. 28
( 35 .)


The fact that the U.S. colonial administration retained the major responsibility for
public policy, even after a national elected legislature was in place, did much to
facilitate the development of locally focused particularistic parties. It was not until
1934, when the Philippines was able to win Commonwealth status, that the provision
of national policies fell to elected politicians. Prior to 1934, national policymaking
was the purview of the colonial government, and as a result parties and elected
representatives were free to engage in other pursuits (Stauffer 1975).
For an opposing view (i.e., that the reports of oligarchs™ deaths were highly exag-
gerated), see Putzel 1993.
Building Party Systems in Developing Democracies

party system. One explanation for the continuity of the party system,
despite the signi¬cant changes that occurred before and during the
Marcos era, is the continuity of key features of the Philippine institu-
tional environment.12
Alongside the historical and sociological factors discussed earlier,
certain features of the Philippines institutional environment encourage
the development of locally focused, noncohesive parties “ namely, a
powerful presidency and the electoral system. These two features have
remained relatively constant across the pre- and postauthoritarian
periods and reinforced, and in some cases ampli¬ed, the effects of
sociological and historical factors.
To begin with, many Filipino scholars blame the establishment of a
strong president for the state of the party system (see, e.g., Grossholtz
1964; Wurfel 1988; Banlaoi and Carlos 1996).13 A powerful presidency
undermines party cohesiveness, frees legislators and parties to focus on
particularistic concerns (leaving national policies in the hands of the
president), and generally discourages the development of a structured
party system. This observation is not unique to the Philippines “ pre-
sidentialism is often associated with weak and noncohesive legislative
parties (Lijphart, Rogowski, and Weaver 1993, 322).14
In addition to the powerful presidency, the nature of the Philippines™
electoral system also helped shape the development of the party system.
Speci¬cally, the electoral systems for the House and Senate give candi-
dates strong incentives to pursue a personal strategy while discounting
the value of party label. Members of the House of Representatives are

The unwillingness of Aquino to capitalize on her popularity to form her own political
party or take over the leadership of an existing party also contributed to the return of
an unstructured party system.
The pre-1972 Philippines presidency was among the strongest in the world. In the
aftermath of the Marcos regime, some of that authority was curtailed, but the
president retains an impressive array of both proactive and reactive powers. In
Shugart™s index of presidential power, the post-1986 presidency rates as “strong”
(Shugart 1999).
However, one must be cautious regarding the direction of causality. While a rela-
tionship exists between strong presidents and unstructured parties the causal arrows
can run both ways. A strong presidency may prevent the rise of a structure party
system, but it may also be employed as an institutional antidote in polities with
unstructured parties (Shugart 1999). In fact, the effort to institutionalize a powerful
executive by the Philippines™ ¬rst President, Manuel Quezon, was in part a reaction to
the perceived shortcomings of the party system (Quezon 1940).
Philippines: Term Limits, Aggregation Incentives, No. of Parties 161

elected from single-seat districts using the plurality rule (SMDP) “ a
system that is often associated with weak parties and locally focused
legislators (Cain, Ferejohn, and Fiorina 1987; Carey and Shugart 1995;
Cox and McCubbins 1993).15 This is certainly the case for the version of
SMDP used in the Philippines. Candidates are not required to obtain the
nomination or endorsement of a political party in order to run for of¬ce.
Candidates may run as independents or run under the banner of more
than one party.16 For their part, party of¬cials often lack strong control
over nomination and endorsement within their own party. Strong can-
didates can usually run under the label of their choosing. In some cases,
strong/wealthy candidates will use a party™s label with or without the
party™s of¬cial endorsement (Wurfel 1988, 96). Districts featuring mul-
tiple candidates from a single party (known as “free zones”) were more
common before martial law (63% of districts), but after martial law free
zones were still found in 20% of the districts nationwide (Kasuya 2001a).
In such situations, intra-party competition can arise with two or more
candidates from the same party running against each other. This lack of
party ballot control undermines candidates™ incentives to pursue party-
centered campaign strategies. In fact, according to Carey and Shugart,
the type of electoral system used in the House generates some of the
strongest incentives to cultivate a personal vote of any electoral system
(Carey and Shugart 1995, 425). This cultivation of a personal vote comes
at the expense of the party label and party cohesion and generally requires
candidates to focus on narrow constituencies.
The method of electing the Senate has also played a part in the
development of an under-institutionalized party system. The Senate

