Dynamics of Voting Credentials in Municipal Elections: Comparative
Study of Kaohsiung and Chicago City Councils
Herlin Chien PhD student
Institute of Political Science National Sun Yat-sen University Tel: 001-886-7-5252000-5551 Fax: 001-886-7-5250053
E-mail: chien888@hotmail.com
Abstract
Study of voting behavior at congressional and senatorial levels, especially in North American region, has been extensively analyzed, within which research results often suggest visible causal relationships between incumbency, racial factor, campaign expenditure, party support, etc, with vote share of candidates. Surprisingly among the broad voting literature, few have derived their potential generalizable findings on the basic analytical unit of local election which shall be invariably the most numerous and frequent type of elections in any existing democratic system. Furthermore, before any generalization of voting behavior, if any, could be drawn, different political systems with various culture bonds shall be carefully looked into.
This paper as a result intends to tackle on these research gaps via two angles: firstly by quantitatively examing voting credentials in city council election turnout for Kaohsiung, the second biggest city located in southern Taiwan with a population of 1,5 million; secondly, by comparing the findings of Kaohsiung with the existing researches on Chicago city council electoral turnout.
Data set for the analysis of Kaohsiung city council voting behavior will be constructed within the time period dating from 1981, the first city council public election after the elevation of Kaohsiung city to the status of Special Municipality, to the year of 2002, marked the 6th election of Kaohsiung city councilor. Variables, which are claimed by previous scholars to have substantial impact on vote share of candidates, such as incumbency, education level, ethnic factor or even sex will be respectively included. Lastly, comparative analysis will be drawn between Kaohsiung City Council voting behavior and its counterpart in Chicago.
Keywords: incumbency, Kaohsiung, Chicago, city council, election, voting behavior
* This paper is prepared for the Democratization, Globalization, and Role of City Council conference held at NSYSU in Kaohsiung city, December 3-4th, 2004, specially delegating to the celebration of 25th anniversary Kaohsiung City Special Municipality Status.
Study of voting behavior at congressional and senatorial level, especially in North American region, has been extensively analyzed, within which research results often suggest visible causal relationships between incumbency (Abramowitz 1991; Cox and Morgenstern 1993), racial factor (Lieske and Hillard 1984; Inglot and Pelissero 1993), party affiliation (Arrington 1978), campaign expenditure (Giles 1985), media support (Graber 1990; 1993), face-to-face political mobilization (Gerber and Green 2000), etc, with vote share of candidates.
A. Incumbency Advantage
Incumbent members of the United States Congress have, according to the quantitative analysis of 24 U.S. states over the period from 1970 to 1986 done by Cox and Morgenstern, been suggested to enjoy electoral advantages over others. Indeed, advantages incumbents have benefited from include the franking privilege1 (Mayhew 1974) to the opportunity to perform casework for constituents (Fiorina 1977) to the use by voters of incumbency as a cue (Ferejohn 1977). Specifically speaking, the advantage of incumbency is rooted in the resources of office, namely official, financial, and reputational. While Holbrook and Tidmarch emphasize the impact of incumbency on electoral turnout, they further hypothesize that occupancy of party and committee leadership positions can be translated into an increased incumbency advantage in state legislative elections.2 In their model, it is statically supported that party leaders and committee chairs have a higher margin of victory than other incumbents in their state election.
At municipal level, the pioneer researches concerning incumbency are works of Prewitt3, in which research has shown that the single most important factor in predicting candidates’ vote 1 On November 8, 1775, The American Continental Congress authorized franking privileges to its members as a
means of informing their constituents. The first U.S. Congress enacted a franking law in 1789. In another word, back in 18th century, members of US Congress can simply use signature instead of postage on mass mailing. Nowadays, franking privilege has been modified significantly; beginning with FY 1999, Members of Congress may use any portion of their official budget for mailings.
2 Thomas M. Holbrook. 1993. “The Effects of leadership Positions on Votes for Incumbents in State Legislative
Elections”. Political Research Quarterly 46:897
3 Prewitt, Kenneth. 1970. The Recruitment of Political Leaders: A Study of Citizen-Politicians. New York: The
share in city council elections is incumbency. Furthermore, a recent survey by the International City Management Association have showed that 84% of city council incumbents were successful in their reelection attempts (DeSantis and Renner 1994, 40), while incumbent reelection success also seems to have increased since the 1960s (Karnig and Walter 1977, 66). Krebs, the Chicago election expert, invariably confirms the positive and significant relationship between political experience and candidates’ percentage of the vote.(Krebs 1998)
B. Racial Factor
In land such as the United States, where racial or ethnic diversity is highly notable, election will be inevitably colored by ethnicity. Prior research on voting behavior, especially in local election suggests the firm tendency of voters to cast ballots for candidates of their own racial group (Lieske and Hillard 1984; Vanderleeuw 1990; Inglot and Pelissero 1993). The alleged phenomenon is further elaborated through two social and political settings. On the one hand, a positive relationship between the socioeconomic position of white voters and white voter support for black candidates as noted by numerous researches. Among all, Hahn and Almy, by observing elections in Los Angeles, suggests an inverse relationship between socioeconomic status and voter’s support for black liberal candidates in biracial politics. “We are moving from “ethnic politics” to “class politics”4 A possibility of a black and blue collar white political coalition in urban election based on mutual economic benefit is emerging.
