The Impact of Free Childcare on Women's Labor Market Outcomes, Children's Development and Health

The intervention is universally free childcare services for 2-5 years old children across the country. Childcare services are offered throughout the scholar year from September to May, five days a week, for eight hours a day. The program has objectives to improve children's cognitive and communication skills and to increase women's labor force participation. The government is the largest provider of childcare services and fully subsidizes public kindergartens. The Ministry of Education, Culture, Sciences, and Sports (MECSS) is in charge of the overall design and implementation of the program.
Childcare services are highly demanded, with less than half of the eligible children are enrolled. To tackle the excess of demand, a lottery was introduced in all public kindergartens with oversubscription in 2016. This enables us to study the impact of universally free public childcare service on maternal employment, children's health and development, using a natural experiment framework.

2017-09-01

2019-05-31

Employment

We examine several outcomes related to employment (current employment, employment in last year, seasonal employment, hourly wage, work hours, and formal employment). Current employment is a binary variable, equals 1 if a household member was employed in the last week for at least one hour or has a permanent job, or was employed in a private business. Employment in last year is a binary variable, equals 1 if household member satisfies the current employment condition plus if he/she employed for any job in the last year. Seasonal employment is a binary variable, equals 1 if household members worked 26 and fewer weeks in the last year. Hourly wage is the natural logarithm of the last month’s hourly wage when the survey is being conducted and zero for unemployed. Work hours are regular weekly work hours of respondents with a permanent job and zero for unemployed. We will also measure formal employment, which in Mongolia is defined by the existence of a written labor contract.

Children's health
Children's development

We look at several outcomes related to children's health (number of cold/flu incidences, infectious disease, medication, and number of hospital visits). We will measure child development with total difficulties score (SDQ). The SDQ is a brief behavioral screening and psychological assessment tool for 2-17-year-olds. The questionnaire exists in several versions to meet the needs of researchers, clinicians, and educators and downloadable from the following link. https://www.sdqinfo.com

This study will evaluate the impact of free public childcare services on women’s labor market outcomes in Ulaanbaatar, Mongolia. Comparable control and treatment groups necessary to estimate the effect of free public childcare services were created using the public lottery run at childcare centers in August 2017.
We will use registration data of childcare centers provided by the Metropolitan Education Department of Ulaanbaatar. For each childcare center, we will collect information on the total number of available slots, the total number of applicants and information on which district the childcare center is located, and whether the childcare center should run a public lottery.

The baseline survey was administered between late August and October 2017. Also, we will use data from two households follow up surveys annually for two subsequent years.

Applicants who want to enroll their children in public childcare centers should complete an online registration process within a month between the end of July and the end of August. Then the public lottery is organized simultaneously at the centers with oversubscriptions, open to the public under the monitoring of the board consists of representatives of parents, teachers, and officials from the local administrative units-the public lottery run by a computer.

Randomization occurred at the level of the individual applicant to the free public childcare centers.

No

N/A

We have planned to include 3,500 applicants for the public childcare program.

The sample size of the baseline survey was set at 3,500 households with 2-year-old children of which we would assign 2,300 to the treatment group and 1,200 to the control group.

Power calculations to determine sample size considered female employment as primary outcome. A sample size of 3,500 would allow us to detect an increase of 7.4 percentage points in employment at the power of 92% for a significance level of 5%. We expect a 'no show' of 15% and a contamination of 13%.