<p>The economic development and industrialization that has taken place in many parts of the world during the past century has brought about a huge increase in economic welfare. During this process, it has repeatedly been debated whether the gains from economic development are shared by everyone or just a few. In the field of economics, vast research has been conducted on this particular subject ever since the 1950’s. The most famous contribution might be said to be Simon Kuznets article, Growth and Income Inequality from 1955 and the ”inverted U”-hypothesis that was formulated on the basis of that article. The essence of the hypothesis is that a country, during its development, moves from agricultural to industrial production. At first, income inequality increases and then, at the end of the process, decreases.</p><p>The aim of this paper has been to investigate the relationship between income distribution and economic development in a particular region, namely the countries of the former Soviet Union, during 1992-2003. Also, we have tested whether Kuznets theory and the “inverted U”-hypothesis hold true for our sample. The investigations method is a survey, which uses secondary data collected from the World Bank’s database of World Development Indicators. Regression-analysis has been employed to conduct cross-sections between 20 countries over 4 periods in Eastern Europe and Central Asia. The number of observations equals to 62. The variables that we use are: the Gini index, GDP per capita and the agricultural sector’s share of GDP.</p><p>The results of the regression do not indicate any resemblance to the pattern of the “inverted U”-hypothesis. The curve we get is that of a “positive U”. Countries with high GDP per capita as well as countries with low GDP per capita have high income inequality. Countries with mediate GDP per capita levels have low income inequality. Our analysis concludes that the countries in our sample might have had a very unique economic development following the fall of the Soviet Union. Only 7 countries partly follow the Kuznets theory if tested individually. Since these seven countries seem to have a reversed development with increasing agricultural sector and the remainder of the countries show increasing income inequality, we reject the “inverted U”-hypothesis and question the ability of Kuznets’ theory to explain income distribution for our sample.</p>

The economic development and industrialization that has taken place in many parts of the world during the past century has brought about a huge increase in economic welfare. During this process, it has repeatedly been debated whether the gains from economic development are shared by everyone or just a few. In the field of economics, vast research has been conducted on this particular subject ever since the 1950’s. The most famous contribution might be said to be Simon Kuznets article, Growth and Income Inequality from 1955 and the ”inverted U”-hypothesis that was formulated on the basis of that article. The essence of the hypothesis is that a country, during its development, moves from agricultural to industrial production. At first, income inequality increases and then, at the end of the process, decreases. The aim of this paper has been to investigate the relationship between income distribution and economic development in a particular region, namely the countries of the former Soviet Union, during 1992-2003. Also, we have tested whether Kuznets theory and the “inverted U”-hypothesis hold true for our sample. The investigations method is a survey, which uses secondary data collected from the World Bank’s database of World Development Indicators. Regression-analysis has been employed to conduct cross-sections between 20 countries over 4 periods in Eastern Europe and Central Asia. The number of observations equals to 62. The variables that we use are: the Gini index, GDP per capita and the agricultural sector’s share of GDP. The results of the regression do not indicate any resemblance to the pattern of the “inverted U”-hypothesis. The curve we get is that of a “positive U”. Countries with high GDP per capita as well as countries with low GDP per capita have high income inequality. Countries with mediate GDP per capita levels have low income inequality. Our analysis concludes that the countries in our sample might have had a very unique economic development following the fall of the Soviet Union. Only 7 countries partly follow the Kuznets theory if tested individually. Since these seven countries seem to have a reversed development with increasing agricultural sector and the remainder of the countries show increasing income inequality, we reject the “inverted U”-hypothesis and question the ability of Kuznets’ theory to explain income distribution for our sample.

