This thesis explores how economic performance, measured as real GDP-growth per capita, affects the vote share of parties which relies on radical ideological platforms. Using a fixed effect model with panel data, based on real electoral outcomes of 18 western European democracies, the result reveals an ambiguous reality. The overall conclusion implies that low growth rates benefit the electoral success of radical-right parties and holds for robustness checks. No solid evidence of the relationship is found on the radical-left side.The result further reveals that the individuals decision to vote radical is relative more affected by the ongoing business cycle trend between the elections rather than sudden changes close to the election day. The relationship also appears to be stronger among nations of southern Europe.

A presente tese é um trabalho de Ciência Política, com um grande esforço de estabelecer diálogos com outras disciplinas, como a Geografia, a Economia e o Marketing, no que diz respeito aos estudos eleitorais. Partindo do pressuposto de que distintos atores políticos possuem diferentes territórios eleitorais, formulamos a hipótese de que existe uma correlação entre a geograficidade e os resultados eleitorais. O termo geograficidade aqui não alude somente àquilo que é objeto da Geografia ou sobre a atividade específica do geógrafo, mas à diferenciação espacial, tendo em vista a relação existencial estabelecida entre o homem e seu território (espaço geográfico delimitado por relações de poder). O próprio título do trabalho, Eleições e Territórios, já evidencia sua abordagem transdisciplinar com a inclusão integrada de dois objetos distintos de estudos, que são considerados por diversas disciplinas. Observados esses elos entre a Ciência Política e tais disciplinas no que diz respeito à política, buscamos desenvolver um estudo mais aprofundado da relação entre eleições e territórios, considerando abordagens e instrumentos variados dessas disciplinas, o que, embora pautado em teorias políticas, destaca outros aspectos e espectros dos casos estudados. / This thesis is a study of Political Science, with a major effort to establish dialogues with other subjects as Geography, Economy, and Marketing, with regard to electoral studies. Assuming that different political actors have different electoral regions, we hypothesized that there is a correlation between geographicity and electoral results. The term \"geographicity\" here do not refers only to what is the object of Geography, or the specific activity of the geographer, but it is the spatial differentiation considering the relationship established between man and his territory (geographical space delimited by power relations). The title of the study, \"Elections and Territories\", already shows its transdisciplinary approach integrated through the union of two distinct objects of study, which are considered by many subjects. Observed these links between Political Science and such subjects concerning politics, we develop a deeper study of the relationship between elections and territories, considering a variety of approaches and tools of these subjects, which, although ruled by political theories, highlight other aspects of the cases studied.

In the literature, electoral accountability has been explored in many ways. Among those are the studies of economic voting examining to what degree government parties are held accountable for the state of the economy. By now, the studies have incorporated variables that reflect how clear is the chain of responsibility for the economic policies. Among those are national level variables, such as the clarity of responsibility index, and party level variables, such as the number of seats a party occupies in a government. This dissertation suggests that the responsibility for the government policies can be obscured by yet another party level variable - party novelty. I define party novelty as the quality that reflects the degree of change within a party in terms of its structure (mergers, splits, etc) and attributes (name, leader, and program) within one electoral cycle. I argue that party change obscures party identity and, thus, affects voters' ability to hold it accountable for the state of the economy. This study explores the concept of party novelty and its effects on voter's party preferences in various economic conditions. I construct the Party Novelty Database (1989, 1994, 1999, 2004, and 2009) and show that party novelty can be measured. Moreover, I demonstrate that party novelty varies in understandable ways, and, most importantly, that party novelty matters. Using the European Election Study and the Euromanifesto Project (1994, 1999, 2004, and 2009) I show that party novelty moderates economic voting, and this effect differs across types of party changes and the timing of change. / Political Science

