The (unexerted) competencies of municipal legislative ﬁ nancial committees in Brazil

Th is article analyses the legislative committee carrying out fi nancial and budgetary oversight at municipalities in Brazil. These committees, similar to the internationally named Public Account Committees, support parliamentarians to analyze, evaluate, and scrutinize budgetary bills, amendments and governor’s accounts covering the entire budgetary cycle. Based on a small-N analysis inductive approach, the study compares budgetary legislative committees from three local parliaments for the term 2013-2016. Such budgetary committees just operate ceremonially and passively, off ering superfi cial recommendations. Th e evidence indicates that the legal mandate to operate and material and human resources are not suffi cient to guarantee the commissions’ effi ciency. Th e collateral undesirable eff ects from coalitional government create incentives to maintain ceremonial commissions. Th e fragile legislative committee’s performance on budgetary issues is a bottleneck to any aspiration to achieve fi scal sustainability in Brazil.


INTRODUCTION
This study focuses on the legislative committee responsible for financial and budgetary oversight at a local government level -the named Public Account Committees (PAC) -as a critical feature for fiscal balance and budgeting oversight in municipalities. Such committees have an advisory competence, supporting councils' legislative activities. Arguably, they can strongly influence the legislative process through their opinions and guidance (Bastos, 1978;Rocha & Barbosa, 2000). However, the obligation to set up such committees does not guarantee its effective performance. Inoperative committees could end up legitimizing the matters under their scrutiny.
In Brazilian legislation, the obligation to set up the PAC of the local legislative branch follows the principle applied to the legislative at the federal level according to the Federal Constitution of 1988. These committees would analyze the budget proposal and its execution by the city government, verifying government accounts regarding the issues that generate economic and financial impacts on the municipal administration. For instance, in Brazil, the PAC monitors the public debt and the budget execution, observing the compliance to the Fiscal Responsibility Law.
The committee operates according to the local regulation proposed by each city council. Its operation ranges from a superficial analysis on the legal compliance and the minimal coherence of the information of the financial statements, up to complex discussions on the fiscal outcomes from projects and government strategies proposed to the city council approval.
Such a lack of technical skills seems to be one of the main reasons for committees' failure, as the councilors themselves compose the committee when appointed by the city council. Those councilors generally have a limited accounting and financial knowledge or background to deliberate on specific matters. In addition to the skills for proper budget oversight and monitoring, there are other barriers from the political system to the committees' effective functioning (Santiso, 2005). The political interference includes the electoral dynamics, the usage of decrees by the executive branch to govern without the legislative appreciation, and the need to construct and maintain coalitions to govern (Santiso, 2005).
Although studies on the Brazilian legislative branch highlight the multiparty system and the dynamics of the coalition government (Pereira & Melo, 2012;Pereira & Mueller, 2000;Praça, Freitas, & Hoepers, 2011;Raile, Pereira, & Power, 2011;Zorzal & Carlos, 2017), in general such literature does not cover the role of the legislature at local level (Castro;Anastasia & Nunes, 2009). More recently, even though a few studies focus on the city council role (e.g., Lopez & Almeida, 2017;Sartorelli & Pires, 2012), still being scarce the attention dedicated to the budgetary surveillance by municipal PACs in Brazil. This paper goes beyond the committees' skills and capabilities as the central barrier to the effective operation of committees. Although such skills are relevant, they are not sufficient conditions for the committee's successful performance. Our findings suggest the dynamics arising from the coalition presidentialism at the municipalities incentive the current coalition to capture the PAC, despite the accounting and financial knowledge carried by its members.
The methodology comprises the comparison of PACs from 3 municipalities covering different organizational contexts for the 2013-2016 term. We conducted interviews with members of these committees and their advisers, triangulated with documents (city council and committees' regulation, minutes of meetings and internal reports and referrals, when available). This study framework proposes a 4-dimensional analysis of the committees, concerning: (i) their composition (and members capabilities), (ii) how it fits into the legislative process, (iii) the content of the committee's referral, and (iv) the transparency on performing their role. The analyses indicate committees' passive acting and, consequently, a weak performance. Such passivity would come not only from the members' lack of technical skills but also to the reduced PAC's activeness due to the coalition dynamics. Lastly, we discuss the potential passivity regarding monitoring and oversight functions.

