This paper examines changes in Social Security that have been proposed as ways to reduce poverty among elderly women. For each option, the paper asks three questions: How much would the option reduce elderly poverty? How much would the option cost? How would the option affect the relationship between payroll taxes and benefits? In general, the proposed changes in Social Security present trade-offs: the broader the target group and the greater the benefit or income guarantee, the larger the expected impact of an option on elderly poverty. But these features would also raise program costs and weaken the link between payroll taxes and benefits. Some options would also provide significant assistance for people who are not poor, while providing little to those most in need.