. 31
( 35 .)


and cohesion, and so the president cannot always count on support from
his or her party. However, the Philippine president is much better
equipped to cobble together a legislative support coalition from across
the various veto gates than is the Thai prime minister. As mentioned
previously, the Filipino president is very powerful. These powers include
proactive powers, such as the ability to shape the legislative agenda,
appointment powers, and control over the pork barrel, as well as reac-
tive powers such as line item and package vetoes. Particularly notable is

In recent years, subnational governments in the Philippines have received larger
shares of government revenues via the International Revenue Allotment (IRA).
Building Party Systems in Developing Democracies

the president™s control of pork and political appointments “ control that
the Thai prime minister has historically lacked (Hicken 2005). In short,
prior to the Thai reforms, power was more horizontally concentrated in
the Philippines than in Thailand, and this generated stronger aggregation
In addition to the higher aggregation payoff in the Philippines, the
probability that the largest party will capture that prize has also
been greater. This was particularly true prior to martial law where
concurrent presidential and legislative elections typically produced two
viable candidates. In effect, this meant that the largest party in the
House had a good chance of capturing the executive “ giving candi-
dates a strong incentive to coordinate across districts under the banners
of the presidential frontrunners. This is in sharp contrast to pre-reform
Thailand where it was frequently the case that someone other than the
leader of the largest party served as prime minister.
To summarize, aggregation incentives in the Philippines typically
outdistanced incentives in pre-reform Thailand. The payoff to aggre-
gation in the Philippines has generally been higher than in pre-reform
Thailand due to a greater degree of horizontal centralization and the
better likelihood that the largest legislative party would capture the
aggregation payoff. Given these stronger aggregation incentives in the
Philippines, we would expect both the in¬‚ation score and the national
effective number of parties to be much lower vis--vis pre-reform
Thailand, as is the case (Table 6.7).
However, institutional reforms in each country led to a recent reversal
of this pattern. As discussed in this chapter, the post-Marcos introduction
of a presidential reelection ban in the Philippines caused a proliferation in
the effective number of presidential candidates and decreased the chance
that the largest legislative party would also control the presidency. The
result has been a sharp deterioration in aggregation and an increase in the
number of political parties. At the same time, constitutional reform in
Thailand increased aggregation incentives causing a dramatic improve-
ment in aggregation and reduction in the number of political parties
(see Chapter 5). The combined effect of both sets of reforms is that the
2001“5 Thai party system came to exhibit better cross-district coordi-
nation and fewer parties than the Philippines (see Table 6.7).35

The 2007 Thai Constitution threatens to undermine aggregation incentives.
Philippines: Term Limits, Aggregation Incentives, No. of Parties 179

6.5 conclusion
Like Thailand, the number of national parties in the Philippines re¬‚ects
both the average size of the district party systems together with the
degree of aggregation between those districts. Both periods of Philippine
democracy have used SMDP for House elections, resulting in a modest
number of parties at the district level. The shift to a multiparty system
since Marcos re¬‚ects primarily a deterioration of aggregation and ag-
gregation incentives. In the ¬rst democratic period, the expected utility
of being the largest party in government was high with both a high
aggregation payoff and a good chance that the largest legislative party
could capture that payoff. When democracy returned in 1986, a new ban
on presidential reelection led to an increase in the number of viable
presidential candidates, a lower probability of capturing the aggregation
payoff, and a corresponding decrease in the incentives to coordinate
across districts. This is consistent with the theory outlined in Chapter 2
and the large-N empirical results discussed in Chapter 3.
Together the experiences of the Philippines and Thailand suggest
that political institutions have a powerful and predictable effect on
aggregation. The rules and institutions in the Philippines were such that
for much of its democratic history the expected utility of becoming the
largest legislative party was quite high “ especially relative to pre-
reform Thailand. As a result, cross-district coordination was much
more extensive. However, institutional reform in both countries during
the 1980s and 1990s has brought about a reversal of this pattern.
Cross-district coordination has deteriorated in the Philippines (while
improving in Thailand) to the extent that the Philippines now has
worse aggregation and more parties than its neighbor.


