Content uploaded by Ignacio Arana
Author content
All content in this area was uploaded by Ignacio Arana on Aug 31, 2014
Content may be subject to copyright.
This article was downloaded by: [University Of Pittsburgh]
On: 30 August 2014, At: 19:57
Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954
Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH,
UK
The Journal of Legislative
Studies
Publication details, including instructions for authors
and subscription information:
http://www.tandfonline.com/loi/fjls20
Budgetary Negotiations: How the
Chilean Congress Overcomes its
Constitutional Limitations
Ignacio Arana Araya
Published online: 11 Aug 2014.
To cite this article: Ignacio Arana Araya (2014): Budgetary Negotiations: How the
Chilean Congress Overcomes its Constitutional Limitations, The Journal of Legislative
Studies, DOI: 10.1080/13572334.2014.943491
To link to this article: http://dx.doi.org/10.1080/13572334.2014.943491
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the
information (the “Content”) contained in the publications on our platform.
However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or
suitability for any purpose of the Content. Any opinions and views expressed
in this publication are the opinions and views of the authors, and are not the
views of or endorsed by Taylor & Francis. The accuracy of the Content should
not be relied upon and should be independently verified with primary sources
of information. Taylor and Francis shall not be liable for any losses, actions,
claims, proceedings, demands, costs, expenses, damages, and other liabilities
whatsoever or howsoever caused arising directly or indirectly in connection
with, in relation to or arising out of the use of the Content.
This article may be used for research, teaching, and private study purposes.
Any substantial or systematic reproduction, redistribution, reselling, loan, sub-
licensing, systematic supply, or distribution in any form to anyone is expressly
Budgetary Negotiations: How the Chilean
Congress Overcomes its Constitutional
Limitations
IGNACIO ARANA ARAYA
Recent research suggests that the Chilean Congress is marginalised in the policymaking
process, especially when setting the budget. This paper argues that previous studies have
overlooked the fact that the legislature uses two amendment tools – specifications and
marginal notes – to increase the national budget and reallocate resources within minis-
tries. This behaviour contradicts the constitution, which only allows Congress to reduce
the executive’s budget bill. To test this empirically, a pooled two-stage time-series
cross-sectional analysis is conducted on ministries for the years 1991–2010. The findings
clarify how the legislature surpasses its constitutional limits and demonstrate that speci-
fications are useful to predict when Congress increases or decreases a ministry’s budget.
Keywords: public budget; executive–legislative relations; constitution.
Introduction
The literature on presidentialism has consistently ranked Chile as having one of
the four most powerful presidential systems in the world (see, for example,
Samuels & Shugart, 2003; Shugart & Carey, 1992; Shugart & Haggard, 2001;
Tsebelis & Alema
´n, 2005). In all these studies, the assumed uncontested budget-
ary powers of the Chilean president help to move it higher in the rankings.
Further, comparative studies suggest that Chilean state powers pay stringent
respect to the law, making them an exception in Latin America. This study dis-
agrees, arguing that previous research has overlooked how the Chilean legislature
bypasses the constitution, which limits the Congress by only allowing it to reduce
the executive’s budget. An examination of the budgetary negotiations shows that
Congress uses two amendment mechanisms – marginal notes and specifications
– to circumvent the rigidity of the charter, a practice that has allowed it to transfer
resources and increase some ministries’ allocations. The analysis of the mechan-
isms that operate during the executive – legislative negotiations questions pre-
vious accounts of budgetary relations in Chile and highlights the limitations of
analysing presidential and legislative powers based only on formal rules.
The budgetary negotiations are highly relevant for the political elite. The
fiscal budget determines the level of public goods for the broader population
and is one of the main tools (along with monetary and exchange rate policy)
that government uses to intervene the domestic economy. Public budgets also
The Journal of Legislative Studies, 2014
http://dx.doi.org/10.1080/13572334.2014.943491
#2014 Taylor & Francis
Downloaded by [University Of Pittsburgh] at 19:57 30 August 2014
condition electoral outcomes because voters assess government’s performance
partially on how they perceive the way that political elites distribute resources
across social and electoral groups (Bueno de Mesquita, Smith, Siverson, &
Morrow, 2004).
A marginal role of Congress in the public budget leads to important policy-
making consequences. As legislatures weakly oversee governmental policies,
the outcomes tend to be technically poor and ill-adjusted to the needs of citizens.
Moreover, the legislature fails to represent the voters’ preferences on the budget.
In the institutional literature, Chile represents one of the most executive-
based models of budget elaboration in Latin America (Alesina, Hausmann,
Hommes, & Stein, 1999; Vial, 2001). Allegedly, Congress and the executive
(‘budgetary actors’ from now on) profess a rigorous respect for the institutional
design (Aninat, Londregan, Navia, & Vial, 2006). A report of the Organisation
for Economic Cooperation and Development (OECD) summarises this view:
‘The current budget process is designed for the Congress to swiftly approve
the government’s proposal with no changes’ (Blo
¨ndal & Curristine, 2004, p. 25).
So far, some studies have described the complex informal relationship
between the executive and Congress in Chile (Carey & Siavelis, 2006; Ferraro,
2008; Siavelis, 2006), and others have partially described legislators’ efforts to
influence the budget (for example, Baldez & Carey, 2001), but none has exam-
ined how and to what extent Congress shapes the public purse. This vacuum
has led scholars to assume that Congress does not try to accommodate the
budget to its preferences, despite the strong incentives to do so, and as happens
in other Latin American countries. Previous research has documented how
Latin American congresses and executives go beyond their legal powers and
directly contradict the law – or the spirit of the law (Hallerberg, Scartascini, &
Stein, 2009).
1
This paper demonstrates that the Chilean Congress also accommodates the
constitutionally based norms to its preferences. The Congress amends the
budget bill voting specifications and marginal notes in a more liberal way than
the constitution allows. While the constitution only allows Congress to reduce
the executive’s budget bill, the legislature uses specifications and marginal
notes to increase it. The assembly prefers to accommodate the budgetary
process than to amend the constitution because the latter is too costly, requiring
a super-majority of three-fifths of Congress.
The specifications are amendments that increase or decrease a ministry’s
budget, and can also modify parts of its content. For example, if the executive’s
bill proposes to spend US$100 million to subsidise the economic activities of an
indigenous group, Congress can increase or reduce the amount, or establish
certain conditions that need to be met before giving the subsidy. Both the execu-
tive and Congress can also include marginal notes in the bill, which specify how
the ministry’s budget will be distributed. Marginal notes can also change the sub-
stantive content of the bill, but they do not change the overall amount of a min-
istry’s budget. For instance, they can determine that a specific amount of the
2THE JOURNAL OF LEGISLATIVE STUDIES
Downloaded by [University Of Pittsburgh] at 19:57 30 August 2014
US$100 million allocated to subsidise an indigenous group has to be spent in a
certain productive activity (for instance, agriculture). Marginal notes may also
list which economic activities cannot be subsidised, but in contrast to specifica-
tions, they cannot modify the US$100 million. On average, Congress annually
votes on 50 specifications and 156 marginal notes and approves 86 per cent
and 68 per cent of them, respectively.
