. 26
( 35 .)


appointed Senates from Thailand™s past. Under the 1978 and 1991
Constitutions, Senators were appointed by the king, and the appoint-
ment then had to be countersigned by the prime minister. In practice,
the prime minister often played an important role in recommending
potential Senators to the king for his endorsement. Under the 2007
charter, the prime minister no longer has a direct role to play in the
appointment process.38 He is not a member of the selection panel, and
he no longer countersigns the selection list. This raises the possibility
that the Senate could once again become a serious reserve domain for
Thailand™s conservative forces.
In the end, the reintroduction of appointed senators should reduce
the leverage future prime ministers will have over the Senate. In fact,
given that the prime minister is now completely cut out of the selection
process, his in¬‚uence over the Senate may be even less than it was
during the late 1980s and 1990s. The fact that the Senate still possesses
only delaying power in legislative mitigates somewhat the effect of
these changes, but on balance the switch to a partially appointed Senate
could potentially undermine aggregation incentives.

National Party List Tier
Following a growing trend (see Shugart and Wattenberg 2000) the
1997 Constitution drafters established a mixed-member or two-tiered
system in Thailand. The 400 seats of the nominal tier were elected from
single-member districts on a plurality basis as described previously.
One hundred additional seats made up the list tier and were elected
from a single nationwide district via proportional representation. (The

Disturbingly, the Senate itself selects many of these appointees in the ¬rst place, at
best muddying the lines of accountability, at worst raising the possibility of a quid
pro quo.
Note the monarch also appears to play no direct role in the selection process, which is
different from the 1978 and 1991 Constitutions.
Building Party Systems in Developing Democracies

2007 Constitution reduced the list tier to 80 seats and divided up the
national district into eight separate regional districts.) Each party was
required to submit a list of candidates for voters to consider, and voters
cast two votes: one vote for a district representative and one for a party
list. Candidates had to choose between either running in a district or
running on the party list. The two tiers were not linked in any way (i.e.,
votes from one tier did not transfer to the other tier).
How might the change to a mixed-member system lead to changes in
aggregation incentives and the number of parties? There is no
straightforward expectation since the addition of a national party list
tier generates competing incentives. On the one hand, the mixed-
member system as used in Thailand gives Thai voters multiple votes,
and we know that multiple votes tend to put upward pressure on the
effective number of parties (Lijphart 1994, 118“24). In addition, the
national party list potentially makes it easier for small, subnational
parties to win seats in parliament. Since seats are awarded on a pro-
portional basis, small parties that in the past could not win, nor even
¬eld enough candidates to meet the electoral requirement, can now win
seats as long as they obtain over 5% of the party list votes. If many more
small parties won seats under the new system, this would increase the
effective number of parties.
Pushing in the opposite direction are the stronger incentives to
coordinate across districts generated by the presence of the national list
tier. The presence of the national list tier is in part an electoral bonus for
linking together to form a large, national party (and to that extent, the
change to regional part lists in 2007 should undermine aggregation
incentives). Parties large enough to be competitive across the nation
should capture more list tier seats than provincial or regional parties
with limited national appeal. Locally strong parties but nationally
weak parties may be able to capture some seats in the nominal tier, but
national parties should dominate the list tier seats both at the national
level and in any given district.
In conclusion, even though the list tier may have important con-
sequences for other aspects of Thailand™s party system (Hicken 2006),
its effect on aggregation is theoretically indeterminate.

Term Limits, Aggregation Incentives, and the
Number of Parties in the Philippines

6.1 introduction
In many respects, the party systems of Thailand and the Philippines look
very similar. In both countries, party labels have historically been weak,
party switching is rampant, and party cohesion is low. Where Thailand
and the Philippines diverge is in the number of parties at the national
level. Recall that in Thailand the average effective number of parties
nationally prior to constitutional reform was 7.2. The corresponding
¬gure for the Philippines over the course of its democratic history is a
more modest 2.6. However, as in Thailand, there is substantial variation
over time in the number of parties. Speci¬cally, in the democratic period
before martial law the effective number of parties at the national level
averaged 2.3. After the fall of Marcos, the number of parties increased to
3.6 on average. Why this large increase? The in¬‚ation of the party sys-
tem post-Marcos has long been a puzzle for scholars of Philippine pol-
itics, and this chapter provides an answer to that puzzle that is superior
to existing explanations. How, too, do we explain the differences in the
size of the Philippine and Thai party systems? Drawing on the theory
from Chapter 2, I explain, ¬rst, why the post-Marcos party system has
been much larger than the pre“martial law party system and, second,
why cross-district coordination differs across the two countries. I dem-
onstrate that differences in the number of parties between Thailand and
the Philippines are primarily a product of aggregation and not variations
in the two countries™ electoral systems.

