Party leader elections in East Asia: Comparative Analysis of Japan and Taiwan Yohei NARITA, Waseda University [email protected]Ryo NAKAI, Waseda University [email protected]Keiichi KUBO, Waseda University [email protected]Paper presented at the ECPR General Conference Bordeaux, 04-07 September 2013. Work in progress: please do not cite without authors’ permission. The present paper is based on the findings of the research project “Data collection and comparative analysis of primaries in democratic countries,” funded by Waseda University, Research Grant for Special Project A (2012A-501). We would like to express our gratitude for their generous assistance.
30
Embed
Party Leader Elections in East Asia: Comparative Analysis ...€¦ · Party leader elections in East Asia: Comparative Analysis of Japan and Taiwan Yohei NARITA, Waseda University
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
The present paper is based on the findings of the research project “Data collection and comparative
analysis of primaries in democratic countries,” funded by Waseda University, Research Grant for
Special Project A (2012A-501). We would like to express our gratitude for their generous assistance.
1
Introduction
Despite the growing interest in the causes and consequences of the party
primaries, there has not been a systematic inquiry of the East Asian cases on this matter.
In order to fill the gap, this paper attempts to make a comparative analysis of Japan and
Taiwan regarding the party leader elections. This paper consists of four parts.
The first introductory section will address the political context, focusing on the
political system and the party system. The second section then examines the rationales
for the adoption of primary elections at individual party level. After the analysis of each
country, an overview of the main strategic incentives for political parties to resort to
inclusive processes of leader selection will be supplied from a comparative perspective.
The third section will be devoted to the analysis of the processes of primary elections in
in Japan and Taiwan. This section will explore the process as a whole, taking into account
both the rules managing the internal elections and the actual functioning of the latter. The
fourth section then examines the impact of the adoption of primaries on political parties,
focusing on the party support among the general public and the membership recruitment.
We would like to make a brief disclaimer on the terminology, the scope of the
present paper and the source of data. First, we use the term “primary” when ordinary party
members have the right to vote for the party leader selection, regardless of the weight of
their votes in the selection process. As discussed below, the actual weight of primary in
the party leader selection varies significantly in the case of Japan, and what have been
practiced by the major political parties in Japan since the 1990s are mostly not typical
primaries. Still, these practices are important organizational reforms in Japanese political
parties to grant ordinary party members a greater influence in the leadership selection
process, and we think that it is valid and interesting to study these cases from the
comparative perspective in the study of primaries. Second, as for Taiwan, primaries have
been conducted not only for the leadership selection but also for the candidate selection.
Due to the theoretical necessity to distinguish primaries for the leadership selection and
the candidate selection, however, we limit ourselves to the analysis of the primaries for
the leadership selection, though the latter will be mentioned briefly in the second and the
fourth section. Third, the empirical data on primaries presented in this paper are based on
the dataset of the party leader selection of major political parties in Japan and Taiwan that
we constructed using the primary sources. The sources of the data will not be presented
in this paper, but the whole dataset will be published later on our personal website, in
which the source of the data will be presented for the each election results of the primary
2
or other type of leadership selection.
1. The Political Context of Japan and Taiwan
This section addresses the general political context of these two Far Eastern
democracies, including their political histories, political systems and party systems. The
two countries differ from each other in terms of the timing of democratization and the
political system, but share commonalities in terms of the party system.
Differences between Japan and Taiwan: timing of democratization and political system
One crucial difference between Japan and Taiwan is the timing of
democratization. Japan was democratized immediately after the WWII under the
occupation by the US, with the promulgation of the Constitution of Japan in November
1946. Since then, the constitution has never been amended and multi-party elections have
been held regularly since then. Taiwan, which was incorporated into the Republic of
China after the WWII, had been ruled under the one-party dictatorship since 1949, when
the Nationalist Party (KuoMinTang: KMT) lost the civil war in the mainland China and
moved its central government to Taiwan, until the beginning of the 1990s, when the
martial law was lifted and the multi-party competitive elections were held. For our paper,
this means that the Japanese political parties have much longer experiences under the
democratic rule and that they have a significantly more cases of intra-party elections to
select party leaders: the first primary was held in Japan in 1978 but only in 1998 for the
leadership selection in Taiwan (the first primary ever was held in 1989 for the candidate
selection, as discussed below).