To understand executive– legislative negotiations, semi-structured interviews
were conducted with 15 key informants involved in the process in Santiago and
Valparaı
´so. Key informants provide information that cannot be obtained from
other sources, given that some practices go against the law and are usually
hidden from public view.
2
The key informants interviewed were selected based on their participation in
the budgetary negotiations. Among the interviewees were two senators, three
deputies and two former ministers who collectively spanned the political spec-
trum. The other eight interviewees included two analysts from think tanks associ-
ated with political parties and six members of the Budget Office (Direccio
´nde
Presupuestos, part of the Finance Ministry); of these six, three were former direc-
tors and the other three were high-level officials.
3
To test the implications of marginal notes and specifications, this paper
reviews the reports of 1991– 2010 that register the budgetary bargaining for 19
ministries plus Congress, the judiciary, and the Office of Comptroller General.
4
These reports usually have hundreds of pages and contain all the budgetary dis-
cussion held in congressional sessions where the amendments are voted upon.
5
The next section describes the institutional setting where the budgetary bar-
gaining takes place. The third section proposes four hypotheses that capture how
the budgetary actors use the amendments to advance their preferences in the
budget. The fourth section presents the results of a pooled two-stage time-
series cross-sectional analysis that tests the hypotheses. The final section dis-
cusses the implications of the Chilean budgetary negotiations for research on
Latin American executive – legislative relations.
Assessing the Executive’s Advantage in the Chilean Public Budget
Budgetary bargaining rules vary across countries. Among presidential systems,
scholars view the US federal budget as being jointly determined by presidents
and Congress (Canes-Wrone, 2006; Kiewiet & McCubbins, 1991). By contrast,
they claim that in most of Latin America the power of the purse is strongly
biased towards the executive, which controls the budget formulation and
execution (Hallerberg et al., 2009; Pelizzo, Stapenhurst, & Olson, 2005). Even
in this context, Chile is an extreme executive-based model of budget elaboration.
The Chilean constitution created by the military in 1980 concentrates almost
all of the budgetary process in the executive, with a strong pre-eminence of the
finance minister and the Budget Office.
6
Article 65 of the constitution gives
the president control over all tax and spending policies, while article 67
BUDGETARY NEGOTIATIONS: CHILEAN CONGRESS 3
Downloaded by [University Of Pittsburgh] at 19:57 30 August 2014
establishes that the executive is the only actor responsible for creating a budget
bill proposal that considers all fiscal expenditures. Article 67 also establishes:
that (1) Congress cannot change the government’s forecasts of future fiscal
income; (2) if Congress has not approved the budget bill within 60 days of the
executive’s proposal being sent, it becomes law; (3) Congress can only reduce
the budget without touching the portion allocated by other permanent laws; but
(4) the legislature cannot increase any budget item or reassign expenditures
among programmes.
A crucial issue is that the annual bargaining covers about 15 per cent of the
total budget; previous laws and governmental contracts condition the other 85 per
cent. The Budget Office sub-director between 2000 and 2010, Sergio Granados,
remarked that the rigidity of the public budget varies slightly by ministry and
year.
7
Although the 15 per cent may seem low, it is not in relative terms.
8
In sum, Congress can only reduce the flexible part of the budget and has
60 days to do so. However, the prerogative to reduce allocations is of little inter-
est to legislators. According to most scholars who have studied executive – legis-
lative relations in Latin America (for example, Hallerberg et al., 2009; Nacif &
Morgenstern, 2002; Scartascini, Stein, & Tommasi, 2010) or elsewhere
(Haggard & McCubbins, 2000; Poterba & Von Hagen, 1999), legislators are
mainly interested in distributing goods and services to their constituencies.
Hence, representatives have material and electoral incentives to increase spend-
ing, not to reduce it.
9
So far, scholars who have discussed the Chilean budgetary process have
assumed that the net effect of the congressional intervention in the public
budget is low and it only reduces the executive’s budget bill. However, this
assumption is inconsistent with congressional behaviour. Table 1 shows the
differences between the executive’s budget bill and the Budget Law approved
by Congress. These differences result from bargaining practices through which
Congress introduces changes to the budget of some ministries. Table 1 documents
that, in seven of 20 years, Congress managed to increase the overall public
budget despite the constitutional prohibition. The net change is between 23.7
and 2.5 per cent of the flexible part of the budget.
To reveal ministry variation, Table 2 shows the public budget disaggregated
by ministry, displaying the annual largest variations. In line with the specialised
literature, Congress is very interested in increasing expenditure: in 13 of 20 years
the largest percentage variation occurs in ministries where the budget increases.
Table 2 also shows that Congress increased the budgets of ministries 60 times
and in 18 of the 20 years under study. According to the constitution, none of the
budgetary increases should have occurred. It shows that, except in 2010, ministry
budgets changed regardless of whether the overall public budget changed. This
occurred because Congress reallocated resources between ministries.
The budgetary actors violate the constitution when Congress votes on amend-
ments. While Congress can only legally reduce the executive’s budget bill, in all
the reports analysed in this article Congress introduced specifications that
4THE JOURNAL OF LEGISLATIVE STUDIES
Downloaded by [University Of Pittsburgh] at 19:57 30 August 2014
increased the ministries’ budgets. Off the record, an interviewee stated that bar-
gaining actors accept this practice as long as the numbers are not excessively
high, something that is decided on a case-by-case basis. But when the negotiation
is about higher values, the executive objects to the increases or, in compliance
with the constitution, proposes an amendment to its own bill. As the government
has no incentive to amend its budget bill, the executive amendments can be con-
sidered as concessions to Congress.
Table 2 also shows that Congress increases the spending in all sorts of min-
istries, irrespective of their relative budget size. Congress has increased up to 89
per cent of the flexible part of a budget (secretariat of government in 1995) and on
average has increased the ministries’ budgets by US$6 million each year.
10
Presumably legislators scrutinise and modify the ministries’ budgets owing to
concerns about their own electoral fortunes. Studies have demonstrated that
Chilean legislators have higher rates of re-election seeking than most others in
the Americas, and that they often succeed (Navia, 2008). Further, recent
studies have described how Chilean legislators actively negotiate with the execu-
tive to distribute pork to their constituencies (Aninat et al., 2006; Toro, 2007).