Building Party Systems in Developing Democracies

This chapter proceeds as follows. I ¬rst brie¬‚y review the history of
the Philippine party system and provide a basic description of its
characteristics. I next apply the aggregation incentives theory to the
pre- and postauthoritarian Filipino party systems to explain the growth
in the effective number of national parties since 1986. I argue that
adopting a one-term limit on the presidency introduced greater uncer-
tainty into presidential elections. The result is delayed aggregation and
the reverse of the concurrency effect usually observed in presidential
elections (Shugart 1995). Finally, I compare the local and national
party systems in Thailand and the Philippines and demonstrate that the
difference in the number of parties nationally is foremost a re¬‚ection of
different levels of aggregation (and aggregation incentives) across the
two countries. For most of its democratic history, aggregation incen-
tives in the Philippines were stronger than those in Thailand, but recent
institutional reforms in both countries have brought about a reversal.

6.2 the history and development of the
philippines party system
The Philippines has one of the oldest democratic traditions in Asia.1
Under U.S. colonial auspices, elections for both national and local
of¬ces were the norm in the Philippines from the early 1900s. After a
brief interruption during Japanese occupation, elections resumed in
1946 in a fully independent Philippines. Elections were a mainstay of
Filipino life until 1972 when President Ferdinand Marcos declared
martial law. After 14 years of dictatorship, democratic government
was restored in 1986. The post-Marcos constitutional drafters chose to
reinstate the pre-Marcos American-style presidential system, with an
elected president, a House of Representatives and a Senate.
What has the Filipino party system looked like during the two
democratic periods since independence? (See Box 6.1.) As I will discuss
in more detail later the post“Marcos party system was different from

Thanks in part to the country™s relatively long electoral history, Filipino elections and
parties have received more scholarly attention than their Thai counterparts. See for
example, Hayden (1950), Grossholtz (1964), Corpuz (1965), Land (1965, 1996),
Liang (1970), Wurfel (1988), Tancango (1992), Carlos and Banlaoi (1996), Banlaoi
and Carlos (1996), Carlos (1998a, 1998b, 1998c); Hartmann, Hassall, and Santos
(2001), and Hutchcroft and Rocamora (2003).
Philippines: Term Limits, Aggregation Incentives, No. of Parties 151

Box 6.1: Four Periods of Filipino Party Development

U.S. Colonial Period: 1901“1946. As the U.S. colonial administration
organized ¬rst local, then national elections, several parties formed to
compete for elected of¬ce. By the 1907 election, the Nacionalista Party
(NP) had become the largest party, and it remained the largest party in
virtually every election from 1907 to 1941. However, the NP did not
go unchallenged. Factional in¬ghting caused regular defections from
the NP, and strong second parties ran against the NP in many elections.
Two-party Period: 1946“1972. After independence, two parties
emerged as the dominant electoral forces “ the Nacionalista Party
and the Liberal Party. These two parties were the largest parties in
every national election from 1946 to 1972. The relatively stable two-
party system masked signi¬cant factional splits within the two par-
ties and frequent party switching or ˜turncoatism™ by politicians.
Marcos Dictatorship: 1972“1986. After President Marcos declared
martial law in 1972, all existing parties were dissolved, including
Marcos™s own NP. In their place, Marcos organized Kilusang Bagong
Lipunan (KBL or New Society Movement). The KBL was the de facto
party of the government and dominated elections during the Marcos
years. Several opposition parties were organized in the late 1970s and
early 1980s. During the 1980s, several of these parties allied together to
form the United Democratic Opposition (UNIDO).
Post-Marcos Period: 1986“. The defeat of Marcos in 1986
brought the return of elected government and most of the democratic
institutions from the pre-Marcos period. The two-party system did
not return, however. Instead, a multiparty system has emerged
marked by frequent party turnover.

