過去對於國會中政黨的研究，多限於探討政黨體制的制度面，或偏重於從國會議員的角度闡述政黨對國會議員行為的影響力，本研究則是試圖突顯政黨身為動態立法過程中的重要政治行動者，必須採取各種行動與策略，整合自身成員的力量，與他黨進行立法聯合互動，以達成特定目標。 本研究主張政黨在進行各種立法聯合時，乃以其選民基礎及其在特定議題上的立場為主要考量，同時在面對實際立法情境時，亦須將因政黨本身資源有限所導致的立法限制，及政黨為表現善意或獲取自身利益而與其他政黨進行立法交換等因素一併加以考量，如此方能更精準地推估立法聯合的可能情況，並完整地描繪出政黨在多元動態的立法過程中所扮演的政治行動者角色。研究結果顯示，政黨在各類法案上確實受上述因素影響而出現不同的立法聯合狀態，如：大陸定位問題類法案即因其涉及的選民基礎及特定議題相當明確，而每呈國新兩黨與民進黨對立的局面。 本研究以政黨的選民基礎與在特定議題上的立場此兩項政黨的基本考量，連結政黨在選舉中的競爭態勢與其在國會中的立法聯合表現，彌補過去研究較少論及國會中政黨競爭的缺憾，有助於對政黨競爭互動的實際情況進行更全面、連貫且深入的瞭解與詮釋。 / Most previous studies on congressional parties concentrate either on institutions of party system, or on the effects that a party could have on individual legislators' behavior, predominantly from the legislators' aspect. However, this research aims to emphasize that the parties are the important political actors in the dynamic legislative process, during which they must adopt some valid actions and strategies, integrate their members' strength, and form legislative coalitions with other parties in order to reach particular political goals. When parties form legislative coalitions, they must consider their bases of constituents and their positions in particular issues. At the same time, they must consider the legislative restrictions caused by their limited resources and the possible legislative exchange with other parties in order to show their benevolence or to accomplish their self-interest goals. The author strongly believes that this is an appropriate method through which people may get a whole picture of the role that the parties play in the complex and dynamic legislative process. The research findings show that the bill categories influenced by the factors stated above as to have different styles of legislative coalitions. For instance, when a bill concerning the status of China is raised, the legislative coalition of the KMT, aligning with the New Party, against the DPP is ineluctable, because their bases of constituents and their positions in particular issues are clear and definite. This research is based on two major considerations of the parties：their bases of constituents and their positions in particular issues. The research furthermore uses these two considerations to connect competitive circumstances of parties in the elections and the legislative coalitions in the legislature. It may help to unravel and interpret the real situations of party competition completely, coherently, and profoundly.

Submitted by Ronildo Prado (ronisp@ufscar.br) on 2017-08-21T14:17:23Z No. of bitstreams: 1 TeseSK.pdf: 4983553 bytes, checksum: 07e162ddc09c2b9535ea3eb301215708 (MD5) / Approved for entry into archive by Ronildo Prado (ronisp@ufscar.br) on 2017-08-21T14:17:33Z (GMT) No. of bitstreams: 1 TeseSK.pdf: 4983553 bytes, checksum: 07e162ddc09c2b9535ea3eb301215708 (MD5) / Approved for entry into archive by Ronildo Prado (ronisp@ufscar.br) on 2017-08-21T14:17:40Z (GMT) No. of bitstreams: 1 TeseSK.pdf: 4983553 bytes, checksum: 07e162ddc09c2b9535ea3eb301215708 (MD5) / Made available in DSpace on 2017-08-21T14:17:47Z (GMT). No. of bitstreams: 1 TeseSK.pdf: 4983553 bytes, checksum: 07e162ddc09c2b9535ea3eb301215708 (MD5) Previous issue date: 2017-03-28 / Não recebi financiamento / This thesis addresses the relationship between mallaportionment and legislative behavior, precisely, the effects of overrepresentation on legislative outcomes, It is well established that overrepresented subnational units benefit disproportionately from resources distributed by the central government. Explanations for this fact are based on a model of legislative bargaining over distributive projects. Therefore, the focus of this investigation is the legislative coalitions. However, this model assumes legislators who are elected in single-member districts, where there is a one-to-one correspondence between the legislator and the district. In these systems, defining the interest of the district is straightforward and the legislator can easily claim credit for defending it. The same is not true in multi-member districts where interests that successfully seek representation are varied and legislators elected from the same district can claim to represent different constituencies. So, the goal of this research is to prove the applicability of the legislative bargaining mechanism in a multi-member district electoral system. The central assumption is that in multi member electoral districts, legislative coalitions cut-across the apportionment cleavage and wash out the impact that overrepresentation might have had under a single-member electoral structure. I test this claim by comparing legislative coalitions in the Brazilian Chamber of Deputies and the