Estudos sobre mobilidade intergeracional de educação sugerem que países subdesenvolvidos apresentam menor mobilidade intergeracional que países desenvolvidos e especificamente para o Brasil, o grau de persistência estimado é ao redor de 0.7, podendo apresentar diferentes graus ao longo da distribuição de educação. Este estudo apresenta uma nova abordagem para a mensuração da mobilidade intergeracional utilizando Regressões Quantílicas. Especificamente, é proposta uma medida de distância entre os quantis condicionais para analisar a mobilidade intergeracional. Como resultado, é obtido um conjunto de matrizes que descrevem o padrão da mobilidade intergeracional em diferentes pontos da distribuição condicional de escolaridade. Utilizando dados para o Brasil, encontra-se que a mobilidade intergeracional tende a ser maior nas caudas da distribuição de escolaridade para filhos e filhas relativo à educação de pais e mães. Comparando filhos e filhas, os filhos tendem a ter menor mobilidade intergeracional que as mulheres relativo à educação de seus pais. Além do mais, a educação das mães tem maior efeito em magnitude do que a educação dos pais tanto para filhos quanto para as filhas. Também se encontrou que a educação dos filhos depende mais da educação do pai e a educação das filhas depende mais da educação das mães, indicando que os filhos tendem a ter educação similar à de seus pais e as filhas tendem a ter educação similar à de suas mães. / Studies on intergenerational educational mobility suggest that underdevelopment countries presents lower intergenerational mobility than developed countries and specifically for Brazil, the estimated degree of persistence is around 0.7 with possible different degrees on the overall distribution of education. This study presents a new approach to measuring intergenerational mobility using quantile regression. Specifically, it is proposed the use of a measure of distance between conditional quantiles to analyze intergenerational mobility. As a result, is obtained a set of matrices which describe the patterns of intergenerational mobility at different points of the conditional distribution of schooling. Using Brazilian Data (PNAD 1996) it is found that intergenerational mobility seems to be higher at the tails of the distribution of schooling for sons and daughters relative to father\'s and mother\'s education. Comparing each other, sons tend to have less mobility than daughters relative to father\'s education. Moreover, mother\'s education has stronger effects than father\'s on both sons and daughters education. It was also found that son\'s education depends more on father\'s education and daughter\'s education depends more on mother\'s education, indicating that sons tends to have education similar to their fathers and daughters tends to have education similar to their mothers.

Zero-inflated outcomes are common in biomedical studies, where the excessive zeros indicate some special but undetectable events. Quantile regression is potentially advantageous in analyzing zero-inflated outcomes due to two reasons. First, compared to parametric models such as the zero-inflated Poisson and two-part model, quantile regression gives robust and accurate estimation by avoiding likelihood specification and can capture the tail events and heterogeneity over the outcome distribution. Second, while the mean-based regression may be misinterpreted for a zero-inflated outcome, the interpretation of quantiles is naturally compatible with the underlying process that such an outcome intends to measure. Unfortunately, uncorrected linear quantile regression is not directly applicable because of two reasons. First, the feasibility of estimation and validity of inference of quantile regression require the conditional distribution of outcomes to be absolutely continuous, which is violated due to zero-inflation. Second, direct quantile regression implicitly assumes a constant chance to observe a positive outcome, but the degree of zero-inflation varies with the covariates in most cases. Thus the conditional quantile function of the outcome depends on the covariates in a nonlinear fashion. To analyze the zero-inflated outcomes by taking advantage of the merits of quantile regression, we propose a novel quantile regression framework that can address all the issues above. In the first part of this dissertation, we propose a two-part model that comprises a logistic regression for the probability of being positive, and a linear quantile regression for the positive part with subject-specific zero-inflation adjusted. Inference on the estimated conditional quantile and covariate effect are not trivial based on such a two-part model. We then develop an algorithm to achieve a consistent estimation of the conditional quantiles, while circumventing the unbounded variance at the quantile level where the conditional quantile changes from zero to positive. Furthermore, we develop an inference tool to determine the quantile treatment effect associated with a covariate at a given quantile level. We evaluate the proposed method and compare it with existing approaches by simulation studies and a real data analysis aimed at studying the risk factors for carotid atherosclerosis. In the second part, based on the proposed two-part model mentioned above, we develop ZIQRank, a zero-inflated quantile rank-score based test to detect the difference in distributions. The proposed test extends the local inference in the first part to a simultaneous one. It is powerful to handle zero-inflation and heterogeneity simultaneously. It comprises a valid test of logistic regression for the zero-inflation and rank-score based tests on multiple quantiles for the positive part with zero-inflation adjusted. The p-values are combined with a procedure selected according to the extent of zero-inflation and heterogeneity of the data. Simulation studies show that compared to existing tests, the proposed test has a higher power in detecting differential distributions. Finally, we apply the ZIQRank test to a human scRNA-seq data to study differentially expressed genes in Neoplastic and Regular cells. It successfully discovers a group of crucial genes associated with glioma, while the other methods fail to do so. In the third part, we extend the proposed two-part quantile regression model for zero-inflated outcomes and the ZIQRank test to analyze longitudinal data. Each part of the proposed two-part model is modified as a marginal longitudinal model (GEE), conditioning on the outcome at the previous time point and its zero/positive status. We apply the model and the test to study the effect of a recommender system aimed at boosting user engagement of a suite of smartphone apps designed for depressed patients. Our novel model framework demonstrates a dominating performance in model fitting, prediction, and critical feature detection, compared to the existing methods.