本研究分析在2005年台北縣長選舉中，民眾是否以回顧性評價或是展望性評估來決定其投票對象。運用2006年「台灣選舉與民主化調查」的研究資料，分析民眾對於過去的家庭經濟情況、社會整體經濟狀況評價及施政表現；以及民眾對於未來家庭經濟、總體經濟狀況評估及主要政黨候選人展望性預期，是否影響其投票行為。 本研究發現：控制民眾的人口學背景以及其他政治態度之後，民眾認為過去整體經濟情況變差者，愈不傾向投給民進黨候選人。認為民進黨候選人愈具備執政能力者則愈傾向投給他。除此之外，民眾的政黨認同也具有重要的影響力。 從本研究的結果可以得知，民眾會運用理性的標準，判斷過過去總體經濟表現與未來哪位候選人較具執政能力而決定其投票對象。顯示執政者應該以總體經濟表現以及優秀的人才，才可以吸引選民持續的支持。 / In this Study, we employ individual-level survey data collected by ‘Taiwan’s Election and Democratization Study” in the 2005 Taipei Magistrate election to examine whether voters apply retrospective voting or prospective voting. We include variables such as personal well-being and national economics performance in the past ,candidate evaluation and expectations on future economic performance to see how these variables might affect their voting behavior. We demonstrate that, after controlling demographic variables and other political attitudes, when voters consider national economy is worse off, he/she voted against the incumbent party. However, voters give greater support to candidates of the incumbent party if he/she believes this candidate is competent to govern Taipei county. Additionally, party identification still exerts a powerful influence upon the individual vote decision. In the study, it is found that people decide their voting behavior is employ their rational calculations to examine sociotropic economic performance and candidate quality. Therefore, citizens are not fools, and the incumbent party has to realize the importance of governance and nominate qualified candidates to govern.

本文討論台灣縣市長選舉中，財政健全程度及收入支出決算狀況會如何影響選舉結果。以台灣21個縣市在1989年至2005年間共5屆的縣市長選舉為實證研究，探討縣市長選舉結果出現政黨輪替的可能因素，在經濟投票及政治意識形態的研究方向以外，就縣市政府財政表現的角度進行分析。實證模型採用Probit model進行估計，本文實證假定為：如果縣市長致力於提升財政健全或是提供符合選民需要的公共支出時，便不容易出現政黨輪替。由實證結果發現，在財政變數方面，選舉年財政健全度及選前一年歲出成長率提高時，有助於執政黨持續執政。然而，選前二年竊案發生率較高的縣市，反而使執政黨被輪替的機率提高。對於與總統相同政黨的縣市長而言，只有在面對較高的選前一年全國及地方失業時，容易受到拖累而出現政黨輪替。競選連任的縣市長具有優勢。桃竹苗區與中彰投區相較其他地區而言，顯然較容易出現政黨輪替。 / By the five campaigns from 1989 to 2005 of the 21 local jurisdictions (includes 16 counties and 5 townships) in Taiwan, this paper discusses how fiscal health, revenue and expenditure conditions affect the election results. We analysis that the possible factors that cause election results to appear party rotation. Except for economic voting and other political factors, we mainly examine if fiscal performance results in party rotation. We adopt the Probit model to setup our empirical model. Our hypothesis is that party rotation possibly happens if local officers do not devote in fiscal health or fiscal policy that voters need. For the fiscal variables, the empirical evidence shows that increasing fiscal health in election year and total expenditures growth rate in one year before election help incumbents persist in political power. However higher crime rate on burglaries and larcenies in two year before election will reduce the probability in party rotation. If the president and incumbents are the same party affiliation, incumbents will be implicated while national and local unemployment rate worsen. Besides, incumbents who run for reelection will be in dominant position. The jurisdictions in the northeast and middle areas of Taiwan compare to the others, it is more possible to occur party rotation.