Legislators from the opposition coalition Alianza por Chile and from the ruling
Table 1: Budgetary Variation due to Congress
Year
Executive’s
Budget Law Total Budget Change (%)∗∗
Bill∗
1991 2,509,002 2,514,309 1.396
1992 3,175,499 3,175,499 0.000
1993 3,866,243 3,880,877 2.498
1994 4,610,964 4,610,964 0.000
1995 5,505,171 5,474,843 23.636
1996 6,267,275 6,265,357 20.202
1997 7,135,565 7,135,565 0
1998 8,053,023 8,053,023 0
1999 8,495,969 8,495,969 0
2000 9,274,596 9,274,709 0.008
2001 10,225,550 10,222,796 20.178
2002 11,340,114 11,336,360 20.219
2003 11,889,979 11,890,246 0.015
2004 12,496,288 12,425,588 23.734
2005 13,102,590 13,102,639 0.003
2006 14,763,933 14,763,971 0.002
2007 17,430,780 17,430,780 0
2008 20,209,839 20,209,839 0
2009 22,906,583 22,922,583 0.461
2010 25,046,832 25,046,832 0
Notes: ∗In Chilean million pesos. (By July 2014, one million Chilean pesos were equivalent to
US$1.817 or English £1.060.)
∗∗This percentage considers only the flexible part of the annual budget (15 per cent of the total
budget).
Sources: Websites of the Direccio
´n de Presupuestos (www.dipres.cl) and the National Congress
(www.congreso.cl).
BUDGETARY NEGOTIATIONS: CHILEAN CONGRESS 5
Downloaded by [University Of Pittsburgh] at 19:57 30 August 2014
coalition Concertacio
´nsupported this view. Opposition Senator Evelyn Matthei
(Independent Democratic Union [UDI]) stated: ‘Some legislators manage to
build special programs in their regions to achieve re-election. It happens all the
time.’
11
The socialist Senator Carlos Ominami confirmed Matthei’s claims: ‘It
is natural that pro-government legislators have more capacity to obtain programs
for their regions than opposition legislators.’
12
According to Matthei, during budgetary negotiations opposition legislators
are unable to detect when the government allocates resources to help Concerta-
cio
´nlegislators to achieve re-election: ‘You cannot know it reading the Budget
Law, because when you approve it you only know the general amount of
money allocated for public policies.’ Matthei claimed that the government
helps Concertacio
´nlegislators via the execution of the budget, concentrating
resources and distributing them in a timely manner to help pro-government
legislators.
Interestingly, the Congress budget was the most changed (14 years out of 20)
and increased in 1991– 2010. Figure 1 shows that in most of the years the
Congress budget incremented significantly. Increases in the congressional
budget allow legislators to dispose of a larger set of human and financial
Table 2: Largest Annual Budgetary Change due to Congress
Year Ministry
Budget Change
(%)∗
Budget
Increases
Budget Size of
Ministries∗∗
1991 Secretariat of Presidency 214.92 1 9
1992 Transport and
Telecommunications
37.49 3 10, 22, (4)
1993 Agriculture 60.85 6 8, 10, 20, (14), 9, 5
1994 Secretariat of Government 13.33 2 16, (3)
1995 Secretariat of Government 88.84 5 10, 14, 21, 13, (4)
1996 Mining 278.14 3 5, 16, 17
1997 Agriculture 14.26 5 9, 13, 21, (14), 4
1998 Congress 20.39 3 (8), 21, 19
1999 Agriculture 38.87 5 16, 13, 20, 18, (14)
2000 Agriculture 52.40 5 10, 4, 21, (14), 7
2001 Congress 17.82 3 (9), 21, 19
2002 Planning 29.24 1 21
2003 Congress 11.95 5 (10), 5, 16, 11, 3
2004 Congress 9.24 2 (10), 21
2005 Congress 30.76 3 (10), 21, 8
2006 Congress 37.03 5 (9), 6, 21, 8, 3
2007 Housing and Urban Planning 23.96 0
2008 Secretariat of Presidency 23.10 1 8
2009 Secretariat of Government 211.09 2 16, 14
2010 0 0
Notes: ∗This percentage considers only the flexible part of the annual budget (15 per cent of the total
budget).
∗∗For each year, the budgets of the 22 ministries were ranked from smallest to largest, with one being
the smallest and 22 the largest. The numbers in parentheses indicate the size of the budget that
experienced the largest budgetary change for that year.
Source: Direccio
´n de Presupuestos (www.dipres.cl).
6THE JOURNAL OF LEGISLATIVE STUDIES
Downloaded by [University Of Pittsburgh] at 19:57 30 August 2014
capital that increases their re-election perspectives. In fact, the reports of the
negotiations reveal that many of the budgetary increases deal with items that
improve the legislators’ electoral perspectives, such as staff expenses, equipment
purchases, and expenditures in goods and services. For instance, in the report for
the 2005 Budget Law (Library of the National Congress [Biblioteca del Congreso
Nacional], 2004, p. 34), Congress included few amendments (contained in 303
words) to its budget, but all the specifications were devoted to increasing budget-
ary allocations (four to the Senate, four to the Chamber of Deputies and three to
the Library of the National Congress).
Executive– Legislative Bargaining
The government sends the budget bill to the Special Budget Joint Committee
(SBJC) every 30 September. The committee reviews and amends the bill for
some weeks before sending it to the Chamber of Deputies, which votes on the
proposed bill and sends it to the Senate. Both chambers can amend the bill, but
the congressional reports show that they rarely do so because they entrust that
responsibility to the committee.
13
Consequently, in this study the committee is
considered to represent Congress in the budgetary process.
The SBJC is composed of 13 deputies and 13 senators who proportionately
represent forces in Congress. During 1991 – 2010, the SBJC was controlled by
pro-government forces. The committee is divided into five subcommittees that
debate the ministries’ budgets for almost four weeks (Blo
¨ndal & Curristine,
2004). Then, the SBJC meets in plenary and votes on the subcommittees’ pro-
posed amendments.
14
When the SBJC includes an amendment, the executive
can make a counter offer, but it rarely does.
Key informants reveal that the executive has incentives to make concessions
to Congress during the negotiations despite its legal budgetary advantage.
15
‘We
have to avoid leaving wounds so deep that may reappear in another negotiation’,
Figure 1: Congressional Budget Percentage of Change 1991– 2010
BUDGETARY NEGOTIATIONS: CHILEAN CONGRESS 7
Downloaded by [University Of Pittsburgh] at 19:57 30 August 2014
said Mario Marcel, a former Budget Office director (2000–6).
16
‘The budget is
not the only bill, so the Finance Ministry has to be alert to several legal projects’,
added Joaquı
´n Vial, who was Budget Office director in 1997 – 2000.
