Last week Esther Duflo, Abhijit Bannerji and Michael Kremer were awarded 2019 Nobel Prize in Economic Sciences for their experimental approach in poverty alleviation. They have demonstrated that simple development interventions experimentally evaluated at small scale can offer promise of success when replicated in similar settings at a bigger scale. This approach indeed helps in resolving an important part of development puzzle – what works in development practice. The rest of the riddle is why it works. Knowing what works is helpful but knowing why it works makes it much more valuable for policy. Field experiments or randomized evaluations help in finding what works, while theory, local knowledge and even unorthodox sources of comprehending human experiences and aspirations work in unison to disentangle the rest of the development riddle. If knowing what works is a boon for development practitioners then a thorough grasp of both ‘what’ and ‘why’ constitutes a policy makers’ dream. Unfortunately, it is absence of the knowledge of both which characterizes development practice in poor countries. An anecdote shows how practically development interventions take place in developing countries. One fine morning, education department in a developing country receives invitation for participation in a meeting with planning bureaucrats and a delegation of a multilateral organization. During meeting, the delegates express an intent to give loan for a specific intervention focusing female education. They offer technical assistance to design a programme as part of deal. A few days later a follow up meeting takes place during which consensus is reached on moving forward. In next meeting design consultants hired by the financier appear on the scene. During next few weeks some in-house stakeholders’ consultations take place primarily between departmental civil servants and the consultants. A few more weeks and consultants present contours of a development programme in a meeting chaired by planning bureaucrats and things move toward formal signing of the agreement between the financier and the recipients of loan. No one goes to the real stakeholders, the intended target population of development intervention. Programme designers pay little attention to local culture, history or institutions. No attention is paid to local human. No experimental evaluation takes place to know the potential effectiveness of programme. Everything comes vertically from supply side. This is what brews development failures while debt burden gets inflated. The outcome of ignoring local realities was endless hurdles in community mobilization despite the best efforts of community mobilizers hired for the programme. Serious delays and deadlocks occurred due to conflicts and non-cooperation among the beneficiaries This is like groping in the dark! I can recall how the programme for improvement of water courses was implemented in Pakistan. The programme design relied on the idea of community participation. The programme beneficiaries were required to share twenty percent of the cost of improvement of the water course irrigating their farms to minimize water losses occurring due to seepage. The users near the head of water course were not as much interested in the programme as were the tail-end users because usually they bore the brunt of losses due to seepage and even pilferage. Local conditions were logically not conducive to the idea of equal monetary participation as head-end users were usually indifferent to requirement of brick-lining the water course. While the stakes of tail-enders were high. The outcome of ignoring local realities was endless hurdles in community mobilization despite the best efforts of community mobilizers hired for the programme. Serious delays and deadlocks occurred due to conflicts and non-cooperation among the beneficiaries. Ultimately, the highstakes of the government rescued the programme from failure. The local district administrators were engaged to compel all water users to pay their share of cost contribution. In this way, a programme designed on voluntarism turned into a coercive power based model. Mere replication of a successful model in one particular setting never implies equally successful transportation to somewhere else. Probably a knowledge of local conditions and testing of intervention at experimental scale might have improved the programme design before scaling up to national level. For developing countries, there is little room for wasting scarce resources on such development programmes whose design lacks higher probability of success. When viewed in this perspective, the contribution of Nobel winning professors in popularizing random evaluations of potential solutions is laudable. Their experimental method helps in knowing what works part of the problem. But still knowing why something has worked is important to generalize the solutions for better scalability, generalization and application elsewhere. If a programme designer can get insight of local knowledge of social and political institutions, it is more likely to find better solution early in a trial. Knowing why certain interventions work need both prior theoretical knowledge as well as an understanding of recipient population. This is the intellectual territory where even hitherto unused unconventional sources can be quite helpful. Professor David Lewis and his colleagues at the London School of Economics and the Brooks World Poverty Institute have even explored the potential of using fiction in guiding development policy. In a paper titled ‘The Fiction of Development: Literary Representation as a Source of Authoritative Knowledge’ they argue that works of fiction can offer useful insights about development processes and sometimes can even do a better job in helping complex development issues. Human experiences and their retellings constitute vital background information essential for tailoring development prescriptions to be experimented. Randomized trials, development theory and unconventional sources of knowing the deep aspirations of people when used together can work as force multipliers. Without getting bogged down in the endless debate of defining superior and inferior forms of knowledge, the development planners and policy makers should harness all of them. In poor countries stakes are really high. It sometimes takes billions of rupees and number of years to see the outcomes of some development programmes. Whatever contributes to success need not be ignored. The programme or project design should be an outcome of careful work as pangs of development failures are often unbearable! If experiments tell what works, local knowledge, theory and unconventional sources fill the gaps in knowing why it works. The writer is a public policy analyst.