Homelessness occurs when individual resources are not enough for the demands of a given environment. Exploring homelessness as a process of resource loss on a continuum of poverty leads to research and explanations concerning how people transition from being housed to being homeless. This study assessed the influence of age, gender, and race along with a set of eleven resource competitiveness variables on the risk of youth becoming homeless. Resource competitiveness variables were: parental income, personal income, possession of a driver's license (DL), live-in partner, parenthood, education and training, annual weeks-employed, substance abuse, and incarceration history. The data came from the Bureau of Labor Statistics' National Longitudinal Survey of Youth 1997 (NLSY97). This sample was restricted to those that were homeless or unstably housed and were between the ages of 18 and 24 (n = 141). Each case was then matched by age, gender, and race to two individuals randomly selected from the remaining NLSY97 sample (n = 282). This resulted in an overall N of 423. A growth model was used to analyze the data longitudinally. Partnership, education and training, DL, annual weeks-employed, and personal income were significantly associated with experiences of homelessness and unstable housing. All were negatively related, except for age, which was positively related to incidents of homelessness and unstable housing. Comparisons across the homeless, unstably housed, and control samples showed incremental changes in nearly all the covariates in this study, in relation to changes in housing status, supporting the importance of studying homelessness as a point on a continuum of resource loss versus a discrete state of being.

This research is a systematic comparison of youth migration experiences between two birth cohorts, using the first ten rounds of two national longitudinal surveys of youth, NLSY79 and NLSY97. Results show both changes and continuities in youth migration patterns across cohorts for ages16-25. Specifically, youth today have a delayed but stronger migration momentum than the late baby boom generation, the dividing point being at age 22. Women are more likely to migrate than men in the recent cohort, but not in the older cohort. Whites migrate considerably more than blacks and Hispanics consistently across cohorts. The likely life events in youth's transition to adulthood are important indicators of youth's migration propensity for both cohorts. Particularly, graduating with a bachelor's degree is the most powerful predictor of youth's migration propensity. Other life events such as getting married; becoming separated, divorced, or widowed; dropping out of college; and losing a job are also significantly associated with youth migration. In general, the effects of these life events on youth's migration propensity are weakened across cohorts, but the importance of having a college degree on migration propensity has been increasing.

It has been common practice in the economics literature to utilize data on observed outcomes and negate what individuals believe or expect will happen in the future. Using responses to a unique set of questions in the 1997 National Longitudinal Survey of Youth (NLSY97) I show that the literature could benefit in several ways by incorporating such data. The leading essay documents a positive association between a student's subjective probabilistic belief that they will complete a four-year college degree and whether or not they attend and complete college. The results indicate the following. First, although overconfident, individuals as young as fifteen are willing and able to answer subjective probabilistic questions concerning education in a cohesive fashion. Second, these expectations are heterogeneous across race, gender, previous academic success, and parent education, and are influential in predicting whether or not they attend and ultimately complete a degree once these characteristics are controlled for. While the magnitude of the effect diminishes when including the standard economic controls, expectations remain significant and play a larger role as the student ages. Parent expectations are also positive and statistically associated with their child's future college success when the student is young but the significance diminishes as the student ages and gathers information related to the costs and benefits of a college degree. These findings indicate that students possess some form of private information that is not being completely captured by the standard variables used by econometricians to predict college attendance and completion. The second essay uses the NLSY97 to examine how students form and update their college completion expectations as they age out high school. I begin by estimating which factors are utilized by students when forming their expectations while in high school. I find that while these students are taking into account several of the relevant factors associated with college success, they also appear to be neglecting the impact that income and ability have on their likelihood of completing college or are over-relying on poor signals. I then test whether or not students update their expectations in a Bayesian fashion. A Bayesian model is developed. The three ways in which Bayesian students should respond to the acquisition of new information are discussed. Four sources of new information are identified and used in the testing. The testing reveals that students who report either a 0% or 100% chance of completing college do not appear to be Bayesian, but those who report within the 0% and 100% bounds do update in a Bayesian fashion. The third essay studies the accuracy and alignment of the individual's expectation that they will complete college. I utilize several unique aspects of the NLSY97 to create a measure of alignment based on the predicted probability that the respondent will eventually complete college and their expectation of doing so while either in high school or of college age. I use this measure to answer the following questions. First, are there any observable differences between those who are aligned and misaligned? Next, do respondents become more aligned as they age and progress out of high school? Last, are those who are more aligned at an early age more likely to reach their outcomes? I find that although the majority of students are overconfident in their belief there are considerable differences in alignment based on several observable characteristics and the availability of information. The alignment of student expectations differ based on parent education, ASVAB percentile, school enrollment, and race. Using two sub-samples of different aged respondents I show that as students age and acquire more information their expectation of completing college becomes more aligned with their estimated probability of completion. I confirm this by examining 700 students who are asked their expectations first in 1997 while in high school then again five years later when they either are in college or the workforce. I conclude by showing that those who are more aligned in either direction with what a model of college completion predicts the more likely they are to eventually reach that outcome. The final essay examines if the private information contained in the student's expectation that they will complete college is associated with future early career earnings. First I note that there are considerable differences in the frequency of reporting, yearly income, hours worked, and hourly wage for those who predict college success and are successful versus those who do not, as well as those who accurately predict that they will not complete college. I then include these expectations in a wage regression and the estimates suggest that when individuals report their college completion expectations between the ages of 15 and 17 they are not associated with future earnings. However, when asked between the ages of 17 and 22 the reported expectations are positively associated with future wages. There is considerable heterogeneity based on gender, whether they reported at one of the three primary heaping points, and the quantile of the wage distribution in which they were located.

Support vector machines (SVMs) are a powerful tool for classification problems. SVMs have only been developed in the last 20 years with the availability of cheap and abundant computing power. SVMs are a non-statistical approach and make no assumptions about the distribution of the data. Here support vector machines are applied to a classic data set from the machine learning literature and the out-of-sample misclassification rates are compared to other classification methods. Finally, an algorithm for using support vector machines to address the difficulty in imputing missing categorical data is proposed and its performance is demonstrated under three different scenarios using data from the 1997 National Labor Survey.

Past research has established a marital wage premium among men, and more recently, among women of the baby boom generation. It is unknown whether: 1) the marriage premium holds among more recent cohorts of men and women, 2) it differs by intensity of work hours among husbands and wives, and 3) cohabiters receive wage bonuses. Using fixed-effects models and data from the 1979-1989 waves of the National Longitudinal Survey of Youth (NLSY79) and the 1997-2010 waves of the National Longitudinal Survey of Youth (NLSY97), this paper compares cohort differences in the gendered marriage premium. While both women and men receive marriage premiums and these premiums are larger for more recent cohorts, men’s premiums are consistently higher and have doubled from the late baby boomers cohort (NLSY79) to the late Generation X (Gen X) cohort (NLSY97). While there was no wage premium for cohabitation among baby boom cohort women, I observe a premium among Gen X men and women. Household specialization matters: while among baby-boomers the marriage premium did not vary by household type, among the Gen X cohort men’s marriage premium is significantly larger among male breadwinner households, and surprisingly, I find marriage penalties for men in female-breadwinner households. Similarly, Gen X female breadwinners and female dual-earners receive the marriage premium while Gen X women in male-breadwinner households experience marriage penalty. In addition, the more highly educated receive larger marital bonuses.