17
The executive usually tries to sidestep a confrontation with Congress because
the legislature uses five tactics to raise the costs of neglecting congressional pre-
ferences. First, Vial and Marcel claimed that if the executive does not incorporate
the proposed amendments in the budget bill, legislators threaten to retaliate by
rejecting other executive bills. Second, legislators threaten to escalate the conflict
through the media to affect the government’s popularity.
18
Third, legislators can
delay the voting on the executive’s bill to earn extra time to bargain and increase
the pressure over the executive. Fourth, the constitution establishes that the fiscal
debt beyond the current budgetary year must be approved by an absolute majority
of both chambers through the Budget Law. Former Christian Democrat SBJC
member and Planning Minister Andre
´s Palma argued that since Chilean govern-
ments usually incur debts that endure beyond their terms, the SBJC threatens to
approve the executive’s bill but reject its debt proposal.
19
Finally, sometimes
Congress uses its constitutional prerogative to reduce a budgetary item to a ridi-
culous amount (such as US$2) to force the government to concede to its
demands.
20
The budgetary negotiations give Congress a unique opportunity to reallocate
resources and increase a ministry’s budget. This paper proposes that the actors
that participate in the negotiations use specifications and marginal notes to rep-
resent their preferences in a way that has predictable budgetary consequences.
Therefore, this paper empirically tests the expectation that Congress uses amend-
ments to ‘accommodate’ the public budget to its preferences and that the execu-
tive makes concessions to the legislature. The results demonstrate that Congress
actively challenges its constitutional limits in the budgetary process, instead of
being passive and marginal, as previous studies claim.
Given the nature of the amendments, the hypotheses proposed are specific to
the Chilean executive– legislative negotiations. However, the results are relevant
to the broader study of legislative practice, especially in Latin America, where
other studies have documented a significant distance between de jure and de
facto rules in budgetary negotiations (see, for example, Hallerberg et al., 2009,
p. 9) and in the broader context of executive – legislative negotiations (Helmke
& Levitsky, 2006).
The executive’s budget bill represents the government’s preferences. While
the government has no motivations to amend its budget bill, the assembly is
interested in adapting the budget to its preferences. Therefore, more amendments
proposed – by any actor – signals that there are more salient issues for Congress.
As legislators prefer to increase a ministry’s budget, the following should
hold true:
H1: The greater the number of amendments proposed to a ministry’s
budget, the larger the increase in the budget of that ministry.
8THE JOURNAL OF LEGISLATIVE STUDIES
Downloaded by [University Of Pittsburgh] at 19:57 30 August 2014
It may seem counterintuitive that the executive proposes amendments to its own
bill, but it does so frequently (it proposed 409 specifications and 364 marginal
notes in 1991–2010) to make concessions to Congress. That explains why the
committee approved nearly 95 per cent of the executive’s proposals. It also
explains why Concertacio
´n’s legislators proposed only 1 per cent of the amend-
ments; they did not need to propose any, because their government did it for
them. Pro-government legislators also allowed the executive to control the pro-
posal of amendments because they do not want to confront the government
given its capacity for retaliation. As Congress is interested in increasing the
budgets, it can be expected that:
H2: The greater the number of specifications introduced by the executive to
a ministry, the larger the increase in the budget of that ministry.
Pro-government legislators can bargain directly with the executive to increase
allocations. However, when Concertacio
´n’s legislators want to decrease expen-
ditures they can minimise the antagonism with the government and maximise
the possibilities of success by jointly proposing amendments with Alianza legis-
lators. Since Alianza legislators were a minority in the SBJC during 1991 – 2010,
they also needed to ally with legislators of the Concertacio
´nto decrease budget-
ary allocations. That explains why when members of both coalitions jointly
proposed amendments in the subcommittees, the SBJC approved 97 per cent of
the specifications and 94 per cent of the marginal notes. In sum, SBJC pro-
government and opposition legislators have incentives to propose amendments
jointly to reduce a ministry’s budget:
H3: The greater the number of joint specifications introduced by legislators
to a ministry, the larger the decrease in the budget of that ministry.
Alianza representatives proposed 20.82 per cent of the specifications and 67.12
per cent of the marginal notes, but the SBJC approved only 13 per cent and 8
per cent of them, respectively. Despite their low success rate, opposition legis-
lators were much more active than pro-government forces in attempting to
amend the executive’s bill. Arguably their goal was to counteract the budgetary
bias towards pro-governmental forces. Opposition legislators are likely to
propose more amendments to a ministry’s budget when their disapproval of it
is stronger. Occasionally they get concessions from pro-government legislators,
who sometimes need to collaborate with the opposition to approve other bills.
Consequently, an increasing number of amendments proposed by the opposition
suggests that Alianza is trying to reduce a ministry’s budget. As the opposition
occasionally succeeds, the following should occur:
H4: The greater the number of amendments proposed by opposition legis-
lators to a ministry, the larger the decrease in the budget of that ministry.
BUDGETARY NEGOTIATIONS: CHILEAN CONGRESS 9
Downloaded by [University Of Pittsburgh] at 19:57 30 August 2014
Of the 440 ministry-years that the legislators could have examined in 1991–
2010, the SBJC amended 74 per cent of them. Formally, the absence of
amendments in 120 budgets means that Congress completely agrees with the
executive’s bill. However, in practice, each year Congress chooses not to
discuss some budgets because it has little time to review them. As Palma
claimed:
21
‘Often large budgets are approved very quickly.’ Consequently, the
committee annually focuses only on the most attractive budgets.
Although the saliency of a ministry’s budget is likely to vary each year
according to the political context, the SBJC should have some recurrent prefer-
ences over them. Four types of ministry should attract the attention of the legis-
lators: first, ministries that allocate more resources to social spending, because
they have a broader social impact and are therefore electorally attractive;
second, ministries with larger budgets allow legislators to oversee and amend
governmental proposals of higher social relevance; third, the committee is
more likely to review ministries that have more agencies, because these minis-
tries target a specific group of voters that can be electorally attractive;
22
and
fourth, legislators should prefer to bargain over ministries that have more sub-
budgets, because they offer more types of pork to distribute.
The political context should also affect the negotiations. First, both govern-
ment and Congress should be tempted to spend more resources in electoral
years to increase their re-election possibilities. Second, the pattern of negotiations
should change in years in which the budgetary negotiators are replaced because
of a new term. Furthermore, the change in the budget of the ministries should
respond to the difference between the number of specifications that increase
the budget of a ministry and the specifications that reduce it.
Data and Methods
A two-stage time-series cross-sectional analysis of all budget items at the minis-
try level between 1991 and 2010 is conducted to examine the use of amendments
during the budgetary bargain. The amendments are classified according to who
proposed them and their rate of approval. Specifications are coded to identify
whether they increase or reduce the executive’s bill amounts, or change the
non-quantitative content of each budget.
