The Sustainable Development Goals have targets for both expansion of schooling to achieve universal completion of primary and secondary schooling, and for learning, to reach universal basic proficiency in reading and mathematics. Yet today not a single developing country has an empirical estimate of how much reaching their schooling goal would contribute to reaching their learning goals. We build a simple, formal, parameterized model of the learning process and calibrate the parameters to replicate observed learning outcomes in developing countries. We then use this model to simulate the progress on global learning goals that would result from achieving universal completion of grade 10. Our simulations suggest that in a “typical” low income country increasing completion of grade 10 from its current level of roughly 30 percent to 100 percent increases cohort learning by only 9 points on a PISA-like scale (mean of 500, standard deviation 100), an effect size of less than one tenth of one standard deviation. More strikingly, in our simulations this massive expansion of enrollment has zero impact on the proportion of youth reaching the SDG targets for learning. The reason for this perhaps counter-intuitive finding is that our simulation model allows for children who fall behind the curriculum to stop learning while in school and assumes that those learning the least dropout first. Therefore, expanding enrollment simply shifts most children from not learning while out of school to not learning while in school. In contrast to the weak impact of expanding enrollment at existing levels of learning, even modest changes to the learning process such as reorienting the curriculum to children’s learning levels can have massive effects. With an improved learning process, achieving universal grade 10 completion has ten times the impact on average scores than under the existing learning process.