s own students.
  • Price: resistance from teacher associations and the coaching industry; spending would have to rise from about 1.4-2% of GDP toward 3.5-4% as revenue allows.

Migrants and diaspora

  • Wants: cheaper and safer migration, security for families, respect, a say.
  • Where they stand: they send $35.6bn a year and have almost no organised voice.
  • What raises their weight: lower recruitment costs, skills certification (nursing, care, technical trades in Japan, Korea and Europe rather than low-wage construction), and investment channels built for them.
  • Price: a short-run loss of talent; and Gulf demand can fall (it has since the Iran war).

Outside partners

  • Wants: repayment, stability, standards.
  • Where they stand: the IMF, World Bank, EU, buyers and Japan can supply deadlines and conditions that a domestic coalition cannot yet impose. The plan should not let outsiders be the author of reform; a package that the government owns is more likely to be carried out than one it can blame on outsiders.43 External relationships also shape the path: India, China, the United States and EU, and the Gulf each have interests in stability and in particular assets, and development happens inside a geopolitical frame.11

What the compact protects and asks, group by group

The ground rules (ground rules; reasoning in the compact across differences) apply to every group alike. One line for each; the reform consequences are in strategy and deals.

Group The compact protects The compact asks
Prime minister and inner circle Fair process and safety if they lose power; a government still able to act Reasons for coercive decisions; no capture of referees; lawful, transparent revision of policy by successors
MPs and local party leaders A fair chance to compete again Disclosure, recusal and open tenders
Party youth and student cadres Lawful student association; fair access to halls and paid work No coercion or collections; individual evidence, not affiliation, decides liability
Opposition parties Participation, oversight roles and peaceful protest Peaceful methods; the same rules they ask of others
Awami League base Civic rights of supporters; fair status review; fair individual proceedings Respect for law; individuals answer for individual conduct
Civil service Merit-based postings, security and fair pay Reasons, deadlines, disclosure and accountability to the same law
Police Operating budgets and protection from political misuse Evidence-based action; remedies for official abuse
Military A predictable budget and a defined role Civilian authority; audit of budgets and commercial interests
Judges and lawyers Independence and secure careers Timely, reasoned decisions; published clearance data
Large defaulters and protected conglomerates Fair hearing and civil settlement on standard terms Disclosure; consequences for wilful non-payment; no immunity from investigation
Exporters Predictable rules and published transition calendars Labour standards and a fair wage process
Small firms, traders, transport owners Protection from extortion whether or not registered Lawful, published fees
Workers, farmers and the public A say, a fair share of gains and relief for transition costs Respect for lawful process while protesting peacefully
Educated youth Fair recruitment and paid placements Open, merit-based competition
Women workers Safe access to work and learning that no gatekeeper can veto Equal rules for all
Religious leaders and communities Conscience, worship, peaceful advocacy and the freedom to teach and persuade No coercion or exclusion; equal rules; neutral grant eligibility; respect for each member’s right to decline and dissent
Teachers and universities Pay tied to certification; autonomy Merit hiring; published results
Migrants and diaspora Verified recruitment costs and complaint channels Lawful channels
Outside partners Predictable rules for lenders and buyers Conditions that are public and applied equally
Victims of past wrongs Preserved evidence, protection and fair, timely remedies Individual proceedings under due process, never reprisals against communities

From today’s payoff to a better one

Each row reads left to right: what the group wants today, how it is paid today, the lever, and the better payoff. These are the same rows as in the profiles, collected on one diagram.

