Content uploaded by Kent Howard Eaton
Author content
All content in this area was uploaded by Kent Howard Eaton on Aug 23, 2021
Content may be subject to copyright.
Political Obstacles to Decentralization: Evidence from
Argentina and the Philippines
Kent Eaton
ABSTRACT
Decentralization has swept across the developing world in recent years.
Although the speed and scope of the shift toward more decentralized practices
is striking, decentralization is neither inevitable nor irreversible. Rather, it
faces enormous political obstacles and can be subject to serious setbacks. This
article accounts for attempts by national politicians to thwart decentralization
in two countries that recently adopted some of the most significant decentral-
izing changes in their respective regions: Argentina and the Philippines. Based
on fieldwork in each country, it suggests that even after the political decision
to decentralize has been made, national politicians may face deep-seated
incentives to preserve centralized control over fiscal policy. In Argentina,
President Carlos Menem partially reversed the previous decentralization of
revenue because fiscally-independent provincial governors were a challenge to
his political interests and capabilities. In the Philippines, legislators attempted
to reverse and then circumvent decentralization since it threatened their status
as brokers claiming personal credit for negotiating fiscal transfers from
the centre. The article identifies an intermediate outcome in both countries,
according to which decentralizing policies are neither entirely reversed nor
implemented as initially designed.
INTRODUCTION
The late twentieth century shift toward more decentralized forms of govern-
ment promises to change political life in developing countries in profound
ways. By shifting power to sub-national governments, decentralization has
the potential to transform some of the most significant actors and relationships,
including the developmental capacity of states, the behaviour of interest
groups and social movements, and the strategic calculations of politicians,
non-governmental organizations, and voters. As an indication of the importance
For helpful comments on an earlier version of this article, I would like to thank John Carey,
Alberto Dõ
Âaz-Cayeros, Eduardo Go
Âmez, Mark Jones, Sylvia Maxfield, David Samuels, Steven
Webb, Eliza Willis and three anonymous referees at Development and Change. The research on
which the paper is based was made possible by support from the Fulbright Foundation, the
Woodrow Wilson International Center for Scholars, and the Center for International Studies at
Princeton University.
Development and Change Vol. 32 (2001), 101±127. #Institute of Social Studies 2001. Published
by Blackwell Publishers, 108 Cowley Road, Oxford OX4 1JF, UK.
of these changes, scholars have articulated a number of different explana-
tions for decentralization, focusing on the possible causal roles played by
economic crisis, the advice of external donors, pressures from sub-national
actors, and attempts by national politicians to reconstruct state±society
relations.
1
In addition to asking why policy-makers decide to decentralize,
debates have focused on the impact of this decision on long-term prospects
for economic development.
2
This article focuses not on the (undeniably important) initial decision to
decentralize, but on what happens next. In scores of developing countries,
the adoption of decentralizing policies is the beginning of the story rather
than the end, setting into motion open-ended struggles between actors at the
national and sub-national levels over the form that decentralization will
take. Not only do we need a better understanding of why political actors
adopt decentralizing policies, we also need to avoid the mistake of assuming
that such policies are implemented as adopted. This is not just for the usual
reason that bureaucrats enjoy discretion in implementing the policies that
politicians legislate, but also because national politicians can and do
continue to use their legislative authority to modify the initial decision to
decentralize. Rather than being a one-shot deal, the decision to decentralize
opens up a new and contentious arena for political struggle.
Understanding the struggles over decentralization that thus emerge requires
us to take seriously the political interests of politicians. In spite of differences
between those following more technocratic or more political approaches to
decentralization, most clearly agree that political factors shape the response
to decentralization.
3
For example, Crook and Manor (1998: 302) warn that
`decentralisation cannot work independently or even against contradictory
forces coming from the social and political structures within which they are
embedded'. According to Samoff, `decentralization is effective only when it
is compatible with the interests of those expected to implement and defend
it' (Samoff, 1990: 523). As Scho
Ènwa
Èlder argues, the overall political setting
in which decentralization schemes occur needs to be made more explicit in
the literature, including attention to bureaucratic obstacles and politically-
motivated resistance from local elites (Scho
Ènwa
Èlder, 1997: 761). Although
he has been criticized for neglecting the political dimensions of decentraliza-
tion, Rondinelli also underscores the salience of political variables by noting
1. For a comprehensive discussion of the possible causes of decentralization, see Manor
(1999).
2. The participants in this debate generally subscribe to either optimistic or pessimistic views
of decentralization, though this distinction obscures important nuances in their argu-
ments. Rondinelli's work is representative of the more sanguine view; see for example
Rondinelli et al. (1989). Those who emphasize the possible negative consequences of
decentralization include Prud'homme (1995) and Slater (1989).
3. For different approaches to decentralization, see the exchange between Rondinelli (1990),
Slater (1990), and Samoff (1990).
102 Kent Eaton
that `attempts to counter centralization are intensely political activities; they
inevitably produce political consequences' (Rondinelli et al., 1984: 2).
This article seeks to contribute to the literature by examining why national
politicians often renege and back-track on their earlier support for decentral-
ization. Based on fieldwork in two countries from different regions of the
developing world, it examines the aftermath of the decision to decentralize,
and shows that decentralization is neither inevitable nor irreversible. Not
only can jealousy on the part of national politicians damage decentraliza-
tion, as Manor has argued, but it can also lead to actual instances of re-
centralization (Manor, 1995: 88). In Argentina, the adoption of revenue and
expenditure decentralization in the late 1980s and early 1990s was followed
by a series of policies that substantially weakened decentralization. As a result
of changing political calculations and electoral results, the very decentraliz-
ing policies that had once advanced the interests of national politicians
threatened these same interests just a few years later. In the Philippines,
legislators passed a historic decentralization act in 1991 and then immedi-
ately initiated partially successful attempts to limit the scope of decentral-
ization and to re-centralize authority over previously devolved services. This
seemingly contradictory and irrational behaviour becomes intelligible only
when the political incentives facing these legislators are taken into account.
THE POLITICS OF DECENTRALIZATION AND RE-CENTRALIZATION
Before proceeding, it is important to establish what is meant by the term
`decentralization'. The literature distinguishes between deconcentration and
devolution as the two main forms of decentralization (Crook and Manor,
1998; Rondinelli, 1989). Deconcentration refers to changes within central
government ministries that re-locate bureaucrats to the sub-national branches
of these ministries. According to Crook and Manor (1998: 6), because central
ministries do not surrender authority over bureaucrats who remain employees
of the central government, deconcentration may actually strengthen the central
state by reinforcing its ability to penetrate into territory where its authority
has traditionally been weak. Devolution, in contrast, transfers authority to
sub-national political actors, `the activities of which are substantially outside
the direct control of the central government' (Rondinelli et al., 1984: 19). In
order for devolution to occur, democratic elections for sub-national actors
are a necessary, but not sufficient, condition (Manor, 1995: 82). Not only do
sub-national actors like mayors and governors need to be independently
elected, they must also be endowed with independent authority over revenues
and expenditures.
The decision to deconcentrate authority within central ministries and the
decision to devolve power to sub-national actors are both the prerogative of
national policy-makers. Support by national politicians for deconcentration
appears straightforward since it is seen as a way of improving their ability to
Political Obstacles to Decentralization 103
command the behaviour of a variety of actors at the sub-national level.
Support for devolution is more puzzling, since by definition it leads to
immediate decreases in the influence that national politicians have over
revenues and expenditures, and concomitant increases in the policy-making
authority of independent actors. Why would politicians surrender power?
In grappling with this question, scholars have generally argued that
national politicians use decentralization to try to fix a variety of long-term,
systemic problems in national political systems. For example, Garman et al.
(1996) argue that presidents in Latin America have supported decentraliza-
tion as a type of public good that can improve overall collective welfare and
reduce the inefficiencies that characterized more centralized approaches to
development in the past. Decentralization has also enabled presidents to
incorporate new actors into the political system and to defuse regional
conflicts (ibid.). According to Manor, in response to the `widespread but
vague notion that centralized structures had fallen well short of adequacy
...national politicians use decentralization to reconnect their regime with
social groups and to sustain or revive their party organizations' (Manor,
1999: 33).
4
Without specifying precise causal mechanisms, others similarly
attribute decentralization to deep crises of legitimacy and political instability
experienced in developing countries from the 1980s onwards (Burki et al.,
1999; Rondinelli et al., 1984). What all these arguments have in common is
the notion that national politicians may cede some of their powers in the
short term in order to bolster their overall positions in the long term. Put
most explicitly by Manor, decentralization can increase political stability
and economic growth in a way that compensates politicians for any loss of
power they may experience in the short run (Manor, 1999: 49).
Decentralization is not always the result of proactive, long-term strategies
by national politicians, however. In many cases, national politicians endorse
decentralization not so much as a matter of choice, but rather because they
are forced to do so by pressures from below. According to Willis et al.