Legislative budgetary institutions
Budgetary institutions are the formal rules that structure the budgetary and fiscal dynamics in a country. The literature has largely debated the design of these budget institutions, either at the central government level in presidential countries (e.g., Gabris & Nelson, 2013;Santiso, 2005), parliamentary countries (eg, Coulson, 2011;Coulson & Whiteman, 2012;Dacombe, 2011;Sandford, 2006;Wehner, 2003Wehner, , 2006Wehner, , 2007 or at the level of subnational governments (eg, Coulson 2011;Coulson & Whiteman, 2012;Dacombe, 2011). These institutions define the role and when (timing) each actor, for instance, executive and legislative branches, play their roles over budget dynamics. They cover, for example, topics as the resource allocation, public policies evaluation or control, and accountability (Kim, 2015). Besides, such institutions establish the competence of legislative financial committees. In Brazil, such institutions are delineated mainly by the Finance Law no. 4,320 (1964), the Federal Constitution (1988), and the Fiscal Responsibility Law.
The interplay between the executive and legislative branches in the budget cycle varies across countries (Santiso, 2005;Wehner, 2003Wehner, , 2006. For example, it varies according to their governance model (Schaap, Daemen, & Ringeling, 2009). Regarding local elections, in some countries, the citizens directly elect the council members, which following indicates the chief-executive-officer for the city. Alternatively, in other countries, the citizens elect the members for the local legislature (i.e., councilors) and also the head of the executive branch (i.e., mayor). The PAC literature is prominent for the first governance model, commonly adopted by Commonwealth countries.
For the second governance model, named of strong mayor (Mouritzen & Svara, 2002), the literature over legislative committees is scarce. Recently the United Kingdom and other European countries have adopted the strong mayor model (Wollmann, 2014). In a nutshell, the mayor elected by the citizens is the head of the local administration and proposes the annual budget to be approved by the legislature. The legislature monitors the budget execution and subsequently approves the mayor's accounts. The few studies on PAC under the strong mayor governance model are precisely from the United Kingdom setting, where recently, the Local Government Act in 2000 transferred the mandate to define and implement public policies from the local councils to a mayor elected by the citizens. From then on, the predominant role of local councilors and budget committees became the so-called overview and scrutiny.
Following the Local Government Act, several studies focused on how to transform the previous legislative commissions' role to the strong mayor model (e.g., Coulson, 2011;Coulson & Whiteman, 2012;Dacombe, 2011;Sandford, 2006). For instance, McGee (2002) presents one of the first committee feature-comparison study. Later, a research committee of the Commonwealth Parliamentary Association (www.cpahq.org) produced the report "Parliamentary Financial Scrutiny: The Role of Public Accounts Committees", with a series of recommendations for legislative bodies (Von Trapp, 2010), mainly on the composition, mandates, and frequency of committee meetings. Hoque (2015) summarized the recommendations: the committees should be relatively small, do not require extensive material resources, full-term members with frequent meetings, chaired by opponents of the government, and that their members have more experience than average councils. Finally, the PAC would hold public hearings and annually meet with the Auditor General to evaluate the audit report received, so only afterward formalize final substantive reports to the parliament.

The effective performance of legislative committees
The mere enactment of regulation would not imply the adequate performance of the legislature, at any government level. Evidence from Uganda and Cambodia, both presidential developing countries, suggests that ensuring budget regulation is not enough to prevent the budget process from being dominated by the executive branch (Kim, 2015). Moreover, although the legislature performance (including its committees) has arguably improved in general terms through Latin America, the developments have a slow pace (Santiso, 2015). One of the recurring reasons for the committee's failure to act is the lack of available counselors' capabilities to carry out budget oversight. Such a context is similar for central governments from Brazil, Chile, Colombia, and Peru (Santiso, 2005).
In Latin America, the effects of the political system for presidential countries frame the executivelegislative dynamics (Santiso, 2005). Some factors from this presidential system affecting the executivelegislative dynamics are the use of the decrees, electoral incentives, and multiparty system (Santiso, 2005). The legislature is supposed to act as a counterbalance of the executive, exercising overview and scrutiny in the budget process, supported by its committees. However, the PAC active proceeding depends on whether the legislature eventually opposes the government. If the government coalition has support from a majority of members in the city council, the propensity in monitoring the executive is weakened (Santiso, 2005).
The coalition presidentialism literature discusses the coalition government side effects in Brazil, affecting the executive-legislative dynamics (Pereira & Melo, 2012;Pereira & Mueller, 2000;Praça et al., 2011;Raile et al., 2011;Zorzal & Carlos, 2017). The evidence shows an "extreme preponderance of the executive" (Pereira & Mueller, 2000), with greater control over the budgetary process. The executive (president, governor, or mayor) uses the release of funds for legislative amendments, as well as secretarial and ministry positions as part of the bargain for legislative support (Pereira & Melo, 2012;Praça et al., 2011;Raile et al., 2011). Consequently, the government can influence the legislative agenda and the composition of committees. Regarding the local government, given that the mayor assigns projects to committees and directs legislative deliberation in plenary (Pereira & Mueller, 2000), one can expect that more the alignment executive-legislature, less the opportunity for PAC members monitor and oversight will be. Zorzal and Carlos (2017) provide an example of how the executive influenced the State PAC of Espírito Santo (Brazilian State on the Southeast region) in the process of analysis of the Budgetary Guidelines. In that case, parties from the mayors' coalition captured the process of indication of members to PAC. Moreover, Sartorelli and Pires (2012) observed that committees stopped questioning the indiscriminate approval of additional credits by the executive branch. For the authors, this is mainly due to the restricted access to the city hall data and due to vagueness in mechanisms enacted by the internal regulation and finally due to lack of resources (material, human, and technological).