On the other hand, partisanship in local election is perceived to play a minimum role. Even if where partisan attachments are relatively strong, race can be an even stronger determinant of voting behavior in local election (Thompson 1984). As concerned with why racial factor is more visible in urban election, rather than national level, Williams provides an explanation: while
national politics has generally been dominated by distributive economic issues, urban politics has been more preoccupied with the distribution of spatial turf, social access, and cultural life styles.5
C. Party Affiliation
In 1952, the leading scholar of Michigan School, Angus Campbell firstly delineates the importance of party identification in relation to voting turnout.6 What has been extensively cited is later the “funnel of causality” (see Fig. 1) model contributed in The American Voter which characterizes partisan orientation as a long term critical psychological cause in influencing voting behavior whereas candidate orientation being the short term cause. The model, explained by R. G. Niemi and H. F. Weisberg, as well puts an essence on time dimension. Indeed, the funnel itself is the “time”. As series of events happen along the way from the open-end of the funnel to the end, i.e. the election date approaches, voter’s choices are undergone a multiple layers of struggles. Although party affiliation or the concept of “unmoved mover” by and large is conceived of as a highly stable orientation, performance of candidates, political sensitive events, etc. will potentially cause a change-over in voter’s election. Especially in modern era where information flow is unbounded and party loyal is less glorified, it has been pointed out that significant numbers of people report changes in their party identification over short periods of time (Fiorina 1981).
D. Campaign Expenditure/Media Support/Face to Face Mobilization
5 Williams, Oliver. 1967. “Life-style values and Political Decentralization in Metropolitan Areas.” Social Science Quarterly 48:299-310.
Although in some cases, campaign expenditure is found to be a less important factor than incumbency or racial factor in voting share, the general positive relation demonstrated between the amount of money candidates spend and the according increase of voting share in election cannot be neglected. Giles and Pritchard also confirm the above findings and support by examing the case of Florida Election of House of Representatives from 1972 to 1980, in which expenditure is found to influence the electoral outcome. As regards to media support and its effect on voting share, Graber in Mass Media and American Politics7 offers an interesting interpretation. The author
unfolds the fog initially by making an analogy between the oracle in ancient Greeks’ Apollo temple and the voice of the people as the voice of God - vox populi, vox dei, in modern democracies. That is in the oracle’s language, ambiguity is always unsolved. To the same token, while voices of voters convey through their votes are seen nowadays as ‘truth’, yet voters are easily manipulated through different interpretations of media coverage; therefore convey the different ‘truth’. Another concern is the favor media coverage often does for incumbents. Inevitably, incumbents would naturally have more news coverage than challengers who are not incumbents. As Table 1-1 depicts, in political attributes, 49 percent of news are for incumbent while only 4 percent given to challengers. This inadequate election news is hard to be avoided since more resources are at the hand of incumbents.
Table 1-1
Themes Incumbent news Challenger news
Campaign organization 22 27
Personal characteristics 10 13
Political attributes 49 4
Issues/Ideology/Group ties 29 12
Source: Peter Clarke and Susan H. Evans, Covering campaigns: Journalism in Congressional Elections (Stanford, Calif.,: Stanford University Press, 1983), 61.
Lastly, face-to-face political mobilization is thought to demonstrate a positive relation with the voter turnout. Gerber and Green from Yale University in 2000 successfully prove the proposal
that declining turnout rates in the United States since the 1960s8 has been the result of decline of personal mobilization, including personal canvassing, telephone calls, and direct mail. Specifically speaking, the experiment they designed studied approximately 30,000 registered voters in New Haven, Connecticut and built upon the hypothesis that personal canvassing mobilizes voters more effectively than other modes of contact that have taken its place. (see Table 1-2) The carefully designed experiment procedure is carried each Saturday and Sunday for our weeks before the election, during which canvassers were sent to contact randomly selected registered voters and be communicated by their mother language, English or Spanish.9 What is worth mentioned is the technique they utilized. Upon consultation with experts of campaign, the messages encouraging people to vote are delivered by nonpartisan appeals. The conversation would begin with the following opening: “Hi. My name is ___. I’m part of Vote New Haven ’98, a nonpartisan group working together with the League of Women Voters to encourage people to vote. I just wanted to remind you that the elections are being held this year on November 3d.” In the findings of the experiment above, Gerber and Green illustrate the fact that face-to-face interaction dramatically increases the chance that voters will go to the polls. If applying such observation into voting share, although whether the same effect can be feasible in voting share is unclear, the former case study case could provide an initial step to trigger further regarding researches on how personal canvassing affect voting behavior and consequently increase the voting share of designated candidates. .
Table 1-2 Results of Previous Voter Mobilization Experiments10
Study Date Type of Election Number of Treatment Effect on
8 Not only a general decrease of voting turnout since 1960s in the US, decline in size of nonpartisan organizations has
also been noted. Putnam found that in the mid-1960s, 2.4 of every 1,000 women over the age of 20 belonged to the League of Women Voters, compared to .79 in 1998.
9 New Haven has a substantial minority population and a significant proportion of non-English speakers. More than
half the canvassers were African American or fluent in Spanish and when possible they were matched to the racial and ethnic composition of the neighborhoods they walked.