Nous proposons dans cette thèse la méthode STREE de construction d'arbres de décision avec des données symboliques. Ce type de données permet de caractériser des individus de niveau supérieur qui peuvent être des classes ou catégories d’individus ou des concepts au sens des treillis de Galois. Les valeurs des variables, appelées variables symboliques, peuvent être des ensembles, des intervalles ou des histogrammes. Le critère de partitionnement récursif est une combinaison d'un critère par rapport aux variables explicatives et d'un critère par rapport à la variable à expliquer. Le premier critère est la variation de la variance des variables explicatives. Quand il est appliqué seul, STREE correspond à une méthode descendante de classification non supervisée. Le second critère permet de construire un arbre de décision. Il s'agit de la variation de l'indice de Gini si la variable à expliquer est nominale et de la variation de la variance si la variable à expliquer est continue ou bien est une variable symbolique. Les données classiques sont un cas particulier de données symboliques sur lesquelles STREE peut aussi obtenir de bons résultats. Il en ressort de bonnes performances sur plusieurs jeux de données UCI par rapport à des méthodes classiques de Data Mining telles que CART, C4.5, Naive Bayes, KNN, MLP et SVM. STREE permet également la construction d'ensembles d'arbres de décision symboliques soit par bagging soit par boosting. L'utilisation de tels ensembles a pour but de pallier les insuffisances liées aux arbres de décisions eux-mêmes et d'obtenir une décision finale qui est en principe plus fiable que celle obtenue à partir d'un arbre unique. / In this thesis, we propose the STREE methodology for the construction of decision trees with symbolic data. This data type allows us to characterize individuals of higher levels which may be classes or categories of individuals or concepts within the meaning of the Galois lattice. The values of the variables, called symbolic variables, may be sets, intervals or histograms. The criterion of recursive partitioning is a combination of a criterion related to the explanatory variables and a criterion related to the dependant variable. The first criterion is the variation of the variance of the explanatory variables. When it is applied alone, STREE acts as a top-down clustering methodology. The second criterion enables us to build a decision tree. This criteron is expressed as the variation of the Gini index if the dependant variable is nominal, and as the variation of the variance if thedependant variable is continuous or is a symbolic variable. Conventional data are a special case of symbolic data on which STREE can also get good results. It has performed well on multiple sets of UCI data compared to conventional methodologies of Data Mining such as CART, C4.5, Naive Bayes, KNN, MLP and SVM. The STREE methodology also allows for the construction of ensembles of symbolic decision trees either by bagging or by boosting. The use of such ensembles is designed to overcome shortcomings related to the decisions trees themselves and to obtain a finaldecision that is in principle more reliable than that obtained from a single tree.

O dilema do prisioneiro (DP) é um dos principais jogos da teoria dos jogos. No dilema do prisioneiro discreto (DPD), dois prisioneiros têm as opções de cooperar ou desertar. Um jogador cooperador não delata seu comparsa, já um desertor delata. Se um cooperar e o outro desertar, o cooperador fica preso por cinco anos e o desertor fica livre. Se ambos cooperarem, ficam presos por um ano e, se ambos desertarem, ficam presos por três anos. Quando o DP é repetido, a cooperação pode emergir entre agentes egoístas. Realizamos um estudo analítico para o DPD, que produziu uma formulação da evolução do nível médio de cooperação e da tentação crítica (valor de tentação que causa mudança abrupta do nível de cooperação). No dilema do prisioneiro contínuo (DPC), cada jogador apresenta um nível de cooperação que define o grau de cooperação. Utilizamos o DPC para estudar o efeito da personalidade dos jogadores sobre a emergência da cooperação. Para isso, propusemos novas estratégias: uma baseada na personalidade dos jogadores e outras duas baseadas na comparação entre o ganho obtido e a aspiração do jogador. Todas as estratégias apresentavam algum mecanismo de cópia do estado do vizinho com maior ganho na vizinhança, mecanismo este, herdado da estratégia darwiniana. Os resultados mostraram que o DPC aumenta o nível médio de cooperação do sistema, quando comparado ao DPD. No entanto, as diferentes estratégias não aumentaram a cooperação comparado à cooperação obtida com a estratégia darwiniana. Então propusemos o uso do coeficiente de agrupamentos, coeficiente de Gini e entropias de Shannon, Tsallis e Kullback-Leibler para classificar os sistemas, em que os agentes jogam o DPD com a estratégia darwiniana, quanto ao nível de cooperação. Como analisamos valores de