Yes / Using International Passenger Survey (2017) data, this paper employs unconditional quantile regression (UQR) to analyse the determinants of tourist expenditure amongst inbound tourists to the United Kingdom. UQR allows us to estimate heterogeneous effects at any quantile of the distribution of the dependent variable. It overcomes the econometric limitations of ordinary least squares and quantile regression based estimates typically used to investigate tourism expenditures. However, our results reveal that the effects of our explanatory variables change across the distribution of tourist expenditure. This has important implications for those tasked with devising policies to enhance the UK’s tourist flows and expenditures.

In the first chapter of this thesis, we investigate the impact of human capital and wage structure on the gender pay in a panel of European countries using a newly available and appropriate database for cross-country comparisons and a comparable methodology for each country. Our first question is : What role do certain individual characteristics and choices of working men and women play in shaping the cross-country differences in the gender pay gap? What is the exact size of the gender pay gap using the “more appropriate” database available for our purpose? Giving that there are mainly only two harmonized data-sets for comparing gender pay gap throughout Europe: the European Community Household Panel (ECHP) and the European Structure of Earning Survey (ESES). Each database having its shortages: the main weakness of the ECHP is the lack of perfect reliability of the data in general and of wages in particular. However the main advantage of this database is the panel-data dimension and the information on both households and individuals. The data of the ESES is, on the contrary, of a very high standard but it only covers the private sector and has a cross-sectional dimension. Furthermore only few countries are currently available : Denmark, Belgium, Spain, Ireland and Italy. We use the European Structure of Earning Survey (ESES) to analyse international differences in gender pay gaps in the private sector based on a sample of five European economies: Belgium, Denmark, Ireland, Italy and Spain. Using different methods, we examine how wage structures, differences in the distribution of measured characteristics and occupational segregation contribute to and explain the pattern of international differences. Furthermore, we take account of the fact that indirect discrimination may influence female occupational distributions. We find these latter factors to have a significant impact on gender wage differentials. However, the magnitude of their effect varies across countries. In the second chapter, we analyse the persistence of the gender pay differentials over time in Europe and better test the productivity hypothesis by taking into account unobserved heterogeneity. Our second question is : What is the evolution of the pay differential between men and women over a period of time in Europe? And what is the impact of unobserved heterogeneity? The researcher here provides evidence on the effects of unobserved individual heterogeneity on estimated gender pay differentials. Using the European Community Household Panel (ECHP), we present a cross-country comparison of the evolution of unadjusted and adjusted gender pay gaps using both cross-section and panel-data estimation techniques. The analysed countries differ greatly with respect to labour market legislation, bargaining practices structure of earnings and female employment rates. On adjusting for unobserved heterogeneity, we find a narrowed male-female pay differential, as well as significantly different rates of return on individual characteristics. In particularly, the adjusted wage differential decreases by 7 per cent in Belgium, 14 per cent in Ireland, between 20-30 per cent Germany, Italy, the Netherlands and Spain and of 41 per cent and 54 per cent in the UK and in Denmark respectively. In the third chapter, we investigate causes of the gender pay gap beyond the gender differences in observed and unobserved productive characteristics or simply the sex. Explanations of the gender pay gap may be the penalty women face for having children. Obviously, the motherhood wage penalty is relevant to larger issues of gender inequality given that most women are mothers and that childrearing remains a women’s affair. Thus, any penalty associated with motherhood but not with fatherhood affects many women and as such contributes to gender inequalities as the gender pay gap. Furthermore, the motherhood wage effect may be different along the wage distribution as women with different earnings may not be equal in recognising opportunities to reconcile their mother’s and earner’s role. This brings us to our third question. Our third question is : What is the wage effect for mothers of young children in the household? And does it vary along the wage distribution of women? This chapter provides more insight into the effect of the presence of young children on women’s wages. We use individual data from the ECHP (1996-2001) and both a generalised linear model (GLM) and quantile regression (QR) techniques to estimate the wage penalty/bonus associated with the presence of children under the age of sixteen for mothers in ten EU Member States. We also correct for potential selection bias using the Heckman (1979) correction term in the GLM (at the mean) and a selectivity correction term in the quantile regressions. To distinguish between mothers according to their age at the time of their first birth, wage estimations are carried out, separately, for mothers who had their first child before the age of 25 (‘young mothers’) and mothers who had their first child after the age of 25 (‘old mothers’). Our results suggest that on average young mothers earn less than non-mothers while old mothers obtain a gross wage bonus in all countries. These wage differentials are mainly due to differences in human capital, occupational segregation and, to a lesser extent, sectoral segregation between mothers and non-mothers. This overall impact of labour market segregation, suggests a “crowding” explanation of the family pay gap – pay differential between mothers and non-mothers. Nevertheless, the fact that we still find significant family pay gaps in some countries after we control for all variables of our model suggests that we cannot reject the “taste-based” explanation of the family gap in these countries. Our analysis of the impact of family policies on the family pay gap across countries has shown that parental leave and childcare policies tend to decrease the pay differential between non-mothers and mothers. Cash and tax benefits, on the contrary, tend to widen this pay differential. Sample selection also affects the level of the mother pay gap at the mean and throughout the wage distribution in most countries. Furthermore, we find that in most countries inter-quantile differences in pay between mothers and non-mothers are mainly due to differences in human-capital. Differences in their occupational and sectoral segregation further shape these wage differentials along the wage distribution in the UK, Germany and Portugal in our sample of young mothers and in Spain in the sample of old mothers. In the fourth chapter, we analyse the combined effect of motherhood and the family status on women’s wage. Our fourth question is : Is there a lone motherhood pay gap in Europe? And does it vary along the wage distribution of mothers? Substantial research has been devoted to the analysis of poverty and income gaps between households of different types. The effects of family status on wages have been studied to a lesser extent. In this chapter, we present a selectivity corrected quantile regression model for the lone motherhood pay gap – the differential in hourly wage between lone mothers and those with partners. We used harmonized data from the European Community Household Panel and present results for a panel of European countries. We found evidence of lone motherhood penalties and bonuses. In our analysis, most countries presented higher wage disparities at the top of the wage distribution rather than at the bottom or at the mean. Our results suggest that cross-country differences in the lone motherhood pay gap are mainly due to differences in observed and unobserved characteristics between partnered mothers and lone mothers, differences in sample selection and presence of young children in the household. We also investigated other explanations for these differences such as the availability and level of childcare arrangements, the provision of gender-balanced leave and the level of child benefits and tax incentives. As expected, we have found significant positive relationship between the pay gap between lone and partnered mothers and the childcare, take-up and cash and tax benefits policies. Therefore improving these family policies would reduce the raw pay gap observed.

