Can Loss Aversion Increase Savings? Experimental Evidence from Tanzania

We experimentally test the role of loss-aversion in savings decisions, framing interest as a lump-sum ``upfront bonus'' that is lost if a savings target is not reached. We partner with a mobile network operator and an NGO in rural Tanzania to study the savings behavior of 1,524 women receiving mobile-money-based cash transfers from the NGO. Participants were randomly assigned to a savings incentive with ``upfront'' interest vs. traditional interest vs. an information-only placebo. Using administrative data on mobile wallet balances, we find that the upfront treatment reduced endline wallet balances by 2 USD (net of the bonus) relative to placebo. The traditional interest treatment, despite offering an above-market interest rate, had no effect on wallet balances. Both savings incentives increased savings for the richest 10% of participants. Overall, our findings suggest that the savings decisions of the very poor may be unresponsive to savings rates or framing