From today's payoff to a better one
Today
Better
Lever
Party leaders
Control judges, police and watchdogs to stay safe
Safe to lose; win on delivery
Caretaker rules, term limit, opposition-led oversight
MPs and local party leaders
Contracts, licences, loan rescheduling to recover election costs
Legal funds; credit for local delivery
Audited public funding; conflict-of-interest rules
Youth and student cadres
Territory, tolls, hall seats, tenders
A wage, a skill, an elected student role
Apprenticeships, hall lists, digital fees, fast courts
Officials and police
Buy the post, recover it in bribes
Higher legal pay; promotion on delivery
Pay tied to asset disclosure; postings by rule; audits
Judges and lawyers
Fees rise with each adjournment
Fees for settling and for fast stages
Stage-based fees, mediation, commercial courts
Large borrowers
Borrow, default, reschedule
Settle, or lose the firm; invest for export finance
Bankruptcy law, one settlement window, then no credit or contracts
Small firms and traders
Pay the collector, stay informal
Formal, protected, able to borrow and grow
Simple phone tax and credit record; protection from extortion for all
Exporters and workers
Buyers positioning ahead of graduation and the end of preferences; wage disputes end on the street
Preferred access; predictable yearly wage rule
2029 compact: tariffs, ports, gas for standards and wage process
Educated youth
Queue for a government post, party link or visa
Skilled private work, or migrate and return
Export jobs, hiring calendar, skilled migration routes
Teachers
Income from private coaching; posts by loyalty
Professional pay tied to certification
Career ladder; school-level results; search committees
Ulama and madrasas
Donations, low pay, influence by mobilising
Better-paid graduates; a stake in growth
Voluntary funded add-ons, imam welfare fund, firm red lines
Migrants
High recruitment fees; low-skill, low-wage jobs
Cheaper, safer, higher-wage migration
Direct employer portals, skills certification, investment channels
Today's route is what each group does now because it pays. The better route is what the lever would make pay.

Who blocks and who could carry each reform

Separately from group profiles, reforms have their own coalitions. This table, from a review of feasibility in 2026-27, shows where blockers and carriers sit. Reforms are never win-win for everyone; a plan has to be direct about who loses and where each coalition comes from.11

Reform Likely blockers Potential coalition Feasibility in 2026-27 (judgement)
Bank asset-quality reviews, resolution, central bank autonomy Politically connected defaulters and bank owners; state-bank management IMF, exporters needing credit, depositors, Finance Ministry seeking revenue Moderate
Tax administration and NBR split NBR officials (who struck in 2025); protected industries Government (to fund the pay scale), IMF, formal firms facing informal competition Moderate-low
Tariff rationalisation, EU trade deal, FDI opening Import-substituting industry; customs rent-seekers BGMEA and exporters, EU, Japan, consumers Moderate (deadline-driven)
Capacity-payment renegotiation, fuel pricing Private power producers (68 of 135 plants are private), LNG importers Finance Ministry, industry, energy analysts Moderate-low
Statutory police, judicial and anti-corruption independence Home Ministry, parts of the bureaucracy, some MPs Jamaat, NCP, students, TIB, bar associations, donors; the governing party’s own insurance interest Low-moderate
Foundational learning Teacher unions (on pay grade), coaching industry World Bank ($300m programme), parents, NGOs High
Women’s work and anti-child-marriage Some religious groups (on legal reform); some employers Garment exporters, buyers, women’s groups, the governing party’s women voters Moderate if framed as jobs, low as family-law reform
Campus depoliticisation Student wings of governing and opposition parties Students (93% in one online poll), faculty Low

The reform coalition also needs a broad social base: the post-2024 student generation, the professional middle class squeezed by both bureaucracy and partisanship, women’s organisations and the women’s economic network around garments, reformist voices within the Muslim community and religious minorities, export-oriented businesses and the diaspora.11

A caution on “humans only respond to incentives”

A common instruction is not to make idealistic assumptions because people always follow incentives. This page mostly agrees and is written that way, with one amendment. In July 2024 students faced bullets with no material reward in view. Status, fairness, dignity and faith move people too, and they are cheaper for the state to offer than money. Several levers above use them: public credit for MPs who deliver, standing for student leaders who win elections, respect for clerics, and the plain satisfaction of seeing large defaulters treated like everyone else. A plan built only on cash payoffs would cost more and work less well.

