Peruvian first and second graders at the bottom half of the income distribution were asked to read a brief passage aloud and to answer simple comprehension questions. Only 25% of the first graders and 41% of the second graders were able to read one or more words in the text at the end of the school year; 75% of the first graders could not read at all. Children who could answer all three comprehension questions correctly read at 77 words per minute on average, whereas children who could answer only one or two questions read much more slowly, at 15 and 41 words per minute respectively. The relationship between reading speed and comprehension is consistent with memory research on reading automaticity. The results are also consistent with other observational studies that suggest widespread illiteracy among poorer students in various countries. Worldwide, actions can be taken so that all children (particularly in countries with phonetically spelled languages) can read fluently by the end of grade 2. A rapid reading test, administered quickly and inexpensively, produced comprehensible and actionable information on student performance. If sampling and instrumentation were refined, the methodology could produce baseline and monitoring data that are locally sustainable and internationally comparable.