METHODOLOGY
The study aims to analyze the potential reasons for the eventual passivity of PACs and relies on 3 case studies (covering municipalities of Ribeirão Preto / SP, Araraquara / SP, and Uberaba / MG) to compare the effect of dependence on resources (human and material) and political incentives. The small sample analysis follows Wehner's (2003Wehner's ( , 2006 recommendations. In such a small sample design, the logic of inference searches for how the variation of case characteristics relates to the primary condition observed (Mahoney, 2000).
The cases were theoretically selected (Eisenhardt, 2007;Graebner, 2007) following an inductive grounded and emergent theory sampling strategy (Patton, 1990;Rapley, 2014). Therefore, the emerging theory guides the selection of cases to lead to an ever-deeper analysis for the most significant possible variability. The case selection looks for three theoretical drivers: (i) the available material and human resources, (ii) the city council effort to oversight and scrutiny, and (iii) the political context at the municipality. Additionally, we considered 3 other essential dimensions: (i) mitigate the size effect of the municipality, in order to avoid influence of urban dynamics or relationships with other spheres of government (for example, we did not consider state capitals), (ii) analyze city councils where the PAC was minimally operative and formalized, so it would be possible to observe the dynamics and quality of the PAC acting, (iii) accessibility for face-to-face interviews.
Notice that the present analysis does not search for generalization. Moreover, the use of cases combined with the inductive/abductive approach (Mahoney, 2000;Reichertz, 2013) allows further analysis of the phenomenon. We are confident about the transferability of context (Rapley, 2014), as there is a regulatory similarity of the settings in which the Brazilian municipal councils operate (e.g., national regulation about the role of councilors, electoral, and budgetary legislation). Therefore, the present qualitative analysis opens opportunities to extend the present analysis to other Brazilian municipalities in future studies. Table 1 shows the central characteristics of the selected cases. Some variables were kept as nominal to reduce comparison subjectivity (Mahoney, 2000). The available material and human resources consider the number of councilors, advisors, and public servants (all city council employees). The required monitoring effort would be proxied by the size of the municipality, and consequently, by the complexity and volume of projects, contracts, departments to be monitored by the councilors.
The size of selected city councils varies enclosing a midsize Brazilian municipality (200,000 to 600,000 inhabitants). For the 2016 municipal elections, the number of councilors per municipality in Brazil ranged from 9 councilors (for 3,641 city councils) to about 50 (for São Paulo and Rio de Janeiro councils). The three selected city councils had elected from 14 to 22 councilors in 2012. For the 2016 legislature, Uberaba elected 14 councilors, and Ribeirão Preto 27 councilors, which represents about 445 similar Brazilian city councils (from 14 to 27 councilors). In Ribeirão Preto, the proportion of graduated councilors was 72% in 2013 (compared to 38% of the neighbor city Araraquara). As another example, the Olímpia city council (50 thousand inhabitants) analyzed by Sartorelli and Pires (2012) had ten councilors at the time In order to adequately capture the legal mandate for the PAC acting, first, we considered the internal regulations for each one of the three councils. The procedures enacted by internal regulation are not often applied by the committee. Therefore, we identified the PAC activity observing: (i) committee meeting agendas, minutes, news, and all material provided by the interviewees, (ii) transcribed interviews, and (iii) content of all referrals obtained from the city council websites or provided by interviewees.