Another difference between Japan and Taiwan is the type of political system.
While Japan has adopted the bi-cameral parliamentary system, Taiwan has adopted the
semi-presidential system.1 The chief executive of Japan is a prime minister, and his2
cabinet is responsible to the Diet. The lower house in Japan, the House of Representatives,
can give a vote of no-confidence motion to the cabinet, and the cabinet can dissolve the
lower house and call snap elections at any time.3 The Emperor of Japan is symbolic in
political procedures. In contrast, the chief executive of Taiwan is a president who is
1 The local Taiwanese often call this system as “dual leadership system (shuang shouzhang zhi)” (Wu 2007: 214). 2 There has never been a female prime minister in Japan or Taiwan. 3 In fact, a legal technical debate remains even now over whether the cabinet can dissolve the House of
Representatives at any time or only after the House of Representatives submits a motion of no-confidence. Even the
Supreme Court has avoided ruling on this debate in the light of acte de gouvernement. At any rate, in practice, the
legislature has been interpreted to allow the cabinet to do this at any time.
3
elected by direct election for a term of four years. The president can appoint the Premier
of Taiwan, who is responsible to the parliament, the Legislature Yuan. The Premier must
resign if the Legislature Yuan passes a motion of no confidence regarding the Premier,
but he can petition the president to dissolve the parliament in these cases.4 For our paper,
this difference means that the leadership selection within the political party can have
different meaning: the leadership selection in Japan almost automatically means the
candidate selection for the chief executive (Prime Minister), but the party leader selection
and the candidate selection for the chief executive (President) can be separated in Taiwan.
This issue will be examined further in the second section of this paper.
Commonalities between Japan and Taiwan: party system
Despite the differences discussed above, Japan and Taiwan also share some
commonalities regarding the party system. First, these two countries had a very strong
and dominant conservative party after the WWII. While Japan was institutionally under
the democratic system, it was dominated by the Liberal Democratic Party (LDP),
established in 1955, and it remained in power for several decades. Although opposition
parties were legally allowed in post-war Japan, the main opposition party, the Social
Democratic Party of Japan (SDPJ)5, was too weak to defeat the LDP due to its internal
ideological conflict and rivalry with the more leftist Japan Communist Party (JCP)
(Kohno 1997). The KMT’s hegemony in Taiwan started in 1949, since when the KMT
had been imposing martial law, which was lifted only in 1987. Thus, after WWII, Japan
and Taiwan experienced a long-term rule by single conservative parties, the LDP and the
KMT. Furthermore, these two dominant parties are also known for the internal division
between factions. This similarity will be interesting when we analyze the rationales for
the adoption and avoidance of primaries.
The political dominance or dictatorship collapsed around 1990 in both countries.
In Japan, the corrupt patron-client culture that supported LDP dominance triggered the
fall of the LDP. Due to the economic backdrop after 1991 and a series of corruption
scandals, the support for the LDP began to wane, and the party split led to the loss of
majority status and establishment of the non-LDP coalition government in 1993. In
Taiwan, the lifting of martial law in 1987 prefaced the democratization of the republic. In
4 These rules were introduced after the amendment of the constitution in 2005. 5 The Japanese name is Nihon shakaito, which literally means Socialist party of Japan, and indeed it is often called
the SPJ, or the JSP – Japanese Socialist Party, in English. However, at the founding convention, it was decided that
their English name is the SDPJ (SDPJ 1986: 17). So we follow the official terminology and use the abbreviation of
the SDPJ.
