This study aimed to complement a Randomized Controlled Trial (RCT) conducted by the Institute for Fiscal Studies from 2015 to 2017 exploring the impact of sanitation microloans in rural Maharashtra (Latur and Nanded districts), combined with information provision and awareness raising. One of the main findings of the RCT was that adding the awareness raising to offering the new sanitation microloan product led to lower sanitation uptake than providing sanitation microloans only, with one in five loan takers who received the awareness constructing toilets compared with one in two who did not receive the loan (Attanasio et al. 2018). The qualitative study undertaken by the French Institute of Pondicherry set out to solve this puzzling finding. Given the time lag between the implementation of interventions and our survey, reconstructing the processes of subscription to the sanitation loan, their conversion into toilets and the use of the toilets proved to be very challenging. In addition to the timing problem, there were two other issues: we had difficulty interviewing the staff of the microcredit organization and the NGO in charge of awareness-raising, even though the survey protocol had considered them to be key informants; the pandemic and lockdown forced us to concentrate data collection in the last three months of the research contract. Ultimately, our answer to the question posed in the terms of reference is necessarily incomplete. Our observations mainly (1) provide insight into the structural barriers to toilet uptake, then give some reasons that (2) explain the diversity of behaviours toward sanitation loans, and their conversion to toilets. Our understanding of the structural factors explaining the reluctance to build toilets, however, allows us to (3) make some general conclusions about the limitations of interventions based solely on credit provision and awareness raising.Information provision and awareness campaigns are popular tools in international development, aiming to raise knowledge on particular topics, such as health, in order to trigger behaviour change, and to generate demand for related goods and services; with the overall aim of improving human capital outcomes. Despite their popularity, potential unintended consequences of such communication activities have to date been little studied. The proposed research is designed to contribute to filling this gap by gaining a deeper understanding of the role of information provision in hindering, or fostering the effectiveness of microcredit for human capital investments. Microcredit – small, usually collateral-free loans given to households in low-income countries – has been much studied and their ability in relaxing poor households’ credit constraints to allow them to invest in entrepreneurial activities and cope with adverse events acknowledged. More recently, microcredit has also been shown to be effective in fostering human capital investments particularly preventive health investments such as bednets, water connections and toilets, supporting the emerging trends of private enterprises increasingly taking a role in solving social and environmental problems; and in growing numbers of investors seeking to make financial investments that produce significant social or environmental benefits. Microcredit programs often ensure high loan repayment through joint liability – groups of borrowers are liable for each other’s loans, and thus monitor each other. However, this feature could influence the take-up and use of microloans for human capital investments, since financial returns to such investments take time to be realised. Acknowledging that poor awareness of the existence of social programs could hinder take-up of these programs, awareness creation campaigns have become an essential component of the roll-out of new policy initiatives. However, such campaigns typically deliver multiple messages, which could inadvertently lead to recipients taking away a message that is contrary to that intended by the program designers, implying that information campaigns could backfire and hinder the adoption of human capital investments. We are in the unique position to research such unintended consequence and their mechanisms by building on an interesting puzzle that emerged from previously conducted research. In particular, we found that providing information alongside microcredit for human capital investment (sanitation), led to a similar increase in loan uptake on average as the provision of sanitation microloans only, but it resulted in fewer clients using the loan for household toilets, contrary to our original hypothesis. We intend to explore three channels that could explain this puzzling finding. First, the delivery of the information campaign – through a different organisation from that providing the sanitation loans – might have how peer monitoring of sanitation loan use occurred, resulting in fewer toilets being built. Second, the information campaign might have increased awareness about the existence of the sanitation loans (and their favourable conditions), attracting households who wanted to borrow for a non-sanitation purpose. Third, the information campaign could have unintentionally discouraged households from using sanitation loans for sanitation, by making them believe, for instance that their own sanitation investments would be ineffective unless accompanied by similar investments from their neighbours. We propose to answer these questions using a mixed-methods approach, combining a theoretical model, with quantitative data collected within the RCT and qualitative research. The key advantage of this combined approach is that it allows us to triangulate the research findings and hence to establish and explain patterns, trends, and mechanisms, in other words dig into the policy relevant questions of ‘what’ and ‘why’.

Approach: (1) Sampling: The first stage of sampling considered the diversity of locations according to the RCT protocol and take up rates and possibly local factors of diversity: - Gram Panchayats (GP) from the RCT control arm, treatment-A arm, treatment B arm - GP were selected according to the lowest, the average, and the highest sanitation uptake in each study arm. A second stage of sampling has been at the level of households (see below) (2) We have compared the perspectives of borrowers, field staff and other key resources persons: - In-depth individual interviews with the MFI/NGO administration in Bangalore - In-depth individual interviews with sanitation loan takers (25 from the villages of Alur, Walag, Betakbiloli, Belur, Andori, Hadtolti, Ramtirth, Gadga, Pimplegaon) - Interviews with the MFI lending group leaders (12) - In-depth interviews with other MFI clients, including those who chose not to take the sanitation loan (12) - MFI/NGO field staff individual interviews, in person (5) and by phone (3), informal discussion (5) - Focus group discussion (FDG) with borrowers (5) the MFI/NGO field staff in Maharashtra (2), masons trained by the NGO (1) - Individual interviews with maistries who are building the toilets – trained by the NGO (2), not trained by the NGO (5) - ASHA (Accredited Social Health Activist) workers (2), Angawadi (Creche) workers (2), ICRP (Inter-community resource person coordinating women’s SHGs) (2), Panchayat members (8), Gram Sevaks (4), Police Patil (conflict resolution officer) (1), School headmaster (1), Caste leaders (1), World Bank Sanitation Consultant (1)