Subjects Turnout*
Gosnell (1927) 1924 Presidential 3,969 Mail +1%
Gosnell (1927) 1925 Mayoral 3,676 Mail +9%
Eldersveld (1956) 1953 Municipal 41 43 Canvass Mail +42% +26% Eldersveld (1956) 1954 Municipal 276 268 220 Canvass Mail Telephone +20% +4% +18% Adams and Smith
(1980)
1980 Special election,
City Council
2,650 Telephone +9%
Miller, Bositis,
and Baer (1981) 1981 Primary 7980
81 Canvass Mail Telephone +21% +19% +15%
Note: The subjects in all experiments were registered voters. *These are the effects given in the tables of the research report. They are not adjusted for contact rates.
II.
Taiwan Voting Credentials
Differing from the observation of American voting behavior, electoral study of Taiwan, although mostly still focuses on national election level, collectively suggests Taiwanese voters are to vote according to firstly candidate orientation (55%), issues orientation (30%) and lastly party orientation (15% yet still under further debate)11
A.
Incumbency AdvantageContrary to voting credentials study in the United States where a sizable researches on incumbency advantage can be found, few Taiwan electoral literatures concerns the issue of incumbency advantage and its relationship with voting behavior.12 Therefore the effect of incumbency on voting share is not confirmed in Taiwan yet. This paper will in the later section conduct a quantitative study on that. Instead, Taiwan voting behavior study is elaborated on a grand scale of candidate orientation, obviously including past experience and accomplishment of candidates, education level, whether he or she is incumbent or not, etc. In fact, since 1990s after 11 Chen, Yi-Yen. 1986. “Review of Voting Behavior in Taiwan”. The Thought and Word 23:557
12 In fact, none incumbency advantage literature in Taiwan electoral study has been found, either in Chinese or
parties in Taiwan exhibited tendency of re-structure and re-grouping, Taiwanese voters in modern era have reconsidered the route of issues orientation or candidate preference, rather than sticking the party identification router as preoccupied previously.13 According to a survey conducted by National Science Council, 26.29% of voters in the chosen sample vote for specific candidate because of his or her past experience and accomplishment while 13.81% answered vote for the good image candidates has demonstrated in the past.14 The other candidate orientation issue deals with the “coat-tails” effect. That is the consequence of one popular candidate in an election drawing votes for other members of the same party. Comparative study of 1998 Taipei City Mayoral Election and City Councilmen election can provide one of the best examples of coat-tail effect. Hsu, by measuring the amount of support delicate to the Mayoral candidate Ying Jiu Ma and the coat-tails effect in relation to the same year city councilmen election. The research concludes that the more confirmation on Dr. Ma’s candidate image, the higher probability that voters would vote along the party line of Ma in Taipei city council election.15
B.
Racial FactorThe ethnic war between mainlander and native Hok-lo Taiwanese rooted immediately after the evacuation of Nationalist from China to Taiwan in 1949. Although language barrier among them poses no problem for communication thanks for the centralized language policy, namely mandarin is regulated by the Nationalist government as the official language in Taiwan. At the election front, especially during competitive campaigns, the racial issue which can stir emotional bias is often been used as a manifesto differing one camp from the other. While pan-blue alliance, including the KMT, New Party, People’s First Party, are holding the flag under the general label of
13 Before 1990s, voters in Taiwan had no choice but vote for KMT. After 1990s, new parties such as the DDP, New
Party, People First Party, etc have emerged and provide voters more options to vote.
14National Science Council project “Voting Behavior and the Evaluation of Electoral Affairs---The Analysis of
Legislative Election in 1998”
mainlander who favor Chinese reunification, pan-green alliance calls for a unification of Taiwanese consists of the DDP and Taiwan Solidarity Union party. However, statistical analysis measuring racial factor in relation to voting behavior did not demonstrate a strong link between the two variables. “The racial factor is not the predominant variable in determining voting share (P=0.398). Neither mainlander voters nor Hok-lo Taiwanese voters ignore the personality and qualification of candidates which are often counting over 40 percents of voters’ decision making process.16 Moreover, data also suggests that age is undermining the leverage power of racial factor. Under frequent exposure of democratization experience and effect of inter-marriages, younger generation of voters in Taiwan exhibit a voting behavior more along the line of candidate than race. Within which, yet, there is a slight difference between voting behavior of young mainlander generation and the one of Hok-lo Taiwanese. Hsu found that the percentage of split ticket voting among the third generation of mainlander is in steady increase while new generation of Hok-lo Taiwanese voters remains loyal to the green-pan with few divided voting behavior.17
C.
Party AffiliationRapid political changes dated in the mid-1980s in Taiwan are marked by the process of democratization and notably the founding of the first opposition party, the Democratic Progressive Party (DDP) in 1986 in defiance of martial law to counterbalance the 30 years-long ruling elites of the Kuomintang (KMT). While party affiliation in the US plays a dominant role in determining voting behavior, the issue of party identification only counts for one out of tenth of voters’ voting credentials. In Taiwan, experiencing from the period of democratization to democratization consolidation, the overall support for Kuomintang is in its decline. In 1999 the KMT received 68.8% of voting share, 42.1% in mayoral election and 21.1% in 1996 presidential election. The
16 Chu, National Science Council Project 17 Hsu, pp133
latest party affiliation data piled in 2001 indicates that 14.5% of voters identified themselves with the KMT, 13.1% with People’s First Party and 34% and 3% with the DDP.18 Furthermore, what is interesting is that through a 10 years period of research project on the emerging of independent voters in Taiwan from 1989 to 1999, Chuang Tien Lien19 critically points to the fact that numbers of independent voters in Taiwan are decreasing in incremental rate. Yet there are indeed more than 30% of voters remain independent in voting preference and do not vote with party identification which could be interpreted as the characteristic of an immature democratic.