nÃo hà / The study extends traditional approaches related to the determinants of labor income in Brazil to analyze them according to quantiles of income within occupational groups selected in 1995, 2002 and 2009. In all three periods, quantile regressions in equations mincerianas are estimated and the results compared with traditional OLS estimates. The estimates differ dramatically and it appears that: i) the return to education increases with the qualifications of office, ii) there is a growing relationship with the return to education and income level within each occupation, but in time gain a year of additional study has been decreasing and iii) although it is verified a reduction of inequality by race and gender in the labor market, there are signs of deterioration in these low-skilled occupations and pay. / O estudo amplia abordagens tradicionais relacionadas aos determinantes dos rendimentos do trabalho no Brasil ao analisÃ-los segundo quantis de renda dentro de grupos ocupacionais selecionados em 1995, 2002 e 2009. Nos trÃs perÃodos, regressÃes quantÃlicas em equaÃÃes mincerianas sÃo estimadas e os resultados confrontados com estimaÃÃes tradicionais por MQO. As estimativas diferem dramaticamente e permitem constatar que: i) o retorno à educaÃÃo cresce com a qualificaÃÃo do cargo, ii) existe uma relaÃÃo crescente do retorno à educaÃÃo com o nÃvel de renda dentro de cada ocupaÃÃo, mas temporalmente o ganho de um ano de estudo adicional vem diminuindo e iii) muito embora seja verificada uma reduÃÃo das desigualdades por sexo e raÃa no mercado de trabalho, existem indÃcios de agravamento destas em profissÃes de baixa qualificaÃÃo e remuneraÃÃo.