4
this process, opposition parties were legalized, including the Democratic Progressive
Party (DPP), which was formed in 1986 and aims for Taiwan’s independence. After the
lifting of martial law, a full national election was held in 1992, which marked the
accomplishment of democratization,6 and the first direct presidential election was held
in 1996. The first government change at the national level happened in 2000, when the
DPP candidate won the presidential elections.
After the major transformation of politics in Japan and Taiwan around 1990, the
party systems of both countries underwent many divisions and mergers. Both countries
seemed to consolidate pseudo/lax two party systems with two big parties and some minor
parties. As for Japan, many MPs who abandoned the LDP set up new political parties,
and the biggest group of them led by Ichiro Ozawa established the New Frontier Party
(NFP). Though the NFP was rather short-lived and soon split into several smaller groups,
some of them, with other ex-LDPs and core members bolted from SDPJ, founded the
Democratic Party of Japan (DPJ) in 1998. The DPJ kept expanding with the support of
labor unions and urban-voters, and finally won the elections in 2009 and took power,
forming a coalition government with smaller parties. We thus regard that the LDP, the
SDPJ, the NFP and the DPJ as major political parties in Japan, which will be analyzed in
this paper. Two other parties – Komeito (also known as the Clean Government Party or
CGP) and the JCP- have consistently held seats for a long time, but never held the intra-
party election until now because of their disciplined and centralized organizations. Table
1 and 2 summarize the electoral performance of major political parties in Japan from 1955
to 2012.
6 See Polity IV (http://www.systemicpeace.org/polity/taw2.htm)
5
Democratized Taiwan’s party system can be summarized as a two-polar system
between the KMT group and the DPP group. In the process of the democratization of
Taiwan, as previously noted, the DPP appeared as a viable rival to the KMT. The first
electoral success of the DPP was the party’s victory by Chen Shui-bian in the Taipei
2012
2009
2005
2003
2000
1996
1993
1990
1986
1983
1980
1979
1976
1972
1969
1967
1963
1960
1958
LD
P61.3
24.8
61.7
49.4
48.5
47.8
43.6
53.7
58.6
48.9
55.6
48.5
48.7
55.2
59.3
57.0
60.6
63.4
61.5
DP
J11.9
64.2
23.5
36.9
26.5
10.4
*
NF
P31.2
SD
PJ
0.4
1.5
1.5
1.3
4.0
3.0
13.7
26.6
16.6
21.9
20.9
20.9
24.1
24.0
18.5
28.8
30.8
31.0
35.5
CG
P6.5
4.4
6.5
7.1
6.5
NF
P10.0
8.8
10.9
11.4
6.5
11.2
10.8
5.9
9.7
5.1
JCP
1.7
1.9
1.9
1.9
4.2
5.2
2.9
3.1
5.1
5.1
5.7
7.6
3.3
7.7
2.9
1.0
1.1
0.6
0.2
DS
PN
FP
2.9
2.7
5.1
7.4
6.3
6.8
5.7
3.9
6.4
6.2
4.9
3.6
Oth
ers
17.3
2.1
1.3
1.3
7.3
0.6
20.9
1.0
2.0
2.2
2.9
1.2
3.3
0.4
0.0
0.0
0.0
0.2
0.2
Ind.
1.0
1.3
3.8
2.3
3.1
1.8
5.9
4.1
1.8
3.1
2.2
3.7
4.1
2.9
3.3
1.9
2.6
1.1
2.6
TO
TA
L480
480
480
480
480
500
511
512
512
511
511
511
511
491
486
486
467
467
467
LD
P: L
iber
al D
emo
crat
ic P
arty
, D
PJ:
Dem
ocr
atic
Par
ty o
f Ja
pan
, N
FP
: N
ew F
ronte
ir P
arty
, SD
PJ:
So
cial
Dem
ocr
atic
Par
ty o
f Ja
pan
, C
GP
: C
lean
Go
ver
nm
ent
Par
ty (
also
kn
ow
n
as K
om
eito
), J
CP
: Ja
pan
Co
mm
un
ist
Par
ty,
DSP
: D
emo
crat
ic S
oci
alis
t P
arty
.