經濟投票研究逐漸重視內因問題，國外的研究對於資料上和方法上均展開革新。相較之下，台灣的經濟投票研究受限於資料和方法，一直無法對認知性與資料性的內因問題進行有效的處理。資料的侷限不是一時能獲得改善的。但方法的反思和修正，正是研究者與方法對話讓方法實用的開始（黃紀 2000b）。處理經濟投票的二個內因問題，本文引用Menard（2010, 145-168）所提出的PALR方法。它保留路徑分析的優點，並讓類別變數能夠在路徑分析方法中被估計；同時結合Heckman因果效應模型的概念，亦能處理資料性的內因問題。 2012年總統選舉「九二共識」和「台灣共識」的交鋒，最後是由主打透過「九二共識」提振台灣經濟的執政黨獲勝。面對前一個任期表現差強人意的執政黨，選民沒有對它進行處罰，反倒再給它執政的機會。這樣的選舉結果讓人聯想是否還有其他因素影響經濟評估。透過PALR希望能一窺2012年總統選舉的全貌。 研究發現，2012年總統選舉存在社會前瞻經濟投票，卻不存在社會回溯經濟投票。政黨認同對社會回溯型經濟評估有直接影響，對社會前瞻型經濟評估則是有間接影響；而且對投票選擇的影響不只存在直接影響力，透過候選人形象評估和社會前瞻型經濟評估，政黨認同對於投票選擇也有相當的間接影響力。這表示選民不僅是有立場地對執政黨2008年到2012年的表現進行評估；對台灣未來的展望同樣存在政黨認同的影響。選民不僅有立場地評估代表二個「共識」的候選人，也有立場地看待台灣未來可能的發展。

在經濟與政治關係日益密切的今日，經濟表現是否會影響選舉結果已成為學者所關切的議題。目前西方國家，特別是美國，針對中央層級選舉經濟投票行為方面之研究已有高度共識，即經濟表現不佳，會對執政黨的選情不利。至於地方經濟表現對州長選舉有無影響的實證結果，則較為分歧。本文將經濟投票理論應用至台灣縣市長選舉，以台灣21個縣市在1989年至2001年4屆的縣市長選舉作為研究對象，探討致使地方首長選舉結果出現政黨輪替的可能變數，已彌補現有文獻僅考量政治層面、以及中央層級選舉之不足。 經由Probit模型估計後發現，地方失業率於選舉年惡化並不會影響縣市長選舉結果；反而是全國失業率對選舉結果有顯著影響。一般而言，總統政黨執政之縣市皆較非總統政黨執政之縣市不易發生政黨輪替，但若全國失業率於選前攀升，則前者所享有的相對優勢會因此而降低，連帶使其被輪替的機率提升。至於另一全國經濟指標物價膨脹率，則不如全國失業率有影響力，即使物價於選舉年上漲，總統政黨執政縣市出現政黨輪替的機率，亦未有隨之提升的態勢。 另外，在政治變數部分，現任者競選連任有利於現任執政政黨於選戰中勝出；府會是否同黨對選舉結果則無顯著影響。而本文依據邊際效用遞減法則提出的假設亦得到證實，即一黨連任屆數與其發生政黨輪替的機率呈正向關係。 最後，關於時間及縣市虛擬變數部分，本文發現，2001年14屆縣市長選舉發生政黨輪替的機率為4屆之首；而21縣市中，則以嘉義縣、屏東縣、新竹市、台中市、台中縣及基隆市，分居最易發生政黨輪替縣市的前五名。 / Substantial scholarly attention has been at the relationship between economic conditions and election outcomes in most western democratic countries, especially in the United States. Most Studies focused mainly on presidential or congressional elections have indicated a solid evidence that the worse economic conditions will be unfavorable to the incumbents to win the elections. Compared with those coincide conclusions, the results about economic voting in state elections are more discrepancy. This paper applies the economic voting theory to analyze the county magistrates and city mayors elections in Taiwan and uses a county-and city-level panel data from 1989 to 2001 to examine the possible factors which cause the party rotation of the county magistrates and city mayors elections. After estimating Probit model, the primary finding is that the local unemployment rates have no impact on election outcomes of the county magistrates and city mayors. Instead, the national unemployment rates have a significant effect on election outcomes. Generally, the probabilities of party rotation of the counties and cities ruled by the president’s party are lower than others. However, this advantage will be damaged as the national unemployment rate is higher in the election year than that in the previous year. As to the inflation rates, the other national economical index, are not as influential as the national unemployment rates. For all rising in the election year, the probabilities of party rotation won’t go up. Additionally, about the political variables, the incumbents are more likely to defeat the challengers and renew their term of office; it has no significant impact on election outcomes whether the incumbent governors and city or county councils are of the same party. The hypothesis which derives from the law of diminishing marginal utility is also proved. The longer the governing party rules the county or city, the higher the probability of party rotation for this county or city. Finally, the probability of party rotation of the 14th county magistrates and city mayors election is the highest from 1989 to 2001. Chiayi County, Pingdong County, Hsinchu City, Taichung City, Taichung County and Keelung City are the top five ones of the twenty-one counties and cities in terms of the frequency of party rotation.

