博士論文審査報告書

全文

(1)早稲田大学大学院アジア太平洋研究科. 博士論文審査報告書論原題名 Original Title. 英訳 In Japanese. 文. 学校教育の投資効果に関する分析－タイにおける実証的検証－. 申. 名. Name 学籍番号 Student ID. 目. Analysis of Returns to Schooling: Empirical Evidence from Thailand. 姓 Last Name. 氏. 題. 請. 者名 First Name. Middle Name. Korwatanasakul. Upalat. 4011S 303 -3. ２０１７年１月. 0.

(2) 1.. Absract of the Dissertation. This dissertation estimates returns to schooling in Thailand during the period of rapid structural transformarion. It estimates the Mincer equation by exploiting the occation of the change in compulsory schooling law in 1978 as a natural experiment. The dissertation finds that the compulsory schooling law played a role in enhancing human capital investment in the eve of the rapid structural transformation in the 1980s, that the returns to schooling based on its IV estimation was around eight percent, while OLS somewhat overestimates (by 20%) such returns, and that returns were higher in urban areas, in services (than in agricultural) sectors and in relatively underdeveloped Northern regions. Those findings are in contrast with most of the recent studies exploiting similar institutional changes from developed countries, where OLS estimates tend to under-estimate returns to schooling with the implication that those school drop-outs (whose behavior is altered by compulsory schooling) tend to have higher returns than those already in school even before the law change. The findings of this study are consistent with the conventional notion of the ability bias, where the OLS estimates of the returns to education are likely to be over-estimated since those with higher ability are more likely to stay in school longer and also earn higher wages. The dissertation argues that such ability bias is more likely to arise in developing countries, rather than in developed countries, possibly because parents could be forced to keep only those (among many) of their children with higher ability in school, thereby reinforcing inequality among children within the household. 2.. Outline of the Dissertation and Summary of the Chapters. The chapter outline of the dissertaiton and the brief summary of the chapters are presented below. Chapter 1.. Introduction. Chapter 2.. Thai Economic and Educational Development. Chapter 3.. Literature Review. Chapter 4.. Theoretical Framework. Chapter 5.. Data and Methodology. Chapter 6.. Instrumental Variable (IV) Estimation using Thai Data. Chapter 7.. Further Analyis: Disaggregated Analysis of Returns to Schooling. Chapter 8.. Conclusion. Following Chapter 1 which summarizes the motivation and objectives of the dissertation, Chapter 2 provides an overview of Thai economic and educational development. The overall average real GDP growth and real GDP growth per capita were around 6.2 per cent and 4.2 per cent, respectively, during 1951-2003. The high GDP growth is generally considered as a result of structural transformation during the pre-boom period of 1951-1986. Thailand went through a structural transformation from an agriculture-based economy to newly industrialised economy. The structure of export and import patterns has also been transformed during the period. 1.