its predecessor in at least one important respect “ the number of national
parties. A relatively stable two-party system had been the norm pre-
Marcos “ with the Nacionalista and Liberal Parties vying for power in
every election. However, it was a multiparty system that materialized
after democracy returned. Although this change is interesting and
signi¬cant, the change from a two-party to multiparty system masks
an underlying constancy on other dimensions of the Philippine party
system. Parties in both periods are characterized by factionalism,
frequent party switching (called turncoatism in the Philippines), and
party labels that generally mean little to voters or candidates. Like
Building Party Systems in Developing Democracies

Thai parties, Philippines parties are generally organized around a
powerful leader, or a temporary alliance of leaders, and tend to be
primarily concerned with distributing the spoils of government to
themselves and their local supporters.
The lack of intra-party cohesion is one of the most notable features of
Filipino parties (Land 1965, 1996; Machado 1978; Banlaoi and Carlos
1996). Parties are generally not uni¬ed actors, but instead are atomized
and/or composed of competing factions or “wings.” Factionalism
emerged early as a property of the party system and continues to be a
feature of many parties. Another indication of the lack of cohesion is the
tradition of party switching, or turncoatism. Party switching is a regular
part of every election in the Philippines, including elections for the
highest of¬ce in the land. Ramon Magsaysay and Ferdinand Marcos left
leadership positions in the Liberal Party to run for president under the
Nacionalista label. In 1992, Fidel Ramos formed LAKAS-NUCD to
support his presidential bid after he failed to win the LDP nomination.
At lower levels of government, party switching occurs both before and
after elections and has important implications for the political economy
of policymaking (Hicken 2002). Party switching in the Philippines is
generally in one direction “ toward the president™s party. The Philippines
president controls valuable resources “ namely pork and political
appointments. As a result, candidates make an effort to align themselves
with the strongest presidential contenders prior to elections. Once the
presidential elections are complete, many who ¬nd themselves in the
parties of losing candidates rush to join the president™s party. Indeed,
within the House of Representatives enough party switching can occur
to change the status of the president™s party from the minority to the
majority party, as happened after the election of Presidents Macapagal,
Marcos, Aquino, and Ramos (Liang 1970; Banlaoi and Carlos 1996;
Land 1996).
Switching to the president™s party in an effort to maximize govern-
mental largess is a fact of life at all levels of government, from members of
Congress to local of¬cials. Table 6.1 presents party switching data from
the 1995 election as an example. The pattern of switching is clearly
evident in the behavior of incumbent House members and governors.
Overall, more than 48% of incumbent representatives switched parties
between the 1992 and 1995 elections, with most switching shortly after
the 1992 election was complete. Of the turncoats, nearly 90% joined
Philippines: Term Limits, Aggregation Incentives, No. of Parties 153

table 6.1. Party Switching Prior to the 1995 Election

Percentage of incumbent House members that switched 48.5
Percentage switched to LAKAS-NUCD 89.9
Percentage switched to other parties 10.1
Percentage of incumbent governors that switched 35.1
Percentage switched to LAKAS-NUCD 88.5
Percentage switched to other parties 11.5
Sources: Author™s calculations from Commission on Elections (1992, 1995).

President Ramos™s LAKAS-NUCD party. This is supportive of Kasuya™s
(2001a) ¬nding that for all House elections (1946“71; 1992“8) an
average of 49.3% of opposition incumbents switched to the president™s
party by the next election.2 The pattern is similar among incumbent
governors. Prior to the 1995 election, 35% of incumbent governors
had switched parties, and of those, 88.5% had moved to the president™s
The high rate of party switching is indicative of weak party labels in
the Philippines. Another indication is the relatively high level of electoral
volatility “ particularly since the return of democracy in 1986. During
the pre“martial law period, the dominance of the Liberal and Nacio-
nalista Parties is re¬‚ected in a relatively low electoral volatility score of
18.5. In comparative terms, this tally places the Philippines on par with
France and Argentina (Table 6.2). Since 1986 the electoral fortunes of
Filipino parties have been much more unstable.3 Electoral volatility for
the 1992, 1995 and 1998 elections is more than double the pre“martial
law ¬gure at 37.3, more than the volatility in Brazil and Thailand.4


. 26
( 35 .)