The objective of this study is to discover the non-linear effect of market sentiment to characteristic factor returns. We run ¡¥Quantile Regression¡¦ to help us extract useful information and design an effective strategy. Based on the quantitative investment method, using the platform of Multi-Factor Model (MFM), we attempt to construct a portfolio and enhance portfolio performance. If the market-sentiment variable increases performance, we could conclude that some characteristic factors in a high sentiment period will perform better or worse in the next period. What is the market or investor sentiment? It is still a problem in the finance field. There is no co-definition or consensus so far. We do our best to collect the indirect data, such as transaction data, price and volume data, and some indicators in other studies, VIX, put/call ratio and so on. Then, we test the proxy variables independently, and obtain some interesting results. The market turnover, the ratio of margin lending on funds/ margin lending on securities, and the growth rate of consumer confidence index have significant effects on some of the characteristic factors. This holds that some market sentiment variables could influence stocks with certain characteristics, and the factor timing approach could improve portfolio performance under examination by information ratio.

This study investigates investment behavior of Taiwan market participants from different aspects of measure, especially with regard to their tendency to forming herding behavior. By applying concepts of Cross-Sectional Absolute Dispersions (CSAD), we find significant evidence of herding behavior in the Taiwan market. Evidences suggest that the herding formation in Taiwan market is strongly influenced by the US market and we should not ignore the impact of globalization. With regard to the issue of financial crises, we find no herding behavior during the 1998 Asian Crisis but partial evidence shows that herding activities may be influenced by crisis during the 2000 Internet Bubble and 2008 Sub-prime Crisis in the Taiwan market. Moreover, all empirical results are reexamined using Quantile analysis to avoid potential bias in estimations. Finally, results from applying herding behavior in portfolio management indicate that investing in stocks of lower liquidity and volatility can reduce the risk of portfolios.

This study mainly explores the effect of market states (price and returns) on the relationship between spot and futures oil prices and targets three important issues: long-run cointegration, causalities, and market efficiency. Based on previous studies exhibiting bi-directional causality between spot and futures oil prices, this study employs quantile regressions to examine the possible feedback effect in their long-run cointegration and their causalities. In particular, it allows for exploring the possible asymmetric responses between spot and futures markets. The empirical results herein find that the long-run cointegrated relationship between contemporaneous spot and futures prices is impacted by the states of the spot markets. Similarly, whether futures oil prices lead spot oil prices is relevant with the states of the futures markets. This study also examines the efficiency of crude oil markets and shows that the efficiency is related to the length of futures contracts. These findings offer some implicative suggestions and strategies.