nÃo hà / A presente pesquisa busca analisar os impactos de variÃveis polÃticas e socioeconÃmicas sobre o sucesso eleitoral de um candidato ao legislativo, a partir de um estudo empÃrico com base na votaÃÃo das eleiÃÃes de 2010 para o cargo de Deputado Estadual no Estado do Cearà e segundo duas regiÃes: RegiÃo Metropolitana de Fortaleza e Interior do Estado. Para alcanÃar os objetivos propostos, sÃo utilizados modelos de regressÃes mÃltiplas estimados por MÃnimos Quadrados OrdinÃrios e RegressÃes QuantÃlicas, com vistas a possibilitar uma maior caracterizaÃÃo da distribuiÃÃo condicional das variÃveis analisadas. Na prÃtica, esta Ãltima tÃcnica permite investigar potenciais diferenÃas no impacto das variÃveis para elevados e baixos Ãndices de votaÃÃo. Dentre os resultados encontrados, se verifica a influÃncia positiva das variÃveis explicativas estatisticamente significantes inseridas nos modelos, em diferentes nÃveis de impacto de acordo com a regiÃo em anÃlise, em especial as variÃveis que tratam das despesas e receitas de campanha, da reeleiÃÃo de candidatos e das ligaÃÃes partidÃrias com o governo estadual e com as prefeituras municipais na Ãpoca da campanha, suscitando assim uma reflexÃo sobre a necessidade de realizaÃÃo de uma reforma polÃtica e eleitoral no paÃs visando proporcionar condiÃÃes mais igualitÃrias aos pleitos. / This research seeks to analyze the impacts of political and socioeconomic variables on the electoral success of a candidate to the legislature, from an empirical study based on the voting of the 2010 elections for the office of State Representative in the State of Cearà and in two regions: Fortaleza Metropolitan Region and Inner State. To achieve the proposed objectives, multiple regression models estimated by OLS and quantile regressions are used in order to enable greater characterization of the conditional distribution of the variables. In practice, this latter technique allows to investigate potential differences in the impact of the variables for high and low rates of voting. Among the findings, it appears the positive influence of statistically significant explanatory variables included in the models at different levels of impact according to the region in question, especially the variables dealing with expenses and campaign revenue, the re-election of candidates and of party connections with the state government and the municipal governments at the time of the campaign, inspiring a reflection on the need to carry out a political and electoral reform in the country aiming to provide more equal conditions for elections.

No description available.

O Brasil é um dos países mais violentos do mundo, e essa realidade se reflete no ambiente escolar. O problema da violência na escola induz nos agentes de educação comportamentos que se contrapõem às metas de melhoria da qualidade do ensino. Por isso, a violência na escola é um dos candidatos a determinante da baixa qualidade educacional encontrada no Brasil. Este trabalho investiga se existe relação entre violência na escola e proficiência dos alunos no Brasil nos anos de 2003, 2005, 2007, 2009, 2011 e 2013. Os resultados das regressões lineares e das regressões quantílicas incondicionais indicam que a ocorrência de um tipo de evento violento adicional está associada a uma redução na proficiência. Eles indicam também que há heterogeneidade na resposta dos alunos à violência na escola: a violência na escola parece afetar mais os alunos da quarta série, os alunos que se encontram na mediana da distribuição das notas e os alunos com melhor condição socioeconômica. Além disso, a violência na escola parece afetar mais a proficiência em Matemática em comparação com a proficiência em Língua Portuguesa, e a violência contra o professor tem efeitos tão ou mais deletérios que a violência entre os alunos. / Brazil is one of the most violent countries in the world, and this reality is part of the school environment. The problem of school violence hinders any possible progress in improving the quality of education. Indeed, school violence is a key element in explaining the low educational quality in Brazil. This research investigates the relationship between school violence and student achievement in Brazil in the years of 2003, 2005, 2007, 2009, 2011 and 2013. The results of linear regressions and unconditional quantile regressions analysis show that an occurrence of an additional kind of violent event is associated with a reduction on student achievement. They also show that there is a high degree of heterogeneity in student responses to school violence: the school violence seems to affect more 4th graders, the students that are in the median of the distribution of the grades and the students with the best socioeconomic conditions. Moreover, school violence seems to affect more grades in Math than Portuguese, and the violence against teachers has effects as or more deleterious than the violence between students.