This concluding chapter is divided into two parts. In the ¬rst section,
I summarize the central arguments and ¬ndings. In the second section, I
identify some of the questions that still remain to be answered and offer
some preliminary thoughts on the implications of various levels of
aggregation for policymaking processes and outcomes.

7.1 summary of key ¬ndings
In this book, I have focused on two dimensions of a country™s party
system “ the number of parties and the degree of the nationalization. I
have attempted to broaden the debate beyond the behavior of voters,
candidates, and parties within electoral districts to include a focus the
coordination of such actors across districts. I argued that aggregation is
a key determinant of both the size of the party system and the degree of
nationalization. Thus it is important to understand what factors shape
the degree of aggregation.
The causal logic of my argument was grounded in the incentives of
party entrepreneurs and candidates for political of¬ce. Aggregation is a
function of the incentives these actors face to ally across districts under
a common party banner. These incentives, in turn, are shaped by (1) the
potential payoff for aggregation and (2) the probability of capturing
that payoff. The incentives for coordinating across districts increase as
the rewards for such coordination rise and the degree of uncertainty
about capturing that reward falls.

Conclusion 181

To the extent the existing literature has explored the determinants
of aggregation incentives, it has focused almost exclusively on the
in¬‚uence of the vertical centralization of power within a political
system (i.e., the distribution of resources and authority between cen-
tral and subnational governments). I demonstrated both theoretically
and empirically that a high degree of vertical centralization is not
suf¬cient to produce strong aggregation incentives. A focus solely on
the distribution of power and resources between national and sub-
national actors misses a key part of the institutional story. To this
vertical dimension I added a second horizontal dimension and showed
that horizontal centralization, or the distribution of power within
the national government, combines with vertical centralization to
affect the size of the aggregation payoff and shape aggregation
I then explored three of the components of horizontal centralization “
bicameralism, party cohesion, and reserve domains. I argued that a
second chamber, party factionalism, and the presence of positions or
policy areas beyond the reach of elected politicians, each have the effect of
dispersing political authority reducing the size of the aggregation payoff
any single party is likely to control. The results of the large-N tests
revealed substantial support for the bicameralism and reserve domain
hypotheses, while party factionalism proved to be an important variable
in the Thai case.
In addition to the size of the payoff, both cross-national and
country-speci¬c evidence suggested that the probability of capturing
that payoff also plays an important role in shaping aggregation
incentives. I found that in parliamentary systems aggregation incen-
tives are stronger where the chance of the largest legislative party™s
capturing the premiership is high, and weaker where there is a high
probability that someone other than the leader of the largest party will
head the government. As hypothesized, I found that this probability
variable interacts with the size of the aggregation payoff to shape
coordination incentives. Aggregation is at its worst where a low
probability combines with a small aggregation payoff.
In presidential systems, I argued that the probability of capturing
executive of¬ce is a function of the number of presidential candidates and
the proximity of presidential and legislative elections. In line with existing
studies, I found that proximate elections lower the number of electoral
Building Party Systems in Developing Democracies

parties but only where the effective number of presidential candidates
is low. A unique contribution of this study was to demonstrate that
proximity and the number of presidential candidates also have an effect
on aggregation. Proximity lowers party system in¬‚ation where there are
relatively few presidential candidates, and the number of presidential
candidates itself has a substantial negative impact on cross-district
coordination. The more presidential candidates there are, the more dif-
¬cult it is for legislative candidates, voters, parties, and donors to identify
and coordinate around the frontrunners. The cost is poorer aggregation.
In addition to the large-N quantitative analysis, empirical support
for my argument came from the Thai and Philippine case studies. For
my analysis of the Thai case, I compiled a unique data set of district-
level election results for seven Thai elections since 1986. I discovered
that the large number of parties that so characterized the pre-1997 Thai
system was primarily a result of poor aggregation. Within individual
districts, actors were typically able to coordinate on a small number of
parties “ in line with what we would expect from Thailand™s electoral
system. Between districts, however, coordination attempts broke
down. I demonstrated that regional differences cannot adequately
account for these coordination dif¬culties “ aggregation was in fact
worse within regions than it was between them. Instead, poor aggre-
gation re¬‚ected the weak linkage present in pre-reform Thailand. Prior
to 1997, rampant party factionalism, an appointed Senate, and other
reserve domains undermined aggregation incentives. The practice of
selecting someone other than the leader of the largest party as premier
also reduced the expected utility of aggregation. Given the weak
aggregation incentives, cross-district coordination in pre-reform
Thailand was poor, and the result was a large number of parties at the
national level. By the 1990s, the lingering reserve domains had begun
to fade away, and Thailand had adopted a rule that enabled the leader
of the largest party to reliably capture the premiership. However, the
1997 constitutional reforms brought even more dramatic changes to
the political-institutional environment. The reforms greatly magni¬ed
candidates™ aggregation incentives. Speci¬cally, the reforms substan-
tially increased the premier™s leverage over internal party factions. This
increased the potential size of the aggregation payoff resulting in
stronger aggregation incentives, better aggregation, fewer parties, and
increased nationalization.
Conclusion 183