D.
Campaign Expenditure/Media Support/Face-to-Face CanvassingVirtually few researches of Taiwan election explore the micro-level variables in voting behavior, such as campaign expenditure, media support or face-to-face canvassing effect since data in Taiwan might not be easily collected and even data is available, the legitimacy and credibility of data can be problmatique. Moreover existing literature on media support varies in its finding. Tsai Chi-Chin in 1977 studied registered voters in Taichung area and the sixth provincial councilmen election and concluded that media is a critical reference while voters make their decisions. Public debates, newspapers, pamphlets, direct mail.20 Lee Chan conducting an electoral survey, a sample group of 2,135, measuring response for 1980 election, found that voters rely primarily on face-to-face canvassing, personal connection and media coverage exhibit the least influence on voting behavior.21 Another research though suggests that the more exposure of a candidate in daily newspapers, the higher voting share and the probability of being elected shall be.22
III.
Research Design and Hypotheses
18 Wu and Hsu, ‘Who are Partisans and Independents?
19 Chunag, Tien Lien. “The Developing and Changing situation of Independent Voters in Taiwan from 1989 through
1999.” pp103
20 Chen, pp574-578 21 Chen, pp577
The above literature review has revealed the gap and differences between the US voting credentials study and that of Taiwan. To dig further in the terrain of comparative voting credentials, this paper is intended to quantitatively study Kaohsiung city council elections for a total of 594 candidates from 1981, the first city council public election after the elevation of Kaohsiung city to the status of Special Municipality, to the year of 2002, marked the 6th election of Kaohsiung city councilor. In a later section, a comparison with its counterpart in Chicago will be drawn and primarily and gratefully contributed by the data mine Timothy B. Krebs, an expert of Chicago urban election at the University of New Mexico, gather from 1979 to 1995. In the vote share model, Krebs concludes that a consistent message which has been found in nationwide election researches is being confirmed as success in city council elections is a function both of being widely known and having one’s candidacy legitimized by outside sources, whereas incumbency advantage is related to spending, media support and party support23. However, whether the above pronounced voting pattern observations of Krebs and others will coincide with the case of Kaohsiung which is setting on a completely distinct political culture remains under close scrutiny in this paper.
To construct a Kaohsiung City Council Election Credentials Data (KCCEC data)
Variables, which are claimed by previous scholars to have substantial impact on voting share of candidates, such as incumbency, party support, racial factor or even education level will be respectively included. The dependent variable is the voting share of each candidate and the election result. As regard to voting share, candidates who received less than 1000 ballots are coded 0; between 1000 - 5000 ballots coded 1 and more than 10,000 coded 2. Candidates who are elected are coded 1 whereas the contrary coded 0. The independent variables include the followings:
23 Timothy b. Krebs, “The Determinants of Candidates’ Vote Share and the Advantages of Incumbency in City
Incumbency and Candidate’s Political Experience
Numerous electoral researches in the United States as extensively discussed in the former section suggest a strong role incumbency plays in voting share. While few studies in Taiwan put attention on this matter, this study will hypothesize the role of incumbency is also an important predictor in candidates’ vote share in Kaohsiung city council elections. The name recognition benefits, abundant resources and personal connections shall give incumbents higher chances of being re-elected. All incumbents are decoded 1, nonincumbents 0. Candidates who held previously high-rank government position are decoded 2. I expect a positive and significant relationship between incumbency and candidates’ percentage of the vote.
Party Affiliation
Throughout the Taiwan electoral history, party affiliation is not yet proven to have significant effects on voting behavior firstly due to the immature democratization process with predominance of Nationalist party in earlier decades and lacking of party loyalty in younger generation in Taiwan. As Dr. Chen Yi-Yen noted in late 1980s, only party affiliation could only explain 15% of voting behavior in Taiwan.24 However if by following the funnel of causality model developed by western political scientists, Niemi and Weisberg, it is suggested that a long term voting share effect party orientation shall exhibit in general election. Therefore, I hypothesize a positive relationship will be exhibited in relation to candidate’s voting share. Candidates gaining the endorsement of pan-blue party are coded 1, pan-green party coded 2 and those stand in independent status coded 0.
Racial Factor/Education Level
In the case of Taiwan, although intermarriages have undermined the degree one can determine the racial status of one individual, racial factor could be strongly tight to the birthplace 24 Chen, pp557
or the original household registers. That is the individual who has the original register in mainland China, he or she will be labeled as mainlander; whereas those register in Taiwan island shall have more chance of being Hok-lo Taiwanese.25 Therefore, in order to capture the racial status of each candidate, those who have original registers in mainland China are coded 0; those in Kaohsiung city are coded 1, the remaining coded 2. However, it shall cautiously note that no data on candidates’ original registers after the third Kaohsiung city council election dated in 1985 can be available. The main reason behind that is the unavailability of original registers, rather than current household registers information of candidates were filed in the election registration booklet. In order to become an eligible candidate in the Kaohsiung city council election, all participants must be holding a Kaohsiung city household register. Therefore, all household registers of candidates are in Kaohsiung city and it is hard to distinguish in this case the racial status of each candidate.
