Applications of Demand Estimation to the Child Care Market

Abstract

This dissertation presents three studies that apply discrete choice demand estimation techniques to policy-focused questions regarding the Minnesota child care market. The first study analyzes provider quality ratings provided through Minnesota's Parent Aware program. The demand system is estimated on provider level data using the method of Berry et al. (1995). Welfare estimates are computed for the value of the ratings to consumers in different locations. Variation in local value of the ratings is driven by density of providers. The value of the ratings is high in most areas with a high concentration of low-income consumers. The second study uses administrative micro-data on subsidized consumers from Minnesota's Child Care Assistance Program, focusing on the role of distance in child care choice. A nested logit model is estimated that simultaneously models the choice of individual provider and the choice between center-based and family providers. The results provide new evidence on the importance of proximity and the geographic scope of child care markets. We present two applications. First, we present a construct for measuring the geographic scope for a policy intervention and analyze how it varies in urban, suburban, and rural areas. Second, we show that differences in what types of providers are available nearby explains most of the difference between White, Black, and Hispanic households in the use of center-based care. The third study uses a simplified version of the same nested logit choice model as a lens to analyze trends in use of family providers. We simulate counterfactuals that measure the importance of different factors, showing that decline in the availability of family providers and changes in the location and demographic composition of the CCAP population are the most important factors explaining the decline in the rate of use of family providers by CCAP consumers.