Using administrative data from France, we document that within the same detailed occupation, industry, and commuting zone, workers who work from home earn on average 12% higher hourly wages than fully on-site workers. Approximately half of this wage premium is accounted for by observable worker characteristics (such as education, gender, and age) and firm characteristics (such as size and productivity). The remaining 6% wage premium largely reflects selection: workers who work from home after the COVID-19 pandemic already earned higher wages before the pandemic.