(1999), when sub-national actors like governors control the career options
of the legislators who represent their districts in national assemblies, they
can pressure legislators to pass decentralizing policies. Brazil is a case in
point: the transition back to democracy in the 1980s served to re-establish
governors as powerful political actors, and they in turn demanded the de-
centralization of fiscal revenues (Samuels and Abrucio, forthcoming; Souza,
1996).
Furthermore, even where decentralization is indeed initiated by central
state actors, this support may reflect short-term, highly strategic political
4. Manor identifies a number of other ways in which decentralization responds to systemic
problems. For example, he argues that decentralization is also attractive to national
politicians because it can provide them with better information about local actors in
contexts where information is costly.
104 Kent Eaton
calculations rather than attempts by politicians to shore up the long-term
resilience of their political systems. The argument that decentralization
solves long-term problems for national politicians often fails to address
adequately the reality that politicians in developing countries frequently face
strong incentives to discount future gains (Ames, 1987). Why exchange
certain losses in the short run for possible gains down the line that could
accrue to future office holders? As the cases discussed in this article suggest,
national politicians may have a variety of reasons to endorse decentraliza-
tion that have little to do with its promise as a substantive solution to long-
term problems. For example, when control of the executive and legislative
branches of central government is divided, the party that controls the
legislature may support decentralization as a means of restricting the power
of the chief executive. Divided government is not a systemic feature of the
political system but a condition that depends on electoral results and that
can appear and disappear in a matter of a few years. As often occurs with
the passage of other economic policies, politicians may agree to support
decentralization not because of its substantive merit, but in exchange for
short-term benefits like spending projects for their home districts. In this
sense, politicians may treat decentralization as just another policy arena,
subject to the usual exchanges and log-rolls, despite the fact that the stakes
associated with decentralization appear to be qualitatively different and
higher than most other policy areas.
If short-term political calculations can result in the decision to de-
centralize, it is less than surprising when these decisions do not stick or are
quickly subject to reversal. Politicians may support re-centralization not
because they come to believe that they were mistaken in their earlier support
for decentralization as a response to systemic problems, but rather because
electoral and political realities have shifted. For example, decentralization
may suddenly look less attractive to a party that gains unified control over
both the executive and legislative branches. Even where national politicians
genuinely supported decentralization for the broad benefits it promises to
deliver, they may continue to face a variety of incentives to meddle in the
activities that have been officially devolved. Another important source of
pressure for back-tracking in the decentralization process is the resistance of
bureaucrats in central government ministries. According to the decentral-
ization literature, chief executives who advocate decentralization have had
to struggle with bureaucrats over its implementation in scores of countries,
including Bangladesh, Kenya, Sudan, Sri Lanka, Morocco, and Tanzania
(Crook and Manor, 1998; Rondinelli et al., 1984: 47).
The following two sections turn to the Argentine and Philippine cases for
insights into the politics of decentralization and re-centralization.
5
Argentina's
experience with decentralization is significant because it represents one of
5. For other comparisons of the Philippines with Latin America, see Fox (1990).
Political Obstacles to Decentralization 105
Latin America's most decentralized systems.
6
The Philippine experience stands
out as one of the most ambitious decentralization attempts in Asia.
7
In both
cases, decentralization triggered important attempts by national politicians
to renege and re-centralize. The comparison of these cases is not meant to
test hypotheses, only to uncover the logic of national politicians' reactions
against decentralizing policies. In both cases, I examine instances of devolu-
tion rather than deconcentration, exploring the question of why powers
were devolved and then partially reclaimed.
ARGENTINA: THE IMPACT OF DIVIDED VERSUS UNIFIED
GOVERNMENT ON DECENTRALIZATION
When Argentines experienced their most recent transition to democracy
in late 1983, the economic and political systems they inherited were quite
centralized. Economic centralization was largely the result of the statist
development models with which Argentina began experimenting in the
1930s and enthusiastically endorsed in the 1940s. Although statism signific-
antly increased the number of policy tools available to government officials
(Cavarozzi, 1994), the expansion of state power accrued almost exclusively
to the central (federal) government, with sub-national governments left to
play diminished roles. From the 1930s, for example, the Argentine provinces
delegated the authority to collect most of their own taxes to the federal
government, an act of centralization called coparticipacio
Ân(henceforth co-
participation) that significantly expanded the revenue powers of officials at
the centre.
Political power has also been concentrated, especially under the six
military-led governments that ruled the country between 1930 and 1983. But
even when presidential, congressional and gubernatorial elections were held
during this fifty-year period, political power remained centralized at the
federal level, largely because of the hierarchical and disciplined nature of
Argentine political parties (McGuire, 1997). Party discipline in Argentina
is reinforced by the control that national and provincial party leaders exert
over the selection of candidates for legislative office (Jones, 1997). If legis-
lators seek to pursue political careers within their parties, it behoves them to
toe the party line on decentralization and other economic policies. Given the
highly centralized economic system discussed above, when the president of
the nation is also president of his party, his ability to command loyalty from
governors and legislators is even more pronounced (Jones, 1997; de Riz and
6. As of 1995, Argentina was Latin America's most decentralized country in terms of sub-
national spending as a percentage of total government spending; see Inter-American
Development Bank (1997: 157).
7. Interview with David C. Nellor, Resident Representative to the Philippines for the
International Monetary Fund, 10 April 1997.
106 Kent Eaton
Smulovitz, 1991). According to Dillinger and Webb (1999), the president's
influence over the behaviour of his party members, including both governors
and legislators, goes a long way in explaining why decentralization did not
undermine macroeconomic stability in Argentina in the 1990s as it did in
neighbouring Brazil, where weak parties deny the president similar influence
over governors and legislators. Others have demonstrated that party discip-
line in Argentina helps the president rein in spending by governors from his
own party (Jones et al., 2000). As the following discussion of decentraliza-
tion reveals, however, party discipline is far from perfect Ð the president
often had to cut special deals with governors in order to get them to
acquiesce to revenue re-centralization.
Revenue Decentralization under Divided Government:
The 1987 Co-participation Law
In the decades that followed the 1934 establishment of an automatic revenue-
sharing system in Argentina, national politicians repeatedly modified the
rules governing the system (Nun
Äez Min
Äana and Porto, 1982; Porto, 1990).
When the 1973 revenue-sharing law lapsed shortly after Argentina's trans-
ition to democracy in 1983, the size and timing of revenue transfers to
individual provinces became the prerogative of the country's new President,
Rau l Alfonsõ
Ân (Carciofi, 1990; FIEL, 1993). As Table 1 shows, although
Alfonsõ
Ân's Radical party held a majority in the lower chamber between 1983
and 1987, the Peronists dominated in the Senate and held a majority of
provincial governorships. Despite the fact that revenue-sharing laws had
been in effect continuously since the beginning of the co-participation
system in 1934, the two parties could not agree on a new revenue-sharing
law and neither had the political power to impose its preferences on the
other. In this four-year period, centralized and discretionary control over
revenue sharing was an important tool for the President, whose attempts to
respond to the country's foreign debt crisis and reverse its economic decline
were aggressively resisted by the Peronists. Because Argentine governors in
their capacities as provincial party leaders can exert significant influence
over legislators from their parties and provinces, Alfonsõ
Ân was able to use his
discretion over revenue transfers to build Peronist support for his policies in
Congress. In 1985, for example, Alfonsõ
Ân used revenue transfers to reward
Peronist governors for encouraging Peronist legislators to approve a series
of important economic measures (Põ
Ârez, 1986: 68).
Against this backdrop, when the Peronists won mid-term elections for the
lower chamber in 1987, they quickly used their increased leverage to legislate
a return to automatic revenue-sharing procedures. By increasing the fiscal
independence of Peronist governors from the Radical President, automatic
transfers made it harder for Alfonsõ
Ân to influence the behaviour of Peronist
legislators in Congress. In addition to substantially limiting the President's
Political Obstacles to Decentralization 107
discretion, legislators also increased the provinces' share in co-participated
taxes to an historic high of 56.6 per cent. Although the President could have
vetoed the law, the political difficulties he faced after the collapse of his
various stabilization programmes made him powerless to resist the shift
back to automatic revenue sharing. Specifically, Alfonsõ
Ân signed off on the
revenue-sharing bill in exchange for Peronist support for tax increases that
he desperately needed to shore up the country's fiscal accounts. The shift to
automatic revenue sharing in Argentina in 1987 is thus a useful example of
how national politicians from one party may use decentralization not so
much to solve systemic problems in the economic or political system, but
rather to check the power of politicians from competing parties.
It is important to note that this 1987 law did not shift taxing powers back
to the provinces; these remained centralized in the federal government.
However, since transfers in the system are automatic, not earmarked for any
specific expenditures, and can be used by elected provincial officials for
whatever purposes they decide, automatic revenue sharing enhances sub-
national autonomy and makes fiscal policy more decentralized.