This dissertation includes two different empirical models about the economic growth and globalization in developed and developing countries from 1970 to 2008. First, we apply the quantile cointegration model provided by Xiao (2009) to examine the non-linear relationship between economic growth and globalization. Our empirical findings provide not only strong evidence that the cointegrating coefficients follow the time-varying process, but also the viewpoint of a long-run approach that overall globalization and their three dimensions act as engines of economic growth. Second, we adopt an advanced panel cointegration method which incorporates multiple structural breaks to examine the growth-globalization relationship. Differing from the weak outcomes of the traditional panel ointegration test without breaks, our findings provide strong evidence that overall globalization and its social dimension are cointegrated with RGDP both in developed and developing samples, and most of the structural break points are discovered in several main events. In addition, in evaluating whether or not the structural breaks affect the RGDP through globalization, we discover that the globalization have a directly positive impact on RGDP but indirectly exhibit negative (positive) impacts on real output via the channel of globalization in developed (developing) samples. Also, Different countries/groups reflect the different outcomes from the common shock of break event under the process of globalization. For the entire performance, the overall globalization brings the most positive effect on the real output in developed samples, and the social globalization is the main factor of promoting the economic development in developing samples.

This study examines the non-linear cointegrated relationship between capture production and the El Nino/La Nina index using the quantile technique proposed by Xiao (2009). According to the annual sample data of 6 Major Fishing Areas in the Pacific Ocean from 1950 to 2008, our empirical findings provide strong evidence that the cointegrating coefficients follow a time-varying process. They also imply that most of these long-run relationships are influenced by potential shocks over time rather than from maintaining a constant effect consistently. Overall, the contributions of this study not only stresses the importance of the quantile property in cointegrated models, but also provides a viewpoint on the long-run approach that the overall El Nino and La Nina act as engines for capture production.

Existing papers have shown that innovation and consumer satisfaction influence the firm performance and stock returns; however, the related papers usually neglect the impacts of market status. This paper extends prior papers by considering the impacts of market status when exploring the relationship among innovation, consumer satisfaction, and firm performance. Empirical results show that in the bull markets innovation and consumer satisfaction do not significantly affect stock returns while in the bear markets stock returns are positively associated with the level of innovation and consumer satisfaction. These results suggest that managers should take market status into consideration when making marketing decisions.

Quantile is a basic and an important quantity of a random variable. In some distributions, their quantiles have closed-form expressions. However, for many continuous distributions, the closed-form expressions of their quantiles do not exist. Yu and Zelterman (2011) and Chang (2004) have proposed an approximation of quantiles. In this paper, we propose an improved method which is combined the Taylor expansion with Newton¡¦s method. Some examples are given to compare the computing time of the method we proposed with the methods in Yu and Zelterman (2011) and Chang (2004).

We investigate the relationship between advertising and research and development (R&D) expenditures with the firm¡¦s systematic and unsystematic risks. Our data covers from January 1981 to December 2007 with more than two thousand publicly listed firms in the New York Stock Exchange. In addition to classical least squares approach, we utilize quantile regression model to examine whether the estimated slope parameters vary across different quantiles of the conditional distribution of the firm¡¦s systematic risk and unsystematic risk. We generate six empirical generalizations. (1) Advertising is significantly associated with lower systematic risk for firms with lower, median and higher systematic risk, but with no significant effects on the firms with extremely low systematic risk. (2) R&D is significantly associated with higher systematic risk for firms with median and higher systematic risk, with no significant effect for those with lower systematic risk. (3) Advertising is significantly associated with lower unsystematic risk for firms with higher unsystematic risk, but with no significant effects for those with median and lower unsystematic risk. (4) R&D is significantly associated with higher unsystematic risk for firms with median and higher unsystematic risk, with no significant effect for those with lower unsystematic risk. (5) Our evidence shows that both advertising and R&D have a stronger effect on firms with higher systematic risk (unsystematic risk) than on those with lower systematic risk (unsystematic risk). (6) Moreover, our evidence suggests that advertising and R&D tests resoundingly support our hypothesis that the coefficients vary across the quantiles.