What is uncertain here

  • All scores are judgement (see above). The two scorings differ most on the ruling party, the opposition, the public and religious institutions.
  • There are no reliable numbers for how many cadres there are or what they earn, for party finance, or for madrasa graduates’ earnings and employment. These unknowns set the cost and the expected result of the youth and madrasa deals.
  • Results from Indonesia, Andhra Pradesh, Pakistan and Georgia were achieved in different political settings. They show that officials respond to monitoring and payment design; they do not show the size of the effect in Bangladesh, which is why the plan proposes pilots (see strategy and deals).

  1. Unverified: no primary source located for several figures (court confession, military welfare-trust size, recruitment fees). Group scores are our judgement. ↩↩↩

  2. Tay and Diener (2011), cited from memory in the analysis; not re-checked. ↩

  3. Jyoti Rahman, The Politics of Synthesis 2.0, Counterpoint, 1 June 2026. ↩↩↩↩

  4. TBS on ASK, violence in the first half of 2026. ↩

  5. Al Jazeera, Is Bangladesh killing reforms introduced after student-led protests?, 22 April 2026. ↩

  6. Ginsburg, Economic Analysis and the Design of Constitutional Courts, 2002. ↩

  7. Prothom Alo, poll on electricity and the Family Card. ↩

  8. Financial Express, TIB: polls largely fair and acceptable (59.9% businesspeople); IANS, business leaders make up 70% of the cabinet; Dhaka Tribune, TIB: 79.46% of MPs are crorepatis. ↩

  9. CPD, Six Months of the Government, 24 August 2026. ↩

  10. Suykens, “A hundred per cent good man cannot do politics”, Modern Asian Studies. ↩

  11. Student-politics poll: The Diplomat, Bangladesh’s student politics (93% in an online poll). Electoral-design, coalition and international points are our assessment. ↩↩↩

  12. TBS, Jamaat’s shadow budget for FY27; Daily Star, policy dialogue. ↩

  13. Dhaka Tribune on IRI, favourability. ↩

  14. Prothom Alo, constitutional amendment committee. ↩

  15. IGC, Woodrudd et al. 2015 working paper. ↩

  16. Counterpoint, Zia Hassan profile. ↩

  17. Daily Sun, TIB household survey. ↩↩

  18. Financial Express, salary rise up to 142%; Dhaka Tribune, ninth pay scale. ↩

  19. TIB, press release. ↩

  20. Olken, Monitoring Corruption, NBER. ↩

  21. J-PAL, Smartcards for smarter payments. ↩

  22. NBER, working paper w20627. ↩

  23. Prothom Alo, Police Reform Commission. ↩

  24. Daily Sun, police services survey. ↩

  25. Daily Star, police station funding. ↩

  26. Dhaka Tribune on IRI, institutions. ↩↩

  27. Just Security, Georgia national police modernization. ↩

  28. LSE, research online record. ↩

  29. Daily Observer, peacekeepers. ↩

  30. ORF, The Bangladesh Army: documenting its corporate interests. ↩

  31. Human Rights Watch, Unkept Promise, 2010. ↩

  32. Jyoti Rahman, The Economy Needs Credibility, Not Optimism, 16 September 2026; the 57-month wage-lag figure: Unverified: no primary source located. ↩

  33. Jyoti Rahman, interview, Prothom Alo, 2 August 2026. ↩↩

  34. Bonik Barta, transport associations. ↩

  35. Dhaka Tribune, labour unrest hits 529 factories. ↩

  36. Dhaka Tribune on IRI, problems named. ↩

  37. Daily Star, female labour force could drive economy further: WB. ↩

  38. Daily Star, madrasa workforce. ↩

  39. Study of imams’ income, EconPapers. ↩

  40. Asadullah and Chaudhury, Holy alliances, Education Economics, 2009; Bano, summary paper. ↩

  41. TBS, Qawmi degree recognition. ↩

  42. TBS, MPO teachers. ↩

  43. Jyoti Rahman, A State of Economic Emergency, Counterpoint, 20 October 2026. ↩

  44. Representation of the People Order, Article 90F (permits registered parties to receive donations and grants within limits); TIB, Institutional governance of political parties, executive summary, 2026. ↩

Go deeper

Where this comes from

Synthesised from 8 research reports, checked against the sources cited above and reviewed before publication. How we work · All sources.