During the past twentieth century, which factors have given rise to the decline of third-party vote shares in American Politics? Hirano and Snyder (2007)’s paper provides a new perspective on this issue. Their research consists mainly of two parts, each of which deals with one baseline model addressing the issue. Their conclusions are pretty clear: contrary to the previous held opinions, the reason should be attributed to the underlying political-economy factors, say Democratic Party’s embrace to the New Deal as a response to the Great Depression after 1929; on the other hand, other systemic affluences like electorate reforms have only trivial contributions, if not complete nothing to do with this trend.
Direct Primary and Australian Ballot:
To evaluate the theory that changes in electoral laws impact candidates’ strategies as well as the voters’ support for third-parties, the authors test two hypotheses. In the first estimation (the first model in Table 1), the result shows that when controlling the fixed effect of third-party support in a given state, there is no significant correlation between either direct primary, or Australian ballot and the third-party support. The second model investigates the relation between the introduction of direct primary and the change of competitors’ realignment with majority parties. The result is similar to the previous one: no statistical significance on nationwide samples. Here come two questions.
Question 1. Instead of calculating the variation of the third-party vote shares from the previous elections, the authors choose the third-party’s vote shares as dependent variable. I am not sure whether a new estimation will show any significant correlation between independent variables and the first-order variation of the dependent variable.
Question 2. The conclusion that there is no correlation between electorate law and the rise or decline of third- party candidates may be subject to contexts. There could be some causal effects between them elsewhere. For example, during Taiwan’s 2000 presidential election, the decision over who’s going to be KMT (the ever ruling party in Taiwan until 2000)’s candidate was through a caucus rather than a direct primary. There was severe divergence on nomination and as a result of that the more popular competitor Song, while not being appreciated by the party’s elites, quitted from KMT and campaigned as an independent candidate. This split of KMT has far-reaching impacts in the long-term trend of political evolution in Taiwan, in that it directly gave rise to the weak-margin victory of Democratic Progress Party in the general election. (The share of population vote for DPP’s candidate Chen, Song and KMT’s candidate Lian are 39.9%, 36.8% and 23.1% respectively.)
Democratic Party’s Absorption
In this part, the authors utilize a skillful method to investigate the relation between the switch of Democratic Party’s position during and after the New Deal age and the decline of third-party support. The relatively small and insignificant difference between the coefficients on pre-period Democratic and Pre-period left third-party implies that there does hardly make any difference between the supports from previously left-party voters and previously Democrats. In other words, almost all votes for the left third-party before New Deal age went to Democratic Party after 1936.
Question 1. The dramatic fall of third-party popularity started from the late 1910’s, more than 10 years before the Great Depression and New Deal (according to Figure1), how shall we account for this lag?
Question 2. in the second half of the twentieth century, the two majority party’s’ positions had varied frequently. For example, the economic policy of Nixon administration had been more regulatory and progressive, while Democratic president Clinton’s economic policy was more pro free market. During 1960’s presidential campaign, the Democratic candidate Kennedy adopted a position more conservative than Republican in foreign policy, challenging that the Republican government had been too soft in the cold war. The foreign policies of both Kennedy and Johnson’s governments during the Vietnam War period were also far from a liberal or left position. If the variations of positions such as the above are not trivial (some of them are also far-reaching), we might expect a corresponding fluctuation of the third-party popularity in American Politics. Therefore, an extension of this research based on updated data will be interesting.
 

 

又是选举投票率，这两个礼拜尽跟这个问题干上了。这次是说campaign对投票的影响。花了两个小时读完哈佛的美女教授Sunshine Hillygus（看这个名字起的就不同凡响）的论文。个人看法是，人长得漂亮和文章的好坏大概没有什么相关性，比如这篇发表在2005年Journal of Politics上的论文：你可以说是strong arguments, 但是大概很难说是convincing arguments.

