We study stable school formation among four students that differ in ability and income. In the presence of ability complementarities and school costs to be shared, we identify the conditions under which a stable allocation is efficient, inefficient, nonexistent, and tell who become peers.

Recent studies of US elite exam schools have yielded the startling conclusion that such schools improve neither educational achievement nor longer-term educational outcomes. Is the same true for exam schools elsewhere? The system in Turkey is ideal for investigating this question. There,...

Recent studies of US elite exam schools have yielded the startling conclusion that such schools improve neither educational achievement nor longer-term educational outcomes. Is the same true for exam schools elsewhere? The system in Turkey is ideal for investigating this question. There,...