As the dependent variable ‘ministry’s budget change’ occurs only when legis-
lators carefully review a ministry, there is a censored selection bias. This bias vio-
lates the Gauss–Markov assumption of zero correlation between the independent
variables and the error term, causing biased and inconsistent estimators
(Kennedy, 2003, p. 282). Heckman (1976) proposed a two-stage estimation pro-
cedure to account for the censored selection effect. In the first stage, the expected
value of the error (called the inverse Mills ratio [IMR]) due to the selection effect
is estimated, and in the second stage the regression with the IMR is rerun as an
extra explanatory variable. Given the structure of the data under analysis and the
selection problem (120 censored agencies), this study follows the Heckman
10 THE JOURNAL OF LEGISLATIVE STUDIES
Downloaded by [University Of Pittsburgh] at 19:57 30 August 2014
procedure. To address the selection bias, a probit regression model is first run
where the dependent variable budget chosen is a dummy that takes the value
of one when the budget of a ministry changes due to the SBJC activity, and
zero otherwise.
The independent variables for the first-stage model are: the number of minis-
try’s agencies, ranging between one (Presidency, Comptroller Bureau and
National Assets) and 39 (Health); the number of ministry’s sub-budgets, which
ranges between one (Presidency, Comptroller Bureau and National Assets) and
44 (Interior); the size of a ministry’s budget, representing between 0.03 per
cent (General Secretariat of Presidency in 1991) and 24.21 per cent (Labour
and Social Security in 2000) of the total budget; and the dummy variable that
measures social spending takes the value of one for the ministries that concentrate
this spending according to the Planning Ministry (Mideplan, 2005) – Education,
Health, Justice, Housing and Urban Planning, and Labour and Social Security –
and zero for the rest.
Table 3 shows the results of the probit, including the model estimated with
marginal effects and the proportional reduction of error. The probit has a high
degree of fit (Prob .chi
2
¼0.000), explains 77 per cent of the cases correctly
and reduces the prediction error from relying merely on the modal category by
20 per cent. The variables that capture the number of agencies and the sub-
budgets are statistically significant at the 0.01 level and the budget size at the
0.1 level.
Marginal effects are computed in a second model to aid an interpretation of
the coefficients. Noticeably, the likelihood that the SBJC will review and
amend a budget decreases by 2.4 per cent when the number of agencies increases
in one unit, holding all other variables to their mean. A plausible interpretation of
this unexpected result is that legislators review budgets with fewer agencies to
produce broader effects when they introduce amendments, instead of targeting
a small constituency.
As expected, the likelihood that the SBJC will review and amend a budget
increases by 3.3 per cent when the number of sub-budgets increases by one
unit, when all other variables are fixed to their mean. Also as expected, the
probability that the SBJC will review a budget increases by 1 per cent
when the size of the budget as a percentage of the total budget increases by
one unit. Unexpectedly, the SBJC does not review the budgets of social spend-
ing agencies more often than others. One possible explanation is that because
previous laws establish a large part of the allocations for those ministries,
legislators may not be interested in touching just a marginal part of those
budgets.
Interestingly, using 1991 as a baseline category, 14 of 19 possible years are
statistically significant and have a positive effect on the likelihood that the
SBJC will review a ministry’s budget. When all variables are fixed to their
mean, legislators are more likely to review the budget in a range between 14
per cent (1993) and 22.7 per cent (in 2009 and 2010). This suggests that the
BUDGETARY NEGOTIATIONS: CHILEAN CONGRESS 11
Downloaded by [University Of Pittsburgh] at 19:57 30 August 2014
Table 3: Budget Chosen by Legislators
Model 1 Model 2 Expected
b
Probit Marginal effects
Variable
Coefficient Coefficient
(Standard Error) (Standard Error)
Agencies 20.089∗∗∗ 20.024∗∗∗
b
.0
(0.030) (0.008)
Sub-budgets 0.119∗∗∗ 0.033∗∗∗
b
.0
(0.027) (0.006)
Budget size 0.037∗0.010∗
b
.0
(0.020) (0.006)
Social spending 20.186 20.053
b
.0
(0.250) (0.074)
1992 0.841∗∗ 0.157∗∗∗
(0.420) (0.049)
1993 0.701∗0.140∗∗
(0.420) (0.057)
1997 1.297∗∗∗ 0.193∗∗∗
(0.439) (0.033)
1998 1.507∗∗∗ 0.204∗∗∗
(0.459) (0.030)
1999 1.128∗∗∗ 0.182∗∗∗
(0.430) (0.037)
2000 2.353∗∗∗ 0.227∗∗∗
(0.588) (0.029)
2001 1.703 0.211∗∗∗
(0.473) (0.029)
2002 1.151 0.184∗∗∗
(0.435) (0.037)
2003 1.049∗∗ 0.176∗∗∗
(0.415) (0.039)
2004 0.961∗∗ 0.169∗∗∗
(0.421) (0.043)
2007 1.718∗∗∗ 0.211∗∗∗
(0.477) (0.029)
2008 1.960∗∗∗ 0.217∗∗∗
(0.515) (0.029)
2009 2.369∗∗∗ 0.227∗∗∗
(0.595) (0.029)
2010 2.377∗∗∗ 0.227∗∗∗
(0.599) (0.029)
Constant 21.169∗∗∗
(0.321)
Log likelihood 2183.908
Prob .chi
2
0.000
Pseudo R
2
0.265
Observations 417
LR chi
2
(22) 132.700
Correctly classified 76.98%
PRE 20.00%
Marginal effect 0.808
Note: The table only reports years that achieved statistical significance. PRE, proportional reduction
of error; LR chi2, Likelihood Ratio Chi-Square test.
∗p,0.1; ∗∗p,0.05; ∗∗∗p,0.01 (two-tailed significance tests).
12 THE JOURNAL OF LEGISLATIVE STUDIES
Downloaded by [University Of Pittsburgh] at 19:57 30 August 2014
SBJC decides each year to review some budgets and not others, perhaps based on
the political context.
Table 4 reports the predicted probabilities for different scenarios. When all
the variables are held at their mean, legislators review a budget 81 per cent of
the time. The number of agencies a ministry has strongly conditions whether
the SBJC will review its budget. The SBJC reviews the budget of a ministry
10 per cent of the time when it has 39 agencies, but reviews the ministry 94
per cent of the years if it has only one agency.
Also of interest, the SBJC always reviews ministries that have 44 sub-
budgets, but the probability drops to 31 per cent when a ministry has only one
sub-budget. Finally, the size of a ministry’s budget relative to the overall
budget does not affect much the likelihood that the SBJC will review it. The prob-
ability ranges between 78 per cent (ministries with a budget share of 0.03 per
cent) and 94 per cent (when the budget share equals 24 per cent).