* T
he
seat
sh
are
wo
n b
y p
roto
-DP
J
Table
1.
Ele
cto
ral R
esu
lts
(Seat
Share
[%
]) in J
apan’s
House
of
Repre
senta
tives
2010
2007
2004
2001
1998
1995
1992
1989
1986
1983
1980
1977
1974
1971
1968
1965
1962
1959
1956
LD
P34.7
34.3
47.5
44.9
40.9
43.7
42.5
43.3
56.7
54.4
54.0
49.8
50.0
52.6
54.8
55.8
56.8
52.8
48.8
DP
J43.8
45.0
33.9
23.9
18.7
NF
P22.2
SD
PJ
1.7
2.1
2.1
3.2
5.2
15.1
28.2
26.6
16.3
17.5
18.8
22.5
24.6
26.5
26.0
29.1
26.4
34.0
32.0
CG
P7.9
8.3
9.9
9.3
8.7
NF
P9.5
8.7
9.5
10.7
10.4
10.0
9.5
8.8
9.6
8.0
6.0
3.6
JCP
2.5
2.9
3.7
8.1
9.1
5.6
4.4
5.6
6.3
5.6
4.8
6.4
7.9
4.0
2.8
1.6
1.6
1.2
0.8
DS
PN
FP
3.6
3.2
4.8
4.8
4.4
4.4
4.0
5.2
4.0
2.8
4.4
Oth
ers
8.7
2.1
0.0
6.1
6.7
4.4
8.3
7.5
4.0
4.0
2.0
1.6
0.0
0.0
0.0
0.0
2.8
4.8
12.8
Ind.
0.8
5.4
2.9
4.5
10.7
4.8
3.6
5.2
2.4
3.2
5.6
5.2
4.0
2.8
2.8
2.8
2.0
3.6
5.6
TO
TA
L242
242
242
247
252
252
252
252
252
252
250
249
252
249
250
251
250
250
250
Table
2.
Ele
cto
ral R
esu
lts
(Seat
Share
[%
]) in J
apan’s
House
of
Councill
ars
LD
P: L
iber
al D
emocr
atic
Par
ty, D
PJ:
Dem
ocr
atic
Par
ty o
f Ja
pan
, N
FP
: N
ew F
ronte
ir P
arty
, SD
PJ:
Soci
al D
emocr
atic
Par
ty o
f Ja
pan
, C
GP
: C
lean
Gover
nm
ent
Par
ty (
also
know
n
as K
om
eito
), J
CP
: Ja
pan
Com
munis
t P
arty
, D
SP
: D
emocr
atic
Soci
alis
t P
arty
.
6
mayoralty elections in 1994. Since then, the DPP has softened its overt advocacy of
separation to build Taiwan as an independent republic and has succeeded in gaining
political support from moderate electorates. The KMT president Lee Teng-hui also
moderated his doctrine of Chinese-reunification by recognizing the present isolated state
of Taiwan, but pro-reunification hardliner James Soong Chu-yu left the KMT and ran for
the presidency in 2000 as an independent and for the 2001 Legislative Yuan election with
his party, the People First Party (PFP). This political incident brought the DPP victory in
both elections as the pro-unification voters split between the KMT and Soong’s faction.
