Is it possible to stimulate women’s income-generating activities by relaxing their financial and human capital constraints? Does involving husbands help or hinder the effort? We examine these questions using a three-arm randomized-controlled trial with 2000 women in Tunisia. Women in the two treatment arms were offered a large cash grant (worth USD768 in PPP terms) and a gender-sensitive financial training. In one of the treatment arms, women were additionally encouraged to bring their male partner to the training. Two years after the program, we show that the treatments stimulated women’s income-generating activities, but only when partners were not involved, and with no downstream effects on women’s agency. Independently of partners’ participation, impacts on household living standards were overwhelmingly positive, suggesting that the program was highly costeffective. Overall, our results highlight the difficulty of stimulating women’s agency in traditional societies, and suggest that involving men in women’s empowerment programs can backfire.