Mandatory pensions are a worldwide phenomenon. However, with fixed contribution rates, monthly benefits, and retirement ages, pension systems are not consistent with three long-run trends: declining mortality, declining fertility, and earlier retirement. Many systems need reform. This book gives an extensive nontechnical explanation of the economics of pension design. The theoretical arguments have three elements: * Pension systems have multiple objectives--consumption smoothing, insurance, poverty relief, and redistribution. Good policy needs to bear them all in mind. * Good analysis should be framed in a second-best context-- simple economic models are a bad guide to policy design in a world with imperfect information and decision-making, incomplete markets and taxation. * Any choice of pension system has risk-sharing and distributional consequences, which the book recognizes explicitly. Barr and Diamond's analysis includes labor markets, capital markets, risk sharing, and gender and family, with comparison of PAYG and funded systems, recognizing that the suitable level of funding differs by country. Alongside the economic principles of good design, policy must also take account of a country's capacity to implement the system. Thus the theoretical analysis is complemented by discussion of implementation, and of experiences, both good and bad, in many countries, with particular attention to Chile and China.