8
Revenue Re-centralization under Unified Government:
The 1992 and 1993 Fiscal Pacts
While divided government helps explain the adoption of decentralizing
policies in 1987, unified control by the Peronist party of both the executive
and legislative branches after 1989 led to the partial re-centralization of
revenues. In the aftermath of his victory in the May 1989 presidential
elections, Carlos Menem correctly identified automatic revenue sharing with
Table 1. Peronist Legislative Seats and Gubernatorial Offices in Argentina:
1983±1991
Period Distribution of
seats by party
in the lower
chamber
Period Distribution of
seats by party
in the upper
chamber
Period Distribution
of provincial
governorships
by party
1983±1985 PJ 43.7%
UCR 50.8%
1983±1986 PJ 45.6%
UCR 39.1%
1983±1987 PJ 12
UCR 7
1985±1987 PJ 40.6%
UCR 51.2%
1986±1989 PJ 45.6%
UCR 39.1%
1987±1991 PJ 17
UCR 2
1987±1989 PJ 42.9%
UCR 46.1%
Source: Jones (1997: 265, 266, 281).
8. There is an important debate in the decentralization literature over whether revenue
transfers should be taken as a form of devolution. For contrasting views, see Manor (1999:
13) and Rondinelli et al. (1984: 44).
108 Kent Eaton
the provinces as a threat to his political ambitions. Despite the constitu-
tional limit to one presidential term, Menem wanted to be re-elected and
calculated that building support for a constitutional amendment and win-
ning re-election depended on achieving fiscal stability through tax reform
(Acun
Äa, 1995). Given this ambition, revenue decentralization presented
Menem with three particular problems. First, automatic revenue sharing
would prevent him from strategically exchanging revenue transfers for
governors' support for his economic and constitutional reforms; second,
since provincial governments had used these transfers to expand the public
sector workforce and strengthen their own patronage networks (Novaro,
1994; World Bank, 1990), the President feared that automatic revenue
transfers would undermine his adjustment programme; third, the automatic
distribution of tax revenues meant that Menem would provide the leader-
ship on tax reform only to have nearly 57 per cent of the proceeds auto-
matically transferred to governors who had borne none of the political
burden associated with increasing tax revenues.
Between 1989 and 1991, Menem and his party in Congress laid the founda-
tion for a new tax system based on a broad, value-added tax. As a result of
these changes, tax revenues increased 152 per cent between 1991 and 1996
(Cetra
Ângulo and Jime
Ânez, 1996: 15; Dura
Ân and Go
Âmez Sabaini, 1993). Once
these changes in tax policy began to bear fruit, Menem negotiated two
different fiscal pacts between the federal government and the provinces that
altered the 1987 revenue-sharing law by reducing the provinces' legal shares
of centrally-collected tax revenues. According to some estimates, the prov-
inces received some US$ 13 billion less between 1992 and 1996 as a result
of the two fiscal pacts (Fundacio
Ân CECE, 1998: 43). Although the 1987
revenue-sharing law had pegged the federal government's share of revenues
at 42.34 per cent, the fiscal pacts of 1992 and 1993 effectively increased its
share of revenues to 54.07 per cent (Porto and Sanguinetti, 1993). These
pacts deserve closer attention because they were the main mechanisms for
the partial reversal of revenue decentralization that took place in the 1990s.
According to the first fiscal pact, which Menem negotiated directly with
governors, 15 per cent of the revenues legally subject to automatic sharing
with the provinces would be deducted and channelled to the national social
security system (Schwartz and Liuksila, 1997). This pact was designed to use
provincial revenues to solve a federal budget problem in the form of large
social security deficits. With legions of angry pensioners poised to physically
take over Congress, Peronist party discipline facilitated the agreement of
this pact, despite warnings by opposition legislators that the executive
branch would not devote all of the deducted revenues to social security.
9
9. Diario de Sesiones de la Ca
Âmara de Diputados de la Nacio
Ân, 19/20 August 1992, pp. 2105,
2106; and Diario de Sesiones de la Ca
Âmara de Senadores de la Nacio
Ân, 2 September 1992,
pp. 2293±325.
Political Obstacles to Decentralization 109
While the pact withheld revenues from the provinces, it also guaranteed
provincial transfers of a minimum of approximately US$ 725 million per
month. However, while the federal government could, at the federal execut-
ive's discretion, send additional revenues beyond this amount, it was not
obliged to do so. In addition, counter to the spirit of decentralization, the
pact allowed for greater federal control over provincial budgets by requiring
the provinces to limit any increases in current expenditures to 10 per cent
and to devote any budget surpluses to reducing debt or to capital expend-
itures.
10
Why did governors accept an agreement that would substantially reduce
the revenues to which they were legally entitled? A number of different
factors help account for this surprising support. As argued by a top official
in the federal Ministry of Economics, governors agreed to the pact only
because tax revenues had increased so dramatically as a result of tax reform
and fiscal stability. In times of crisis, or even normal times, the federal
executive would most likely have been unable to convince the provinces to
forego their legal revenue shares.
11
But there is also a more political reason
for the governors' support of the pact, in which short-term political cal-
culations are central. Although the governors in theory could have joined
ranks and refused to negotiate, it is clear that Menem strategically took
advantage of the incentives facing individual governors to cut special deals
with him for their provinces. Thus, as part of the pact, Menem created
a special fund to finance fiscal disequilibria in the provinces, made up of
revenues that would also be deducted from the automatic revenue-sharing
system. The amount that each province would receive from this fund was
determined in one-to-one negotiations with the President. For example, the
fact that the Radical governor of Chubut province offered early and public
support for the pact reflected Chubut's position as one of three provinces
receiving the largest cut from the disequilibria fund (US$ 3 million per
month).
12
Provincial governors who initially criticized the pact but
eventually signed it (such as Mario Moine of Entre Rõ
Âos) received a smaller
cut, and provincial governors who refused to sign (such as Rolando
Tauguinas of Chaco) received none of the special funds.
13
Thus Menem
effectively replaced the distribution criteria legislated in the 1987 revenue-
sharing law with what amounted to little more than political deal-making.
In August 1993, exactly one year after the first fiscal pact, Menem signed
a second fiscal pact with governors from sixteen of the twenty-three prov-
10. Diario de Sesiones de la Ca
Âmara de Diputados de la Nacio
Ân, 19/20 August 1992, p. 2113.
11. Interview with Carlos Ferna
Ândez, director of Fiscal Relations with the Provinces in the
Federal Ministry of Economics, 7 May 1996.
12. Diario de Sesiones de la Ca
Âmara de Diputados de la Nacio
Ân, 19/20 August, 1992, pp. 2105,
2106.
13. Diario de Sesiones de la Ca
Âmara de Diputados de la Nacio
Ân, 19/20 August, 1992, pp. 2105,
2106; and Pa
Âgina 12, 9 August 1992.
110 Kent Eaton
inces. At the heart of the second fiscal pact was an attempt by the President
to bring about reforms in provincial tax systems that mirrored the stream-
lining reforms Peronists had implemented at the national level (Repu blica
Argentina, 1994, 1995: 621±6). On the one hand, this pact could potentially
serve the purpose of decentralization by improving the revenue-generating
potential of provincial tax systems.
14
On the other hand, the pact made
provinces more dependent in the short run on transfers from the centre,
increased the scope for discretion in the distribution of tax revenues, and
represented policy-making for the provinces by the federal government.
In order to build support for the second pact, the President played to the
short-term political interests of governors.
15
Most importantly, he altered
and extended various aspects of the minimum revenue guarantee established
in the first fiscal pact. Specifically, the federal government would no longer
withhold revenues in excess of the minimum guaranteed transfers for those
provinces that signed the pact. It also increased the minimum revenue guar-
antee established in the first fiscal pact from US$ 725 million to US$ 745 million
monthly, but only for those provinces that agreed to implement the terms of
the pact.
16
Thus the pact continued and accelerated the process of erosion in
the provinces' fiscal policy-making autonomy.
As with the first pact, the President encouraged provinces to sign up by
using a variety of different tools. First, his Economy Minister increased the
pressure on provincial governments to sign by pledging to lift the tax on
minimum assets for companies located in provinces that had signed the
pact.
17
Second, the President pledged help in a variety of forms for prov-
inces that were inclined not to sign.
18
In Chaco, for example, Governor
Ronaldo Tauguinas from the provincial party Accio
Ân Chaquena signed the
pact when the national government agreed to resolve long-standing debts
that the national and provincial governments owed to each other with a
balance favourable to Chaco (Confederacio
Ân General Econo
Âmica, 1993:
173). Subsequently, Tauguinas came out in favour of constitutional reform
and Menem's re-election as President. In neighbouring Formosa, though
Peronist Governor Vicente Joga did not belong to Menem's wing of the
party, he agreed to sign the fiscal pact in exchange for a series of benefits to
the province, including a free-trade zone and public works projects, and in
14. Interview with Hugo Garnero, Economy Minister of the Province of Santa Fe
Âand former
Subsecretary of Fiscal and Economic Relations with the Provinces in the National
Ministry of Economics, 22 July 1996.
15. Interview with Jose
ÂOctavio Bordo
Ân, former Governor of Mendoza and presidential
candidate of FREPASO, 7 August 1996.