Since this election, the KMT have allied with the PFP as a “pan-blue coalition” to avoid
vote splitting, and they fielded a united candidate for each presidential election in the
2000s. In competition with this coalition, the DPP also formed a so-called “pan-green
coalition” with the Taiwan Solidarity Union (TSU), which was formed by ex-KMTs who
are pro-independence. In general, the presidential system tends to foster the fragmentation
of party systems, because politicians need to make broader coalitions for wider support
in order to win a presidential election (Golder 2006). The case of Taiwan follows this
argument. In this way, the party system in Taiwan has been a two party (coalition) system
between the pan-blue led by the KMT and pan-green led by the DPP. Thus, these two
major parties will be analyzed in the present paper. Table 3 and 4 summarize the electoral
performance of major political parties and their presidential candidates in the elections
Second, as for the two major parties in Japan (the LDP and the DPJ), they tend
to adopt a mixed system for the first round, in which some points are distributed to the
candidates via primary and other points are allocated based on the votes of the incumbent
parliamentarians. One vote in the primary counts much less (for example, 1 points for
every 10,000 votes in the primary, as in the LDP internal elections in 1995 and 1999)
compared to one vote cast by an incumbent parliamentarian (usually one man, one point,
but in some cases, one man, two points – as in the DPJ internal elections in 2010 and
2012). This system gives a significant room for maneuver to limit the impact of the results
of the primary election by changing the share of the points distributed via primary in the
first round. Indeed, the LDP frequently changed the internal rules to change the share of
the primary in the first round. The DPJ also have changed the rules related to the share of
the points allocated via primary. Table 9 summarizes main features of the rules regarding
18
the internal leadership elections in Japan, including the share of the primary (i.e. the share
of the points distributed via primary out of total points) in the first-round elections. The
actual share of the primary varies from 20% to 60% in the LDP after 1991 and is around
20-30% for the DPJ. This tendency, however, is not shared by the opposition parties which
never took power, namely the SDPJ and the NFP. The SDPJ adopted a simple closed
primary. The NFP once adopted an open primary, in which everybody who is above 18
years old and pays the participation fee (1,000 yen, or around 10 US$) can participate in
the primary.
Table 9. Features of the rules for the primaries in Japan
Source: created by the authors.
19
These two features suggest that the politicians in the major parties (the LDP and
the DPJ) tend to prefer limiting the impact of the primary elections for the selection of
party leaders. How about the actual degree of participation and competition in the
primaries? Does the smaller share of the primary lead to the lower level of interests among
the eligible voters and the lower degree of participation? How real is the competition
between the candidates in the primary elections? Let us examine these issues below.
Table 10. Participation in the primaries in Japan and Taiwan
Source: Created by the authors.
Participation and Competitiveness
First, let us examine the level of participation in the primaries. Table 10
summarizes the turnout in the primaries in Japan and Taiwan. The first noticeable feature
is the extremely high turnout for the first two primaries held in the LDP in 1978 and 1982
(both around 90%), much higher than subsequent internal elections in the LDP. Given
that these were the only instances in the LDP history where the share of the primary in
the first round was 100%, one may conclude that the weight of the primary may explain,
at least partly, the degree of interests of the eligible voters and the degree of participation
(or the mobilization by the politicians). However, if we compare the results of more recent
primaries in Japan, one can also find that the turnout varies considerably within the same
party even when the share of the primary is roughly the same (for example, 2009 and
20
2012 in the LDP and 2010 and 2012 in the DPJ). Thus, it is also clear that the weight of
the primary alone cannot fully explain the variance in the turnout. Third, if we compare
Japan and Taiwan, one cannot find that the turnout in Taiwan, in which the share of the
primary is always 100% and no run-off elections are held, is considerably higher than that
in Japan, where the share of the primary is much smaller and run-off elections are held.
Therefore, the cross-national comparison also suggests that the turnout (degree of
participation or the mobilization) is not necessarily higher when the primary has more
weight in the leader selection process.
Table 11. Competitiveness in the primaries in Japan and Taiwan
Source: Created by the authors. As for the primaries in the LDP from 1991 to 2008, we could not obtain the
number of votes gained by the candidates, so the ENC is calculated based on the points gained by the
candidates in the primary (denoted by *). Number marked by yellow denotes the incumbent candidate.
21
Let us now examine the degree of competition. Because we are dealing with the
internal elections where the number of position under competition is just one (party
chairpersonship), we calculated the effective number of candidates (ENC), following the
suggestions by Kenig (2009). This index is the same as the effective number of parties,
which takes into consideration the absolute number of candidates as well as their relative
strength in the elections. Table 11 presents the number of candidates, the ENC and the
share of votes of the three leading candidates in the internal elections in Japan and Taiwan.