We conducted fourteen interviews: nine councilors (members and chairmen of the three selected committees), two legislative advisors, and three senior public servants from the city council (the general controller, director of the legislative department, and chief financial officer). The interviews were mostly held at the city hall facilities. In some cases, respondents were reassessed for additional clarifications. The interviews were recorded after the voluntary consent and transcribed. Whenever the respondent hesitated to consent, the authors did not insist on requesting the recording.
The interview protocol included the following themes: (i) conditions for PAC active performance (available human and material resources), (ii) active performance on legislative control, including the issuing of opinions through formal referrals ; (iii) the level of transparency and usefulness of the committees according to the perception of other councilors, and (iv) any perceived barriers to action. Regarding PACs acting, we dedicated particular attention to the assessment of fiscal risks and the remedies to such risks mentioned in the Budgetary Guidelines Law, given the relevance for fiscal sustainability.
The referrals are the primary delivery of the committees. Such referrals would support the vote of the other councilors in the plenary. The committees would be the first specialized reviewer of the budgetary matters at the city council. Eventually, for some themes, the PAC referrals are built upon the 'preliminary opinion' (report on the mayor's accounts) prepared by the Court of Accounts. The relevance of the referral is highlighted in the 2005 Brazilian Senate "Councilors' Handbook" and in the "Comptroller General's Guidebook". The subjects the PAC must oversight a range of themes from the abstract ones -such as the recomposition of the actuarial balance of social security funds -to those closer to their ordinary daily routine, as the impact of launching new health care centers on current expenditure. Arguably, in order to vote, the councilor would need a qualified opinion, an understanding of the concepts, metrics and eventually prepare specific calculus considering data from City Hall financial statements.  The three selected city councils have different levels of resources employed in favor of the committees. While the Uberaba city council employed the entire annual legislative grant and provides 13.7 advisors per councilor, with a payroll 1.67 times that of Araraquara (considering salaries of advisors and civil servants, except salaries of councilors), the other two councils payback a residual of their annual legislative grant. Although Ribeirão Preto has more extensive administrative staff and higher salaries, it does not imply PAC effectiveness.
It is also observed that councilors in Uberaba and Ribeirão Preto allowed an overestimation of revenue in the bill of the annual budgetary law, from 12% to 66% in the period. This overestimation could be a result of an inability to anticipate current revenues, but such an explanation would fit in the 2014 fiscal year, as the Brazilian political and financial crisis deepened that year. However, it is less justifiable for 2015 and 2016 (given that the crisis was arguably already underway). Notice that any over or underestimation of revenue distorts the annual budget sent to the legislature (Azevedo & Aquino, 2016), and should be identified and notified by the local PAC. Even if the PAC cannot avoid overestimation (in the event of a drop in ex-post revenue generated by external factors), the committees can point it out and analyze its effects on the budget.
For the three cases, the 2013-2016 mandate was characterized by the city councils' support for the elected mayor. Extreme cases of influence by the mayors' coalition would mean interference in the legislative process. In such a scenario, the councilors from parties in opposition would face reduced incentives for pointing out problems in budget execution and other misleading acts of the mayor. Such problems would be politically barred by the council's presidency not selecting the issue for voting or by the vote of the majority of council members supporting the mayor. This context prevailed in the Ribeirão Preto city council for the entire 2013-2016 term, until a local corruption scandal involving the mayor, lawyers, outsourced firms and city counselors, named Operation Sevandija, dismissed 9 (out of 22) city councilors in September 2016, and also arrested the Mayor and then the head of city council in December 2016.