(3) In terms of educational development, there was a rapid expansion of the education system in the primary education, esperically during the 20 year period between 1960 and 1980. In contrast, the demand and supply for secondary and tertiary education grew at a much slower pace. An improvement of overall education quality could also be observed as the student-teacher ratios kept declining over time across different levels of education. Examining the rates of return to education in this period helps us better understand the process of Thai economic development as well as the interplay between the rates of return to schooling and economic development in Thailand during 1980s to 1990s. Chapter 3 provides a review of the relevant literture. The chapter starts by revisiting the Mincer equation. Second, the chapter reviews previous literature regarding methodological issues in estimating the rates of return to schooling. The third section surveys previous studies of the rates of return to schooling with a particular focus on studies utilising compulsory education law change as an instrumental variable. Fourth, previous studies on the sample selection are briefly discussed. While the estimted rate of returns to schooling is one of the most frequently investigated empirical relationships in empirical economics literature, there has been a long debate on the exact magnitude of such rates due to the possibility of endogeneity problems in estimating the Mincer equation. The chapter reviews relatively recent studies utilizing quasi-experimental approaches. The review is mainly focused on the studies using the changes in the compulsory schooling laws as natural experiments in developed countries, which the main empirical analysis of the thesis closely follows. Chapter 4 briefly presents a general theoretical framework of the analysis of returns to education. It mainly focuses on the Mincer’s (1974) human capital earnings function. The chapter also discusses the recent criticisms raised aginst the Mincer model by Heckman, et al. (2006). Despite its relative simplicity, however, the Mincer model has been found to fit the actual age-wages profiles quite well by many authors (Card, 1999). This dissertation also finds that the Mincer equation fits the Thai data reasonably well. Chapter 5 discusses the data, methodology, and identification strategy employed in this study. This study exploits an opportunity of quasi-experiment in Thailand which occurs from a change in compulsory education law to estimate the rates of return to schooling. The main estimation method is the IV estimation and the data used is the pooled cross-sectional Labour Force Survey (LFS) data from 1986 to 2012. Chapter 6 presents the main empirical results. The results include the estimates from the first stage least square regression, the reduced form, the OLS regression (years of schooling on log wage), and the IV regression. First, the first stage estimation results show that the compulsory education law change led to six additional years of schooling which roughly corresponds to the first year of the upper secondary school (grade 10). Second, the reduced form estimation results show that the change in compulsory education level had a large effect on the monthly wages, leading to approximately 36 per cent to 50 per cent increase. Third, the IV estimation results indicate that one additional year of schooling leads to approximately 8 per cent increase in monthly wages. Comparing the OLS estimates with those of IV, the OLS estimates are greater than those of the IV by around 3 per centage points. This indicates the upward bias (overestimation) in the OLS regression, which is consistent with the conventional argument of ability bias in returns to schooling. 2.

(4) The ability bias as found in this study can mainly be explained in terms of the inequality of income and educational opportunity during the early period of social and economic development where more-able child within poor households are selected for schooling due to resouece costraints. Even though there could be a tendency that individuals with higher ability are more likely to attain more years of schooling in rich countries, the ability bias seems to be more important in the case of Thailand and, possibly, in developing countries in general. Chapter 7 reports the results of estimating returns to schooling with disaggregation by different demographic and geographical characteristics and by the sector of employment. First, the analysis finds that the urban rate of returns exceeds that of rural areas due to availability of more sophisticated jobs that need skilled workers in the urban labour market and higher quality of schools in urban areas. Second, in terms of regional disaggregation, the higher returns to schooling in least developed regions, the Northeast and the North regions, are observed. This could be explained by a regional migration and a movement of demand for and supply of educated labour. The dominant effect of the interregional migration flows from the Northeast and the North regions to Bangkok, the Central, and South regions leads to an increase in the supply of educated labour and a reduction of returns to schooling in those labour-absorbing regions. In contrast, outmigration reduces the supply of educated labour in the Northeast and the North regions; therefore, returns to schooling increase in those regions. Thirdly, the rate of returns to schooling is higher in the service and the manufacturing sectors, while the lowest returns to schooling is found in the agricultural sector. An excess demand for high skilled labour in the early stage of economic development creates rents or education premium in the manufacturing and the service sector to a small group of more educated labour to enjoy. Finally, the differential returns to education along demographic dimensions of disaggregation, such as by gender and by cohort, are found to be of minor magnitudes. Chapter 8 summarizes the main findings from the study with additional discussions of policy implications, current limitations of the study and an adenda for future research. The policy implications include the potentially large effects that compulsory schooling law could have for enhancing human capital investment in the relatively early stage of economic development and potential importance of educational policies catered to different geographical areas and economic sectors as implied by the results of the disaggeregated analyses in this study. Some limitations of this study include the limitations of the data coverage by the Labor Force Survey which only covers wage employment and excludes self-employment, as well as the rather puzzling finding of the returns to education being higher in the relatively underdeveloped North and Northeast regions which is not fully explained. Addressing those issues is among the future research agenda. 3.. Discussions of Oral Defense Examation. The examination committee conducted the oral defense examination on December 19, 2016 after carefully reading the dissertation. The examination lasted approximately 90 minutes. The major comments received from the examiners included the following: . The summary chapter was too short and needed to be longer with a summary of findings, policy implications, limitations of the study and futher research agenda. . The discussions in the introductory chapters needed to be more tightly connected with the main analysis of the study, including more relevant descriptive statistics 3.

