Even though the problem of grade repetition is high on the policy agenda of virtually every developing country, extremely little is known about either the causes or the educational effects of repetition. The general concern about grade repetition derives from the budgetary and social implications of having large numbers of repeaters taking up scarce positions in schools. This concern notwithstanding, fundamental disagreements about the nature of the problem have clearly inhibited the development of sensible policies. This research, relying on unique panel data for students in northeast Brazil, considers how the schooling system and individual students interact in determining enrollment patterns in primary schools. This investigation into underlying student and school behavior lays the groundwork for analysis of alternative policies. Discussions of repetition are often subsumed in larger discussions of "wastage"-the combination of repetition and dropout rates.' This combination is unfortunate in many ways because the two phenomena are quite different in both their causes and their consequences. Nevertheless, they tend to be intertwined both in the data and estimation and in the policy debate. Indeed, the combination of high dropout and repetition rates has been identified as one of the main failures of the Brazilian education system, in part because the rates appear above those in other countries. A brief review of the available data and more general issues sets the stage for the analysis of grade repetition. Reflecting the low level of information about the issues, disagreement arises immediately about the magnitude and form of the wastage problem. Commonly available and cited data reveal important discrepancies even among estimates of the level of aggregate wastage and its