RESULTS
4.1. The legislative committees operating conditions, as regulated Table 2 presents information on the city council's PAC internal regulation. The regulation in general covers: (i) rules to select members, (ii) how the PAC interacts within the legislative process, (iii) the expected content of the PAC referrals, and (iv) the transparency of PAC procedures. Following the recommendations by Commonwealth Parliamentary Association in "The Role of Public Account Committees", and also accordingly to Hoque (2015), we identified the frequency of PAC meetings, whether the society was welcome to attend the meeting, and whether there was any provision for the counselors from opposition to the Mayor chair the committee. Besides, we identified the requirement to issue substantial reports regularly and immediately publicize its referral about the 'preliminary opinion' (report on the mayor's accounts) prepared by the Court of Accounts. As the table shows, the recommendations vary for the cases.
The PAC's legal mandate to act is formally defined by the internal regulation of each city council. These regulations, updated in 2014 and 2015, are similar regarding their frame but differs on specific remedies and preventions. Therefore, they arguably reflect a renewed understanding by the local legislature on how such commissions should operate, amplifying, or reducing the PAC mandate. The internal regulations define how and how many members compose the committee, for instance, the members should represent the parties in parliament, and their term coincides with the mayor and parliament mandate. Nevertheless, none of the cases regulates the appointment for the presidency of the committee, neglecting the Commonwealth Parliamentary Association recommendation for a senior counselor from an opposition party to chair the committee. Such silence leaves the committee presidency designation free to the mayor's coalition to capture the board.
The three councils' internal regulations set the deadline for the committee to appoint the rapporteurs and to issue referrals for each item delegated to the PAC analysis. The remedy intends to ensure that the PAC will not eventually delay the legislative process. However, Ribeirão Preto and Uberaba city councils do prevent the plenary voting taking place without the PAC referral, Araraquara does not. Ribeirão Preto council also prevents the ordinary PAC meeting concurs to the plenary section, therefore favoring society attends both meetings; additionally, its internal regulation defines a frequency of PAC meetings, inducing a committee institutionalization.
The PAC referral is mandatory in Ribeirão Preto and Uberaba, but any of the cases the expected minimum content for the referral is appropriately defined. Comparing to the capitals of São Paulo state and Minas Gerais state, the regulation of São Paulo council (revised in 2013) and Belo Horizonte council (revised in 1996) dedicated an entire section to PAC referral requirements. The referrals are defined as "the pronunciation of the committee, based on their judgment, on a proposal subject to its examination"(art. 85, Internal Regulation of the City Council of Belo Horizonte), or "the official pronouncement of the committee on any matter under its scrutiny" (art. 74, Internal Regulation of the City Council of São Paulo). Both capitals express what the report should include, for instance, the subject under examination and the rapporteur's conclusion by expressing his opinion on "the desirability of approving or rejecting the matter in total or partially" or adding recommendations or eventually amending the proposal. Finally, the referral must bring the member's signature supporting or rejecting the rapporteur position.
The city councils' regulation regarding the transparency of committees' activities -ideally -would include publicizing the schedule of the committee's meetings in order to engage civil society and the minutes recording deliberations and the referrals. Such remedies would increase social accountability over the PAC recommendation and on the votes of each committee member. As observed, only Ribeirão Preto council presents a minimal level of transparency regarding the PAC activities. Therefore, if Ribeirão Preto council regulation defines a broader mandate to its PAC in comparison to the other two councils, such mandate may be blocked by the mayors' coalition or not used in practice. Notes: The table describes the regulation, but not necessarily the routines and processes reflect such regulation. "Yes" in cells means the item is explicitly mentioned by the regulation, and "-" means the opposite, or that nothing is mentioned about the item. (1) In Ribeirão Preto city, any counselor can require (formally written and justified) a PAC audience if it was not pushed to offer an opinion on a specific issue. However, in the same city, the regulation defines that the referral may be not required for urgent cases undersigned by the majority of counselors. In Uberaba, If the PAC does not issue the referral, another committee constituted by the presidency will issue a referral in a six days period. In Araraquara, the internal regulation states: "the issue, with or without PAC referral, will be automatically listed on plenary voting agenda".

