Evaluating the effects of governmental regulations on South Korean private cram schools

Abstract

Using two nationally representative datasets, this paper examines how the imposition of a curfew on private cram schools affects the consumption of private tutoring services as well as the time use patterns of Korean high school students. To identify the impact of the curfew, this study capitalizes on inter-city/province variation in private cram school regulatory ordinances and estimates a difference-in-differences estimator between the years of 2009 and 2012. The findings indicate that the curfew was ineffective in reducing the level of household expenditure or the total number of hours spent on private tutoring. However, imposition of the curfew was found to be associated with increases in time spent on sleeping as well as in internet usage for non-academic purposes. Interestingly, changes in time use patterns were not found among students in the highest socio-economic status group indicating that the curfew policy may have certain unintended consequences. Policy implications are discussed.

Keywords::

• private tutoring
• regulation
• time use
• difference-in-difference
• South Korea

Acknowledgement

The work by Rosa Minhyo Cho was supported by the National Research Foundation of Korea Grant funded by the Korean Government [NRF-2013S1A3A2055042]. The work by Jaesung Choi was supported by Sungkyun Research Fund, Sungkyunkwan University, 2014.

Notes

^1. Information available at the Korean Statistical Information Service website (http://www.kosis.kr/) accessed on 2 January 2014.

^2. Although some school districts in rural areas and small cities do not implement the random assignment policy, currently the policy is applied to 70% of academic high school entrants, including those in Seoul (the capital) and the six major metropolitan areas.

^3. Korea is divided into nine provinces (Chungbuk, Chungnam, Gangwon, Gyeonggi, Gyeongbuk, Gyeongnam, Jeonbuk, Jeonnam, and Jeju), one special city (Seoul), and six metropolitan cities (Busan, Daegu, Daejeon, Gwangju, Incheon, and Ulsan) for purposes of local self-governance. In July 2012, an additional special autonomous city (Sejong) was added to the original division of 16 (and thus a total 17) to accommodate for the creation of a special administrative district from parts of Chungnam and Chungbuk provinces.

^4. This information is obtained from the Korean Educational Statistics Service web portal (http://kess.kedi.re.kr) sponsored and maintained by the Korean Educational Development Institute (KEDI). Information accessed on 20 March 2014.

^5. The regression model using the SPT sample is estimated including all 0 observations for respondents reporting no participation in any kind of private tutoring. Our results do not substantively change if we exclude them from the sample. Results are available upon request.

^6. The year fixed effects are especially important in the context of Korea since many policies are implemented by the central government uniformly across all cities and provinces. For example, the internet “Shutdown Law” which barred children ages 16 and under from playing online games between midnight and 6am was put into effect in November 2011 uniformly across all regions in Korea. This specific law has important implications toward our findings on the amount of time spent on the internet for non-academic purposes among high schools students during the sampling period.

^7. US dollar is equivalent to 1068 KRW (checked on 18 March 2014). This implies that 10,000 KRW roughly amounts 9.36 USD.

^8. Information available at the Korean Statistical Information Service website (http://www.kosis.kr/) accessed on 18 March 2014.

^9. For both the SPT and KYRBWS, we have missing values only for parental education. In the SPT, the proportion of observations with missing values is 7.7% for mother's education and 8.5% for father's education. In the KYRBWS, the proportion of observations with missing values is 2.2% for mother's education and 4.0% for father's education. When imputing the missing values for our independent variables, we included the dependent variables in our model specification.

^10. This roughly amounts to 2,464–2,482 USD (at 20 December 2014 official exchange rate).

^11. The 2011 sample of the KYRBWS was collected in September 2011, two months prior to the passage of the Shutdown Law. This implies that estimates of the amount of time spent using the internet in the KYRBWS should not be affected by the regulation until 2012. The KYRBWS collected reports of internet usage time in 2012 on the month of June.

^12. We also conducted sensitivity analyses on these estimates. When using Seoul or the remaining eight provinces as the comparison group, the relative contribution of going to bed earlier on the total minutes of sleep becomes larger and remains statistically significant at the 1% level (82–98%). Results are available upon request.