 

 

 

 

 

 

 

 

 

 

这是第一周美国政治的作业。我知道写得烂；写的时候就一边咬牙切齿地骂。但也不能完全怪我，因为第一阅读的材料是指定的；第二规定作业不得超过一页纸的篇幅。按照我一贯罗嗦的行文风格，能够在一页纸涉及两三个问题已属难能。美国政治这个学科里的情况经常是，无论什么题目，都已经有n多人在做，而且挖掘的深度（至少从所用的技术层面来说）常常超乎你的预期。好比挖石油，你还在地表勘探，人家已经先打到了地下一，二千米（我不知道究竟应该打到多少米）。当然这样能不能保证打出油是另一回事。

 

Following Riker and Ordeshook (1968), a lot of rational choice models of turnout have adopted the calculus of voting model (Aldrich, 1993), with the form of P(i)B>C. Given the voting cost level C as constant, the focus goes to the generalization of P(i)B. Since the value of benefit B is reasonably bounded, the probability of a vote being pivotal must be significantly bigger than 0. This case is more plausible within a strategic interaction environment rather than within the conventional decision-theoretic framework. Palfrey and Rosenthal (1983) finds that the multi-equilibrium exits in such generalized game situation, which implies that the election must be very competitive for the purpose of achieving the high turnout (thus higher pivotal probabilities). Since here the P(i) is voter’s subjective probability, we should not neglect the effect of uncertainty upon turnout. As Palfrey and Rosenthal (1985)shows, with uncertainty, when the population magnitude increases, the turnout converges to zero even if the perceived competition is equivalent. Feddersen and Pesondorfer(1996, 1999) adopt another perspective. They assume that C is zero and thus change the question from “why voting” to “why abstaining”. The gist is that there are a fraction of independents (non-partisanship) who’re uninformed. In other words, P(i)B>C may not hold because B is negative for the uninformed. The turnout will be low when uncertainty makes it ex ante optimal for them to abstain. It is doubtful that Feddersen and Pesondorfer have accurately addressed the paradox of voting since they eliminate the cost so that the problem itself does not remain for voters with given preference. Similarly, the treatment of adding a consumption benefit D may not be promising, in the sense that the original paradox has been cancelled. As John Aldrich (1993) puts it, by adding a term D, rational choice models “add nothing to the explanation of how preferences shape actions”.

Now we turn to empirical literatures. The rational choice models mentioned above imply several relevant points. The implications include, but not limited to: competitive election deduces higher turnout; the large electorate population reduces turnout; uninformed voters tend to abstain more often and education (which makes voters more informed) improves turnout. Most predictions are supported by empirical analysis on turnout. However we must first discriminate the effects of individual influences from contextual factors upon turnout. Leighley and Nagler(1992) argue that some individual influences vary between different years. For example, they find that from 1972 to 1984’s presidential election, the influence of being black upon voting increase large and significantly (from 0.112 to 0.287, p<.05); simultaneously women have become more likely to vote relative to men, other things being equal. The authors seem to suggest that such changes are brought by the evolution of institutions, while their models also show that the influences of five contextual factors upon turnout are NOT significant, with the exception of closing, and party competition in non-southern regions (which is partially consistent with the implication drawn from Palfrey and Rosenthal’s model). We are not sure whether these significant influences are really at individual level, or that they are determined by other underlying systemic/institutional factors. Lassen (2005) reevaluates Wolfinger and Rosenstone (1980)’s implication which emphasizes the importance of being informed. Besides, other previous researches emphasize the more contextual influences, like the feeling of decreased political efficacy (Shaffer 1981), and the negative effect of economic disadvantage on voting, as opposed to the previously held mobilization effect (Rosenstone 1982). Different empirical literatures have suggested that both individual and systemic factors are important, or the marginal effects of individual factors are context-dependent. Thus we might need a unified theory of voting behavior which incorporates individual choice and context influences (including group identification, mobilization, etc.), and uncertainty.

 

在度过了相对舒适的一个礼拜后，明天就要开始上课。虽说一直做学生，明天的课却是差不多一年多来我第一次正儿八经的听课！OMG，想想这一年来我都做了些什么。考试，申请，然后跑到这里来，大概算是transaction cost吧。

星期二要把人上傻了：9:00－12:00，经济增长；12:00－2:00，Math Review；4:00－6:00，美国政治。。

还是给自己一点鞭策吧。吃喝玩乐的事情放在一边。给自己订个目标。两门核心课程的理想状态是看完4本书，20篇论文，当然如果加上notes…。数学当然是希望按部就班地看。鲁丁的书不太适合自学，读起来有点折磨，但写得真漂亮。

希望把目标定高一点，就算打个折扣也还能过得去。我们在弗吉尼亚的好邻居Chuck老伯说，早起的鸟有虫吃；就算吃不到虫子，学个松鼠拣点松果吃也行啊。