The legislative committees operating conditions, as observed
The internal regulation for the three selected councils is not sufficient to guarantee autonomy and effective acting for their PACs. The committees do not hold their own programmatic agenda; they issue superficial referrals and present low transparency and timid engagement of the society (Table 3).  and 2016Araraquara = 50,85 / 52,45 / 51,20 / 48,40;and Uberaba = 37,19 / 38,97 / 38,94 / 36,82. (2) The preliminary opinions by the Courts of Accounts, about the mayors' annual accounts up to 2013, were ready to be analyzed by the legislative branch in 2015. In 2015, Ribeirão Preto city council appreciated the Courts of Accounts' preliminary opinion up to 2010; three years were not considered yet (2011,2012,2013).

Agenda autonomy and the legislative process
The committee may be urged to act on the legislative process providing referrals on a specific issue. Upon receiving a demand, the committee chairperson nominates the rapporteur to analyze and prepare the arguments (pros and cons) to support the vote by committee members. The referral following goes to the presidency of the city council, who may circulate the referral to councilors as soon as s/he includes the issue in plenary voting agenda. Secondly, the committees can also act proactively on budgetary analysis and oversight. For example, the PAC could appreciate the warnings, audits, and recommendations produced by the Court of Accounts regarding the performance and actions of the executive branch. Such warnings of non-regularities about the mayor's acts and accounts may foreshadow the Courts of Accounts' preliminary opinion on the mayor's annual accounts (Verbruggen, Christiaens, & Milis, 2010).
The evidence suggests the three passive committees, acting just when demanded. The committees have low autonomy as their work agenda is pushed by the legislative process (content and deadlines) and by the presidency agenda. The councils of Uberaba and Araraquara vote budgetary and financial issues at the plenary sections without a referral issued by their PACs. Those PACs do not even participate in the legislative process. Ribeirão Preto committee always issues a referral; however, following an "urgent regime" to be listed to the next plenary agenda preventing an extensive debate on the issue by the PAC. Moreover, in Ribeirão Preto the rapporteur and the content of the PAC opinion issued by the referral are controlled by the mayor, according to the interviewees. One of the PAC presidents exposed the political barriers to action: "Once there was even an attempt on the part of the councilors to intervene in the executive's activity, exerting greater supervision, but the structure is too stagnated, and we suffered strong pressure, finally without providing effective supervision" (PAC president).
The committees also passively act concerning the appreciation of amendments on the budgetary law, as the PACs start to consider the amendment proposed by the executive branch only when the proposals reach the city council. The committees did not apply a ceiling on budgetary changes (relocation between governmental functions or programs). Such passive behavior is similar for Araraquara and Ribeirão Preto, where the city hall exceeded the ceiling on personnel expenses.
Finally, the city councils do not meet the legal deadline (60 to 120 days) for providing referrals on the preliminary opinion by the Courts of Accounts about the Mayors' accounts. According to the Secretary-General of the State Court of Accounts of São Paulo, such delay is recurring. For instance, from 2002 to 2010 (9 annual accounts), several municipalities from São Paulo state accumulated annual accounts without proper referrals. In the three analyzed cases, the delay is five years for Araraquara and Ribeirão Preto as they did not prepare referrals for 2010 Mayor's accounts and four years for Uberaba, as it missed referral for 2011 mayor' accounts.

Human and material resources
According to Santiso (2005), the consultants and specialized support services provide technical skills to councils. Such capabilities could be achieved through hiring expert advisors to convert part of the available financial resources into specialized support teams. No technical competence is required for parliamentary candidacy; therefore, there are few councilors with a background in accounting or related disciplines. It was observed in all three cases. Moreover, there is a lack of advice from experts and an overload of activities. For example, in Uberaba, the 14 councilors split positions into 19 standing committees. Therefore it is not expected technical referrals sufficiently detailed and complete. According to one of the interviewed advisers, the councilors " […]do not receive any guidance or understanding from party directorates about municipal accounting or finance practices" (Councilor's advisor).
The combined effect of task overload and lack of training highlights the barriers to PAC acting at the smaller councils.
City councils in which financial resources are not a constraint could allocate some amount of their funds for hiring specialized advisors. Such support would alleviate the negative impacts of work overload and councilors' lack of technical skills. However, these actions were observed in neither of the three cases. Although the legislative branch distributes funds to councilors for hiring advisors, advisors specialized in finance and accounting are unusual. Moreover, committee members do not perceive the lack of skills as a critical issue, which suggests a restricted perception of the role of the PAC.
In the three cases, the PAC does not get facilities and equipment -e.g., office rooms, administrative staff, and dedicated computational devices (see Table 3). There are no institutionalized routines and processes. Besides, they do not accumulate data or previous experiences carried out by the PAC and the city council. Consequently, there is no organizational history, operating manuals or committees' standard procedures. Therefore, the three PACs are quite similar to Olímpia City Council, analyzed by Sartorelli and Pires (2012).
In such a context, the PAC organizational learning is limited. For instance, the PACs do not register minutes or activities consistently, or just superficially. Therefore, it lacks proper intellectual capital systematization. The knowledge expertise is symbolically represented by the "seasoned councilman" who chairs or serves for a long time on the committee. The individual members' identity blurs the identity of the committee, as the PAC identity is not disentangled from its members, mainly the seasoned councilman. The PAC dynamic eventually adopts the frequency, routines, activity, and depth of analysis imposed by the current PAC presidency or by the most experienced members. Sometimes it does not even meet the minimum requirements of the regulation.
In addition to reduced capacity for overview and scrutiny, such a context of weak specialization can also negatively impact the committees' legitimacy. As mentioned by members at the Araraquara PAC, the councilors do not adequately consider the opinion of the committee as "the committee members are not technical in the area, [so] the councilor does not consider the opinion of the committee" (PAC member).

