Essays on Socio-Economic Inequality and Development in India and China
: A Historical & Institutional Perspective

Student thesis: Doc typesDoctor of Philosophy


This PhD dissertation is a collection of three independent chapters linked through the long-run dynamics of institutions, culture, human capital and economic inequalities. Chapter 1 is joint work with Li Yang. This chapter compares the evolution of human capital accumulation and the development of modern education institutions in China and India between 1900 and 2018. It builds a novel education time series on various educational outcome measures like enrollment, graduates, teachers and expenditure. The long period allows us to understand the changing educational policies and outcomes with changing political institutions. The main differences between these two countries in expanding educational systems are related to - China’s bottom-up vs India’s top-down, China’s diversifying approach, and China’s prioritizing quantity vs India’s prioritizing quality. The first difference is that China focused on expanding primary-level mass education (1900-70) first, later shifted to develop middle-level and, in recent decades, is expanding its tertiary-level education. Due to colonial influence combined with a more heterogeneous society, India ignored primary-level mass education for a long time. It resulted in higher average years of education in China since the 1960s-born cohort. The paper also finds China has developed a robust vocational education system and produces a higher share of engineering graduates from non-vocational tertiary education. To provide the degree of differences, at the tertiary level, half of the students go through the vocational track compared to a meagre 10% in India. The share of engineers every year the Chinese education system produces has remained close to 30-35% (of the total non-vocational graduates) in the entire 20th century compared to less than 10% in India. India produces a large percentage of humanities and social science graduates. We conjecture that China’s type of human capital was more apt for developing the manufacturing sector than India lacked. The second part of the project systematically studies the education-wage inequality linkage. The development of education in India resulted in not only higher educational inequality but also a higher rate of return resulting in a more substantial impact on wage inequality. Chapter 2 first measures the wealth inequality in India from 1961-2018 and later explores the determinants of the inequality dynamics. I combine data from wealth surveys (NSS-AIDIS) and millionaire lists to produce wealth inequality series. I find a substantial rise in wealth concentration in the post-liberalization period when the growth rate was very high. It is in line with recent research using income. The most conservative estimate shows that the top 10% wealth share rose from 45% in 1981 to 61% in 2018, while the top 1% share rose from 13% to 33%. Two important stylized facts - land is the most valuable asset of household wealth (60% of the total wealth value) in all strata, and higher caste groups own much more wealth than their population share. The historical distribution of land in the hands of the upper caste (during colonial times or even before) still seems to determine wealth inequality. The paper shows this relationship rigorously by computing land inequality at the village level for 374k villages (universe of villages from ten large states of India). I find the share of the Scheduled Caste (or Dalits, the lowest caste group) population is strongly correlated to the village-level land inequality, controlling for institutional, geographical and demographic factors. It highlights the strong imprint of history on the current-day wealth inequality. I explore another plausible mechanism behind rising inequality - increasing homogamy between wealthy individuals. I prepare a novel dataset on married couples and estimate the correlation between husband and wife’s education, income and occupation (unfortunately, this data does not contain wealth, but these are a good proxy for wealth). I find an increasing education assortativity in society. Though I do not causally establish the relationship, the patterns across different caste groups (which could be treated independently of each other given 95% of the marriages are within castes) perfectly match the level of wealth inequality within each caste group. Chapter 3 is joint work with Sutanuka Roy. The setup is the largest state of India, called Uttar Pradesh, with a 200 million population, and the state has a history of communal conflicts (Hindu-Muslim riots). The paper uses fixed effects approach to causally establish that exposure to these conflicts’ environment during early childhood (age 0-6 years) has a long-term persistent impact till adulthood. It studies judges’ decision-making in pre-trial detention cases where decisions are often made on limited information. The paper shows that judges are more likely to deny bail (i.e. send a defendant to jail) if they were exposed to conflict environments in their early childhood. It shows that the effects are not driven by selection into a judicial occupation or judges’ ability differences but rather by changes in preferences for strong law and order. Three pieces of evidence help in arriving to this conclusion: absence of inter-group hostility behavior among judges, stronger effect if the exposure happens during 3-6 years of age, which is a crucial time for preference development and the result is driven by judges experiencing effective state intervention, i.e. high state-imposed lockdowns and low casualties in the conflict.
Date of Award26 Sept 2022
Original languageEnglish
Awarding Institution
  • University of Namur
SupervisorGuilhem Cassan (Supervisor), Thomas Piketty (Co-Supervisor), Lorenzo Trimarchi (Jury), Clément Imbert (Jury) & Nishith Prakash (Jury)

Cite this