[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT AUTHOR'S REQUEST.] Essay 1 analyzes the married couples' retirement decision using the PSID data. I employ the proportional hazard model to examine the factors that influence the retirement decision of husband and wife, and focus on examining the correlation of husband and wife's retirement status. This essay finds that an individual is more likely to retire if his or her spouse has retired. The retirement hazard is higher if an individual is in worse health. The worse health status also affects the spouse's retirement hazard, but the spouse effect is asymmetry. With the wife in worse health, the husband's retirement hazard decreases. With the husband in worse health, the wife's retirement hazard increases. I also find that the greater the social security income or pension, the higher the retirement hazard. But for the spouse effect, the husband's social security income or pension has impact on the retirement schedule of his wife, while I find no significant impact of wife's retirement benefit on husband's retirement timing. Essay 2 explores the transitions of health status using PSID data from 1984 to 2011 with the ordered logit model and the Cox proportional hazards model. The result shows that the impact of current health status on future health status is relatively large. A worse current health status would lead to a smaller probability for health deterioration, but it is less likely to be in a good health status in the future. There is strong health persistence. Social economics factors' impact on latent health status is also significant, although the magnitude is relatively small. Higher income level and education level would decrease the likelihood of health deterioration, and individuals with high income and high education would be more likely to be in better health status. When comparing different occupations, white-collar job is less associated with health deterioration, and this type of worker is more likely to be in better health status. Essay 3 applies the competing risks model to estimate the movement of corporate credit ratings using WRDS COMPUSTAT data. The credit rating variable is the Standard and Poor's long-term domestic issuer credit rating. The explanatory variables contain measures of leverage, liquidity, current profitability and future profitability. I estimate the impacts of these financial ratios on the upward and downward of credit rating. In addition, I estimate samples before and after the 2008 subprime crisis to study the influence of financial crisis on the credit rating. The result shows that firms with a higher liquidity are more likely to be upgraded and less likely to be downgraded. The impact of liquidity is weaker after the crisis. I find that when the current level of profitability increases, the firm is more likely to be upgraded than to be downgraded. The effect of current profitability is larger after the crisis. Firms with higher leverage ratio are more likely to be upgraded and less likely to be downgraded. And the effect of leverage is similar before and after the crisis.
On-Body Sensor Positions Hierarchical Classification