Referrals' content
The referrals -when issued by the analyzed PAC-only superficially analyze the expected fiscal hot themes. The primary referral about the Budgetary Guidelines proposal by the executive used to be issued looking to accomplish a formal requirement. An interviewed counselor states that the referral "is organized just to comply with the legislation. " Another councilor justifies that "the executive branch proposes the budget bill [Budgeatary Guidelines] without opportunities to suggest changes. " In the same municipality, another councilor states that "the committee has no prerogative to propose any sort of referrals for the City council". The evidence fits Dacombe (2011).
The referrals do not address relevant issues, such as: (i) the deficit carry-over, (ii) financial surplus or deficit, , (iii) adequate use of ring-fenced resources, (iv) use of extraordinary non recurrent revenues to pay operational and compulsory expenses, (v) the use of additional credits, (vi) long-term debt, (vii) the compliance to the 'golden rule' (art. 12 of the Budgetary Guidelines Law and art. 167, item III of the Federal Constitution) at the Budgetary Guidelines, (viii) the social security sustainability. For example, in Uberaba and Ribeirão Preto, PACs did not proactively comment on the overestimation of revenue when in the process of analysis of the Budgetary Guidelines. Consequently, there was no debate regarding the Budgetary Guidelines. One interviewed councilor configures the compliance orientation, referrals are formally issued, but "only to 'promote legality'[as it is mandatory to the process]". Another PAC counselor mentions that "the committee conducts a superficial analysis of the public accounts, to afterward propose the public hearings".
The absence of such analysis creates a severe concern for municipalities, for instance, the increasing of the deficit carry-over, reducing transparency, and pushing a "parallel budget" without legislative scrutiny as the City Hall overuses unverified unpaid commitments (see Aquino & Azevedo, 2017).

Transparency and active participation
PACs' transparency is lower for the three cases. Despite mandatory, in Ribeirão Preto, the schedule for PAC's meetings and the associated minutes are not always disclosed. In Araraquara, the minutes only superficially mentioned the main subject of the meeting and the members who attended. Although it does not publicize PAC referrals, one can access them upon request. However, the lack of transparency reduces the chance that citizens appropriately attend public hearings. In sum, we observed an interviewee's consensus on public hearings not promoting citizen participation and not assisting legislative control. In Ribeirão Preto, for instance, an interviewee at the City council was not aware of the formal requirement for PAC meetings to be open to citizens.

Councilors' attention to fiscal sustainability
The committees are short-term oriented, particularly when they are urged to express an opinion on the fiscal balance (current revenues and expenses). There is no focus on long-term aspects; our interviewees converge on the perception that short-term obligations and the yearly fiscal balance get the full city councils' attention.
In general, counselors and advisors are not sensitive to the municipal debt. For instance, the pension funds for the public servants in Uberaba and Ribeirão Preto operate bellow the radar, being considered to debates only if contested in public complaints. In Ribeirão Preto, in 2015, the City Hall was accused of illegally withdrawing financial reserves from the pension fund in 2013, implicating the cancelation of its certificate by the federal government until the payback of reserves. Such irregularity was identified two years later its occurrence -and not by the City council, but by the federal secretary of Social Security. In Uberaba, the pension fund is not debated by PAC members or by councilors in general.
The Courts of Accounts frequently issue warnings about fiscal irregularities, for instance, on the minimal expenses for education and health. However, such warnings do not pose PAC more attentive in Ribeirão Preto. From 2009 to 2015, Ribeirão Preto accumulated only two positive preliminary reports from the Court of Accounts on the mayors' annual accounts (2009,2011), but recurrently accumulate warnings about fiscal irregularities.

Mayor's coalition preponderance
The PAC activity is ceremonial and weakly transparent for the three selected cases. One interviewed councilor describes the committee as "simply a formality, it serves only to comply with the regulation". Such inactivity is a consequence of lack of skills, resource misallocation (human and material) (Santiso, 2005;Sartorelli & Pires, 2012), and political incentives (Lopez & Almeida, 2017;Santiso, 2005).
The mayors' coalition surpasses the PAC oversight and scrutiny. The literature on coalition governments has shown that legislators may face reduced incentives for opposing to the executive branch (Pereira & Melo, 2012;Praça et al., 2011;Raile et al., 2011;Santiso, 2015;Zorzal & Carlos, 2017). Arguably, although councilors bargain amendments to the budget, once the executive branch accomplishes their demands, they promptly support the coalition agenda. Consequently, it reduces the propensity to enforce issues against the government's interests, as the city council presidency may demotivate PAC's contrary positioning. Finally, governments counting on a stable coalition, with broad support from the city council, the PAC may have limited power to exercise its scrutiny role, and it is not expected an open opposition. However, if the coalition is unstable, the PAC may be urged to change its positioning pro-government.
For all analyzed cases, the PACs did not count on specialized advisors. Neither in Ribeirão Preto, getting the best regulation among the three cases. Although the most substantial amount of resources allocated for councilors to hire advisors, there was no discussion on how to strengthen PAC or even initiatives to hire skillful advisors.
At Ribeirão Preto, why was the PAC unable to joint specialized resources to improve its oversight and scrutiny capacities? Several corruption scandals were widely reported by local and national media at the end of the 2013-2016 term in Ribeirao Preto. Our interviews in 2015 captured a context in which the opposition represented a minority in City council, despite the scandal had not leaked the local political crisis was boiling. The City council, according to an interviewed councilman, operated "as an extension of the executive", which considered the PAC "at the service of the mayor". Another councilor stated that the PAC "was used to legitimize the interests of the government", and continued arguing that the PAC "was captured, approving all the projects of the executive interest, at the required pace". This context lasted from 2013 to mid-2016, and all over this period, the City council presidency blocked the PACs' oversight. Nevertheless, from September 2016, the mayor lost support as a consequence of the scandals.
The City council presidency may employ three different measures aiming to intimidate PAC oversight. First, as previously mentioned, allocating smaller or restricted material resources to support staff. Concerning Ribeirão Preto, we observe two other forms based on political blockage exercised by the City council presidency: (i) controlling the voting dynamics, either selecting the matter that will be voted by the counselors at the plenary section or influencing the votes and (ii) interfering on PAC's composition and controlling its agenda.