This table demonstrates that, in some cases, the level of competition is indeed
quite low. For example, in two primaries held in the KMT in Taiwan, the margin between
the first and second candidates was fairly large (44.8 percentage points in 2005 and 74.4
in 2007). In the DPP primary in 2010, the margin was as large as 80.6 percentage points.
This low level of competition can also be observed among the Japanese parties, especially
in the primaries in the SDPJ, and also in some primaries in the LDP. This does not mean,
however, that all primaries lack serious competition. Quite on the contrary, some
primaries both in Japan and Taiwan were marked with fierce competition. Some of the
notable examples in this regard are as follows: Masayoshi Ohira and Takeo Fukuda in the
LDP in 1978; Junichiro Koizumi and Ryutaro Hashimoto in the LDP in 2001; Ma Ying-
jeou and Wang Jin-pyng in the KMT in 2005; Tsai Ing-wen and Koo Kwang-ming in the
DPP in 2008. One of the most oft-cited examples is the primary in the LDP in 2001, when
Junichiro Koizumi vowed to “destroy the LDP,” challenged the established leadership
within the party and made a landslide victory in the primary. The degree of
competitiveness thus varies significantly across primaries both within and between parties
in both countries.
Why are some primaries competitive while others are not? Is there any regularity
regarding the degree of competitiveness in the primaries? One hypothesis that came to
the mind of the authors is the presence or absence of the incumbent candidate. If the
incumbent is seeking reelection in the primary, this incumbent candidate may enjoy
considerable advantages vis-à-vis other candidates, and this may lead to the lower degree
of competition, i.e. the higher degree of concentration of votes in one candidate. In order
to examine whether this hypothesis can be supported empirically, we examined whether
the margin between the first and second candidates in the primary differs considerably
between the primaries with and without the incumbent. In Table 11, the number of the
vote share marked by yellow color denotes that the candidate is incumbent. Out of total
26 primaries, 10 were with incumbent candidates, and the average of the margin between
22
the first and second candidates is 44.2 percentage points, while that in the primaries
without incumbent is 39.9 percentage points. This data suggests that there is no significant
difference between the primaries with and without incumbent candidates (we conducted
a t-test and the difference was not statistically significant). At this moment, we cannot
find any factor systematically associated with the degree of competitiveness of the
primary in Japan and Taiwan: it seems to depend simply on the political context of each
primary.
4. Consequences of primaries
This section briefly analyzes consequences of primaries in Japan and Taiwan.
Two dependent variables are examined here, namely (1) membership of the political
parties and (2) party support for the political parties.
Impact on party membership
First, let us examine what kind of impact primaries had on the party membership.
Because the introduction of primaries for the leadership selection gives more internal
decision-making power to the grass-roots members, which means more incentives to
mobilize internally and be enrolled in the parties, it may have some positive impact on
the party membership. If this hypothesis is correct, one should observe an increase of
party membership when the primaries are adopted. Can we observe this increase in the
cases of Japan and Taiwan?
The case of the LDP in Japan provides some empirical support for this hypothesis.
When the primary was introduced for the first time in the LDP in 1978, the party
membership jumped from around 300,000 in 1977 to around 1.5 million in 1978
(Tsurutani 1980: 849). However, this increase was not because of the incentives for the
potential members to be enrolled to actively participate in the primary. Rather, it was
because the adoption of the primary gave incentives to the LDP parliamentarians to
increase the number of party members under his/her control to support his/her faction
leader who was running for the party presidency. According to Tsurutani, many LDP
politicians were “said to have registered and even paid party membership dues for whole
bloc of local voters, sometimes without the latter’s knowledge and, in a few instances, a
number of non-existent voters as well” (Tsurutani 1980: 853). An impressive high turnout
of the first primary in the LDP as presented above, therefore, may reflect not so much a
degree of active participation of the voters but rather a high degree of control exerted by
23
the LDP politicians on the voters at that time.
