Previous research has established that people bid more for houses in high-performing school districts. But what is it about school outcomes that drive house prices: the parents, the peers, or the school inputs? We study the extent that house values are affected by each of the components of an education production function. Based on 123 school districts and 26,000 house transactions, we find the primary component of school outcomes that is capitalized into house prices is the amount of parental inputs. In the explanation of variations in house prices, variations in parental characteristics are eight times more important than similar variations in school inputs, and twelve times more important than variations in peer groups. This result suggests that land values in a particular community will be increased more by attention to zoning laws that influence the mix of renters to homeowners and the type of households entering a community compared to investing in additional public school inputs.