The development sector has never had access to more evidence. Decades of impact evaluations, systematic reviews, evidence gap maps, and data systems have produced a substantial body of knowledge (see 3ie’s Development Evidence Portal) about what works, for whom, how, and under what conditions. And yet this evidence often fails to reach the decisions that matter most. Why?
From 3ie’s work on evidence-informed policymaking, delivered through the FCDO Research Commissioning Centre, we have developed a framework that explains how evidence use depends on the alignment of three things: demand for evidence, supply of evidence, and what might be called the “missing middle” — the translation, synthesis, brokerage, and interpretation that connects evidence to actual decisions. Understanding this missing middle and how to invest in it is the focus of our ongoing research program. In this blog, I want to focus on three related observations.
1. Not all data serves the same purpose
One of the most important distinctions that gets lost in development conversations is this: different decisions require different kinds of evidence. We often speak about “using data” as if all data serve the same purpose. They do not.
A scorecard indicator can be extremely valuable for accountability and performance monitoring. It can tell us whether we are moving. But it cannot tell us whether we chose the right destination, or whether there is a better, more efficient route available. Those are fundamentally different questions, requiring fundamentally different kinds of evidence.
Decision-makers will require different evidence at different points in the policy cycle such as:
- To identify priorities: they need diagnostic evidence about problems and constraints.
- To choose interventions: they need evidence on effectiveness — impact evaluations and, preferably, systematic reviews of bodies of evidence.
- To allocate resources: they need information on costs and cost-effectiveness.
- To improve implementation: they need operational and delivery data.
At 3ie we have worked with the Millennium Challenge Corporation (MCC) to develop the PIR methods menu to help development evaluators and practitioners choose the right evidence approach(es) for their program phase, context, and needs.
2. Evidence use is not fully internalized
The second observation is one that the development sector has been slow to fully internalize: evidence use does not happen automatically. It requires institutions to deliberately create the conditions for it.
At 3ie, we have been working with partners through the Global Evidence Commitment — including Foreign Commonwealth & Development Office (FCDO), Inter-American Development Bank (IDB), MCC, Development Bank of Latin America (CAF), Norwegian Agency for Development Cooperation (Norad), Kreditanstalt für Wiederaufbau (KfW Development Bank), German Institute for Development Evaluation (DEval) and Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ) — to better understand what distinguishes organizations that consistently use evidence from those that merely aspire to. This work produced the TRIPS framework, which identifies five levers available to institutions seeking to improve their culture of evidence use: Training, Resources, Incentives, Processes, and Signals from leadership.
One finding from our most recent assessment stands out: incentives remain the weakest part of the system. Leadership talks about the importance of evidence. Organizations invest in data systems, monitoring, evaluations, and knowledge platforms. But performance reviews, promotion criteria, and operational processes rarely incorporate rewarding staff for using evidence well. The signals are there. The incentives are not.
Coming to what good practice looks like. Some institutions have started embedding evidence requirements into project approval processes, ensuring that proposals must explain how they draw on existing evidence before resources are committed. Others are beginning to incorporate evidence use into competency frameworks and staff performance discussions, making it part of what good management looks like, rather than an optional extra. These may sound like small changes, but they create an organisational reflex that values learning, adaptation, and outcomes.
3. Outcomes must be reflected in the incentives of the wider system
The third and most structural observation concerns the incentives of the development ecosystem itself. I wrote about this in a paper a few years ago called ‘creating a market for outcomes’. Much of development finance still rewards activities, disbursements, and outputs. We talk a great deal about outcomes. But relatively little funding is actually contingent on achieving them.
Results-based financing has moved the field part of the way in this direction. But a broader question remains. If a development institution such as the World Bank positions itself as a knowledge partner — helping countries choose effective policies and investments — should it not also share some responsibility for achieving those outcomes? At present, countries typically bear almost all the risk. Loans are repaid regardless of whether the intended outcomes materialise.
What would happen if development partners and countries shared responsibility for outcomes more explicitly? The hypothesis is that we would see a much stronger outcomes reflex throughout the system, not only among implementing agencies, but also among those who design projects, commission evidence, advise governments, and allocate resources.
Towards a coherent agenda
Taken together, these three observations point to a coherent agenda. Matching evidence to the right decision point. Building the institutional infrastructure that makes evidence use a reflex rather than an afterthought or add-on. And aligning the incentives of the wider financing system with the outcomes that justify the investment in the first place.
