Master of Science / Department of Statistics / Weixing Song / In the classical Tobit regression model, the regression error term is often assumed to have a zero mean normal distribution with unknown variance, and the regression function is assumed to be linear. If the normality assumption is violated, then the commonly used maximum likelihood estimate becomes inconsistent. Moreover, the likelihood function will be very complicated if the regression function is nonlinear even the error density is normal, which makes the maximum likelihood estimation procedure hard to implement. In the full nonparametric setup when both the regression function and the distribution of the error term [epsilon] are unknown, some nonparametric estimators for the regression function has been proposed. Although the assumption of knowing the distribution is strict, it is a widely adopted assumption in Tobit regression literature, and is also confirmed by many empirical studies conducted in the econometric research. In fact, a majority of the relevant research assumes that [epsilon] possesses a normal distribution with mean 0 and unknown standard deviation. In this report, we will try to develop a semi-parametric estimation procedure for the regression function by assuming that the error term follows a distribution from a class of 0-mean symmetric location and scale family. A minimum distance estimation procedure for estimating the parameters in the regression function when it has a specified parametric form is also constructed. Compare with the existing semiparametric and nonparametric methods in the literature, our method would be more efficient in that more information, in particular the knowledge of the distribution of [epsilon], is used. Moreover, the computation is relative inexpensive. Given lots of application does assume that [epsilon] has normal or other known distribution, the current work no doubt provides some more practical tools for statistical inference in Tobit regression model.

Recently, variable selection in high-dimensional data has attracted much research interest. Classical stepwise subset selection methods are widely used in practice, but when the number of predictors is large these methods are difficult to implement. In these cases, modern regularization methods have become a popular choice as they perform variable selection and parameter estimation simultaneously. However, the estimation procedure becomes more difficult and challenging when the data suffer from outliers or when the assumption of normality is violated such as in the case of heavy-tailed errors. In these cases, quantile regression is the most appropriate method to use. In this thesis we combine these two classical approaches together to produce regularized quantile regression methods. Chapter 2 shows a comparative simulation study of regularized and robust regression methods when the response variable is continuous. In chapter 3, we develop a quantile regression model with a group lasso penalty for binary response data when the predictors have a grouped structure and when the data suffer from outliers. In chapter 4, we extend this method to the case of censored response variables. Numerical examples on simulated and real data are used to evaluate the performance of the proposed methods in comparisons with other existing methods.

[pt] A utilização de derivativos por empresas não financeiras vem se desenvolvendo com o intuito de protegê-las de riscos indesejados, especialmente aqueles advindos do mercado financeiro. Entretanto, o resultado da utilização do hedge pode, não necessariamente, agregar valor para as empresas que as utilizam.A literatura financeira internacional identifica ao menos cinco benefícios que podem ser obtidos através do uso do hedge, são eles: 1) a redução da possibilidade de falência ou estresse financeiro, 2) a redução valor esperado de impostos a serem pagos, 3) redução dos custos dos agentes da empresa (empregados, fornecedores, acionistas, clientes e outros), 4) a redução do custo de endividamento e 5) assegurar a continuidade dos investimentos através da redução dos riscos à geração operacional da empresa.Este trabalho procura verificou para um grupo singular de empresas brasileiras não financeiras, através de regressões não lineares que as empresas realizam o hedge especialmente para se protegerem de exposições cambiais e para garantir os investimentos futuros. Com este resultado, pode-se observar que o uso de derivativos para gerenciamento de risco por empresas brasileiras está em desenvolvimento. O presente estudo agrega à literatura correlata utilizando um método para relacionar os benefícios do hedge com sua posição nominal. / [en] The use of derivatives among non-financial companies has been developed with the purpose of protecting the firms from unworthy risks, especially those generated by the volatility in the financial markets. However, the hedge can bring consequences that differ from those expected by the higher management.The financial literature identifies at least five benefits from hedge that can be value maximizing: 1) reduction of costs associated with financial distress or bankruptcy, 2) reduction of the expected tax liabilities, 3) reduction of the costs of the stakeholders, 4) reduction of the debt costs or increase the debt capacity and 5) protect the future investment and growth opportunities through administration of the risks associated with the operational cash flow.This dissertation tries to verify throughout a series of non- linear regressions if a set of Brazilian non-financial companies reaches any of the benefits proposed by the financial literature. The conclusion shows that there is evidence that firms in the sample use interest rate hedge to protect the future investment opportunities.