médias configuracionais, tais classificadores não foram eficientes ao classificar os sistemas. Isso é consequência da existência de distribuições de extremos nos resultados que compõem as médias. As distribuições de extremos suscitaram uma discussão acerca da definição do regime de cooperação no dilema do prisioneiro. Discutimos também as consequências de utilizar apenas valores médios nos resultados ignorando seus desvios e as distribuições. / Prisoner\'s dilemma (PD) is one of the main games of game theory. In discrete prisoner\'s dilemma (DPD), two prisoners have the options to cooperate or to defect. A cooperator player does not defect his accomplice, while a defector does. If one player cooperates and the other defects, the cooperator gets jailed for five years and the defector goes free. If both cooperate, they get jailed during one year and if both defect, they get jailed during three years. When this game is repeated, cooperation may emerge among selfish individuals. We perform an analytical study for the DPD, that produced a formulation for the evolution of the mean cooperation level and for the critical temptation values (temptation values that promote abrupt modifications in the cooperation level). In continuous prisoner\'s dilemma (CPD), each player has a level of cooperation that defines his/her degree of cooperation. We used the CPD to study the effect of the players\' personality on the emergence of cooperation. For this, we propose new strategies: one based on the players\' personality and two others based on the comparison between the player\'s obtained payoff and the desire one. All strategies present some mechanism that copies the state of the neighbor with the highest payoff in the neighborhood, mechanism inherited from the Darwinian strategy. The results showed that the CPD increases the average cooperation level of the system when compared to DPD. However, different strategies do not increased the cooperation compared to cooperation obtained with the Darwinian strategy. So, we propose the use of cluster coefficient, Gini coefficient and entropy of Shannon, Tsallis and Kullback-Leibler as classifiers to classify systems, in which the individuals play DPD with Darwinian strategy, by the cooperation level. As configurational averages were analyzed, such classifiers were not efficient in classifying the systems. This is due to the existence of distributions with extreme values of the results that compose the means. Distributions with extremes values emerged a discussion about the definition of the cooperation state in the prisoner\'s dilemma. We also discussed the consequences of using only average results in the analysis ignoring their deviations and distributions.

O objetivo primordial deste trabalho é analisar se a democracia é uma instituição política que produz resultados econômicos menos desiguais do que os regimes autoritários. A importância deste tema reside no fato que a própria promoção da democracia na agenda da política internacional tornou-se fundamental por inúmeras razões entre as quais sua suposta propensão em reduzir estas disparidades econômicas. Em primeiro lugar apresentamos no Capítulo 1 um balanço da discussão teórica e empírica a partir da qual constatamos que, a despeito do senso comum de que a democracia está relacionada a uma cidadania mais igualitária, os seus efeitos sobre a desigualdade ainda são discutíveis. Mesmo existindo um razoável consenso teórico de que os regimes democráticos devem, de alguma forma, produzir uma melhor distribuição de bens, os resultados empíricos são inconclusivos e contraditórios. Em seguida, diante de tal impasse empírico, propomos no Capítulo 2 uma reformulação da argumentação na qual entendemos que os efeitos da democracia sobre a desigualdade devem ser reinterpretados. A principal contribuição da tese reside na constatação, tanto teórica quanto empírica, de que estes efeitos são heterogêneos e interagem com o próprio nível de desigualdade, e, por conseguinte, é equivocado o suposto de que esses efeitos são homogêneos e independentes do contexto sócio-econômico da desigualdade. No Capítulo 3 apresentamos os dados e os conceitos de democracia e desigualdade. Assumimos que democracia se caracteriza como o regime político no qual os líderes competem entre si por meio de eleições e verificamos se os seus efeitos variam ao longo da própria distribuição de desigualdade econômica mensurada pelo coeficiente de GINI. Para tal análise, realizamos uma série de modelos de regressão quantílica, a metodologia adequada para avaliar o debate sobre a heterogeneidade versus homogeneidade dos efeitos. O argumento teórico, a partir do qual elabora-se a hipótese dos efeitos heterogêneos, refere-se à necessidade de uma convergência entre os interesses eleitorais dos partidos o lado da oferta e as clivagens sobre as quais uma potencial maioria dos eleitores tem interesse em ser atendido o lado da demanda por políticas públicas e plataformas. Isto posto, é 