The provision for a mixed-member system was included in the 1987 Constitution, but
a law fully implementing the measure was not passed until 1995 and not used in an
election until 1998. The party list seats make up to 20% of the total House and are
allocated using proportional representation. Both political parties and sectoral
organizations can compete for the seats, save the ¬ve largest parties from the previous
election, which are barred from competing. To obtain a seat, parties (or sectoral
organizations) must receive at least 2% of the party list votes. For every 2% of the
vote, a party is awarded a seat, with an upper limit of three seats in the list tier. During
the 1998 elections, only 13 parties passed the 2% threshold and so many party list
seats were un¬lled. The remaining seats were ¬lled by appointed representatives from
groups that fell below the threshold. The rule was subsequently changed to require
that un¬lled seats be distributed among parties above the 2% threshold, but below
the three-seat cap (Hicken and Kasuya 2003).
See the earlier discussion of guest and joint candidacies.
Building Party Systems in Developing Democracies

consists of 24 seats, with 12 seats contested every 3 years. Senators are
elected from a single nationwide district using the block vote electoral
system “ the same system used in Thai elections. Each voter casts up to 12
votes “ each for a distinct candidate. Seats are awarded to the 12 senators
with the highest vote totals. As in the Thai case, the block vote encourages
senatorial candidates to eschew party strategies in favor of personal
strategies. Senate elections are ¬rst and foremost personality contests,
and senators generally possess little in the way of party loyalty. Multiple
votes allow voters to split their votes among senatorial candidates from
different parties “ something Filipino voters frequently take advantage of.
In every Senate election since 1957, voters have returned candidates from
more than one party. (If voters were casting votes on the basis of party
label, they would cast all of their votes for candidates from the same
party, and all 12 seats would go to a single party.)17 The vote shares of
copartisans also vary widely, another indication that voters split their
votes among different parties. In the post-1987 Senate elections, a party™s
top vote-getter has received as many as 2.7 times the number of votes as
other victorious copartisans.18
Another contributing factor is the write-in ballot used in Filipino
elections. Voters are required to write in the name of each of their chosen
candidates for every elected of¬ce. Given that local and national elections
are synchronized, this can mean that voters must write-in up to 40 names
on election day.19 This cumbersome ballot structure provides voters with
ample opportunities to split their votes among many different parties,
thus undermining the value of party label. This was not always the case.
Shortly after independence, the election code was revised to allow for
party voting. Rather than writing individual candidates™ names, voters
could write in the name of a party and the ballot would be “deemed as a

This assumes that all parties present a full slate of senatorial candidates and that
voters cast all 12 votes. In reality, the largest parties almost always run a full slate,
while many smaller parties present only partial slates. In addition, voters may not cast
all of their votes.
Author™s calculations from Commission on Elections (1992, 1995, 1998, 2001,
For this reason, the distribution of sample ballots to voters becomes extremely
important. Prior to elections most candidates distribute sample ballots containing
their name and the names of candidates for other of¬ces. Tellingly, it is not un-
common for these sample ballots to contain the names of candidates from more than
one party. Candidates often include popular candidates from other parties running in
other races on their sample ballot in a bid to bolster their own electoral prospects.
Philippines: Term Limits, Aggregation Incentives, No. of Parties 163

vote for each and every one of the of¬cial candidates of such party for the
respective of¬ces” (Revised Election Code of 1947, Article XI, Section
149, No. 19). A 1951 amendment to the Election Code eliminated the
party voting option and, subsequently, split ticket voting became the
norm in elections (Wurfel 1988, 94).
To summarize, the Philippine electoral system is one that discourages
the development of an institutionalized party system. In the House,
SMDP with weak control of nominations undermines the value of party
label. Similarly, the Senate™s block vote system privileges personal over
party strategies. Finally, the write-in ballot gives voters ample oppor-
tunity to split their votes between different parties.

