Recent panel data sets suggest that in some places perhaps half of the poor are not poor all the time, and also as much as 80 percent of "poverty severity" may be due to large fluctuations through time. Some of these dynamics are due to life-cycle events, but much of it represents "damaging fluctuations." These cause immediate damage, and may trigger responses leading to chronic poverty or intergenerational poverty. Moreover, just the possibility of damaging fluctuations, even if they do not occur, may generate more risk-averse behavior, which hampers growth and (because it is more likely to be common among poorer people) increases inequality. The paper discusses the concept of damaging fluctuations, contrasting it with the more common, and related, terms of "risk" and "shocks." Six categories of damaging fluctuations are identified: disease or injury, violence, natural disaster, harvest failure, terms-of-trade deterioration, and reduced access to income-earning work. The paper presents evidence of how the poor differ from others in their exposure, vulnerability, and aversion to damaging fluctuations, and how these lead to more poverty.