Agenda control or vote orientation at the plenary section
These forms of control stem from the presidency of the city council acting at the plenary section. Concerning agenda control, the presidency may not address the preliminary report by the Court of Accounts on the Mayor's annual accounts, leaving it out of the spotlight. Consequently, as there is no deadline to vote such a particular matter, the PAC is not incentivized to give an opinion, even if its members are willing to. This topic may not be listed on the agenda. In a few cases, the presidency is required to fix the voting agenda within a specific time, such as the deadline to deliberate and vote the executive bill for the Multiyear Plan and the Budgetary Guidelines. Moreover, we observed an orchestrated effort at the Ribeirão Preto city hall to block the PAC referral, influencing the coalition vote. If, eventually, the referral opposes the coalition interests, the coalition members will vote against the referral. According to an interviewed councilor in Ribeirão Preto, "allied councilors are more susceptible to accepting the amendments required by the executive". Any request for more critical oversight is denied at the plenary section.

PAC composition and agenda and referral control
If aligned with the mayors' coalition, then the PAC president may alienate its referral to the coalition's interests. For instance, in the case of Ribeirão Preto, the Mayor's coalition appointed the PAC president, he also acted as the coalition leader at the City council. Two mechanisms were employed by the City Council presidency to control the PACs' referral impact. According to respondents: (i) the 'urgency regime" imposed by the City Council presidency gave PAC only one day to analyze a specific matter and issue its referral -therefore a full debate by the committee was unfeasible; and (ii) to allocate the case to himself (PAC president) or to other coalition member as be rapporteurs. Consequently, s/he precluded opposite opinions to the coalition's interests. As a result, it isolated any PAC member who represented the opposition and could generate some independent oversight.

CONCLUSIONS
Our analysis provides evidence that resources and formal regulation are not sufficient conditions for PACs' oversight and scrutiny. The executive-legislative interplay may become mutually protective under coalition governments. Therefore, PAC scrutiny on the executive branch will be timid. The mayors' coalition, aligned with the City council presidency, tends to block the committee's proactivity. In the case of Ribeirão Preto, besides the agenda control and the blockage of voting at the plenary section (which would nullify any questioning from the committee), two additional mechanisms were employed to capture the committee: the nomination of the committee chairman -who used his discretionary power to alienate the PAC referral -and the 'urgency regimes' to prevent accurate scrutiny by the committee.
The interviewed PAC members are short-term oriented with a narrow perception of committees' potential and competencies. Moreover, evidence corroborates with their expected lack of financial skills and the lack of technical support from the City Council. Consequently, the PAC is mainly inoperant regarding its primary functions of oversight and scrutiny. The remaining councilors are not exposed to a more qualified opinion on the social and economic impact of the fiscal or financial matter at hand. Thus, the councilor's decisions are biased by the political drivers and all sorts of interpretations and distortions about the discussed matters.
These findings are associated with middle-sized municipalities. However, our warnings can be even more intense in smaller municipalities, given there are fewer councilors and resources available to advisory teams. Moreover, other issues may arise, such as the daily interaction between councilors, mayors, and senior civil servants at local administration. Such a context broadly represents the small urban areas across the country. Therefore, Courts of Accounts could induce PAC activism, for instance, easing the engagement of councilors, enforcing simplified extracts of previous referrals, and offering online training.