How do individuals form their perceptions of economic matters? In this thesis I present three chapters which aim to demonstrate that individuals are influenced by their immediate environment in shaping their perceptions of the national economy or their economic status. In Chapter 1, I use cohabitation between young adults and their parents as a proxy for the economic difficulties experienced by individuals aged between 18 to 34. Using survey data from 32 countries, I show that parents' daily exposure to their adult child's difficulties negatively influences their perception of the national economy as well as the performance of the government. Chapter 2 deals with perceptions of an individual's economic status. I show that rising housing prices lead to economic anxiety among renters because of a fear of gentrification in their locality, as well as economic barriers to become home owner. My results are drawn from two survey experiments, one conducted in the United States, the other in Montreal. Finally, in Chapter 3 I take a more traditional approach to investigate the link between the local economic context and perceptions of the national economy. For this last chapter, I propose to reconsider the local economic context by taking into account not only the unemployment rate at the place of residence but also the place where adult individuals spend a large part of their time, i.e. their work. My results show that while the level of local unemployment is indeed correlated with the perception of the national economy, taking into account the average unemployment rate at the destination weakens this correlation. On the other hand, a global measure that takes into account the residential area as well as where individual's use to go on a daily basis is better correlated with the perception of the national economy. / Comment les individus forment-ils leur perception des questions économiques ? Dans cette thèse, je présente trois chapitres qui visent à démontrer que les individus sont influencés par leur environnement immédiat dans la formation de leurs perceptions de l'économie nationale ou de leur statut économique. Dans le chapitre 1, j'utilise la cohabitation entre les jeunes adultes et leurs parents comme indicateur des difficultés économiques rencontrées par les individus âgés de 18 à 34 ans. À l'aide de données de sondages provenant de 32 pays, je montre que l'exposition quotidienne des parents aux difficultés de leur enfant adulte influence négativement leur perception de l'économie nationale ainsi que leur perception de la performance du gouvernement en place. Le chapitre 2 traite des perceptions de la situation économique d'un individu. Je montre que la hausse des prix de l'immobilier entraîne une anxiété économique chez les locataires en raison de la crainte d'un embourgeoisement dans leur localité, ainsi que des obstacles économiques à l'accession à la propriété. Mes résultats sont tirés de deux expériences par sondage, l'une menée aux États-Unis, l'autre à Montréal. Enfin, dans le chapitre 3, j'adopte une approche plus traditionnelle pour étudier le lien entre le contexte économique local et les perceptions de l'économie nationale. Pour ce dernier chapitre, je propose de reconsidérer le contexte économique local en prenant en compte non seulement le taux de chômage du lieu de résidence mais aussi celui du lieu où les individus adultes passent une grande partie de leur temps, c'est-à-dire leur travail. Mes résultats montrent que si le niveau de chômage local est effectivement corrélé avec la perception de l'économie nationale, la prise en compte du taux de chômage moyen à la destination affaiblit cette corrélation. En revanche, une mesure globale prenant en compte la zone résidentielle ainsi que le lieu de déplacement quotidien des individus est mieux corrélée avec la perception de l'économie nationale.