O Brasil é um dos países mais violentos do mundo, e essa realidade se reflete no ambiente escolar. O problema da violência na escola induz nos agentes de educação comportamentos que se contrapõem às metas de melhoria da qualidade do ensino. Por isso, a violência na escola é um dos candidatos a determinante da baixa qualidade educacional encontrada no Brasil. Este trabalho investiga se existe relação entre violência na escola e proficiência dos alunos no Brasil nos anos de 2003, 2005, 2007, 2009, 2011 e 2013. Os resultados das regressões lineares e das regressões quantílicas incondicionais indicam que a ocorrência de um tipo de evento violento adicional está associada a uma redução na proficiência. Eles indicam também que há heterogeneidade na resposta dos alunos à violência na escola: a violência na escola parece afetar mais os alunos da quarta série, os alunos que se encontram na mediana da distribuição das notas e os alunos com melhor condição socioeconômica. Além disso, a violência na escola parece afetar mais a proficiência em Matemática em comparação com a proficiência em Língua Portuguesa, e a violência contra o professor tem efeitos tão ou mais deletérios que a violência entre os alunos. / Brazil is one of the most violent countries in the world, and this reality is part of the school environment. The problem of school violence hinders any possible progress in improving the quality of education. Indeed, school violence is a key element in explaining the low educational quality in Brazil. This research investigates the relationship between school violence and student achievement in Brazil in the years of 2003, 2005, 2007, 2009, 2011 and 2013. The results of linear regressions and unconditional quantile regressions analysis show that an occurrence of an additional kind of violent event is associated with a reduction on student achievement. They also show that there is a high degree of heterogeneity in student responses to school violence: the school violence seems to affect more 4th graders, the students that are in the median of the distribution of the grades and the students with the best socioeconomic conditions. Moreover, school violence seems to affect more grades in Math than Portuguese, and the violence against teachers has effects as or more deleterious than the violence between students.

Le sujet principal de cette thèse est de proposer de nouvelles méthodes d'extractions de signaux avec applications en finance. Par signaux, nous entendons soit un signal sur lequel repose une stratégie d'investissement; soit un signal perturbé par un bruit, que nous souhaitons retrouver. Ainsi, la première partie de la thèse étudie la contagion en volatilité historique autours des annonces de résultats des entreprises du Nasdaq. Nous trouvons qu'autours de l'annonce, l'entreprise reportant ses résultats, génère une contagion persistante en volatilité à l’encontre des entreprises appartenant au même secteur. Par ailleurs, nous trouvons que la contagion en volatilité varie, selon le type de nouvelles reportées, l'effet de surprise, ou encore par le sentiment de marché à l'égard de l'annonceur. La deuxième partie de cette thèse adapte des techniques de dé-bruitage venant de l'imagerie, à des formes de bruits présentent en finance. Ainsi, un premier article, co-écrit avec Matthieu Garcin, propose une technique de dé-bruitage innovante, permettant de retrouver un signal perturbé par un bruit à variance non-constante. Cet algorithme est appliqué en finance à la modélisation de la volatilité. Un second travail s'intéresse au dé-bruitage d'un signal perturbé par un bruit asymétrique et leptokurtique. En effet, nous adaptons un modèle de Maximum A Posteriori, couramment employé en imagerie, à des bruits suivant des lois de probabilité de Student, Gaussienne asymétrique et Student asymétrique. Cet algorithme est appliqué au dé-bruitage de prix d'actions haute-fréquences. L'objectif étant d'appliquer un algorithme de reconnaissance de formes sur les extrema locaux du signal dé-bruité. / The main objective of this PhD dissertation is to set up new signal extraction techniques with applications in Finance. In our setting, a signal is defined in two ways. In the framework of investement strategies, a signal is a function which generates buy/sell orders. In denoising theory, a signal, is a function disrupted by some noise, that we want to recover. A first part of this PhD studies historical volatility spillovers around corporate earning announcements. Notably, we study whether a move by one point in the announcer historical volatility in time t will generate a move by beta percents in time t+1. We find evidences of volatility spillovers and we study their intensity across variables such as : the announcement outcome, the surprise effect, the announcer capitalization, the market sentiment regarding the announcer, and other variables. We illustrate our finding by a volatility arbitrage strategy. The second part of the dissertation adapts denoising techniques coming from imagery : wavelets and total variation methods, to forms of noise observed in finance. A first paper proposes an denoising algorithm for a signal disrupted by a noise with a spatially varying standard-deviation. A financial application to volatility modelling is proposed. A second paper adapts the Bayesian representation of the Rudin, Osher and Fatemi approach to asymmetric and leptokurtic noises. A financial application is proposed to the denoising of intra-day stock prices in order to implement a pattern recognition trading strategy.