16. Anales de Legislacio
Ân Argentina, vol. LIII-D, p. 4240.
17. ClarõÂn, 18 August 1993.
18. The governors were especially susceptible to the President's offer to help them with their
debts. Partly as a result of the first fiscal pact, which cut revenue transfers, the provinces
used their guaranteed future revenue transfers as loan guarantees and became heavily
indebted in domestic financial markets.
Political Obstacles to Decentralization 111
exchange for a personal reward for Joga whom Menem named Intervenor of
the Peronist party in neighbouring Corrientes province (ibid.: 207).
One of the outcomes of these two fiscal pacts in Argentina was increased
complexity of the revenue distribution system (Saiegh and Tommasi, 1998,
1999). As the literature on decentralization has noted, complex and elab-
orate administrative rules are the enemy of decentralization because they
reinforce the discretionary power of bureaucrats and politicians at the centre
(Manor, 1999: 59; Rondinelli et al., 1984: 58). According to complaints lodged
by governors subsequent to the signing of the two fiscal pacts, complex rules
have made it possible for the federal government to transfer to the provinces
less than the guaranteed minimum level to which it agreed.
19
Further, although
the 1994 constitutional reform required legislators to negotiate a brand new
revenue-sharing law by the end of 1996, the Peronists instead voted in 1996
and 1998 to simply extend the current, highly discretionary system.
20
Thus the
evidence clearly suggests that control over revenues became more central-
ized between 1991 and 1999, despite the important revenue decentralization
that occurred in 1987.
Expenditure Decentralization and its Aftermath
The expenditure side of decentralization in Argentina has proven to be nearly
as volatile as the revenue side. Rather than actually re-centralize expenditure
responsibilities, national politicians continued to meddle in these services
after their decentralization. A brief discussion of education policy shows
that the initial agreement to decentralize authority over schools has not been
implemented as initially designed.
In 1991 and 1992, the federal government transferred responsibility for
secondary schools and health care to the provinces (Murillo, 1999; World
Bank, 1990, 1993). Rather than reflecting a desire for decentralization as a
public good, President Menem's support for the devolution of responsibility
for services reflects short-term political interests. Menem proposed the
transfer of secondary schools and hospitals not as a means of reforming and
improving education or health care, but rather to solve the fiscal problems
of the federal government. This is reflected in the fact that the transfer of
schools was initially proposed as part of the 1992 budget, with no previous
substantive consultation about how decentralization should take place.
21
By
transferring services but not additional revenues, Menem designed decentral-
ization as a way of reducing federal budget expenditures and forcing the
19. Interview with Rau l Baglini, President of the Budget and Finance Committee, Chamber of
Deputies from 1985±1989, 11 August 1989; and ClarõÂn, 3 April and 13 April 1996.
20. Ambito Financiero, 7 August 1996, p. 5.
21. Diario de Sesiones de la Ca
Âmara de Diputados de la Nacio
Ân, 5/6 December 1991, p. 5343.
112 Kent Eaton
provinces to spend the additional revenues generated by tax reform on what
was, until then, a federal budget responsibility. According to Menem's
proposal, the federal government would immediately begin withholding
from the revenue-sharing system an amount equal to the cost of running
secondary schools.
22
It would only release these revenues to the provinces
once the specific terms of the transfer had been worked out with each
province on a case by case basis. This sequence maximized the President's
bargaining position in the negotiations.
In the years subsequent to the transfer of secondary schools, President
Menem continued to meddle in this area. For example, in December 1995,
he issued Decree 892/95 that would tighten federal control over how prov-
inces spend revenue transfers on social expenditures, including education,
health and social welfare. According to the decree, the federal executive
could immediately suspend transfers and sanction responsible provincial
officials who use the funds for current expenditures. Thus, Menem regained
supervisory control over approximately US$ 2.2 billion of expenditures that
had officially been transferred to the provinces.
23
In 1997, Menem intro-
duced a bill that would expand federal control over provincial school
systems by requiring them to implement federally designed administrative
and labour reforms.
In 1998, in response to widespread demonstrations in the capital city of
Buenos Aires by teachers unhappy with the very unequal salaries paid by the
different provincial school systems, Menem proposed a 1 per cent tax on
cars, boats and yachts to increase teachers' salaries. Peronist legislators
passed this tax in 1999. According to Peronist Deputy Ana Marõ
Âa Mosso,
the tax was unconstitutional since it duplicated the pre-existing provincial
tax on cars and served to crowd out an important provincial tax base.
24
This
increase was designed to finance two extra annual payments to teachers,
but only in provinces that agreed to implement the terms of the Federal
Education Law to the satisfaction of the executive branch.
25
Although his
own Finance Minister argued against this tax increase, the President saw
an opportunity to claim credit for increasing the salaries of workers who
provide services now legally under the control of provincial officials. Thus,
despite the sudden but clear shift toward decentralized expenditures in 1991
and 1992, national politicians continued to try to assert their authority in
policy areas that were formally the prerogative of sub-national officials.
22. See the discussion of Article 14 of Law 24,049 in Diario de Sesiones de la Ca
Âmara de
Senadores de la Nacio
Ân, 28 November 1991, pp. 4534±51.
23. ClarõÂn, 15 December 1995: 26.
24. Interview with Ana Marõ
Âa Mosso, Peronist Deputy (1997±2000) and Economy Minister of
the province of Mendoza (1996±7), 12 August 1999.
25. ClarõÂn, 25 April 1998.
Political Obstacles to Decentralization 113
THE PHILIPPINES: THE AMBIVALENT RESPONSE OF LEGISLATORS
TO DECENTRALIZATION
Revenue and expenditure decentralization has also been a dynamic area
in the Philippines in the last decade. In 1991, the passage of the landmark
Local Government Code ushered in a new era by devolving authority over
revenues and expenditures to villages, municipalities, cities, and provinces.
As in Argentina, in the aftermath of decentralization national politicians
have jealously guarded their prerogatives and supported re-centralizing measures.
After putting the 1991 law into historical perspective and analysing the logic
of national politicians' support for fiscal decentralization, this section then
evaluates how legislators responded to decentralization in the implementa-
tion phase.
In the four decades following independence in 1946, national politicians
in the Philippines considered but declined to adopt a variety of proposals
that would have devolved authority over revenues and expenditures to local
governments. In their place, legislators passed bills that merely deconcen-
trated responsibilities within central government agencies, including the
Local Autonomy Act of 1950 and the Decentralization Act of 1967 (Manasan,
1992; Tapales, 1991, 1993). Throughout this period, devolution consistently
proved less attractive to national politicians than deconcentration, which
simply redistributed power within the central government bureaucracies over
which Philippine legislators and presidents exert much influence (Abueva,
1965). In the words of one politician, the Philippines remained a country
where local governments had to appeal to national politicians to get
windows fixed in City Hall.
26
In order to understand this historic opposition to decentralization, a few
aspects of the Philippine political system are relevant. Most importantly,
political parties in the Philippines are profoundly weak and poorly discip-
lined (Coronel, 1998; Montinola, 1999). According to Carey and Shugart's
ranking of electoral incentives, representatives to the lower House of
Representatives in the Philippines face some of the most particularistic
incentives of any democracy (Carey and Shugart, 1995). Since candidates
and incumbents in House races can run under any party label they choose to
adopt, national and sub-national party leaders control neither access to the
ballot nor the subsequent behaviour of legislators in Congress. In the
absence of such control, electoral success in the Philippines depends on
establishing personal reputations and delivering particularistic benefits to
local constituents. Given these electoral incentives, centralized control over
revenues and expenditures enables legislators to negotiate fiscal transfers
and `pork barrel projects' with central bureaucrats and to then claim
26. Interview with Margarito Teves, Member of the House of Representatives, 5 March 1997,
Quezon City.
114 Kent Eaton
personal credit for making these possible (Stauffer, 1975; Wurfel, 1988).
27
Considering these incentives, it is little wonder that Philippine legislators
long favoured centralized control.
Presidents also had cause to resist genuine devolution. Because of weak
parties and individualistic behaviour by legislators, presidents of the Philip-
pines have had a very difficult time getting their policy agendas through the
legislature (Doronila, 1992; Rocamora, 1996). Traditionally, presidents have
depended on their formal authority over the release of pork barrel funds as a
power resource that allows them to purchase the support of legislators for
substantive policy change (Gutierrez, 1994; Hutchcroft, 1998). As a govern-
ing tool, the effectiveness of the president's control over the release of
these funds would be seriously undercut by the broad devolution of revenues
and expenditures. Local governments would become much less dependent
on pork barrel funds, decreasing their political leverage and utility to the
president.