(5) . Definitions of informal sectors and the way rural and urban samples were drawn in the data were unclear.. . More interpretations needed to be included for the main findings, such as the large impact of compulsory schooling on the years of schooling and the geographical differences in the returns to schooling. . Some suggestions for additional interpretations of the main empirical results and for the major contribution of the study. 4.. Evaluation and the Result of Examination. The examiners find that this dissertation makes a number of important contributions to the research on the returns to education in developing countries. There are three major contributions. First, this study is one of the very few studies in developing countries that utilize an incidence of the change in the compulsory schooling law as natural experiment to estimate returns to schooling, and finds the results that are quite different from those based on similar approaches from developed countries. This study finds the importance of the conventional ability bias in estimating returns to schooling, in contrast with such studies in developed countries. Second, this study conducts a detailed analysis of returns to schooling with disaggregations by demographic, geographical and eonomic sectors of employment. It finds some important dimensions of heterogeneity in returns to schooling which, in turn, could potentially inform policies for other developing countries in similar stages of economic development as those of Thailand during the priod of analysis. Third, this dissertation makes a diligent use of multiple rounds of labor force survey data from Thailnad to construct a cohort panel dataset. The data yield a large enough smaple size to conduct a detailed disaggeregated analysis with reasonably sufficient statistical power. While such attempts to construct a cohort panel involves tedious manipulations of large scale data, the analytical results of this study exemplifies the fact that such attemps are well worth the effort. Considering the results of careful assessment of the submitted dissertation and making sure that the necessary corrections were made in the revised dissertation in response to the examiners’ comments, the Committee members came to a unanimous conclusion that KORWATANASAKUL, Upalat, the author of the submitted dissertation, be granted a Ph.D.. 4.

(6) 申請者名： Upalat Korwatanasakul 博士論文審査委員会主査. Chief Examiner：. 氏名 N a m e : 黒田一雄㊞(Signature) 所属 A f f i l i a t i o n : 早稲田大学大学院アジア太平洋研究科職位 Ti t l e : 教授学位 Degree: PhD 取得大学 C o n f e r r e d b y ： Co en e ll Un iv e r s ity 専門分野 Specialty: Educat ion 副査. Head Deputy Examiner:. 氏名 Name: 不破信彦㊞(Signature) 所属 Affiliation: 東京大学公共政策大学院職位 Ti t l e : 教授学位 D e g r e e : PhD 取得大学 C o n f e r r e d b y ： Un iv e r si ty o f Ca l ifo rn i a , B e rk el ey 専門分野 S p e c i a l t y : Ag ricu ltura l and Resou rc e Economics 副査. Deputy Examiner:. 氏名 Name: 鍋嶋郁㊞(Signature) 所属 A f f i l i a t i o n : 早稲田大学大学院アジア太平洋研究科職位 Ti t l e : 准教授学位 Degree: PhD 取得大学 C o n f e r r e d b y ： Un iv e r s ity o f Ca l ifo rn ia , Da v i s 専門分野 S p e c i a l t y : Economics 副査. Deputy Examiner:. 氏名 Name: 弦間正彦㊞(Signature) 所属 A f f i l i a t i o n : 早稲田大学大学院社会科学研究科職位 Ti t l e : 教授学位 Degree: PhD 取得大学 C o n f e r r e d b y ： Un iv e r s ity o f Min n e so ta 専門分野 S p e c i a l t y : Economics. 201７年 1 月. ２６日. 5.

(7)