Does lasting behavior change require knowledge change? Evidence from savings interventions for young adults

Is financial knowledge change necessary for lasting behavior change? Or, akin to Friedman’s billiard player, can behavior persist “as if” such knowledge is held? We randomize 240 Ugandan young-adult clubs to financial education, savings account access, both, or neither. Each education arm, but not the account-only arm, increases members’ financial knowledge and trust at one-year. At five-years, knowledge effects essentially disappear and trust effects weaken. However, savings, wealth and income increase for each treatment at both one and five years, suggesting multiple viable paths to statistically indistinguishable average outcomes and that textbook knowledge change is unnecessary for lasting impacts.