Bangladesh has just put a 35–38% fuel shock through an economy where real wages were already negative, and is about to put a second one through before the year ends. What that does to prices will be in the CPI print. What it does to households — what they stop eating, whose schooling stops, who they blame, whether they think it was fair — will not be recorded anywhere unless somebody records it now.
This page is the honest version: what each amount buys, what it returns, what is already built, and what is still missing. If something here does not add up, that is worth telling me.
Fuel subsidy reform fails politically about 70% of the time — only around 30% of reforms survive twelve months, on Mahdavi, Ross and Simoni’s 2025 count. The literature explains that mostly through perceived fairness, but almost all of the household-level evidence comes from Nigeria, Iran, Indonesia and Ecuador. Bangladesh, with 170 million people and a live reform in progress, is close to absent from it. Its own 2022 shock produced protests, fare disputes and commentary — and almost no systematic household sociology.
The pay scale was gazetted 19 September; the hike took effect on the 21st. That splits the population by occupation two days apart: ~33 lakh public employees indexed to inflation, ~7.5 crore private and informal workers not. Relative deprivation is normally very hard to identify cleanly. Here it arrived dated and gazetted, and it is currently unclaimed.
The government has to choose in the next quarter between another hike, a duty cut, and targeted compensation through the Family Card rail. P5 answers the exact question on that table — whether people prefer universal cheap fuel or targeted cash, and what makes them accept either. Evidence delivered in December is usable. Evidence delivered in 2028 is history.
A completed phone interview costs $4–6. A discourse paper costs a few thousand dollars because the corpus is already built. This is not an expensive programme; it is a cheap one with a hard deadline, which is an unusual combination and the reason a small, fast funder can matter more here than a large, slow one.
Costs are built up from per-interview and per-day rates, not set as a target. Where a range is given, the low end is a smaller sample, not worse work.
Turns a built corpus into a publishable paper.
What comes out: One paper on how blame and fairness are framed across the 2022 and 2026 shocks, plus a public frame-share series that every other strand can cite.
The one purchase that expires.
What comes out: A genuine pre-shock baseline. After the next hike, this measurement cannot be made at any price — there is no “before” left. Every later wave is worth more because this one exists, and worth much less if it does not.
P3 — the channel nobody is watching.
What comes out: The acreage response to Tk 135+ diesel against a frozen procurement price — which is to say, an early read on 2027 rice prices, rice imports and the foreign-exchange pressure that follows.
Where the shock is actually absorbed.
What comes out: An account of how an illegal energy economy is actually governed, and a policy note on metering and legalisation — the lever that converts 500 MW–1 GW of theft into billed revenue without destroying the livelihood underneath it.
Baseline through the Boro harvest.
What comes out: Two to three papers — coping and sacrifice ordering, relative deprivation, the moral economy of fuel pricing — and a de-identified public dataset. Nothing comparable exists for Bangladesh’s 2022 shock.
A conjoint experiment, n≈2,000, varying fuel price, compensation size, targeting rule, delivery channel and who pays. Small if embedded in a P1 wave. It answers the question the government is actually deciding this quarter, and it adds Bangladesh to an international literature — Iran 2010, Indonesia, Nigeria 2023 — that currently explains why fuel subsidy reform fails without much South Asian evidence in it.
The second column is the one worth reading.
| Risk a funder would ask about | Where it stands | Honest residual |
|---|---|---|
| Is the analysis any good? | Published in full, with a complete audit log behind it: every search, every figure, its source, and every derivation. Derived numbers are marked est. and cross-checked against each other — the implied diesel throughput reconciles with the government’s own claimed annual saving. | Not peer-reviewed. Built rapidly, by one person, in the days around the hike. |
| Can the team actually execute? | P4 is evidence rather than assertion: a 1,996-article bilingual corpus was collected, a pipeline built, an exploratory analysis run and a validation harness written — unfunded, inside a week. | No track record of running field survey operations. That is precisely why a survey partner and an institutional host are the first asks. |
| Is the design fundable elsewhere later? | P1 is built to standard panel specifications: FIES food insecurity, sector stratification, an embedded randomised framing arm, panel fixed effects. It maps directly onto IGC, ESRC and WIDER expectations. | Most of those funders require a PhD-holding PI or an eligible institution, and their cycles are slower than October. |
| Will anything come out of it? | P4 can publish on its own. P1 W0 alone supports a descriptive brief. Each study is designed to stand alone rather than depend on the rest of the programme being funded. | The panel’s value compounds with waves; a single wave is a good brief, not a good paper. |
| Where would the money sit? | The intended route is an institutional host — BIGD, SANEM, CPD or a university department — holding and disbursing funds under its own financial controls, with the survey firm contracted directly. | No host is secured yet. Until one is, the realistic instruments are commissioned work or a partner’s internal funds, not a grant. |
| Ethics and safety? | Written into every design: Bangla consent, verbal consent recorded on phone waves, no names stored alongside illegal activity, encrypted diaries, no published charging-point locations, no partisanship stored with identifiers, respondents paid. | No IRB approval yet, because IRB comes with the host. Fieldwork does not start before it. |
Surveyed 21 September 2026. Every entry was checked against the programme’s own page; none of the calls had a confirmed open deadline in late 2026.
| Programme | Size | Eligibility | Fit |
|---|---|---|---|
| IGC Bangladesh is a partner country | £20–30k small up to £125k full | PI must hold or be pursuing a PhD; small grants run through the country team | Strong fit for P1 and P5, gated on the PI requirement |
| SANDEE (ICIMOD) | $20–30k | Must be employed at a university or research institute | Good fit for P3 — diesel versus solar irrigation is squarely environmental economics |
| Wenner-Gren Post-PhD | up to $25k | Independent scholars eligible; needs an anthropology PhD or equivalent appointment | The natural home for P2, and unusually open to non-institutional applicants |
| ESRC / UKRI via international co-lead | scheme-dependent | Needs a UK-based lead and a Bangladesh co-lead at an eligible organisation | Viable if a UK collaborator co-leads; slow relative to the window |
| FES Bangladesh | commissioned | Concept note; has commissioned CPD and RAPID on RMG wages | The most realistic route inside four weeks, alongside a partner’s internal funds |
| GDN awards | $10–30k | Researchers from low- and middle-income countries | 2026 theme fits poorly; worth watching the 2027 theme |
Happy to walk through the instrument, the cost build-up, the sampling, or the parts that are weakest. The designs, the corpus and the derivations are all public already — there is nothing to reveal in a call that is not already on this site.