Estimates of Quality-adjusted Cost Function in the Okinawa's Child Care Industry: Evidence from Micro-level Data (in Japanese)

It is not well known that Okinawa suffers from a severe shortage of child care services, even among non-urban areas in Japan. This study is the first to take advantage of some unusually rich micro-level data to address to quantitative assessment of quality of child care and estimates of quality-adjusted costs function by types of management in Okinawa. Our empirical results on quality of child care are as follows. The "test score" approach uses several indexes to evaluate quality and to compare the total scores by different management. We observe that public centers have higher scores in the structural index for labor or capital inputs. However, this is not the case for the developmental psychological index, parents' convenience index, and other indexes. In sum, private licensed-centers earn the most, followed by public and non-licensed centers. On the other hand, the "quality of workers" approach adjusts for the non-random allocation of workers to measure the quality of child care closely related with the quality of employees. The results based on the treatment effect model demonstrate that the superiority of public centers disappears after adjusting for a sample selection bias. The estimates of the quality-adjusted cost function show that public centers are apparently more inefficient than non-licensed centers, while there is no statistically significant difference in efficiency between private-licensed and non-licensed centers. In Okinawa, the child care market is divided into two types; licensed centers to provide higher quality at higher costs, and non-licensed facilities to provide lower quality at lower costs, which creates higher efficiency. This nature of different child care allows consumers to use different types of child care facilities.