Moreover, education level of each candidate is also carefully examined. While candidates who do not hold a college degree are coded 0, bachelor degree holders coded 1 and master degree and above are coded 2. At this point, it is unclear what kind of relationship shall the education level play in candidate’s voting share. After carefully analyzing the KCCEC data, hopefully the myth can be unwrapped.
Number of Competitor/media support
Concerning number of competitors, Holbrook and Tidmarch (1993) found that competition is measure by the number of opponents one faces on election day. As this number increases, it is expected that candidates’ vote share will decrease. However, since Kaohsiung city council election practices the multiple members in single district system, number of competitors are roughly equal in each election district; therefore, this paper do not think the claim that number of competitors
25 Not all citizens whose original registers are in Taiwan are Hok-lo Taiwanese, some of them could be from family of
shall have an effect on voting share is feasible in this case. To the token of media support, obviously in the advent of multimedia equipments in virtually every family, the positive effect of media support in relation to voting behavior is undeniable. Yet, due to time constrain and lack of concerning data collections, this paper will not as well discuss through interpretation of data on how media support shall relate to voting share.
IV.
Analysis
Candidate’s Voting Share Analysis
Table 1 demonstrates the result of multiple regression models of all Kaohsiung city council election candidates for each election year. Except data for Education (candidate’s educated level) in 1981, 1985, 1989 election and that of Racial Factor in 1994, 1998, 2002 are not available, the model in general performs extremely well. Most party of the model can explain approximately 40% of the variation in candidates’ voting share (1985 - 44.5%; 1994 – 38.7%; 1998 – 38.8%; 2002 – 46.4%) whereas model in 1981 and 1989 might suffer from the defect of missing data and therefore only account for less than 7% of variation. As the paper has hypothesized initially, incumbency and candidates’ political experience display a significant and positive effect on candidates’ voting share.26 Notably in 1994, 1998 and 2002, incumbency produces an increase in these candidates’ voting share by over 42.9%, 55.2% and 51.3% respectively at the significant level of 0.01. Party affiliation also matched the earlier hypothesis, confirming the important role of party play in voting behavior. Over 50% of candidates’ vote share increased is the result of party identification, disregard which party one affiliated to. (1994 – 52.6%; 2002 – 50%; 1985 – 35.4%; 1998 – 36.1%)
26 Please note that there are some missing data for Incumbency in 1981, 1985 and 1989, which might cause the lower
How racial factor and education level of candidates affect the quantity of vote one received is not statistically demonstrated at significant level. Missing data and the inability to distinguish and locate effectively mainlander and Hok-lo Taiwanese as two large ethnic groups in Taiwan might be the two primary reasons that the role of race is mitigated in voting behavior. However, interestingly noted, education is a significant predictor in voting share and the more a candidate is educated, the higher the alleged vote share will be obtained. At most in 1998, 22.1% of votes could be increased due to the high education level candidates hold.
Lastly, overall sex exhibits negative correlation with candidates’ voting share. Yet with or without surprise, sex accounts for 34.3% of vote increase in 2002. One of many explanations might be the effect of inclusion of female governor strategy introduced by President Chen’s new government in 2000 in which more women have been included into the Chen administration ever than any other previous political eras. Leaded by the first female Vice-President in history – Annette Lu, 2000 also marks the first time Cabinet members were occupied by one fourth of female representatives. However, since sex variable in 1998 and 1994 does not exhibit positive or indicate great influence on candidates’ voting share (-.058 and .042), only an observation rather than an generalizable trend can be made at this point before further elections in the future being studied.
Table 1.
Candidate’s Voting Share, Kaohsiung City Council Elections, 1981-2002
Independent Variables 1981 1985 1989 1994 1998 2002 Incumbency .257 (n = 40) .154 (n=42) .255 (n=43) .429* .552** .513** Party Affiliation .140* .354* .006 .526* .361** .500** Racial Factor .018 ﹣.136 .137 --- --- ---Education --- --- --- .117 .221* .127 Sex ﹣.140 ﹣.328* ﹣.093 .042 ﹣.058 .343**
N 81 70 94 129 105 114
R .412 .707 .404 .637 .641 .695
R Square .170 .499 .163 .406 .411 .483
Adj. R Square .075 .445 .075 .387 .388 .464
F. Value 1.786 9.226 1.850 21.188 17.455 25.440
Note: *Significant <0.05 **Significant <0.01
Election Result Analysis
If not counting candidates’ voting share, instead use purely election result of whether being elected or the contrary as dependent variable, slight variants can be noted as one can refer to Table 2. Incumbency once again as hypothesized plays a visible role in election turnout, mostly notably to the foremost in 1998 approximately 60% of winning chance has been induced simply by the status of incumbent at significant level of 0.01. Party affiliation to the same token concurs with the prediction of being positively related to election result. However, the influence of party is stronger in determining candidates’ voting share than election result. (refer to table 1 to 2) Racial factor for the same reasons given earlier cannot demonstrate its clear influence. Education in 1998 produced a 22.4% increase of wining opportunity confirms the same observation in the relation between voting share and education. (Table 1 1998 education .221*) Invariably, sex in 2002 exerts the most influential weight on election turnout, accounting for 10.8% chance, if comparing with all other election year. Once again, since only in 2002 sex displays a positive correlation to election result whereas the rest correlated negatively. No generalization can be noted at this moment before further election, 2006 city council election for instance, being observed.
