Duverger’s Law and Village elections in Taiwan

Contrary to Duverger’s Law, most winners in Taiwan’s village and li Single-Member-District elections are independent, and the proportion goes larger and larger. In Fig 1, the black line is the proportion of voteshare that independent candidates in village election received with years. Why? Is partisanship not effective on attracting vote in this election?


My new research article emphasizes the key factor is the amount of electorates in districts. Taiwan’s village and li (里, a small unit like village) were designed by colonial Japanese government, and maintained by martial ruling KMT government after 1949. The original goal of village and li is to control and censor people, so the size cannot be too big. After 1953, KMT government used village and li as the basic unit for electoral fraud and mobilization, and a ceiling is set for realigning village and li if needed.

With time goes by, some village grows very large (with almost 30000 citizens) while others are small (with minimum 24 residents). Due to this variation, I argues that the campaign strategy of village and li chief candidates must change conditioning on the number of electorates in district. The special historical background of village and li in Taiwan enables researchers to test whether the effect of partisanship on voteshare and candidate’s strategies are mediated by the number of electorates in district, controlling for the level of government and electoral system.

For small district, candidate might not want to conflict with voter’s partisanship or suffer from coattail effect from higher office, so they choose to run election independently; they can make the best use of their personal social network and resource to change enough vote, a. k. a. clientalism. For larger district, candidates lack the capacity to run election independently, so they ask partisanship as a cost-saving information cue to mobilize vote.

I use 2010 Taiwan village and li elections results to reveal that

(1) number of electorate negatively correlates to single-candidate district and independent incumbent reelection.
(2) Number of electorate positively correlates to larger proportion of candidates from two major parties, and that candidates with partisanship enjoy larger vote share.
As can be seen in Fig 2, the green and blue columns are the estimated effect of candidate with partisanship of the two major parties, compared to independent candidates. And X axis is the number of electorates in districts. With the size of district goes up, candidates with partisanship enjoy more votes in average.

Empirical results suggest that vote-seeking candidates choose campaign strategy conditioning on the number of electorates in district: campaigning in small district relies on personal network and resource, while in larger district needs partisanship as cost-saving information cue.

Regression result also suggests that candidates benefit from middle age, male, and incumbency.

Besides, descriptive statistics show that in Taiwan’s village and li elections the incumbent reelection rate is about 60%, 30% of districts is single-candidates and 50% two-candidates, 70% of candidates are independent, and, especially, in several districts numerous candidates from the same party competed against others.


1 thought on “Duverger’s Law and Village elections in Taiwan

  1. Pingback: 村里長選舉中的政黨標籤效應 | 菜市場政治學

Leave a Reply

Your email address will not be published. Required fields are marked *