Endogenous Income Insecurity and the Welfare States




Against the backdrop of fiscal austerity and pressure to cut social spending in Western Europe, the world has witnessed the development of welfare states in Asia in last twenty years. However, while all new democracies in Asia pursue universal health cares, universal pension programs were developed at a much slower speed. This project addresses the following questions: Why do we observe the emergence of welfare states in Asia? Moreover, why do Asian countries unanimously pursue their development in a way that only favors universal health cares?

Conventional wisdom has long suggested that the peculiarity of Asian welfare regimes is the “family-centered” welfare system featuring minimal state roles. The implication is that the demographic shift toward a nuclear family structure triggers the demand for bigger states’ roles in social protection.    

My research refutes this family-centered account and proposes another mechanism through which social policy demand emerges in Asia: economic vulnerability. Drawing on a growing literature from behavioral economics, I argue that economic vulnerability (measured as the probability of falling under the poverty line) affects individual preference formation on social policy far more strongly than actual rates of poverty.  More specifically, high levels of economic vulnerability make individuals desire more certain, immediate benefits rather than uncertain future benefits. Such short-term and certainty-biased behavioral tendency, in turn, makes vulnerable households prefer health care to old-age protection due to the short time-horizon associated with health care benefits. In Asia, where the informality of the labor market makes a large share of the population vulnerable to poverty, this large vulnerable population, in turn, this helps account for the cross-policy variation we observe in today’s Asian welfare states regimes.   


My dissertation tests these claims in Indonesia through a triangular research design. Indonesia has been extending its social security system to the larger population. In such reform, the health insurance program achieves universal coverage for all citizens in 2014, whereas the pension program only extends coverage to formal sector workers.

The first component of the project utilizes the longitudinal Indonesia Family Life Surveys (IFLS) to demonstrates that non-wage workers (informal sector workers) experience higher economic vulnerability and, hence, are more certainty and present-biased in their daily behaviors. The project leverages the longitudinal feature of the survey and traces the same households before and after the Asian Financial Crisis. This research analyzed the data with the Hierarchical Linear Model and showed that households engaging in informal economic activities became more fragile financially after the crisis regarding the probability of falling under the poverty line. The increasing economic vulnerability, in turn, leads to households’ short-term oriented behaviors.

The second step is to build the connection between economic vulnerability and individual preferences for various social policies. Because no social policy opinion data exist in Indonesia, I designed and conducted an original public opinion survey to fill this gap. After two summers of fieldwork, my survey was fielded in Indonesia with a nationally representative sample in June 2015. One significant advantage of the survey is that it is designed in a way to force respondents to reveal true policy preferences. With the utilization of the Logit Model, the paper showed two results: First, individuals with higher vulnerability show higher supports for health care. Second, the level of economic vulnerability cuts along the line of wage and non-wage workers, making an insider-outsider distinction in Indonesia.          

The last element in the research design tests the generalizability of the linkage between high economic vulnerability and preference biases toward health care. To that end, I compile an original dataset in which I match cross-national data from the International Social Survey Programme (ISSP) with hand-coded data on the time lapse between universal health care programs and pension programs from different countries. I pool the data at the national level and adopt the Event History Model to fit the data structure. The expectation is to show that a country passes its universal health care program much faster (than the pension program) when the majority of its citizens are subject to high economic vulnerability.  


In the developing world, one key challenge facing policy makers is to extend social protection to all citizens. To design a sustainable policy, it is of importance to meet citizens’ needs. This research project sheds new lights on when and why people in the developing countries demand more social protection. It contributes in disentangling the role of citizen preference on social policies and the social foundations of social policies in developing countries.