Table 2.
Independent Variables 1981 1985 1989 1994 1998 2002 Incumbency ---(n = 40) ---(n=42) ---(n=43) .569** .606** .392** Party Affiliation .227* .304* .041 .364** .179 .450** Racial Factor .149 ﹣.112 ﹣.101 --- --- ---Education --- --- --- .061 .224* ﹣.004 Sex ﹣.113 ﹣.093 ﹣.085 ﹣.006 ﹣.049 .108 N 81 70 94 129 105 114 R .313 .553 .163 .632 .637 .581 R Square .098 .306 .027 .400 .405 .338 Adj. R Square .063 .231 ﹣.006 .380 .381 .313 F. Value 2.794 4.085 .817 20.636 17.033 13.893
Note: missing data reported in Incumbency for 1981, 1985, 1989 election *Sig. <0.05 **Sig. <0.01
Aggregate Analysis
Overall, the Kaohsiung city council election regression model sustains findings which are hypothesized yet not tested or confirmed previously. Throughout Taiwan electoral literature, very few studies have dealt with the issue of incumbency and its influence in either election of national or local level. As to the survey conducted by National Science Council, 26.29% of voters in the chosen sample vote for specific candidate because of his or her past experience and accomplishment, the survey result is further strengthened and supported by the exceptional performance of incumbency variable in both determining candidates’ voting share and the ultimate election turnout. Furthermore, as Taiwan only recently undergoes from the transitory democracy to the democracy consolidation period within the past two decades or less, while the shadow of a strong party remains, party affiliation in local election still occupy a crucial role, despite the existence of independent voters without a straight ticket among whom might focus more on
candidates’ performance and political experience. The influence of voting credential of racial factor is dismissed in the grand picture of this data set due to insufficient and inability to distinguish races between mainlander and Hok-lo Taiwanese in Taiwan. Sex variable in general exhibits negative correlation with election turnout while education can assert a more critical influence in voters’ decision.
Table 3.
Kaohsiung City Council Elections, 1981-2002
Independent Variables
Voting Share Election Result
Incumbency .454** .471**
Party Affiliation .342** .240**
Racial Factor ﹣.137* ﹣.006
Education .140* .090
Sex ﹣.043 ﹣.020
Note: *Sig. <0.05 **Sig. <0.01 N= 594
V.
Chicago and Kaohsiung city council election comparison
Variants in electoral regulation and political culture are thought to alter voters’ behavior in distinct manner. Before getting into the heart of comparing Chicago and Kaohsiung city council election results, a good understanding of Chicago city political culture might be required.
Brief Chicago City Political History
Chicago city with current population of more than nine millions has granted the status of city in 1837 as the population reached 4000. Ever since, a ward system27 is being used in city council
27 Big cities in the US such as Chicago, Boston and New York all elect city council members to represent a geographic
district, or ward. Advocates of the ward plan say it’s a better way to address the needs of specific neighborhoods and minority groups. In Chicago, for example, when a resident has a problem with a street light on his block he can call his alderman. Indeed, a ward system would make each council member accountable for resolving problems in their section of the trouble-plagued city.
election. State law stipulates that ward boundaries are to be adjusted every ten years so that population shifts can be taken into consideration and equality of representation be based thereon. Over the years political division of the city has been as table 4 displays. Number of wards increases as the city expands in population, at present time, the Chicago City Council consists of 50 aldermen, namely city councilmen which is a tradition followed since 1923, one representing each city ward. The council meets regularly to discuss legislation, orders and ordinances whose subject matter includes traffic code changes, the city's relationship with utility providers, taxes, and much more.
Table 4 Chicago Ward History
Number Date Authority of wards
March 4, 1937 City Charter of March 4,1837 6
1847 General Assembly 9
1857 General Assembly 10
1863 Revised charter of 1863 16
March 10, 1869 General Assembly 20
April, 1875 General Incorporation Act of April, 1875 18
1888 General Assembly 24
1889 General Assembly 34
1900 General Assembly 35
1923 General Assembly 50
Note: Since the adoption of the 50-ward system there have been six redistrictings, in 1931, in 1947, in 1961 and in 1970. Since 1970, 1981 and 1992. Court ordered partial redistrictings have also occurred in 1973, 1974, 1982, 1983, and 1986.