Given this political reality, support by President Aquino for decentraliza-
tion and its passage by Congress in 1991 is politically intriguing. After all,
the Local Government Code promised to curtail the policy-making authority
of both presidents and legislators in important ways (Brillantes, 1994; Pan-
ganiban, 1995). First, the Code replaced the previous system of negotiated
transfers to sub-national governments with automatic revenue-sharing
procedures. Because revenue transfers no longer depend on sub-national
officials lobbying national politicians, this change has enhanced local
autonomy. Second, the Code stipulated that 40 per cent of the tax revenues
collected by the central government would be automatically shared with
local governments. In the previous system of discretionary transfers, the
share of local governments in tax revenues averaged only 12 per cent be-
tween 1986 and 1991 (Manasan, 1992: 24). Table 2 documents the growth in
revenue transfers to local governments from 1990 to 1994. Moreover, the
Code allows local governments considerable leeway in how they use their
additional revenues, not explicitly tying them to particular expenditures.
Third, in addition to giving sub-national governments a greater and steadier
share in national tax revenues, the Code also decentralized expenditure
responsibilities in a number of areas. Specifically, it transferred expenditures
27. The term `pork barrel' refers to local infrastructure projects in legislators' home districts
for which they are able to secure financing from the central government. In the period
reviewed here, the two major budgetary vehicles for pork barrel projects in the Philippines
were Countrywide Development Funds (CDF) and Congressional Insertion Allocations
(CIA). CDF monies are distributed evenly among all legislators, who receive the right to
identify local projects eligible for the funds up to a certain amount Ð approximately
US$ 750,000 for Senators and US$520,000 for House members in calendar year 1996. CIA
funds are not distributed evenly among legislators but instead are controlled by members
of the Appropriations Committees of each chamber who, in budget deliberations, support
higher budgets for bureaucrats in exchange for a say in how these extra funds are spent in
their districts. For a classic study of pork barrel politics, see Ferejohn (1974).
Political Obstacles to Decentralization 115
formerly provided by the national Departments of Agriculture, Education,
Culture and Sports, Environment and Natural Resources, Health, Public
Works and Highways, Transportation and Communication, and Social
Welfare (Pimentel, 1993). Given the historic scope of these changes, the
Code can be considered the most important policy departure of Aquino's
entire administration.
Why should national politicians support a reform that transferred such
authority to sub-national officials? The key to explaining the political logic
of the 1991 decentralization is the role played by different electoral incent-
ives, beginning with the lack of electoral incentives facing President Aquino.
A non-traditional politician and political widow who was uninterested in
remaining in power beyond her six-year term, Aquino was committed to
the `no re-election' clause of the 1986 constitution, which was written by
individuals she appointed and which included several measures that were
designed to avoid repetitions of Marcos' successful attempts to perpetuate
himself in power. Aquino was therefore unlike most presidents in that she
was not personally threatened by the loss of power to the national govern-
ment that the Code would effect. Furthermore, she considered decentral-
ization to be the linchpin of her administration and a reform that would
facilitate the transition to democracy, the main legacy of her government.
Aquino's efforts on behalf of decentralization contrast sharply with the
centralizing behaviour of the traditional politicians who preceded her in
office, and explain why many consider the passage of the Local Government
Code to be something of a historical accident.
28
The response of legislators to decentralization is more complicated, in
part because of the imposition of term limits on legislators in the 1986
Constitution. As a consequence of these term limits, nearly ninety legislators
were barred from running for re-election in the 1998 congressional races.
On balance, the term limits increased support for decentralization, giving
Table 2. The Impact of Automatic Revenue Sharing on Transfers to Local
Governments in the Philippines: 1990±94 (in constant 1994 pesos)
Type of Local
Government
1990 1991 1992 1993 1994
Provinces 2,031,174,331 2,697,481,707 4,571,136,402 8,445,600,000 11,498,994,198
Cities 2,191,470,949 2,742,969,221 4,559,895,793 8,445,600,000 10,239,577,496
Municipalities 3,054,601,475 4,046,837,742 7,127,522,550 12,484,800,000 10,325,288,074
Villages n/a n/a 3,942,000,000 7,344,000,000 8,689,140,232
TOTAL 7,227,246,755 9,487,288,670 20,200,554,745 36,720,000,000 46,753,000,000
Source:Department of Interior and Local Government Digest, 19 September 1994: 4, 5.
28. Interview with Ben Diokno, Aquino's Budget Undersecretary, 31 January 1997.
116 Kent Eaton
legislators reason to consider running for local office in the future. A few
legislators even abandoned Congress to run for local office before term
limits took effect, including one of the Code's main sponsors, Hilarion
Ramiro, who resigned from his seat to become governor of South Cotobato
province.
29
Understanding legislators' support for decentralization is also complic-
ated by the complex and varied relationships that bind legislators to the sub-
national politicians who would be empowered by decentralization (Franco,
1997). In many cases, legislators have cause to resist giving local executives
automatic shares of revenue and expanded expenditure authority because
these individuals are likely to be their political rivals. Local executives with
autonomous fiscal resources would be much more threatening as opponents
should legislators choose to contest local offices in the future. At the same
time, local executives could use their enhanced powers to contest incumbent
legislators for national office.
30
In this sense, decentralization confronted
some legislators with a dilemma: while they wanted strong sub-national
governments in the event that they were to hold these offices in the future, in
the interim they wanted to protect their roles as valued brokers of fiscal
benefits from the centre.
Many national legislators, however, were saved from this dilemma because
they maintained solid and strategic alliances with sub-national politicians.
31
For
example, when the spouses, children, parents, relatives, and political puppets
of national legislators occupy local offices, as they often do in the Philip-
pines, the devolution of revenue and expenditure authority to local officials
does not substantially threaten the political interests of legislators (Gutierrez,
1994; Lacaba, 1995). Although decentralization would negatively affect the
power of these legislators at the national level, it would not prevent them
from influencing how the local officials under their control use their de-
volved powers. The prevalence of districts in which traditional family clans
and political bosses dominate all important elected offices has led many
observers to note the dangers inherent in policies that shift revenues and
expenditure responsibilities to what are in reality authoritarian enclaves at
the local level (Blair, 2000; Rocamora, 1995; Sidel, 1995).
32
Reflecting the ambivalence of many legislators toward decentralization,
the Local Government Code sat in Congress for more than four years, from
29. Interview with Rosario Manasan, Philippine Institute for Development Studies, 29 January
1997.
30. For a discussion of the often fraught relationship between national and sub-national
politicians, and its impact on democratization, see Franco (1997) and McCoy (1994).
31. I am very grateful to an anonymous referee for drawing this distinction to my attention
and for helping me evaluate its consequences for legislators' attitudes toward decentral-
ization.
32. For a discussion of decentralization and authoritarian enclaves in Latin America, see Fox
(1994).
Political Obstacles to Decentralization 117
1987 to 1991. Only when the campaign for the 1992 presidential election
began to get underway did House Speaker Ramon Mitra become a powerful
advocate of decentralization in the hopes of securing the all-important
`anointment' as Aquino's chosen successor.
33
Legislators finally approved
the Code in 1991, with an eye to gaining valuable electoral support in the
1992 congressional elections from both the House Speaker and the
President. Legislators were careful, however, to design the Code in ways
that would protect their interests.
Designing the Local Government Code
In agreeing to decentralize revenues and expenditures, members of the
House of Representatives tried to protect themselves against their greatest
fear about decentralization, namely that local executives might use their
enhanced powers to build independent support networks that challenged the
interests of legislators. First, legislators in the House inserted into the Code
a requirement for the Audit Commission to review and audit the fiscal
accounts of each sub-national government.
34
Based on these annual reviews,
legislators could then withdraw powers and revenues from a local executive
in the event of abuses. This provision would allow legislators to monitor
troublesome local officials in their districts. Second, the House strengthened
mechanisms to recall local officials, including the auditors appointed by
local executives. Thus national legislators could fund recall drives targeted
against local executives whom they found threatening.
35
Third, in order to
protect the employment of bureaucrats who owed their jobs to national
legislators, the House insisted that sub-national governments absorb and
pay the salaries of all government employees devolved from the central
government, including those who would now be paid more than local
executives themselves.
36
Finally, though the Senate wanted the Code to take
effect immediately, the House insisted that the implementation date be
moved back to January 1992 in order to reduce the ability of local officials
to use automatic revenue transfers in campaigns against national legislators
in the 1992 elections.
37
The relative burdens placed by decentralization on different types of
sub-national governments also reflect the particular electoral concerns of
33. Interview with Romulo Neri, Executive Director, Congressional Budget and Planning
Office, 30 January 1997.
34. Journal of the House of Representatives, 26 November 1990, p. 33.
35. Journal of the House of Representatives, 13 December 1990, p. 426.
36. See the complaints of Governor Roberto Pagdanganan, head of the League of Governors
(Pagdanganan, 1992, 1996).