Due to the social environment transition and the health care reform, hospital¡¦s transformation has made the high cost nurse resource of the medical organization issue. While facing salary pressure and nursing shortages, nursing leadership has taken an important role in stabilizing/establishing a positive work environment and maintaining good health care quality and job performance. In the past, personality and leadership have been proved to relate to job performance, but few studies show the relationship between these three variables. This study investigates which dimensions of the Five-Factor Model of personality of the head nurse (HN) are related to job performance. This study also analyzes the HN and investigates whether leadership style moderates personality-job performance relations. This study carried out a survey research and secondary date analysis in three regional hospitals of Kaohsiung-Pingtung area. The sample included 35 HN and 174 nurses who worked with their HN for over 6 months. t-test was used to examine the difference of personalities and leadership style in different demographic variables. In addition, the Tobit regression model explained significant portions of variance in these criterions. Results support the hypothesis that openness and extraversion are positively related to job performance. Results also support the hypothesis that consideration is appeared to moderate relationships between openness and job performance. This study show that personality influences job performance. Moreover, it shows that the leadership style could be the moderator between personality and job performance. This study suggested that personality might be a crucial factor in selection and recruiting of head nurses. In addition, providing training in leadership will facilitate the job performance. This study suggests that future studies should increase the sample size in terms of decision making units as well as random selection from different hospital levels.

To the best of my knowledge, there has not been any academic study about capital structure of Turkish REITs so far. This study attempts to fulfill this gap in the literature by analyzing the capital structure choices of Turkish REITs which are listed in Istanbul Stock Exchange (ISE) over the period of 1998 - 2007. The key contribution of this study is to understand whether the firm specific, institutional and country specific factors that affect the capital structures of all institutional firms including REITs in developed and developing countries are also applicable to the Turkish REITs sector. The data analysis demonstrates that Turkish REITs employ little long term debt in their capital structure and there exists strong short term debt dominance in the sector. Employing Tobit regression and panel data models, it is concluded that capital structure determinants that are significant in developed and developing countries are also significant in Turkish REITs&amp / #8217 / debt financing choices. However, we observe inconsistency in the sign and significance of some factors which give a way to understand the different institutional and country specific factors of Turkish real estate market and Turkish REITs.

The development of effective personnel assignment methodologies has been the focus of research to academicians and practitioners for many years. The common theory among researchers is that improvements to the effectiveness of personnel assignment decisions are directly associated with favorable outcomes to organizations. Today, companies continue to struggle to develop high quality products in a timely fashion. This elevates the necessity to further explore and improve the decision-making science of personnel assignments. The central goal of this research is to develop a novel framework for human resource assignments in skill-based environments. An extensive literature review resulted in the identification of the following three areas of the general personnel assignment problem as potential improvement opportunities: determining assignment criteria, properly evaluating personnel capabilities, and effectively assigning resources to tasks. Thus, developing new approaches to improve each of these areas constitute the objectives of this dissertation work. The main contributions of this research are threefold. First, this research presents an effective two-stage methodology to determine assignment criteria based on data envelopment analysis (DEA) and Tobit regression. Second, this research develops a novel fuzzy expert system for resource capability assessments in skill-based scenarios. The expert system properly evaluates the capabilities of resources in particular skills as a function of imprecise relationships that may exist between different skills. Third, this research develops an assignment model based on the fuzzy goal programming (FGP) technique. The model defines capabilities of resources, tasks requirements, and other important parameters as imprecise/fuzzy variables. The novelty of the research presented in this dissertation stems from the fact that it advances the science of personnel assignments by combining concepts from the fields of statistics, economics, artificial intelligence, and mathematical programming to develop a solution approach with an expected high practical value.

The most recent research in police domestic violence has shown that officers may perpetrate domestic violence at a higher rate than the general population, 28% versus 16%, respectively (Sgambelluri, 2000). Traditional police sub-culture has been identified, in several studies, as contributing to higher work stress, and using force on the job (Alexander et al., 1993; Drummond, 1976; Johnson et al, 2005; Kop and Euwema, 2001; Sgambelluri, 2000; Wetendorf, 2000). This research, however, has not fully examined the link between adherence to the traditional police sub-culture and officer involvement in domestic violence. This study attempts to identify whether officers who adhere to the aspects of the traditional police sub-culture are more likely to use violence against their intimate partner using two types of domestic violence-physical assault and psychological violence-as well as examine gender's moderating influence on police domestic violence and traditional police sub-culture. Using a survey created from existing scales, 250 officers were contacted within several departments in Central Florida, of these, 90 officers responded. Using Tobit and Logistic Regression the study found that officers who adhere to aspects of the traditional police subculture are more likely to engage in psychological domestic violence. There was no relationship found between traditional police culture and physical domestic violence. A thorough discussion of the results and future research directions is also included.