Source: Chicago public library source http://www.chipublib.org/004chicago/timeline/poldivision.html
Deductively speaking, two characteristics of Chicago city council election are noted. First of all, the hallmark of Chicago politics is its infamous political machine, the Cook County Democratic Party Organization, which has lost its power in 1990 yet still asserts influences in political map of Chicago until present day. Immediately after it lost the bid to elect its slated candidate for Cook County board president, the election of populist black reform leaders, such as
Mayor Harold Washington28, the first African-American mayor of Chicago city and the emergence of Hispanics revealed. Secondly, due to the fact that Chicago is one of the most ethnic conscious cities and ethnic diverse cities in the US where ethnic labels are still the message as where your ancestors came from, the so called “rainbow theory” of Chicago machine politics shall be kept in mind. Rainbow theory (Merton 1968) contests that the Chicago machine is a functional body that pursued political incorporation of many ethnic groups in the political party. In return for loyalty to political party, machines delivered a variety of social services to ethnic immigrants, in addition to jobs, friendship and opportunities for social and economic advancement.
Chicago & Kaohsiung City Council Election Comparison
A similar candidate voting share dataset on Chicago City Council as compared to Kaohsiung city council elections was assembled by Timothy B. Krebs at the University of North Carolina. With the dataset Krebs provided, the following analysis can be drawn as table 5 demonstrates. Concerning the candidate’s vote percentage, analysis confirms the claim that single most important factor in predicting candidates’ vote share in city council elections is incumbency (Prewitt 1970). In each election year, incumbency accounts for over 20% increase of vote percentage. Although a lesser degree of influence incumbency assert upon voting share if compared to the result of Kaohsiung city council election with at most 55.2% increase (refer to Table 1), both elections, despite different political culture, display a strong role of incumbency in municipal level election.
Furthermore, minority status (.061, .051, .068 ) does not correlate significantly with voting share in Chicago city council election for at least two reasons. First of all, very few nonmajority race candidates run in elections due to large amounts of racial segregation in the city and the 28 Miranda, Rowan A. “Post-Machine Regimes and the Growth of Government.” Urban Affairs Quarterly, Vol. 28 No.
present of super-majority wards. Secondly, the rainbow theory mentioned earlier, majority race incumbents will still seek support from minority group by providing extra care and allocation of resources in their preference. For instance, white incumbents have historically and currently represent the majority black Fifth Ward from 1983 to 1995. Thomas Murphy currently represents the majority black Eighteenth war on the city’s southwest side.29 As for the case of Kaohsiung city council, although whether there is also a “rainbow effect” is subjected to further investigation, the racial factor only represents a mild predictor of candidates’ voting share from 1981 to 1989. (.018, .136, .137) What is not available in Kaohsiung city council election data, yet in Chicago data is the campaign spending which is a positive and significant predictor of vote share in each election.
Table 5.
Candidates’ Vote Percentage, Chicago City Council Elections, 1979 – 95, by Election Year
Independent Variables 1979 1983 1987 1991 1995 Note Incumbency .212** .264*** .280*** .300*** .280*** Campaign Spending .025*** .019*** .021*** .017*** .025*** Minority Status -.052 .061 -.003 .051 .068 Adj. Square .64 .65 .73 .76 .74 Number of Case 79 134 133 159 126
Note: Entries are unstandardized regression coefficients +p =.10; *p=.05; **p=.01; ***p=.001
Table 6 displays results when removing incumbents from the candidate list. As Chicago is often referred to as the Machine City as marked by the machine of Democratic Party organization. The role of Democratic support is examined. The result shows that candidates who received the backing of the Democratic organization in their wards can expect an increase in their share of the vote in the range of 10-19%. Besides, if one looks more carefully, from 1979 to 1995, the influence of Democratic support is in its decline which matches the fact that Democratic Party 29 Krebs, pp928
began to lose its influence since 1990. Similar observations can be noted in Kaohsiung city council elections where party affiliation plays an even greater role in determining voting share, at most up to 50% increase of vote and most interestingly noted is that the importance of the role of party increases along time from 1981 (.140*) to 2002 (.500**). Political experience and campaign spending also have significant influences on elections outcomes.
Table 6.
Nonincumbent Candidates’ Vote Percentage, Chicago City Council Elections, 1979 – 95, by Election Year Independent Variables 1979 1983 1987 1991 1995 Note Democratic Support .127*** .191*** .112*** .101*** .029 Political Experience .024 .005 -.008 .037** .065*** Campaign Spending .019*** .011*** .016*** .014*** .016*** Adj. Square .73 .73 .53 .57 .55 Number of Case 61 104 100 121 94
Note: Entries are unstandardized regression coefficients +p =.10; *p=.05; **p=.01; ***p=.001
Conclusion
All in all, distinct voting behaviors shall display throughout elections at national level and local level. This paper firstly, by studying the case of Kaohsiung city council elections and Chicago elections, concurs with the observation of many American political scientists that incumbency is the single most important factor in predicting candidates’ vote share in city council elections. Despite cultural differences, incumbency in Kaohsiung city council elections plays a significant role in terms of increase of candidates’ vote share. Secondly, as to the issue of race and its role played in voters’ decisions, no significant correlation between racial factor and voting share has been found in both Kaohsiung and Chicago city council elections turnout. Especially in the case of Kaohsiung, after 1989 city council election, candidates’ racial status is hardly identified to a degree raising the doubt that whether race can be counted as an issue in Taiwan or simply as a
tool of election manifesto. Thirdly, the funnel of causality model proposed by Niemi is still feasible yet party affiliation in the chosen two cities only differentiates in the degree party influence on voter’s behavior. Due to time constrain and data availability, many missing data in KCCEC such as race or media support have caused the limited explanatory power of the current analysis.