The second-stage model tests the theory advanced in the paper. In this central
model the dependent variable captures the percentage of change that a ministry’s
budget experiences between the executive’s budget bill and the Budget Law due
to congressional activity.
23
The variable contains the changes of 22 ministries
during 1991–2010, and ranges between 211.8 per cent (Mining in 1996) and
13.5 per cent (General Secretariat of Government in 1995).
Values of the independent variables were calculated from the annual congres-
sional reports that record the executive – legislative bargaining. The annual
number of specifications presented ranges between zero and 34; the approved
specifications that the executive proposes vary between zero and 23; the opposi-
tion-initiated marginal notes range between zero and 27; the approved specifica-
tions jointly proposed by opposition and pro-government legislators vary
between zero and 23; and the percentage of the opposition’s proposed specifica-
tions rejected ranges between zero and 100.
The control variables capture contextual disturbances. The variables that
capture elections take the value of one when they occur (1993, 1999, 2005 and
2009 for presidential elections; 1993, 1997, 2001, 2005 and 2009 for congres-
sional), and zero for the other years. The variables that capture new legislatures
and governmental administrations are also dummy variables. Finally, the variable
Table 4: Predicted Probabilities of Observing the Legislators Reviewing the Budget
Scenario Predicted Probability Standard Error
All variables at their mean 0.811 0.025
Ministry with 39 agencies 0.104 0.135
Ministry with 1 agency 0.937 0.036
Ministry with 44 sub-budgets 1.000 0.003
Ministry with 1 sub-budget 0.314 0.098
Ministry budget size at 24.212 0.939 0.053
Ministry budget size at 0.030 0.778 0.030
BUDGETARY NEGOTIATIONS: CHILEAN CONGRESS 13
Downloaded by [University Of Pittsburgh] at 19:57 30 August 2014
that captures the difference between the number of specifications that increase the
budget of a ministry and the specifications that reduce it ranges between 217 and
nine.
The model is estimated with fixed effects and random effects, and the
Hausman test is performed to discriminate between both estimators.
24
The null
hypothesis of the Hausman test is that the random effects assumption that there
is no correlation between independent variables and unit effects hold. Since
Prob .chi
2
¼0.98, the null hypothesis can be rejected and fixed effects
should be preferred. The post-estimation Wald test for group wise heteroskedas-
ticity in the residuals is also conducted. The result suggests that the fixed effects
estimation should be combined with robust standard errors, clustered by the
cross-sectional units (ministries).
25
Table 5 presents the results of the fixed
effects controlling for robust standard errors in the second stage of the model.
Table 5: Ministry’s Budget Change
Variable
Coefficient
Expected
b
Hypothesis(Standard Error)
Presented specifications 0.104∗∗∗
b
.0H1
(0.033)
Presented marginal notes 20.023
b
.0H1
(0.016)
Executive specifications 0.045
b
.0H2
(0.050)
Joint legislative specifications 20.101∗∗
b
,0H3
(0.045)
Presented opposition specifications 20.186∗∗
b
,0H4
(0.082)
Presented opposition marginal notes 0.026
b
,0H4
(0.016)
Presidential election 0.087
(0.195)
Legislative election 20.038
(0.154)
New legislature 20.046
(0.174)
New government 0.052
(0.299)
Increase– decrease specifications 0.139∗∗∗
(0.042)
IMR 0.787
(0.562)
Constant 20.165
(0.240)
RHO 0.123
SIGMA 1.331
Overall R
2
0.1797
Within R
2
0.1217
Observations 297
Prob .F0.000
Note: ∗p,0.1; ∗∗p,0.05; ∗∗∗p,0.01 (two-tailed significance tests).
14 THE JOURNAL OF LEGISLATIVE STUDIES
Downloaded by [University Of Pittsburgh] at 19:57 30 August 2014
The within R-squared is 12.2 per cent, suggesting that the within effects in
ministries is high and that the model includes relevant variables in explaining
it. The rho – the correlation between the errors in the two equations – is 12.3
per cent, so we can reject the null that rho equals zero. This result supports the
argument that not taking into account the process by which legislators choose
to review and amend certain budgets each year can lead to biased coefficients
and potentially incorrect substantive inferences.
The model shows that, in line with H1, specifications are associated with
budgetary increase. This result confirms that legislators are mainly interested
in increasing the budgets, not in reducing them. Each additional specification
proposed increases the budget of a ministry by 0.1 per cent at the 0.01 level
of statistical significance. Since each year the budgetary actors propose
nearly 50 specifications, in any year the specifications proposed explain 5
per cent of a ministry’s budgetary change. The amounts involved are desir-
able for re-election seekers: even 0.1 per cent of a ministry’s budget rep-
resents nearly US$11.5 million of the largest 2010 budget (Labour and
Social Security). In contrast to specifications, marginal notes are not associ-
ated with budgetary change despite them being related to the saliency of the
negotiations.
Contrary to what was expected in H2, the specifications that the executive
introduces to a ministry’s budget do not predict budgetary change. However,
the absence of statistical significance does not mean the expectation that the
executive mainly incorporates legislative demands in their specifications is
wrong. It may happen that those demands do not tend only to increase the
budget, but also to change part of its content.
In line with H3, each additional joint specification that SBJC legislators
include tends to decrease a ministry’s budget by 0.1 per cent. This result is stat-
istically significant at the 0.05 level and confirms the expectations that pro-gov-
ernment forces ally with the opposition when they want to decrease allocations.
On average, each year the SBJC includes 18 joint specifications, which is associ-
ated to a change of 1.8 per cent in a ministry’s budget.
As expected in H4, the specifications that the opposition presents predict
budgetary decrease at the 0.05 level of statistical significance. Each additional
specification that the opposition presents decreases the budget of a ministry by
0.2 per cent. As the opposition presents an average of 21 specifications each
year, annually they are associated with a decrease of 4.2 per cent of a minis-
try’s budget. However, as in H1, marginal notes are unrelated to budgetary
change.
Interestingly, the control variables show that the pattern of executive – legis-
lative negotiations does not change in electoral years or when there is a new gov-
ernment or Congress. This null effect suggests that the negotiations are highly
stable despite the fact that some bargaining members are replaced or the apparent
incentives that the budgetary actors have to pursue a more aggressive bargaining
to increase their electoral possibilities.
