Governance under pressure has hidden mechanics
IPD studies the evidence, institutions, and expert networks that make public reasoning survive chaos.
Good governance under chaos
Most policy research is a collection of separate studies on separate topics. We are organised around a single question — and every piece of work is a lens on it.
A number of pressures usually treated as Bangladesh’s particular afflictions may in fact be the early form of a condition wealthier, more stable societies are drifting toward. We hold this as a hypothesis to be tested — not a truth to be asserted.
If that is right, a society already navigating this combination is a leading indicator — and Bangladesh shifts from a struggling case to be aided into a source of knowledge the world may need.
“A fifteen-year clean panel is the cheapest durable moat an institute can own. Narrative can be matched; the data cannot.”
Six divisions. One shared standard.
Each division keeps its own domain depth, but all answer the same question — that is what makes us one institute rather than six. Active at founding, the rest opening as leads join.
Methods & Measurement
It owns the Longitudinal Governance Panel and carries our rigor ambition explicitly: causal identification, pre-registration, replication, and open data and code. Method and data are part of the brand, not a back office.
A route to real influence — for the known and the unknown alike.
Bangladesh’s genuine experts are systematically denied institutional identity. We provide what the current system does not: fellowship titles, bylines, a professional home, and the experience of being treated as the intellectual one actually is.
Student member
University research societies — beginning with BUET and Dhaka University — as the first stage.
Fellow
Methods training, a byline, and a first cohort of mentored research.
Project researcher
Original, publishable work scoped to a fundable project.
Division contributor
Shaping a division's agenda and mentoring those who follow.
Division lead
Own a field, set its research agenda, lead its fellows.
We judge arguments by evidence anyone can examine — not by who makes them.
We are non-partisan as a matter of method: we will criticise or commend any government, party or institution against one standard — the quality of the public reasons offered for what it does. This is not a posture. It is the source of our independence, and the only thing that will let the collective outlast any administration, any funder, and any one of us.
How the method works ›
We default to open
Open data, open code, open methods, community-driven research — because evidence anyone can examine requires materials anyone can access.
Governance designed against capture
Contribution-weighted membership, mandatory founder term limits, a sunset on founder control, and a standing independent board seat. We make governance dynamic — not economic ownership.
A strict, one-way wall
Commercial entities may fund the collective; the collective's reputation, access and research are never instruments of any commercial interest. The wall is absolute.
The reward is ownership.
This is early. The resources are modest, the work is real, and the credibility we build does not yet exist — you would be lending yours to help create it. Founding members shape the mission, the standards and the culture in a way no later joiner can.
Division leads
Own a field, set its agenda, lead its fellows.
Board & advisors
Independent oversight that can constrain the founders.
Founding fellows
Do the work that sets the standard, and mentor others.
Institution-building
The decisive work of turning an initiative into an organisation.