A positive impact of the primary can be observed for other parties. For example,
the “last-minute registration” occurred in the SDPJ before the voters’ list was finalized
for the chairperson election in December 1981: the party members increased by around
3,000 for one month in November 1981, considerably more than the typical figure of 500-
1,000.9 As for the NFP, it was reported that the party membership was around 200,000
as of June 1995, but it reached 491,000 when the NFP finalized the voters’ list of party
members10. As for the DPJ, it is estimated that the party membership was around 15,000
in 1997 and around 40,000-50,000 before the new millennium, but it increased
dramatically to around 300,000 in 2002 when a primary was held for the first time
(Uekami 2011a: 14). It is also reported that, even before 2002, the introduction of primary
at the prefectural level for the party leader selection led to a significant increase of party
members in that prefecture.11
These facts, however, do not necessarily mean that the trend of party membership
is correlated with the implementation of primaries. For example, as for the LDP, while no
primary was held in the 1980s after 1982, the party membership was on the increasing
trend, reaching its peak in 1991 with more than 5 million members.12 Even though
primaries have been held more frequently in the 1990s and even more frequently in the
new millennium with increasing weight in the selection process, the party membership
has kept decreasing since 1991.
How about the political parties in Taiwan? Unfortunately, for the KMT, we could
not find a relevant data to make an analysis in this regard. There are some indications,
however, that the introduction of primaries led to the increase of party members. For
example, Huang (1996: 114) points out that the adoption of the primary for the candidate
selection in 1989 “significantly contributed to the further expansion of the KMT
membership.” Fell (2010: 943) also points out that the adoption of primary for the
candidate selection led to the “Pan Blue merger,” i.e. merger of the politicians and parties
with similar political orientations. Though Fell does not mention the party membership,
this merger should have led to the increase of party members.
9 Asahi Shinbun, 11 December 1981. 10 Asahi Shinbun, 19 August 1995; Asahi Shibun, 7 December 1995. 11 In Fukushima prefecture, for example, the party membership increased from 331 on 6 June 1999, when the
prefectural office of the DPJ decided to conduct a primary for the party leader selection, to 847 on 13 June 1999, i.e. 2.6 times increase in just one week. See, Asahi Shinbun, 15 September 1999. 12 Because those who are eligible for the party president election are the party members who paid the membership
fee for the last two years, the number of eligible voters for the primary and the number of party membership does not
match, with the latter exceeding the former even though the former includes party supporters and other individuals.
24
As for the DPP, we cannot observe any particularly positive effect of primaries
on the party membership. Table 12 demonstrates the time-series data of DPP party
membership from 1986 to 2009. Primaries were held in 1998, 2006, 2008, 2010, 2012,
but as far as we can see, these years are not marked by a particularly large increase. The
year 1998 marked an increase by 19.9 percent, but given the overall upward trend of party
membership since 1996, this figure is not particularly high. The year 2006 was marked
by an increase by 2.5 percent, when the party membership reached its peak, while the
year 2008 saw a decrease by 6.2 percent. We can see that there is a clear long-term trend:
the party membership kept growing until 2006 since its inception, and started decreasing
since then. To analyze a precise effect of the conduct of primaries on the party
membership with this clear long-term trend requires a sophisticated methodology, which
is beyond the scope of the present paper.
Table 12. Party membership of the DPP, 1986-2009
Source: Website of the DPP. http://www.dpp.org.tw/history.php?data_type=%E9%BB%A8%E5%93%A1
Impact on party support
Let us now examine the impact of primaries on the level of party support in Japan
and Taiwan. In this regard, the literature is divided. On the one hand, some emphasize the
mutual de-legitimation between candidates and aggressive campaigns (Peterson and
Djupe 2005) and divisiveness of the primary (Johnson et al. 2010; Wichowsky and
Niebler 2010), implying that primaries have negative impact on the party performance in
the elections. On the other hand, others point to the fact that primaries can give a public
image of being open and “democratic” and give ordinary members and potential
supporters an incentive to be actively engaged in the party activities (Scarrow et al. 2000).