BUDGETARY NEGOTIATIONS: CHILEAN CONGRESS 15
Downloaded by [University Of Pittsburgh] at 19:57 30 August 2014
Conclusion
This analysis has demonstrated that the Chilean legislature successfully manages
to circumvent the stringent budgetary constitutional restrictions and uses the
amendments to reallocate resources within ministries (through marginal notes)
and increase the ministries’ budgets (through specifications). Between 1991
and 2010, Chilean legislators increased the budgets of ministries 60 times,
despite the constitution only allowing reduction of allocations. Furthermore,
SBJC legislators have redistributed hundreds of millions of dollars within minis-
tries each year. These results challenge previous studies that have depicted the
budgetary role of the Chilean Congress as irrelevant and the budgetary actors
as outliers in Latin America owing to their alleged rigorous adherence to the law.
Following Helmke and Levitsky (2006), specifications and marginal notes
can be considered as ‘accommodative’ informal institutions. The authors
define informal institutions as ‘socially shared rules, usually unwritten, that are
created, communicated, and enforced outside officially sanctioned channels’
(Helmke & Levitsky, 2006, p. 5).
26
Although specifications are loosely described
in the Senate Rules coded in the 1980 constitution,
27
the budgetary actors use
them and the marginal notes in a non-written, non-legal fashion. The liberal
use of the amendments is based on the will of the bargaining parties and the gov-
ernment can reverse its unconstitutional use at any time without expecting any
formal sanction. Helmke and Levitsky (2006, p. 15) claim that actors who
dislike outcomes generated by the formal rules but cannot change them or
openly violate them often create accommodative informal institutions that
combine effective formal rules and divergent outcomes. Apparently, Chilean
legislators prefer to use the amendments in a non-constitutional way instead of
amending the constitution, which entails a bigger challenge.
This study has broader implications.The findings suggest that the legalistic
classifications of presidential and legislative powers should be revised because
they are mainly based on formal powers that do not take into account actual
executive– legislative practice.
As legalistic classifications could all be biased, scholars need to study how
state powers routinely relate to each other to uncover the true balance of
power between them. Budgetary rules are embedded in a broader policymaking
process in which actors and their incentives cannot be ignored if the goal is to
understand institutional outcomes. Political actors usually challenge institutional
arrangements when the formal provisions do not allow them to achieve their
goals and the rules are ambiguous or difficult to change. To uncover congres-
sional behaviour, researchers need to examine relevant issues such as the level
of rule enforcement in the country, how enforceable is the rule being assessed,
the motivations that actors have to bypass the rules, and the costs incurred in
such behaviour.
Chile has been widely depicted as a regional leader in terms of institutional
effectiveness, which implies that formal institutions are tightly enforced. This
16 THE JOURNAL OF LEGISLATIVE STUDIES
Downloaded by [University Of Pittsburgh] at 19:57 30 August 2014
paper has demonstrated that an analysis based only on formal rules is misleading
in understanding budgetary negotiations. Arguably, the distance between de jure
and de facto rules should be greater in developing countries with feebler insti-
tutions than that of Chile. In fact, it is widely accepted that in most of Latin
America auditing agencies do not behave as they should, while some numerical
rules regarding spending and deficits are not enforced (Hallerberg et al.,
2009, p. 9).
Note on Author
Ignacio Arana Araya is a PhD candidate in Political Science at the University of
Pittsburgh, PA, USA; email: iga3@pitt.edu
Acknowledgements
The author is very grateful to Anı
´bal Pe
´rez-Lin
˜a
´n, Scott Morgenstern, George
Krause, Cassilde Schwartz and two anonymous reviewers for their comments
and suggestions to improve this manuscript.
Notes
1. For instance, the congresses of Ecuador and Colombia increase their discretionary spending, real-
locating the fiscal debt among budget items.
2. Previous studies have used the key informant technique to analyse the informal relationship
between the assembly and the bureaucracy in Chile (for example, Ferraro, 2008; Siavelis, 2002).
3. All the key informants interviewed were asked about the budgetary negotiations, their role and
goals in them, and their assessment of all of their counterparts in the negotiations. All the inter-
views were conducted without terms of confidentiality between 2004 and 2009.
4. For simplicity, in the rest of the study all the budgets analysed are referred to as pertaining to
ministries.
5. The reports were downloaded from the Library of the National Congress website (www.bcn.cl),
where it is possible to trace the history of the law. In this study I traced the history of laws 19.012,
19.103, 19.182, 19.259, 19.356, 19.430, 19.486, 19.540, 19.596, 19.651, 19.702, 19.774, 19.842,
19.915, 19.986, 20.083, 20.141, 20.232, 20.314 and 20.407.
6. Constitutional designers reinforced the executive to legitimise the concentration of powers that
characterised the Pinochet dictatorship.
7. Phone interview, 4 June 2009. Unfortunately, there is no public record that could allow making
cross-ministry comparisons over time to test this claim.
8. The rigid part of the Chilean budget is lower than its equivalent in Peru (95 per cent), Ecuador (92
per cent), Colombia (91 per cent) and Costa Rica (88 per cent), and higher than in Honduras (82
per cent), Bolivia (81 per cent) and Guatemala (66 per cent) (Cetra
´ngolo, Jime
´nez, & Ruiz del
Castillo, 2010).
9. Besides these reasons, members of Congress may be interested in having more influence in the
public budget to advance their ideological preferences and to enhance their role as members of
an institution that oversees the executive.
10. The amounts are estimated using the January 2012 prices as a reference.
11. Personal interview, Santiago, Chile, 15 May 2006.
12. Personal interview, Santiago, Chile, 18 May 2006.
13. In case of controversy between both chambers, a new special committee has to unite criteria to
approve the public budget, but so far it has never occurred.
BUDGETARY NEGOTIATIONS: CHILEAN CONGRESS 17
Downloaded by [University Of Pittsburgh] at 19:57 30 August 2014
14. The amendments are approved by a majority of the present members in both the subcommittees
and the SBJC.
15. The government does not have incentives to anticipate congressional preferences in its bill
because it can amend it during the negotiations. Furthermore, the government knows that each
year the SBJC does not review and amend all ministries’ budgets owing to time constraints,
and therefore cannot anticipate which ministries the SBJC will amend.
16. Personal interview, Santiago, Chile, 18 March 2006.
17. Personal interview, Santiago, Chile, 11 August 2005.
18. Personal interview with Rube
´n Catala
´n, technical member of the staff of the SBJC. Valparaı
´so,
Chile, 13 April 2006. Ominami also emphasised this point.
19. Personal interview, Santiago, Chile, 17 January 2006.
20. The executive can still block this tool supplementing expenditures through the Provision for
Committed Financing established in the Financial Administration Organic Decree Law No.
18,899 (1975), although it rarely does.
21. Personal interview, Santiago, Chile, 17 January 2006.
22. The agencies are the ministries’ administrative subunits. In Chile, they are called servicios.
23. The dependent variable in the second stage model was regressed on the independent variables of
the first model and they were not correlated.
