From its inception, the HRS (Health and Retirement Study), together with the companion AHEAD study (Asset and Health Dynamics of the Oldest Old), has been designed to provide data for the community of scientific and policy researchers who study the economics and demography of aging. This paper describes the evolution of the HRS from its origins as a longitudinal study of persons born in 1931-1941 who were 51-61 years of age in 1992 to its current design as a `steady state' sample of older Americans which, in cross-section, is representative of the entire U.S. population over age 50 and follows respondents longitudinally until they die. I discuss how the conceptual framework that underlies the original HRS design has lead the community of researchers to embrace the steady state design described above. In particular, I summarize a set of distinct, but interacting theoretical ideas whose origins lie in several disciplines that have guided us in determining the content and design of the survey. I then describe how the data collected in the HRS will allow these theories to be tested and briefly describe findings from recent research using the HRS, including some reported in this volume, with a special focus on understanding the very strong relationship between health and wealth that has been found in the HRS.