Most Haitians are more familiar with lottery wagers than any other financial transaction and lack access to savings products. Using a lab-in-field experiment in Portau-Prince, we asked 306 participants to allocate a fixed budget across consumption, a real-world lotto product, a real-time traditional savings product, and a lotto-linked savings (LLS) product that returned 60-100% of the principal with lotto credit in lieu of interest. We find that the introduction of an LLS product increased total upfront savings by 22%, an increase roughly equivalent to that induced by raising the traditional interest rate from 5 to 20%. This savings response increased to 30% when the LLS product returned more principal and less lotto credit. An LLS product with a lower expected return was equally effective at increasing savings as one that had the same expected return as the savings product, suggesting it is the presence, not the extent, of the lotto component that drives the savings response. The LLS-induced increase in savings was financed by large reductions in lotto spending, slightly smaller reductions in traditional savings, and yet smaller reductions in consumption, which increased subjects’ overall expected returns in the experiment. Individuals who allocated more to
the lottery pre-LLS and who overweight small probabilities increased savings most in
response to the LLS.