Senate between 1989 and 2014 and examined the pattern of transferences of federal funds of the central government to subnational units of the country. While both Brazilian houses are symmetric in their power, elections for the former are based on proportional representation electoral systems in large districts and for the later on majority rules in one or two-member districts. I performed three types of descriptive statistical analyses: voting unit analysis; dyadic analysis and winning coalitions analysis. The results showed that the relation between the higher volume of fiscal transferences and overrepresentation is stronger in the Senate. In turn, it was demonstrated that, although the effects of overrepresentation on the coalitions is also higher in the Senate; in general, the impact of malapportionment in the congress is discreet. Thus, denying the expectations disseminated by the literature, the legislative bargain for distributive projects is not, necessarily, the underlying mechanism to distortion in the discretionary distribution of social resources which favors the overrepresented units. / A tese trata do impacto da sobrerrepresentação no comportamento legislativo com o foco na formação de coalizões. Está bem estabelecido que as unidades subnacionais sobrerrepresentadas são desproporcionalmente beneficiadas por recursos distribuídos pelo governo central. As explicações para este fato são baseadas no modelo distributivista de barganhas legislativas. Entretanto, o modelo considera legisladoras e legisladores eleitos em distritos uninominais. O objetivo desta pesquisa foi testar a aplicabilidade do modelo de barganhas legislativas por projetos distributivista em um sistema eleitoral de distritos plurinominais e; analisar o peso da desproporcionalidade de representação distrital nas coalizões legislativas. A hipótese central foi que os efeitos da sobrerrepresentação distrital dependem do tipo de sistema eleitoral. Em sistemas eleitorais de distritos uninominais há a correspondência de um para um entre legisladora (legislador) e o distrito e a definição dos interesses do distrito é, relativamente, mais clara. Nesse caso, a(o) parlamentar pode facilmente reivindicar o crédito por defendê-los e esperar obter reconhecimento de seu distrito eleitoral por causa disso, isto é, obter retorno em forma de votos. O mesmo não é verdadeiro para sistemas eleitorais de distritos plurinominais, onde os interesses bem sucedidos na busca por representação são variados e as(os) legisladores que se elegeram por um mesmo distrito podem reivindicar a representação de diferentes círculos eleitorais. Sistemas eleitorais de distritos plurinominais incentivam a formação de coalizões que atravessam distritos eleitorais. Esta condição rompe com a representação enraizada nos limites dos distritos eleitorais e atenua o impacto que a sobrerrepresentação teria sob uma estrutura eleitoral de um só membro. Testei esta afirmação comparando as coalizões legislativas na Câmara e no Senado do Brasil entre 1989 e 2014 e, também, examinando o padrão das transferências de fundos federais do governo central para as unidades subnacionais do país. Ambas as casas legislativas brasileiras são simétricas quanto ao poder, mas as eleições para a primeira são baseadas em sistema de representação proporcional em distritos de grande magnitude e para a segunda sobre a regra majoritária em distritos de um ou dois membros. Realizei três tipos de análises estatísticas descritivas: análise da coesão do voto; análise de votos iguais de duplas de parlamentares e análise da participação em coalizões vitoriosas. A hipótese secundária é que legisladoras (legisladores) de distritos sobrerrepresentados têm menor propensão a votar conjuntamente e são menos cruciais para o resultado das votações legislativas, do que quando elas e eles são analisados levando em conta outros atributos, tais como filiação partidária e coalizão do governo. Os resultados mostraram que no Senado a relação entre maior volume de transferências fiscais e a sobrerrepresentação é mais forte. Mas, embora no geral, os efeitos da sobrerrepresentação sobre as coalizões seja também maior no Senado, o impacto da representação desproporcional nas votações nas duas casas legislativas é discreto em relação aos outros fatores examinados. Entre eles a participação na coalizão do governo se destaca como determinante do comportamento legislativo nas votações. Portanto, contrariando as expectativas divulgadas na literatura, a barganha legislativa por projetos distributivos não é o mecanismo subjacente à distorção na distribuição discricionária de recursos fiscais que favorece unidades sobrerrepresentadas.