6.3 accounting for change in the ¬lipino
party system
While in many respects the Philippines™ party system today looks similar
to the pre“martial law norm, this constancy masks changes in aggregation
and aggregation incentives over time. In the last chapter, I discussed the
way in which constitutional reforms altered aggregation incentives in
Thailand. Constitutional reform also led to change in aggregation
incentives in the Philippines. In this section, I explore the way in which
aggregation has varied across the Philippines™ democratic history, focus-
ing speci¬cally on the pre“ and post“martial law party systems. Recall that
from 1946 to 1972 the Philippines had regular, democratic elections.
President Ferdinand Marcos brought an end to democracy by declaring
martial law on September 21, 1972. After 14 years of the Marcos regime,
democratic government returned to the Philippines in 1986 in dramatic
fashion. After Marcos™s attempt to steal the 1986 snap presidential elec-
tion, millions of Filipinos jammed the streets of EDSA avenue in a show of
people™s power. For three days, the people stared down Marcos and the
military, culminating in the end of the Marcos™s dictatorship and a return
to democratic government.
Many features of the pre-1972 party system returned along with
democracy. The Nacionalista and Liberal Parties were reborn after
Marcos, though at only a fraction of their former strength. Many of the
individual and family faces prominent during the earlier democratic
period reemerged as party leaders after 1986 and, as discussed in the
preceding section, post-Marcos political parties were just as weak as
Building Party Systems in Developing Democracies

table 6.4. Aggregation and the Number of Parties before and after
Martial Law

Country (election year) ENPavg Average ENPnat Average I
Philippines I (1946“69) 2.0 2.3 9.8
Philippines II (1992“98) 2.3 3.6 32
Sources: Author™s calculations from COMELEC (various years); Hartmann et al.
(2001); Kasuya (2001b).

their predecessors. However, at least one characteristic of the earlier
democratic period did not return “ the stable national two-party system.
From 1946 to 1969, the average effective number of national parties
(ENPnat) was 2.3. Since 1987, ENPnat averages 3.6 (see Table 6.4),
reaching a high of just under 5 in the 1992 election.20 This rise in the
number of parties has been the subject of much scholarly attention in the
Philippine literature (see for example Kimura 1992; Kasuya 2001b). I
will ¬rst demonstrate that the rise in the number of parties nationally is
primarily the result of deteriorating aggregation post-Marcos. I will
then review some of the existing explanations for the rise in the number
of parties and argue that they are either incorrect or incomplete. Finally,
I will show how changes to the 1987 constitution undermined the
incentives to coordinate across districts during concurrent elections.
6.3.1 The Deterioration of Aggregation Post-Marcos

What is the source of the growth in the number of national parties post-
Marcos? As Table 6.5 makes clear, this growth is not a result of many
more parties winning seats at the district level after 1986. From 1946 to
1969, the average effective number of parties at the district level was 2.0.
After 1986, ENPavg increased only slightly to 2.3, an increase of less than
11%.21 Indeed, a large change in ENPavg would be surprising given that

Data are from the 1992, 1995, and 1998 elections. The 1987, 2001, and 2004
elections are excluded due to the lack of comprehensive data on candidate and party
vote shares.
ENPavg for the Philippines is actually the average effective number of candidates
(ENCavg). Because House elections use single-seat districts the ENCavg should be
nearly equal to ENPavg. The exception is where, due to the Philippines lax nomination
requirements, more than one candidate declares for the same party. Where this is the
case ENCavg will be slightly higher than ENPavg.
Philippines: Term Limits, Aggregation Incentives, No. of Parties 165

table 6.5. ENPres, ENPnat, and ENPseat Compared (Concurrent

Year ENPres ENPnat ENPseat
1946 2.0 3.3 2.9
1949 2.4 2.4 2.1
1953 1.7 2.6 2.3
1957 3.4 2.1 1.5
1961 2.0 2.0 1.7
1965 2.2 2.3 2.1
1969 1.9 2.1 1.5
Average: 1946“69 2.2 2.4 2.0
1992 5.8 5.0 3.5
1998 4.3 3.1 2.7
2004a NA
3.2 1.7
Average: 1992“98 5.1 4.0 2.6
In 2004, the formal K-4 electoral alliance is counted as a single party.
Sources: Author™s calculations from COMELEC (various years); Hartmann et al.
(2001); Tehankee (2002).

the electoral rules for the House remained virtually unchanged between
the two periods. This small increase in the average size of the local
party system cannot account for the 58% increase in the size of the
national party system. In short, intra-district coordination failures are not
primarily to blame.
Cross-district coordination is another story. Prior to martial law,
aggregation between districts was extremely good. The same two parties
were the frontrunners in most districts nationwide. Thus the average


. 28
( 35 .)