In order to examine the validity of these arguments, we examined the level of party
support before and after the party leader selection with and without primaries and see if
we can find a difference regarding the magnitude of party support change depending on
the type of selection process.
25
Table 13. The level of party support before and after the leader selection in Japan
Source: created by the authors, based on Jiji Press (1981); Jiji Press (1992); Jiji Press (1992-2013).
Table 13 presents the level of party support on monthly opinion poll in the
previous and subsequent time of the party leader selection and the magnitude of change
(increase by percentage) of the party support for all parties in Japan. For example, for the
LDP in 24th April 2001, when Koizumi won the party presidency after the dramatic
primary, the percentage of those who expressed the support for the LDP in the opinion
poll increased from 21.4 percent at the beginning of April to 25.5 percent at the beginning
of May – increase by 19.2 percent. On average, the magnitude of change in the level of
party support when primaries were held was an increase by 9.6 percent, while that without
primaries but with multi-candidate elections (voted only by the MPs) was an increase by
26
4.62 percent and that without any contest was an increase by 4.80 percent. This seems to
support the argument that the introduction of primaries has a positive impact on the level
of party support (and thus the electoral performance of the party). However, this positive
effect cannot be observed for other smaller parties, such as the SDPJ, DPJ and NFP. These
results seem to confirm the findings of Umeda and McElwain (2011), who analyzed the effect
of primaries on the party support for two major parties, the LDP and the DPJ, and concluded that
primaries had a positive impact on the party support for the LDP but not for the DPJ.
Unfortunately, we cannot make a similar analysis for the case of Taiwan, as we
do not have an equivalent data, taken so frequently and regularly, of the party support or
the party identification. However, some authors do suggest that introduction of primaries
did have a positive impact for the political parties. For example, Fell (2010) argued that
the conduct of the first multi-candidate primary in the KMT and the election of popular
candidate Ma Ying-jeou as party chairman had a significant impact on the increase in the
electoral support for the KMT. Table 14 demonstrates the share of those who expressed
the party identification in the opinion poll from 2004 to 2005. This demonstrates that the
share of those who express the party identification with the KMT increased considerably
after the election of Ma as party chairperson in July 2005. Even though the level of party
identification did not change immediately after MA’s election (in August 2005 it remained
almost the same as May 2005), it increased by 15 percentage points – or 47 percent
increase – in December 2005, when the local elections were held and Ma pledged to
resign his chairmanship if the party failed to win half the seats.
Table 14. Party Identification in Taiwan, 2004-05
Source: Fell 2010: 938.
Can we observe a similar effect for other cases of primaries? Table 15
demonstrates the annual data of the share of those who expressed the party identification
from 1997 until 2013. The multi-candidate competitive primaries were held in 2005 and
2007 in the KMT and in 1998, 2006, 2008, 2010 and 2012 in the DPP. The magnitude of
change in the level of party identification was 47.2 percent increase in 2005 and 4.2
27
percent decrease in 2007 for the KMT, while it was 27.3 percent increase in 1998, 13
percent decrease in 2006, 5.5 percent increase in 2008, 26.2 percent increase in 2010, 3.2
percent increase in 2012 for the DPP. Overall, the conduct of primary led to the increase
in five out of total seven cases. One may thus conclude that primaries tend to lead to the
increase of the party supporters in that year. Needless to say, however, this is only a rough
estimation, and a more precise and methodologically valid estimation of the effect of
primaries on the level of party support/party identification should be pursued elsewhere,
with various control variables taken into consideration. This is beyond the scope of this
paper.
Table 15. Party Identification in Taiwan, 1997-2013.
* denotes either "neutral" or "non-response".
Source: Election Study Center, National Chengchi University, important political attitude trend distribution.