Essay 1: (with Henrik Jordahl and Panu Poutvaara) We present and test a theory of prospective and retrospective pocketbook voting. Focusing on two large reforms in Sweden, we establish a causal chain from policies to sizeable individual gains and losses and then to voting. The Social Democrats proposed budget cuts affecting parents with young children before the 1994 election, but made generous promises to the same group before the 1998 election. Since parents with older children were largely unaffected we use a difference-in-differences strategy for identification. We find clear evidence of prospective pocketbook voting. Voters respond to campaign promises but not to the later implementation of the reforms. / Essay 2: This essay presents a detailed analysis of voters' response to municipality and regional level unemployment and economic growth, in Swedish general elections from 1985 to 2002, using data on 284 municipalities and 9 regions. The preferred specification suggests that an increase in regional growth or a reduction in regional unemployment by one percentage point is associated with an increase in the support for the national government by about 0.6 and 1.0 percentage points. Changes in unemployment and growth at the municipality level seem to have muchsmaller effects on government support. / Essay 3: One prediction from cognitive dissonance theory is that the act of voting makes people more positive toward the party or candidate they have voted for. Following Mullainathan and Washington (2008), I test this prediction by using exogenous variation in turnout provided by the voting age restriction. I improve on previous studies by investigating political attitudes, measured just before elections, when they are highly predictive of voting. In contrast to earlier studies I find no effect of voting on political attitudes. This result holds for a variety of political attitudes and data from both Sweden and the United States. / Essay4: (with Niclas Berggren and Henrik Jordahl) We conduct an extensive robustness analysis of the relationship between trust and growth by investigating a later time period and a bigger sample than in previous studies. In addition to robustness tests that focus on model uncertainty, we systematize the investigation of outlier influence on the results by using the robust estimation technique Least Trimmed Squares. We find that when outliers (especially China) are removed, the trust-growth relationship is no longer robust. On average, the trust coefficient is half as large as in previous findings.

Au cours des années 1970 et 1980, quelques politologues se sont lancés dans la conception de formules prévisionnelles basées sur des indicateurs politico-économiques afin d’anticiper les résultats de courses électorales. Cette discipline s’est considérablement développée aux États-Unis où l’on compte actuellement un grand nombre de modèles ayant pour objectif de prédire l’issue des élections au Congrès ou le sort des candidats à la présidence. Bien qu’un certain nombre de modèles aient vu le jour pour la France et le Royaume-Uni au cours des dernières années, le Canada, à l’instar de la majorité des démocraties, n’a reçu jusqu’à maintenant que bien peu d’attention. Ce mémoire vise par conséquent à développer un modèle ancré dans une théorie du vote capable de prédire suffisamment à l’avance la part du vote populaire récolté par la formation ministérielle lors des scrutins fédéraux canadiens. Pour ce faire, nous avons procédé à l’élaboration d’une formule de régression par la méthode des moindres carrés ordinaire exploitant les résultats des élections qui ont pris place depuis 1953. Cette formule est composée de cinq variables, à savoir la différence entre les taux de chômage canadien et américain trois mois avant la tenue du vote, le logarithme du nombre de mois passés au pouvoir par le parti sortant, un indicateur dichotomique concernant la substitution du premier ministre à proximité d’une élection, le nombre d’années d’expérience politique du premier ministre par rapport à son (sa) principal(e) adversaire et un facteur relatif à l’origine provinciale des leaders. / During the 1970s and 1980s, a small group of political scientists started to develop forecasting equations based on political and economic indicators to predict election results. Election forecasting is now a thriving discipline in the United States, where a large number of different models are being used to forecast the outcome of congressional elections or the fate of presidential candidates. Although forecasting models have been developed for France and the United Kingdom over the past years, Canada, like most other democracies, has received very little attention. The goal of this thesis is to develop a theoretically-driven model that can be used to predict the popular vote share of the incumbent party in Canadian federal elections with sufficient lead time. To this end, we devised an ordinary least squares regression model using the results of elections going back to 1953. This model is composed of five variables; the difference between the unemployment rates in Canada and the United States three months before the vote, the natural logarithm of the number of consecutive months the incumbent party has been in office, a dichotomous variable related to the substitution of the Prime Minister near an election, the number of years of political experience gained by the Prime Minister in relation to his/her main opponent, and a factor related to the province of origin of party leaders.