9 necessário discutir as condições que estimulam as lideranças políticas a utilizarem o problema da desigualdade econômica como argumento eleitoral e as condições nas quais surge uma demanda dos cidadãos por redistribuição via ação estatal. Somente nas sociedades mais desiguais tanto os partidos políticos têm interesse em ofertar políticas redistributivas, quanto tende a surgir no seio da cidadania uma demanda por redistribuição por parte de uma maioria de eleitores. No Capítulo 4 comprovamos empiricamente que os efeitos da competição democrática em sociedades mais desiguais são diferentes seus efeitos em sociedades mais iguais; e estes efeitos estão em direção à maior redução da desigualdade apenas nas sociedades mais desiguais. Os resultados são robustos às mais diferentes especificações dos modelos estatísticos, dados e formas de mensuração, tanto de democracia quanto de desigualdade, em diferentes cortes temporais e horizontes históricos de análise. Inclusive quando estendemos o recorte temporal para antes do pós-2ª Guerra Mundial utilizando dados que abrangem o período de surgimento dos primeiros regimes representativos democráticos no século XIX, a veracidade das hipóteses dos efeitos heterogêneos e de que há maior contundência da democracia em direção à redução da desigualdade nas sociedades mais desiguais permanece. Por fim, além deste problema teórico e empírico de crucial importância, também controlamos a análise para a potencial relação recíproca entre democracia e desigualdade. Enquanto parte da literatura discute os potenciais efeitos igualitários da democracia, outra importante literatura debate se o aumento da desigualdade aumenta ou reduz a probabilidade de um país tornar-se ou manter-se democrático. Posto isto, apresentamos uma lista de variáveis instrumentais para estimar validamente os efeitos da democracia sobre a desigualdade independente da relação entre desigualdade e democracia / The primary aim of this study is to analyze whether democracy is a political institution that produces less unequal economic outcomes than authoritarian regimes. The importance of this issue lies in the fact that the very promotion of democracy in the international political agenda has become essential for many reasons, including its supposed propensity to reduce economic disparities. First, at Chapter 1 we overview the theoretical and empirical discussion from which we find that despite the common sense that democracy must be related to a more egalitarian citizenship, its effects on inequality is still debatable. Even with a reasonable theoretical consensus that democracies must somehow produce a better distribution of goods; the empirical results are inconclusive and contradictory. After that, facing such empirical impasse, we propose at Chapter 2 a reformulation about the rationale to explain and analyze the effects of democracy on inequality. The main contribution of this thesis lies in both the theoretical and the empirical claim that these effects are heterogeneous and should interact with the level of inequality and, therefore, the assumption that these effects are homogeneous and independent of the socio-economic context of inequality is wrong. In Chapter 3, we present the data and concepts of democracy and inequality. We assume that democracy is characterized as a political regime in which leaders compete through elections and we test whether the effects vary along the distribution of economic inequality measured by the Gini coefficient. To do that, we conducted a series of quantile regression models, appropriate to evaluate the alternative hypothesis whether the effects are heterogeneous or homogenous. The theoretical argument, from which we elaborate the hypothesis of heterogeneous effects, refers to the need for a convergence between the electoral interests of the parties - the supply side - and the political cleavages on which a majority of voters have potential interest being played - the demand side for other public policies and platforms. Hence, it is necessary to discuss the conditions that lead the political leadership to use the problem of economic 11 inequality as an electoral argument and the conditions under which a demand by citizens for redistribution via state action rises. Only at the most unequal societies the political parties have an interest in offering redistributive policies, as well as there is a higher propensity for a redistribution demand by a majority of voters. In Chapter 4, we proved empirically that the effects of democratic competition at more unequal societies are different from the effects of democracy in more equal societies; and these effects tend to be greater toward inequality reduction only at more unequal societies. These results are robust to different statistical model specifications, data and measurement methods, about both democracy and inequality, and to the use of different time horizons. Even when we extend the time frame of the analysis to the period