I estimate the long-run earnings impacts of for-profit and non-profit charter high schools in Sweden, which as of 2023 enroll nearly half of all high school students in urban areas. Unlike in many other settings, there are no schools operating outside of the public system: all schools rely on equal public funding, cannot charge top-up fees, and are subject to the same regulation. Using a combination of quasi-experimental and value-added methods, I find that charter schools lower earnings by 2% on average — comparable to the returns to half a year of schooling in similar settings. My results suggest that for-profits drive these losses by hiring less-educated, lower-paid teachers, non-profits by emphasizing arts and humanities programs. In a discrete choice framework using rank-ordered school applications, I show that students’ preferences are weakly related to schools’ earnings impacts and instead center on location and program offerings, which explain most of the charter market share.