Wind power becomes one of the most promising energy sources in the electricity generation sector in Texas over the past decade by declining levelized cost of wind energy. However, recent studies show that the wind farms in Texas are relatively less productive. Hence, this study aims to find out reasons of inefficiencies by constructing a two-stage performance assessment of wind farms in Texas. In the first stage of analysis, comprehensive input- and output-oriented Data Envelopment Analysis (DEA) models are applied to evaluate productive efficiencies of the 95 large utility-scale wind farms by using pre-determined three input and two output variables. The sensitivity analysis is provided for the robustness of the DEA models with different combinations of input and output variables of the original model. The slack analysis and projection data are obtained for inefficient wind farms to find out optimal input-output variables. Tobit regression models are conducted for the second stage of the analysis to investigate the reasons of inefficiencies. DEA results indicate that half of the wind farms were operated efficiently in Texas during 2016. 13 wind farms were performed at the most productive scale size, ten wind farms should reduce their operational size to improve production efficiency, and 72 wind farms have the notable potential to increase their production efficiency by expanding operational sizes with modern wind turbine technologies. The sensitivity analysis shows the importance of each input-output variables. Tobit regression models indicate that elevation of the site, rotor diameter, hub height, and brand of the turbine have significant contributions to the relative efficiency scores of the wind farms, and the age of turbine has a negative impact on the productive efficiency of the wind farms.

Wind power is one of the most promising renewable energy sources that has gained enormous attention, especially in the electricity generation sector over the past decade in the United States. In this study Data Envelopment Analysis (DEA) is implemented to quantitatively evaluate the relative efficiencies of the 236 large utility-scale wind farms. Input- and output-oriented CCR (Charnes, Cooper, and Rhodes) and BCC (Banker, Charnes, and Cooper) models are applied to pre-determined three input and three output variables. The sensitivity analysis is conducted for the robustness of DEA by introducing seven new models with the various combinations of input and output variables of the original model. Tobit regression models are developed for the second stage of the analysis to investigate the effects of specifications of the wind turbine technologies. DEA results indicate that two-thirds of the wind farms are operated efficiently. On average, 70% of the wind farms have a considerable potential for further improvement in operational productivity by expanding these wind farm projects, 24% of them should reduce their operational size to increase their productivity level, and 6% of them are operating wind power at the most productive scale size. Nonparametric statistical tests show that the most efficient wind farms are located in Oklahoma because of the relatively high wind speed resources. Tobit regression model indicates the selection of the brand of the wind turbine has a significant contribution to the productive efficiency of the wind farms. The results of this study shed some light on the current efficiency assessments of the 236 large utility-scale wind farms in the United States and the future of wind energy for both energy practitioners and policy makers.

The average global surface temperature increased by 0.85 °C since 1850 because of irrepressible increase of the concentration of greenhouse gases (GHG). Electricity generation is the primary source of GHG emissions in the United States. Hence, renewable energy sources, which produce a negligible amount of GHG emissions, have gained enormous attention, especially in the electricity generation sector over the past decade. Wind power is the second largest renewable energy source to generate electricity in the United States. Therefore, in this study, a two-stage Data Envelopment Analysis (DEA) is developed to quantitatively evaluate the relative efficiencies of the 39 state's wind power performances for the electricity generation. Both input- and output-oriented CCR (Charnes, Cooper, and Rhodes (1978)) and BCC (Banker, Charnes, and Cooper (1984)) models are applied to pre-determined four input and six output variables. The sensitivity analysis is conducted to test the robustness of the DEA models. Tobit regression models are conducted by using the DEA results for the second stage analysis. The DEA results indicate that more than half of the states operate wind power efficiently. Tobit regression indicates that early installed wind power was more expensive and less productive relative the currently installed wind power. Findings of this study shed some light on the current efficiency assessments of the states and the future of wind energy for both energy practitioners and policy makers.