Once again, this research can be served as the initial step in understanding the importance of various voting credentials in different countries in which not only electoral rules but also culture might vary. Future researches with a more completed dataset on similar matter are in great need as to further our knowledgeable territory of the voting behavior at local or municipal level.
References
Source in English
Abramowitz, Alan. 1991. “Incumbency, Campaign Spending, and the Declined of Competition in U.S. House Elections.” Journal of Politics 53:34-56.
Acock, Alan and Robert Halley. 1975. “Ethnic Politics and Racial Issues Reconsidered: Comments on an Earlier Study.” Western Political Quarterly 28:737-738
Arrington, Theodore S. 1978. “Partisan Campaigns, Ballots, Voting Patterns: Charlotte.” Urban Affairs Quarterly 14:253-61.
Campbell, Angus and Robert Kahn. 1854. How People Elect a President. Ann Arbor: University of Michigan Press.
Cox, Gary W., and Scott Morgenstern. 1993. “The Increasing Advantage of Incumbency in the U.S. States.” Legislative Studies Quarterly 18:495-514.
DeSantis, Victor, and Tari Renner. 1994. “Term Limits and Turnover Among Local Officials.” In Municipal Year Book, 1994. Washington, DC: International City/County Management Association. Ferejohn, John A. 1977. “On the Decline of Competition in Congressional Elections.” American Political
Science Review 71 (1): 166-176.
Fiorina, Morris P. 1981. Retrospective voting in American National Elections. New Haven: Yale University Press.
Gerber, Alan S. and Donald P. Green. 2000. “The Effects of Canvassing, Telephone Calls, and Direct Mail on Voter Turnout: A filed Experiment.” The American Political Science Review 94:653-663. Graber, Doris. A. 1990. Media power in politics (3rd Ed). Washington, DC: Congressional Quarterly. Graber, Doris A. 1993. Mass Media and American Politics. Congressional Quarterly Inc.
Hahn, Harlan, and Timothy Almy. 1971. “Ethnic Politics and Racial Issues: Voting in Los Angeles.” Western Political Quarterly 24:719-30.
Holbrook, Thomas M.., and Charles M. Tidmarch. 1993. “The Effects of leadership Positions on votes for Incumbents in State legislative Elections.” Political Research Quarterly 46:897-909.
Inglot, Tomasz, and John P. Pelissero. 1993. “Ethnic Political Power in a Machine City: Chicago’s Poles at Rainbow’s End.” Urban Affairs Quarterly 28:526-43.
Karnig, Albert K., and Oliver Walter. 1977. “Municipal Elections: Registration, Incumbent Success, and Voter Participation.” In Municipal Yearbook, 1977. Washington DC: International City Management Association.
Krebs, Timothy B. “The Determinants of candidates’ Vote Share and the Advantages of Incumbency in City Council Elections.” American Journal of Political Science 42: 921-935
Lieske, Joel, and Jan William Hillard. 1984. “The Racial Factor in Urban elections.” Western Political Quarterly 37:545-63.
Mayhew, David R. 1974. “Congressional Elections: The Case of the Vanishing Marginals.” Polity 6:295-317.
Merton, R.K. 1968. Social Theory and Social Structure. Rev. ed. New York: Free Press.
Miranda, Rowan A. “Post-Machine Regimes and the Growth of Government”. Urban Affairs Quarterly 28:397-422
Prewitt, Kenneth. 1970. The Recruitment of Political Leaders: A Study of Citizen-Politicians. New York: The Bobbs-Merrill Company, Inc.
Putnam, Robert C. 2000. Bowling Alone: The Collapse and Renewal of American Community. New York: Simon and Schuster.
Thomas M. Holbrook. 1993. “The Effects of leadership Positions on Votes for Incumbents in State Legislative Elections”. Political Research Quarterly 46:897-909
Thompson, Edward, III. 1984. “Race and the Chicago Election.”Journal of Ethnic Studies 11 (Winter):1-10.
Vanderleeuw, James. 1990. “A City in transition: The Impact of Changing Racial composition on Voting Behavior.” Social Science Quarterly 71:326-38.
Williams, Oliver. 1967. “Life-style values and Political Decentralization in Metropolitan Areas.” Social Science Quarterly 48:299-310.
Source in Chinese:
Chen, Yi-Yen. 1986. “Review of Voting Behavior in Taiwan”. The Thought and Word 23:557-585 Chu, Yun-Han, Fu Hu, Yung-Tai Hung and Chen, Te-Yu. “Voting Behavior and the Evaluation of
Electoral Affairs---The Analysis of Legislative Election in 1998” National Science Council: CEC001-03
Chunag, Tien Lien. 2001. “The Developing and Changing situation of Independent Voters in Taiwan from 1989 through 1999”. Journal of Electoral Studies 8:71-115
Hsu, Sheng-mao. 2001. “Ticket Splitting: The 1998 Taipei City Mayoral Election. Journal of Electoral Studies 8:117-158.
Wu, Chung-li and Wen-pin Hsu. 2003. “Who are Partisans and Independents?—Determinants of Party Identifications of Taiwan’s voters in 2001. Political Science Review 18:101-139