24. The test discriminates between both estimators examining whether the unobserved heterogeneity
exists and should be handled as a random variable averaged across all icross-section units, or
instead be modelled explicitly by the inclusion of dummies for each cross-sectional unit save
one (N21).
25. The null hypothesis is that there is homoskedasticity, and since (Prob .chi
2
¼0.000), the test
suggests that robust standard errors are necessary to account for heteroskedasticity.
26. By contrast, formal institutions ‘are created, communicated, and enforced through channels that
are widely accepted as official’ (Helmke & Levitsky, 2006, p. 5).
27. The Senate Rules (available at www.senado.cl) do not differentiate specifications from the
marginal notes.
References
Alesina, A., Hausmann, R., Hommes, R., & Stein, E. (1999). Budget institutions and fiscal perform-
ance in Latin America. Journal of Development Economics, 59(2), 253–273.
Aninat, C., Londregan, J. B., Navia, P., & Vial, J. (2006). Political institutions, policymaking
processes, and policy outcomes in Chile (Latin American Research Network, Working Paper
R-521). Washington, DC: Inter-American Development Bank.
Baldez, L., & Carey, J. M. (2001). Budget procedure and fiscal restraint in post-transition Chile. In
S. Haggard & M. D. McCubbins (Eds.), Presidents, parliaments, and policy (pp. 105–148). Cam-
bridge: Cambridge University Press.
Blo
¨ndal, J. R., & Curristine, T. (2004). Budgeting in Chile. OECD Journal of Budgeting, 4(2), 7– 45.
Bueno de Mesquita, B., Smith, A., Siverson, R. M., & Morrow, J. D. (2004). The logic of political
survival. Cambridge, MA: The MIT Press.
Canes-Wrone, B. (2006). Who leads whom? Presidents, policy, and the public. Chicago, IL: Univer-
sity of Chicago Press.
Carey, J. M., & Siavelis, P. (2006). Election insurance and coalition survival: Formal and infor-
mal institutions in Chile. In G. Helmke & S. Levitsky (Eds.), Informal institutions and democ-
racy: Lessons from Latin America (pp. 160 –177). Baltimore, MD: Johns Hopkins University
Press.
Cetra
´ngolo, O., Jime
´nez, J. P., & Ruiz del Castillo, R. (2010). Rigidities and fiscal space in Latin
America: A comparative case study. Macroeconomics of Development series, 97, 1– 33.
Ferraro, A. (2008). Friends in high places: Congressional influence on the bureaucracy in Chile. Latin
American Politics and Society, 50(2), 101–129.
Financial Administration Organic Decree Law No. 18,899. (1975). Santiago, Chile.
Haggard, S., & McCubbins, M. D. (2000). Presidents, parliaments, and policy. New York, NY:
Cambridge University Press.
Hallerberg, M., Scartascini, C., & Stein, E. (2009). Who decides the budget? A political economy
analysis of the budget process in Latin America. Cambridge, MA: Harvard University Press.
18 THE JOURNAL OF LEGISLATIVE STUDIES
Downloaded by [University Of Pittsburgh] at 19:57 30 August 2014
Heckman, J. J. (1976). The common structure of statistical models of truncation, sample selection and
limited dependent variables and a simple estimator for such models. Annals of Economic and
Social Measurement, 5(4), 475– 492.
Helmke, G., & Levitsky, S. (Eds.). (2006). Informal institutions and democracy: Lessons from Latin
America. Baltimore, MD: Johns Hopkins University Press.
Kennedy, P. (2003). A guide to econometrics (5th ed.). Cambridge, MA: The MIT Press.
Kiewiet, D. R., & McCubbins, M. D. (1991). The logic of delegation: Congressional parties and the
appropriations process. Chicago, IL: University of Chicago Press.
Library of the National Congress. (2004). History of Law No. 19.986. Retrieved from www.bcn.cl.
Mideplan. (2005). Gasto pu
´blico social regional y paı
´s perı
´odo 1990– 2004. [Regional social public
expenditure and country 1990–2004]. Santiago: Divisio
´n de Planificacio
´n Regional.
Nacif, B., & Morgenstern, S. (2002). Legislative politics in Latin America. Cambridge: Cambridge
University Press.
Navia, P. (2008). Legislative candidate selection in Chile. In P. Siavelis & S. Morgenstern (Eds.),
Pathways to power: Political recruitment and candidate selection in Latin America
(pp. 92–118). University Park: Pennsylvania State University Press.
Pelizzo, R., Stapenhurst, R., & Olson, D. (2005). The role of parliaments in the budget process.
Washington, DC: The World Bank Institute.
Poterba, J. M., & Von Hagen, J. (1999). Fiscal institutions and fiscal performance. Chicago, IL:
University of Chicago Press.
Samuels, D. J., & Shugart, M. S. (2003). Presidentialism, elections and representation. Journal of
Theoretical Politics, 15(1), 33– 60.
Scartascini, C., Stein, E., & Tommasi, M. (2010). How democracy works: Political institutions,
actors, and arenas in Latin American policymaking. Cambridge: Inter-American Development
Bank.
Shugart, M. S., & Carey, J. M. (1992). Presidents and assemblies. Cambridge: Cambridge University
Press.
Shugart, M. S., & Haggard, S. (2001). Institutions and public policy in presidential systems. In
S. Haggard & M. D. McCubbins (Eds.), Presidents, parliaments, and policy (pp. 64–104). Cam-
bridge: Cambridge University Press.
Siavelis, P. (2002). Exaggerated presidentialism and moderate presidents: Executive–legislative
relations in Chile. In S. Morgenstern & B. Nacif (Eds.), Legislative politics in Latin America
(pp. 79–113). New York, NY: Cambridge University Press.
Siavelis, P. (2006). Accommodating informal institutions and Chilean democracy. In G. Helmke &
S. Levitsky (Eds.), Informal institutions and democracy: Lessons from Latin America (pp. 33– 55).
Baltimore, MD: Johns Hopkins University Press.
Toro, S. (2007). Conducta legislativa ante las iniciativas del Ejecutivo: Unidad de los bloques polı
´ti-
cos en Chile. Revista de Ciencia Polı
´tica, 27(1), 23– 41.
Tsebelis, G., & Alema
´n, E. (2005). Presidential conditional agenda setting in Latin America. World
Politics, 57(3), 396– 420.
Vial, J. (2001). Institucionalidad y desempen
˜o fiscal: Una mirada a la experiencia de Chile en los 90
[Institutionality and fiscal performance: A look at the experience of Chile in the 90 s]. Serie Estu-
dios Socio/Econo
´micos, 5, 1– 39.
BUDGETARY NEGOTIATIONS: CHILEAN CONGRESS 19
Downloaded by [University Of Pittsburgh] at 19:57 30 August 2014