37. Journal of the House of Representatives, 11 September 1991, p. 8; and Business World,
29 January 1992, p. 1.
118 Kent Eaton
legislators. While provinces and municipalities combined receive 57 per cent
of the revenue transfers and shoulder 92.5 per cent of the cost of devolved
functions, highly urbanized cities and villages bear only 7.5 per cent of the
costs but receive 47 per cent of the transfers (Emboltura, 1994a). This
skewed distribution reflects the reality that governors are a more common
threat to legislators than are mayors. Legislators represent districts that are
either coterminous with the boundaries of a single province, or are one of
many districts within a single province. In either case, legislators have cause
to worry about fiscally-independent governors. While mayors pose similar
threats to some representatives, there are fewer big city mayors than there
are governors and most legislators represent districts without highly
urbanized cities in them. Thus, by making provinces responsible for more
services than they could pay for with automatic revenue transfers, Congress
ensured that governors would remain dependent on legislators who could
subsequently offer their services as brokers of additional revenues from the
centre in a personalized fashion.
Attempting to Re-centralize Devolved Services
Unsurprisingly, given the scope and depth of the changes introduced by
decentralization, several glitches and setbacks marked the first few years of
the Code's implementation. Faced with some of the confusion and delays
associated with the initial implementation of the Code, the House Committee
on Local Government immediately began hearings on the decentralization
process.
38
In these hearings, many legislators argued that administrative
functions should be retained by national bureaucratic agencies, with only
operative responsibilities devolved to local governments. Within a year of the
introduction of the Code, over thirty bills had been introduced to amend it
(Hofilen
Äa, 1993). In 1992, at the behest of legislators, the Budget Depart-
ment deducted approximately US$ 1.5 million from the local governments'
legal share of tax revenues to finance a special fund to ease the transition to
decentralization. National government bureaucrats, with the help of legis-
lators, were to decide which localities would benefit from this fund and how
(Macabuhay, 1992).
That decentralization was particularly threatening to members of the
House of Representatives is seen in their responses to some of the Code's
early implementation problems. In December 1992, the House unanimously
approved a measure that would permanently exclude the Department of
Health from the national agencies whose functions were to be devolved to
local governments (Emboltura, 1994b). After a national survey revealed the
38. See `Government decentralization can't be implemented', Manila Chronicle, 21 November
1992, p. 15.
Political Obstacles to Decentralization 119
negative impact that decentralization was having on the quality of health
services, the Senate passed a measure in August 1994 that would temporarily
suspend the transfer of health services to local governments for a period of
three years.
39
The two chambers eventually agreed to a compromise bill that
appropriated additional revenues for the Department of Health to ease the
transition period. Through their close relationships with bureaucrats in the
Health Department, legislators would retain influence over the exact distri-
bution of these funds. The same compromise bill also established a Health
Devolution Review Panel, to be staffed by individuals who were hostile to
decentralization, including the president of the alliance of health workers.
The panel had the authority to re-centralize hospitals on a case-by-case
basis, opening up various opportunities for legislators to intervene in these
negotiations.
40
In 1995 and 1998 the House continued to discuss a variety of
bills that would reverse decentralization by re-nationalizing many of the
health and agricultural services that the Code had devolved.
41
In keeping
with his defence of decentralization as a broadly popular measure that cir-
cumvented the traditional politicians who opposed his economic reforms in
Congress, President Fidel Ramos consistently voiced his opposition to these
proposed reversals.
The health sector was not the only area in which legislators revealed their
need to remain personally involved in decentralized services. For example,
Representatives encouraged the Department of Agriculture to keep open its
central research office in violation of the Code, rather than turn over
research functions to local governments (De Leon, 1992). While the most
dramatic attempts to reverse decentralization failed over the course of the
1990s, legislators and their bureaucratic agents have clearly instigated
a process of `creeping re-nationalization'.
42
As part of this process, the
national government has retained many services in localities that should,
according to the Code, have been devolved. As one governor said, `there is a
perception among local leaders that Representatives, while giving lip service
to local autonomy, are reluctant to surrender some of their powers to local
governments' (Cruz, 1993). According to another governor, the House of
Representatives in particular intended the Code not to succeed in order `to
maintain the highly centralized government' (Pagdanganan, 1992).
39. Senate Bill No 1173. For evidence that local officials are not using revenue transfers for
devolved services, see Manasan (1996) and `Local Chief Executives Warned on Misuse of
IRA', Department of Interior and Local Government News Digest 3(25): 1 (1±7 April 1996).
40. House Bill No 3331, 25 July 1994.
41. `Committee on health discusses devolution, renationalization', Committee News,12
December 1998.
42. Interview with Felipe Medalla, Dean of the School of Economics, University of the
Philippines, 4 February 1997.
120 Kent Eaton
Circumventing the Code through Augmentation Funds
Faced with the President's strong defence of decentralization, legislators sub-
sequently turned their attention to `augmentation funds' to complement Ð
some would say duplicate Ð the provision of devolved services by local
executives. In wanting to continue to claim credit for expenditures that had
been transferred to local governments, national legislators authorized addi-
tional revenues from the national budget to fund these same services. In fact,
their ability to do this was used by the sponsors of the Local Government
Code to bolster support for the Code's passage in Congress in 1991.
43
Accord-
ing to Representative Ciriaco Alfelor, the Local Government Committee that
he chaired in the House successfully places additional funds in the budget
every year to augment revenue transfers to sub-national governments.
44
In
1995, for example, the Committee placed an additional US$ 6.6 million
in the budget for health services. These augmentation funds are distributed
to specific local governments according to a classificatory scheme that is
decided by Committee members themselves, which effectively allows them
to take personal credit for additional transfers.
45
In addition to augmentation funds, legislators have responded to decentral-
ization in a variety of other ways. Firstly, evidence suggests that legislators
have increased their reliance on pork barrel funds since the passage of the
Code. Against the backdrop of a 250 per cent increase in the amount of
national tax revenues sent to local governments between 1992 and 1997,
legislators have included more and more pork barrel funds in the budget
(Republic of the Philippines, 1997: 179; Saspa, 1997). According to Senator
Ernesto Herrera, Chair of the Finance Committee, legislators began to use
their pork barrel funds in ways that compete with, rather than complement,
the way that local officials spend their shares of national tax revenues.
46
Although politically expedient and not illegal, this practice clearly under-
mines rational public budgeting. Secondly, legislators have repeatedly
attempted to divert tax revenues from the automatic revenue-sharing system
to special infrastructure funds over which they would retain greater control.
47
Finally, the fact that legislators are no longer the brokers of discretionary
fiscal transfers heightens their interest in using the tax system for the
43. As Representative de Pedro argued in sponsoring the bill, `Congress may in effect
authorize the national government to continue providing these same facilities and
services', Journal of the House of Representatives, 11 September 1991, p. 15.
44. Interview with Congressman Ciriaco Alfelor, Chair of the Local Government Committee,
4 March 1997.
45. According to Alfelor, in 1997 the Oversight Committee studied a bill that would pre-
deduct special monies from the revenues shared with all local governments in order to send
additional revenues to fourth, fifth and sixth class (i.e. lesser developed) municipalities.
46. Interview with Senator Ernesto Herrera, 3 March 1997.
47. Record of the House of Representatives, 11 November 1993, Second Regular Session, vol. 3,
p. 143.
Political Obstacles to Decentralization 121
purposes of particularism by claiming credit over other benefits such as tax
exemptions. Those legislators who fear local officials as political rivals now
have greater incentives to resist the broadening of national tax bases, since
legislating tax exemptions serves the dual purpose of rewarding supporters
and reducing the size of revenue transfers to local officials. Together, these
different behaviours suggest that national legislators in the Philippines
remain wary of the potential impact of decentralization on their own career
goals.
CONCLUSIONS
The adoption of decentralizing policies throughout the developing world
in recent decades suggests the operation of powerful, systemic pressures,
including democratization, economic liberalization and the preferences of
external donors. While these cross-border pressures certainly set the stage
for decentralization, however, short-term political calculations determine the
form that decentralization actually takes, and there is considerable variation
in this respect. Exploring the shifting attitudes of national politicians toward
decentralized practices is a necessary step in understanding variation across
time and space in different countries' experiences with decentralization. The
two cases studied here suggest a deep ambivalence of national politicians
toward decentralization, and their abiding fears that the long-term gains
promised by advocates of decentralization may fail to outweigh short-term
losses. These two cases also suggest that, while decentralization has made
the study of sub-national politicians one of the most pressing and exciting
research agendas in comparative politics, national politicians remain critical
for understanding how decentralization unfolds.
In Argentina, the Peronist party used revenue decentralization when it
was in opposition as a tactical measure that would reduce the power of the
Radical-controlled executive branch over Peronist governors and legislators.
When the Peronists captured the presidency, support for revenue decentral-
ization diminished and party discipline paved the way for a return to
revenue-sharing procedures characterized by increased presidential discre-
tion. In the first fiscal pact, Menem encouraged governors to forego their
legal shares in tax revenue by guaranteeing them a minimum level of revenues.
In the second fiscal pact, Menem then made this guarantee conditional on
progress toward a series of federally-defined provincial economic reforms.
The net result has been growing discretion in the fiscal relationship between
the federal and provincial governments. Discretion is anathema to the spirit
of genuine decentralization but highly valuable to a president whose political
interests and capabilities are challenged by fiscally independent governors.