O Brasil é um dos países mais violentos do mundo, e essa realidade se reflete no ambiente escolar. O problema da violência na escola induz nos agentes de educação comportamentos que se contrapõem às metas de melhoria da qualidade do ensino. Por isso, a violência na escola é um dos candidatos a determinante da baixa qualidade educacional encontrada no Brasil. Este trabalho investiga se existe relação entre violência na escola e proficiência dos alunos no Brasil nos anos de 2003, 2005, 2007, 2009, 2011 e 2013. Os resultados das regressões lineares e das regressões quantílicas incondicionais indicam que a ocorrência de um tipo de evento violento adicional está associada a uma redução na proficiência. Eles indicam também que há heterogeneidade na resposta dos alunos à violência na escola: a violência na escola parece afetar mais os alunos da quarta série, os alunos que se encontram na mediana da distribuição das notas e os alunos com melhor condição socioeconômica. Além disso, a violência na escola parece afetar mais a proficiência em Matemática em comparação com a proficiência em Língua Portuguesa, e a violência contra o professor tem efeitos tão ou mais deletérios que a violência entre os alunos. / Brazil is one of the most violent countries in the world, and this reality is part of the school environment. The problem of school violence hinders any possible progress in improving the quality of education. Indeed, school violence is a key element in explaining the low educational quality in Brazil. This research investigates the relationship between school violence and student achievement in Brazil in the years of 2003, 2005, 2007, 2009, 2011 and 2013. The results of linear regressions and unconditional quantile regressions analysis show that an occurrence of an additional kind of violent event is associated with a reduction on student achievement. They also show that there is a high degree of heterogeneity in student responses to school violence: the school violence seems to affect more 4th graders, the students that are in the median of the distribution of the grades and the students with the best socioeconomic conditions. Moreover, school violence seems to affect more grades in Math than Portuguese, and the violence against teachers has effects as or more deleterious than the violence between students.

This thesis examines inter-industry wage differentials in the Czech Republic, using the European Union - Statistics on Income and Living (EU-SILC 2009) survey as our primary data source. Findings show that even after controlling for large number of workers and jobs characteristics wage differences based on industry affiliation still persist. The variation of the inter-industry wage differentials amounts to approximately 5 percent with the maximum wage level difference of 25 percent between the financial sector and agriculture. By applying two distinct methodologies we tested the hypothesis that the inter-industry wage differentials are actually caused by higher concentration of workers with better unmeasured abilities in higher-paying industries. Neither of the two methods rejected the unobserved ability hypothesis. Finally, our analysis also shows that the inter-industry wage differentials can be to a certain extent attributed to rent-sharing and different labour turnover costs across sectors.