For the Philippines, I also used district-level election data to analyze the
extent of aggregation in Philippine elections since independence. I found
that aggregation was very good in pre“martial law elections but that it
deteriorated sharply after the return of democracy in 1986. This deteri-
oration is the primary cause of the demise of the two-party system since
Marcos. In the initial democratic period, the expected utility of being the
largest party in government was high with both a high aggregation payoff
and a good chance that the largest legislative party would capture that
payoff. However, when democracy returned in 1986, a new ban on
presidential reelection led to an increase in the number of viable presi-
dential candidates, a lower probability that the largest legislative party
would also capture the presidency, and a corresponding decrease in the
incentives to coordinate across districts. The reelection ban also explains
the relatively unusual pattern I observed in post-Marcos elections “
namely, poorer aggregation and more parties in concurrent elections,
better aggregation and fewer parties in midterm legislative elections.

7.2 unanswered questions and research
The multicountry analysis together with the experiences of the
Philippines and Thailand support the claim that political institutions
have a powerful and predictable effect on aggregation via their effect
on aggregation incentives. However, as always, the questions left
unanswered and issues left unresolved present ample opportunities for
further research. This book is no exception.
First, how does aggregation actually unfold on the ground? What
are the mechanics involved? On the one hand, we might envision
aggregation as a bottom-up process, driven mainly by the alliance
choices of local candidates or subnational political elite. On the other
hand, the recent Thai experience suggests that aggregation might also
be a top-down affair “ with political entrepreneurs taking the lead in
organizing cross-district coordination. This suggests several questions
worthy of future research. When aggregation occurs what conditions
shape whether it is a top-down or bottom-up process? Does the process
of aggregation affect the way in which the resulting parties are orga-
nized internally? Is there a relationship between the type of process that
predominates and the stability and endurance of cross-district electoral
Building Party Systems in Developing Democracies

alliances? Does the process by which aggregation occurs affect the
likelihood of democratic consolidation? Pursuing answers to these
questions is part of my future research agenda and is likely to require
the use of multiple methods “ from formal modeling of voter, candi-
date, and elite behavior (see Morelli 2001) to careful ¬eldwork
studying how aggregation unfolds in speci¬c country contexts.
Second, does aggregation in¬‚uence dimensions of the party system
other than nationalization and party system size? For example, aggre-
gation is potentially an important determinant of the degree of party
system institutionalization. Mainwaring and Scully (1995) discuss four
criteria for party system institutionalization. The last of these criteria
deals with party organization and includes the notion that parties should
be “territorially comprehensive” (Mainwaring and Scully 1995, 5).
Party system institutionalization need not imply that support for parties
must be equally distributed across the nation, but it does imply that
parties reject strictly local or regional strategies in favor of a more na-
tional focus. Party system institutionalization may be weaker, then,
where parties fail to move beyond local strongholds and compete across
districts or regions nationwide.
Finally, the question of greatest interest to me in terms of a future
research agenda is the way in which aggregation affects policymaking “
both processes and outcomes. If one is to understand the dynamics of
policymaking, it is useful to begin with three sets of questions.
 Who are the actors that make policy decisions?
 What are their interests? (To whom do they respond? What is the
nature of their constituency?)
 What are their capabilities? (How able are they to implement their
preferred policies? What constraints do they face?)


. 31
( 35 .)