While the Philippines has also witnessed some back-tracking on the
decision to decentralize after 1991, efforts there to reverse and subvert key
aspects of the Local Government Code have been led by legislators rather
122 Kent Eaton
than presidents. For legislators whose relatives occupy elected local offices
in their home districts, decentralization has created few headaches. The
record suggests, however, that many legislators have cause to worry about
how local officials are using their expanded powers. This is seen in the way
that legislators amended the Local Government Code when they endorsed
it, in the various attempts to re-centralize services, and in their strenuous
efforts to augment non-automatic transfers and thus remain indispensable
`back home'. As a result, in neither the Philippines nor Argentina were the
decentralizing policies that were initially endorsed by national politicians
actually respected in implementation by those same politicians.
This raises a number of broader questions for the literature on decentral-
ization. First is the question of whether there are limits to the abilities of
national politicians to back-track on their support for decentralization.
According to Crook and Manor, because decentralization in most countries
is a top-down affair rather than the result of grassroots pressure from below,
the interests that benefit from decentralization are rarely organized enough
to defend it against reversals. They find no evidence of preventive action at
the grassroots in response to the erosion or destruction of decentralized
authorities (Crook and Manor, 1994; Manor, 1999). In the cases reviewed in
this article, however, there do appear to be limits on re-centralization. In the
Philippines, President Ramos cited the widespread popularity of decentral-
ization when he vetoed congressional legislation that would have re-centralized
some devolved services. In Argentina, in order to cut provincial shares of tax
revenues, the President had to offer in return guarantees that he would
transfer a minimum amount of revenues to the provinces. In both countries
there is an intermediate outcome in which the decision to decentralize is
neither fully reversed nor implemented as designed.
Second, this discussion raises questions about how waves of decentraliza-
tion and re-centralization might relate to the larger shift in economic
development models. According to many scholars, the adoption of state-led
industrialization strategies in the mid-twentieth century led to centralized
patterns of control over revenue and expenditure while the shift to more
liberal economic strategies in the later part of the century facilitated more
decentralized approaches (Burki et al., 1999; Manor, 1999). Others suggest
that waves of decentralization and re-centralization cannot be attributed to
support for different development models. One survey of decentralization in
developing countries, for example, reports that relatively short cycles of
centralization and decentralization have occurred in developing countries
ever since the 1950s (Rondinelli et al., 1984: 28). While more research is
needed on this question, the cases evaluated here suggest that political
struggles over the control of revenues and expenditures may have less to do
with substantive debates over development strategy than with short-term
and highly dynamic political calculations.
Third, the Argentine and Philippine experiences can be mined for insights
about the impact of decentralization on the transparency of policy-making
Political Obstacles to Decentralization 123
and on the accountability of policy-makers. One of the promises of de-
centralization is that by shifting policy authority to local officials who are
closer to citizens, it will become easier for voters to monitor the uses and
abuses of this authority (Musgrave, 1965). While this is an important goal,
in the two cases reviewed here, national politicians responded to decentral-
ization in ways that compromised the goals of transparency and account-
ability by increasing the complexity of relations between national and sub-
national governments. In Argentina, Peronists implemented a series of highly
complicated fiscal pacts that obscured the balance of authority between the
different levels of government. In the Philippines, by re-centralizing control
over some services, and by legislating funds that duplicate automatic
transfers, national legislators have made it difficult for citizens to determine
which of their elected representatives are responsible for actual policy out-
comes. That the resulting confusion does not bode well for transparency and
accountability is less a criticism of decentralization than of the ways that
national politicians often respond to its implementation.
REFERENCES
Abueva, J. (1965) `Social Backgrounds and Recruitment of Legislators: Administrators in a
Developing Country: The Philippines', Philippine Journal of Public Administration (9): 10±29.
Acun
Äa, C. (1995) La Nueva Matriz PolõÂtica Argentina. Buenos Aires: Ediciones Nueva Visio
Ân.
Ames, B. (1987) Political Survival. Berkeley, CA: University of California Press.
Blair, H. (2000) `Participation and Accountability at the Periphery: Democratic Local Govern-
ance in Six Countries', World Development (28)1: 21±39.
Brillantes, A. (1994) `Redemocratization and Decentralization in the Philippines: The
Increasing Leadership Role of the NGOs', International Review of Administrative Sciences
(60): 575±86.
Burki, S., G. Perry and W. Dillinger (1999) Beyond the Center: Decentralizing the State.
Washington, DC: The World Bank.
Carciofi, R. (1990) `La desarticulacio
Ân del pacto fiscal: una interpretacio
Ân del sector pu blico
argentino en las u ltimas de
Âcadas'. Documento de Trabajo No 36. Santiago: Naciones Unidas,
Comisio
Ân Econo
Âmica para America Latina y el Caribe.
Carey, J. and M. Shugart (1995) `Incentives to Cultivate a Personal Vote: A Rank Ordering of
Electoral Formulas', Electoral Studies 14: 417±39.
Cavarozzi, M. (1994) `Politics: A Key for the Long Term in South America', in W. Smith,
C. Acun
Äa and E. Gamarra (eds) Latin American Political Economy in the Age of Neoliberal
Reform, pp. 127±55. Miami, FL: University of Miami North-South Center.
Cetra
Ângulo, O. and J. Jime
Ânez (1996) `El conflicto en torno a las relaciones financieras entre la
nacio
Ân y las provincias. segunda parte: desde la Ley 23.548 hasta la actualidad'. Serie de
Estudios No 10. Centro de Estudios para el Cambio Estructural.
Confederacio
Ân General Econo
Âmica (1993) El libro azul de las provincias: Las economõÂas
regionales en la transformacio
Ân econo
Âmicas de la repu
Âblica argentina. Buenos Aires: Instituto
de Investigaciones Econo
Âmicas y Financieras de la C.G.E.
Coronel, S. (ed.) (1998) Pork and Other Perks: Corruption and Governance in the Philippines.
Manila: Philippine Center for Investigative Journalism.
Crook, R. and J. Manor (1994) Enhancing Participation and Institutional Performance: Demo-
cratic Decentralization in South Asia and West Africa. London: Overseas Development
Administration.
124 Kent Eaton
Crook, R. and J. Manor (1998) Democracy and Decentralisation in South Asia and West Africa.
Cambridge: Cambridge University Press.
Cruz, B. (1993) `Lower House Body Mulls Local Government Code Changes', Business World
25 January: 2C1.
De Leon, H. (1992) `DA's Regional Offices Formed in Violation of Local Government Code',
Business World 13 August.
Dillinger, W. and S. Webb (1999) `Fiscal Management in Federal Democracies'. World Bank
Policy Research Working Paper No 2121. Washington, DC: The World Bank.
Doronila, A. (1992) The State, Economic Transformation, and Political Change in the
Philippines, 1946±1972. Singapore: Oxford University Press.
Dura
Ân, V. and J. C. Go
Âmez Sabaini (1993) `Lecciones Sobre Reformas Fiscales en Argentina:
1990±1993'. Serie Polõ
Âtica Fiscal No 68. CEPAL/PNUD.
Emboltura, A. (1994a) `Searching for Equity: The IRA Debate', Legislative Features 2(1): 5±8.
Emboltura, A. (1994b) `Recentralization or Suspension: A Step in the Right Direction?',
Legislative Alert (July±September): 23±27.
Ferejohn, J. (1974) Pork Barrel Politics: Rivers and Harbors Legislation, 1947±1968. Stanford,
CA: Stanford University Press.
Fox, J. (1990) `The Challenges of Rural Democratization in Latin America and the Philippines',
Journal of Development Studies 26(4): 1±18.
Fox, J. (1994) `Latin America's Emerging Local Politics', Journal of Democracy 5(2): 105±116.
Franco, J. (1997) `Elections and Democratization in the Philippines'. PhD Dissertation,
Brandeis University.
Fundacio
Ân CECE (1998) Federalismo fiscal en Argentina. Buenos Aires: Fundacio
Ân centro de
estudios para el cambio estructural.
Fundacio
Ân de Investigaciones Econo
Âmicas Latinoamericanas (FIEL) (1993) Hacia una nueva
organizacio
Ân del federalismo fiscal en la argentina. Buenos Aires: Ediciones Latinoamericanas.
Garman, C., S. Haggard and E. Willis (1996) `Decentralization in Latin America', paper
presented at the Annual Meeting of the American Political Science Association, San
Francisco (29 August±1 September).
Gutierrez, E. (1994) The Ties that Bind: A Guide to Family, Business, and Other Interests in the
Ninth House of Representatives. Pasig: Philippine Center for Investigative Journalism.
Hofilen
Äa, Chay (1993) `Confusing Code', Sunday Chronicle, 25 July: 5.
Hutchcroft, P. (1998) Booty Capitalism: The Politics of Banking in the Philippines. Ithaca, NY:
Cornell University Press.
Inter-American Development Bank (1997) Latin America After a Decade of Reforms. Wash-
ington, DC: Johns Hopkins University Press.