before World War II - using new data that covers XIX century, the veracity of the hypotheses about the heterogeneous effects and that these effects of democracy toward the reduction of inequality are larger at the most unequal societies remains intact. Finally, beyond this theoretical and empirical issue of crucial importance, we also control the analysis for potential reciprocal relationship between democracy and inequality. This is because while much of the literature discusses the potential effects of egalitarian democracy, another important literature debate discusses whether greater inequality increases or reduces the probability of a country become or remain democratic. Hence, we present a list of valid instrumental variables to estimate the effects of democracy on inequality independent of the relationship between inequality and democracy

This paper examines the effect of governance on inequality, specifically if improvements in the World Bank’s Worldwide Governance Indicators affect inequality as measured by two Gini coefficients: Market Gini, before taxes and redistribution, and Net Gini, after taxes and redistribution. The data for the Gini measurements was taken from the Standardized World Income Inequality Database (SWIID) and the data for the Worldwide Governance Indicators was taken from the World Bank. Data for fifteen (15) years, from the start of the Worldwide Governance Indicators until 2013, was combined with data from SWIID for the same years. In all, data from one hundred fifty-six (156) countries with a full set of six (6) indicators for the years that had at least one corresponding Gini measurements were used in this study: in total one thousand seven hundred and forty-seven (1747) observations. In a pooled OLS regression, controlling for growth with the variable GDP per Capita expressed as a per cent (%) change on an annual basis, the individual indicators gave the following results, where a positive sign indicates increased inequality and vice versa: Control of Corruption and Regulatory Quality showed a positive sign for both Gini measurements. Rule of Law, Government Effectiveness, Political Stability and the Absence of Violence/Terrorism, gave a negative sign for both Gini measurements. Voice and Accountability showed a positive sign for Market Gini and a negative sign for Net Gini. The fact that an improvement in Control of Corruption increased inequality both before and after taxes and redistribution was unexpected and should be further researched.

The aim of this thesis is to investigate possible improvement of scoring models prediction power in retail credit segment by using structural models estimating the future development of behavioral score. These models contain the informa- tion about past development of the behavioral score by parameters which take into account the sensitivity of clients' probability of default on individual market and life changes. These parameters are estimated by Markov Chain Monte Carlo methods based on score history. Eight different types of structural models were applied to real data. The diversification measure of individual models is compared using the Gini coefficient. These structural models were compared with each other and also with the existing scoring model of the credit institution which provided the underlying data. 1

Obtaining variances for the plug-in estimator of the Gini coefficient for inequality has preoccupied researchers for decades with the proposed analytic formulae often being regarded as being too cumbersome to apply, as well as usually based on the assumption of an iid structure. We examine several variance estimation techniques for a Gini coefficient estimator obtained from a complex survey, a sampling design often used to obtain sample data in inequality studies. In the first part of the dissertation, we prove that Bhattacharya’s (2007) asymptotic variance estimator when data arise from a complex survey is equivalent to an asymptotic variance estimator derived by Binder and Kovačević (1995) nearly twenty years earlier. In addition, to aid applied researchers, we also show how auxiliary regressions can be used to generate the plug-in Gini estimator and its asymptotic variance, irrespective of the sampling design. In the second part of the dissertation, using Monte Carlo (MC) simulations with 36 data generating processes under the beta, lognormal, chi-square, and the Pareto distributional assumptions with sample data obtained under various complex survey designs, we explore two finite sample properties of the Gini coefficient estimator: bias of the estimator and empirical coverage probabilities of interval estimators for the Gini coefficient. We find high sensitivity to the number of strata and the underlying distribution of the population data. We compare the performance of two standard normal (SN) approximation interval estimators using the asymptotic variance estimators of Binder and Kovačević (1995) and Bhattacharya (2007), another SN approximation interval estimator using a traditional bootstrap variance estimator, and a standard MC bootstrap percentile interval estimator under a complex survey design. With few exceptions, namely with small samples and/or highly skewed distributions of the underlying population data where the bootstrap methods work relatively better, the SN approximation interval estimators using asymptotic variances perform quite well. Finally, health data on the body mass index and hemoglobin levels for Bangladeshi women and children, respectively, are used as illustrations. Inequality analysis of these two important indicators provides a better understanding about the health status of women and children. Our empirical results show that statistical inferences regarding inequality in these well-being variables, measured by the Gini coefficients, based on Binder and Kovačević’s and Bhattacharya’s asymptotic variance estimators, give equivalent outcomes. Although the bootstrap approach often generates slightly smaller variance estimates in small samples, the hypotheses test results or widths of interval estimates using this method are practically similar to those using the asymptotic variance estimators. Our results are useful, both theoretically and practically, as the asymptotic variance estimators are simpler and require less time to calculate compared to those generated by bootstrap methods, as often previously advocated by researchers. These findings suggest that applied researchers can often be comfortable in undertaking inferences about the inequality of a well-being variable using the Gini coefficient employing asymptotic variance estimators that are not difficult to calculate, irrespective of whether the sample data are obtained under a complex survey or a simple random sample design. / Graduate / 0534 / 0501 / 0463 / aahoque@gmail.com

The role of inequality in development has been the subject of long-standing debates in academic and policy circles. Notwithstanding disagreements about exactly how the two are linked, conventional wisdom agrees that inequality is an objective 'fact' that can be measured free from ideological considerations. New data detect trends towards higher inequality, weaker economic positions for those at the bottom, and a concentration of wealth at the very top of the distribution in most regions. Inequality studies as currently practiced are ill-equipped to accommodate the empirical changes and the resulting theoretical implications. Putting an end to over half a century of mainstream consensus assuming that inequality would automatically recede with developmental progress, the discipline needs rethinking. My thesis proposes a new research agenda for studying inequality that is not only able to integrate these empirical developments, but which also challenges what has been taken for granted: that inequality just is, independently of context, time and observer. Instead, it proposes that along with its objective existence, inequality is a relational phenomenon subjectively experienced relative to a particular context. In five interconnected Sections, my dissertation challenges conventional views of how inequality looks, how it is seen, and what can be done about it, especially in developing countries. The study focuses on the ways in which inequality is perceived, and how it is perpetuated. After an introduction to the subject in Section I, Section II investigates how inadequate measurement perpetuates inequality, proposing a new indicator that shows that inequality is largely defined in the extreme ends of the income distribution. Section III examines the reproduction of inequality at the bottom, contrasting minimum wage policies over recent decades in Mexico with those of other countries in Latin America. In light of a political economy resistant to change, Section IV scrutinizes Mexican elites, exploring how inequality is perceived from the very top of the income distribution, how this affects policy-making and, subsequently, measured inequality levels. Section V concludes by outlining the theoretical and practical implications of my findings.

The aim of the presented work is to investigate possible improvement of scor- ing models prediction power in retail credit segment by using structural models estimating the future development of behavioral score. These models contain the information about past development of the behavioral score by parameters which take into account the sensitivity of clients' probability of default on in- dividual market and life changes. These parameters are estimated with Markov Chain Monte Carlo methods based on score history. Eight different types of struc- tural models were applied on the real data. The diversification measure of indivi- dual models is compared using the Gini coefficient. These structural models were compared with each other and also with the existing scoring model of the credit institution which provided the underlying data. 1