Essays on human capital, expectations and behaviors

TitleEssays on human capital, expectations and behaviors
Publication TypeThesis
Year of Publication2012
AuthorsJia, M
AdvisorDickens, WT
UniversityNortheastern University
CityBoston, MA
Accession Number1238265098
KeywordsDemographics, Expectations, Health Conditions and Status, Net Worth and Assets, Risk Taking

Chapter 1: While previous theoretical analysis suggests that racial differences in death rates might play an important role in explaining the black-white education gap in the U.S., there is little empirical research to test this implication. This paper estimates the extent to which differences in expected mortality risks prior to entering college can explain differences in adult educational attainment in the 2000s, using data from the 1997 National Longitudinal Survey of Youth (NLSY). This study finds that the impact of mortality is not as important as suggested by prior research. Specifically, of the total black-white education gap (roughly 1.12 schooling years), only about 0.05 years or less can be attributed to differences in mortality expectations. As this study confirms, the role of self-reported mortality expectations in explaining black-white education gap is small, and the impacts of death expectations from actual death rates on education are statistically insignificant for reference groups. Chapter 2: The second chapter examines whether individuals are likely to alter personal health-related behaviors once they increase their subjective longevity expectations. To determine if there is a relationship between health behaviors and longevity beliefs, I test one of implications of the Cutler-Glaeser (2009) smoking decision model, which suggests that nonsmokers whose expected survival probabilities have increased are unlikely to start smoking. This study uses data from the Health and Retirement Study (HRS), which is conducted every two years, from 1992 to 2010 (Waves 1-10). Specifically, the HRS data show that a certain share (2.13%) of nonsmokers at Wave t-1 whose subjective expected longevity beliefs increased across two waves did start smoking at Wave t. This small percentage is close to the fraction of new smokers who have steady or decreased survival beliefs (1.99% and 2.19%, respectively). This finding also holds true for other behaviors including heavy drinking, obesity, and physical inactivity. Thus, the findings I present based on the HRS data contrasts with the Cutler-Glaeser model. Chapter 3: Using scores from the Armed Forces Qualification Test (AFQT), Herrnstein and Murray (1994) reported that intelligence can be a powerful predictor of a range of outcomes related to social behaviors (e.g., incarceration, marriage, out-of-wedlock birth, low birth weight and poverty). In contrast, a recent study found that measured intelligence using the same AFQT scores plays a considerably smaller role on an important socioeconomic indicator, namely, hourly wages as measured from 2000 to 2010. My third paper attempts to replicate the Herrnstein and Murray study using a different data set, the 1997 National Longitudinal Survey of Youth to look into several behaviorally-related social outcomes. The main finding is that, in general, the role of AFQT scores in predicting social behaviors has not substantially changed over the last 20 years. I provide a few possible explanations for this finding.

Endnote Keywords


Endnote ID


Citation Key6362