Jones, M. (1997) `Evaluating Argentina's Presidential Democracy: 1983±1995', in S. Main-
waring and M. Shugart (eds) Presidentialism and Democracy in Latin America, pp. 259±99.
New York: Cambridge University Press.
Jones, M., P. Sanguinetti and M. Tommasi (2000) `Politics, Institutions, and Fiscal Perform-
ance in a Federal System: An Analysis of the Argentine Provinces', Journal of Development
Economics 61: 305±33.
Lacaba, J. (ed.) (1995) Boss: Five Case Studies of Local Politics in the Philippines. Pasig:
Philippine Center for Investigative Journalism.
Macabuhay, Maria Luisa (1992) `P36M Set aside for Devolution Process', Daily Globe,
29 August: 6.
Manasan, R. (1992) `Intergovernmental Fiscal Relations, Fiscal Federalism, and Economic
Development in the Philippines'. Working Paper Series No 92-04. Philippine Institute for
Development Studies.
Manasan, R. (1996) `Patterns of Budget Allocations Using Social and Human Priority
Expenditure Ratio'. Philippine Institute for Development Studies (mimeo).
Manor, J. (1995) `Democratic Decentralization in Africa and Asia', IDS Bulletin 26(2): 81±88.
Manor, J. (1999) The Political Economy of Democratic Decentralization. Washington, DC: The
World Bank.
Political Obstacles to Decentralization 125
McCoy, A. (ed.) (1994) An Anarchy of Families: State and Family in the Philippines. Quezon
City: Ateneo de Manila University Press.
McGuire, J. (1997) Peronism without Pero
Ân: Unions, Parties, and Democracy in Argentina.
Stanford, CA: Stanford University Press.
Montinola, G. (1999) `Politicians, Parties, and the Persistence of Weak States: Lessons from the
Philippines', Development and Change 30(4): 739±74.
Murillo, M. V. (1999) `Recovering Political Dynamics: Teachers' Unions and the Decentraliza-
tion of Education in Argentina and Mexico', Journal of Interamerican Studies and World
Affairs 41(1): 31±57.
Musgrave, R. (ed.) (1965) Essays in Fiscal Federalism. Washington, DC: The Brookings Institution.
Novaro, M. (1994) Pilotos de tormentas: crisis de representacio
Ân y personalizacio
Ân de la polõÂtica
en argentina. Buenos Aires: Ediciones Letra Buena.
Nun
Äez Min
Äana, H. and A. Porto (1982) `Coparticipacio
Ân Federal de Impuestos: Distribucio
Ân
Primaria', Jornadas de Finanzas Pu
Âblicas 15: 4.1±4.73.
Pagdanganan, Roberto (1992) `LGUs Blame Congress for Local Code Mess', Philippine Daily
Inquirer 9 August: 26.
Pagdanganan, Roberto (1996) `Small Towns Going Broke Over Devolution Policy', Philippine
Graphic 6(47) (April 29).
Panganiban, E. (1995) `Democratic Decentralization in Contemporary Times: The New Local
Government Code of the Philippines', Philippine Journal of Public Administration 39(2):
121±38.
Pimentel, Aquilino (1993) The Local Government Code of 1991: The Key to National Develop-
ment. Metro Manila: Cacho Publishing House, Inc.
Põ
Ârez, P. (1986) Coparticipacio
Ân federal y descentralizacio
Ân del estado. Buenos Aires: Centro
Editor de America Latina.
Porto, A. (1990) Federalismo fiscal: el caso argentino. Buenos Aires: Editorial Tesis.
Porto, A. and P. Sanguinetti (1993) `Decentralizacio
Ân Fiscal en America Latina: El Caso
Argentino', Serie Polõ
Âtica Fiscal No 45. Santiago de Chile: Comisio
Ân Econo
Âmica para
America Latina y el Caribe.
Prud'homme, R. (1995) `The Dangers of Decentralization', The World Bank Research Observer
10(2): 201±220.
Republic of the Philippines (1997) Budget of Expenditures and Sources of Financing: Fiscal Year
1997. Manila: Department of Budget and Management.
Repu blica Argentina (1994) Cambios estructurales en la relacio
Ân nacio
Ân-provincias. Buenos
Aires: Ministerio de Economõ
Âa y Obras y Servicios Pu blicos.
Repu blica Argentina (1995) El federalismo fiscal a partir de la reforma constitucional. Buenos
Aires: Asociacio
Ân Mutual Federal de Empleados de la Direccio
Ân General Impositiva.
de Riz, L. and C. Smulovitz (1991) `Instituciones y dina
Âmica polõ
Âtica: el presidencialismo
argentino', in D. Nohlen and L. de Riz (eds) Reforma institucional y cambio polõÂtico,
pp. 121±76. Buenos Aires: CEDES, Editorial Legasa.
Rocamora, J. (1995) `Classes, Bosses, Goons, and Clans: Re-imagining Philippine Political
Culture', in J. Lacaba (ed.) Boss: Five Case Studies of Local Politics in the Philippines,
pp. vii±xxxi. Manila: Institute for Popular Democracy.
Rocamora, J. (1996) `Unnatural Disasters', Politik 2(4): 46±48.
Rondinelli, D. (1989) `Decentralizing Public Services in Developing Countries: Issues and
Opportunities', Journal of Social, Political, and Economic Studies (14)1: 77±97.
Rondinelli, D. (1990) `Decentralization, Territorial Power and the State: A Critical Response',
Development and Change 21(3): 491±500.
Rondinelli, D., J. Nellis and G. S. Cheema (1984) `Decentralization in Developing Countries:
A Review of Recent Experience', World Bank Staff Working Papers No 581. Washington,
DC: The World Bank.
Rondinelli, D., J. McCullough and R. Johnson (1989) `Analysing Decentralization Policies in
Developing Countries: A Political Economy Framework', Development and Change 20(1):
57±87.
126 Kent Eaton
Saiegh, S. and M. Tommasi (1998) `Argentina's Federal Fiscal Institutions'. Buenos Aires:
Centro de Estudios para el Desarrollo Institucional.
Saiegh, S. and M. Tommasi (1999) `Why is Argentina's Fiscal Federalism so Inefficient?
Entering the Labyrinth', Journal of Applied Economics, 2(1): 169±209.
Samoff, J. (1990) `Decentralization: The Politics of Interventionism', Development and Change
21(3): 513±30.
Samuels, D. and F. Abrucio (forthcoming) `The New Politics of the Governors: Subnational
Politics and the Brazilian Transition to Democracy', Publius: The Journal of Federalism.
Saspa, J.P. (1997) `The Stinking Pork War', Philippines Free Press 18 January: 14.
Scho
Ènwa
Èlder, G. (1997) `New Democratic Spaces at the Grassroots? Popular Participation in
Latin American Local Governments', Development and Change 28(4): 753±70.
Schwartz, G. and C. Liuksila (1997) `Argentina', in Teresa Ter-Minassian (ed.) Fiscal
Federalism in Theory and Practice, pp. 387±422. Washington, DC: International Monetary
Fund.
Sidel, J. (1995) `Coercion, Capital, and the Post-Colonial State: Bossism in the Postwar
Philippines'. PhD Dissertation, Cornell University.
Slater, D. (1989) `Territorial Power and the Peripheral State: The Issue of Decentralization',
Development and Change 20(3): 501±31.
Slater, D. (1990) `Debating Decentralization Ð A Reply to Rondinelli', Development and
Change 21(3): 501±12.
Souza, C. (1996) `Redemocratization and Decentralization in Brazil: The Strength of the
Member States', Development and Change 27(3): 529±55.
Stauffer, R. (1975) The Philippine Congress: Causes of Structural Change. Beverly Hills, CA:
Sage Publications.
Tapales, P. (1991) `Devolution and Empowerment: The Local Government Code of 1991 and
Local Autonomy in the Philippines', in P. Tapales and N. Pilar (eds) Public Administration
by the Year 2000: Looking Back into the Future, pp. 7±32. Quezon City: College of Public
Administration, University of the Philippines.
Tapales, P. (1993) Devolution and Empowerment: The Local Government Code of 1991 and Local
Autonomy in the Philippines. Quezon City: University of the Philippines Press.
Willis, E., C. Garman and S. Haggard. (1999) `The Politics of Decentralization in Latin
America', Latin American Research Review 34(1): 7±56.
World Bank (1990) Provincial Government Finances. Washington, DC: The World Bank.
World Bank (1993) Argentina: From Insolvency to Growth. Washington, DC: The World Bank.
Wurfel, D. (1988) Filipino Politics: Development and Decay. Ithaca, NY: Cornell University
Press.
Kent Eaton is Assistant Professor of Politics and International Affairs at the
Woodrow Wilson School at Princeton University (contact address: Center
for International Studies, Bendheim Hall, Princeton, NJ 08544, USA). His
work focuses on the politics of economic reform in new democracies; he has
articles forthcoming in the Journal of Latin American Studies and the Latin
American Research Review.
Political Obstacles to Decentralization 127