This article divides the era of performance-based budgeting (PBB) into three periods: rise and development in the 1990s, continuation and peak in the 2000s and level-off afterwards. The article examines the performance of PBB on four aspects: development of performance measures, its use in budgeting and management, its utility across the business cycle, and its usefulness for budget players. the author devises an analytical framework that builds on publicness as the distinctive feature of government, draws on the principal-agent model to emphasize issues of accountability, and brings in incentive mechanisms theory to show difficulty in performance measurement. The author surveys 11 representative states and find: that developing performance measures is a gradual process, that PBB functions better in management than in budgeting, that PBB is used widely in boom years but not in bust years, and that PBB is used selectively by legislators but more widely by the executives.