Why budget accountability fails ? The elusive links between parliaments and audit agencies in the oversight of the budget

RESuMO: Os parlamentos e as agências de auditoria têm papéis fundamentais e complementares na fiscalização do orçamento e na execução de responsabilização do governo. No entanto, a relação entre os parlamentos e agências de auditoria é um dos elos mais fracos da cadeia de responsabilidades, gerando uma lacuna de responsabilização no processo orçamental. Este artigo analisa as interações entre os parlamentos e agências de auditoria na supervisão das finanças públicas durante os últimos estágios do processo orçamentário. usando proxies para avaliar a qualidade desses vínculos, tais como o acompanhamento às constatações da auditoria e a descarga de governo, ele mostra disfunções importantes nas interações entre os parlamentos e os organismos de controle devido a uma combinação de limitações de capacidade técnica e desincentivos de economia política. Ele sugere que o funcionamento eficaz do sistema de freios e contrapesos no orçamento público depende criticamente da agilidade das ligações entre as instituições de prestação de contas. Como tal, a falta de prestação de contas do orçamento é devida a disfunções sistêmicas no sistema, mais do que ao fracasso de uma instituição agindo isoladamente. Além disso, a eficácia da arquitetura de accountability horizontal depende dos incentivos de economia política que moldam o processo de orçamento, que são geradps pelas interações entre a escolha de regras de projeto e orçamento institucional, com o grau de competição política e regras eleitorais.

Brazilian Journal of Political Economy 35 (3), 2015 • pp.601-621 ABSTRACT: Parliaments and audit agencies have critical and complementary roles in the oversight of the budget and the enforcement of government accountability.Yet, the nexus between parliaments and audit agencies is one of the weakest links in the accountability chain, generating an accountability gap in the budget process.This articles analyses the interactions between parliaments and audit agencies in the oversight of government finances during the latter stages of the budget process.using proxies to evaluate the quality of those linkages, such as the follow-up to audit findings and the discharge of government, it shows important dysfunctions in the interactions between parliaments and audit agencies due to a combination of technical capacity constraints and political economy disincentives.It suggests that the effective functioning of the system of checks and balances in public budgeting critically hinges on the agility of the linkages between accountability institutions.As such, the failure of budget accountability is due to systemic dysfunctions in the systems of accountability, rather than the failure of an individual accountability institution acting in isolation.In addition, the effectiveness of the horizontal accountability architecture depends on the political economy incentives shaping the budget process, which are generated by the interactions between the choice of institutional design and budget rules, with the degree of political competition and electoral rules.KEYWORDS: public budgeting; budget process; political economy; accountability; parliaments; audit agencies; institutional design; electoral rules; political competition.JEl Classification: H61; H83; O54; P51.

MIND THE ACCOuNTABIlITY GAP IN THE BuDGET PROCESS
Constant experience shows us that every man invested with power is apt to abuse it [...] it is necessary from the very nature of things that power should be a check to power Charles de Montesquieu, Spirit of the Laws, 1748, XI,4.
The emerging literature on the political economy of the budget process has focused on the dynamics of executive-legislative relations in the early phases of the budget process.However, less attention has been paid to the institutions and incentives shaping these interactions in the latter stages of the budget process.In these latter stages, parliaments and audit agencies have a critical and complementary role in the scrutiny of government finances, the control of budget execution and the enforcement of government accountability (Wehner, 2006).Yet, the nexus between parliaments and audit agencies is one of the weakest links in the accountability chain of the budget process (Stapenhurst et al., 2008;Santiso, 2006Santiso, , 2008)).It is also one of the least studied by the literature on public financial management (Allen et al., 2013;Cangiano et al., 2013) and the political economy of the budget process (Wehner and De Renzio 2013a;Hallberg et al 2009).There is indeed an important gap between the potential and actual effectiveness of this link in the architecture of accountability institutions.This link is seldom as effective as it could and should be, which generates an "accountability gap" in the management of public finances in the downstream phase of the budget process (Santiso, 2009).
Parliaments and audit agencies are critical actors in the oversight of public spending and the ex-post control of budget execution.The role of parliaments in the budget process has been imbued in controversy but a consensus gradually emerged that anchoring fiscal responsibility requires restraining the budgetary powers of parliaments (Von Hagen, 1992;Stein et al., 2006;Stapenhurst, 2008).This is due because legislative budgeting is faced with an intractable collective action dilemma referred to as the "common pool" challenge that inhibits cooperation and inter-temporal bargains (Ostrom 1990;Hallerberg et al., 2009).Individual parliamentarians have an incentive to spend more and tax less for electoral purposes and they lack incentives to internalize budget constraints, worsening fiscal deficits.As Halleberg et al., (2009:299) note, "legislators therefore vote for more spending than they would if they considered the full tax effects of their spending".As a result, fiscal rules that strengthen the budget powers of the executive in its relationship with parliament are expected to lead to more fiscal restraint and better budget outcomes (Filc and Scartascini, 2007).Most countries have adopted centralized budget institutions and fiscal responsibility rules that provide for a stronger role of the executive in budget decision-making.They established hierarchical budgetary processes dominated by the ministries of finances so as to anchor fiscal discipline.
Nevertheless, excessive government discretion in public budgeting also carries risks.Centralized budgetary processes dominated by the ministries of finance are not, per se, a guarantee against budget opacity or fiscal imprudence.The recent public debt crisis in developed countries has led to a regain of interest in the role of parliaments in checking government and more rigorously overseeing fiscal policy and the budget process.Parliaments have sought to regain greater influence in the budget process (Schick, 2008;Wehner, 2006), not necessarily by increasing their budget powers but also by strengthening their technical capacity to exercise the budgetary powers they have (lienert, 2013;Santiso and Varea, 2013;Santiso, 2006Santiso, , 2004) ) 1 .There is also a renewed interest in strengthening independent fiscal councils and parliamentary committees to better check and oversee fiscal policies (Hemming and Joyce, 2013).
Similarly, the role of audit agencies in the oversight of public spending and the control of budget execution has received greater attention, as part of the global trend to strengthen accountability in the budget process.This stems from a dual concern for both increasing transparency in the management of public finances and improving the performance of public spending (Santiso, 2009).Independent audit agencies are indeed key actors in the architecture of the budget and the system of checks and balances.They are tasked with promoting transparency, efficiency, effectiveness and accountability in public administration.Their purpose is to oversee the management of public finances, ensure compliance with financial regulations, and hold govern-ment to account for the results achieved and the manner in which it manages public resources.Through their compliance audit, they certify the legality of public spending; through their financial audits, they certify the alignment of public spending with the budget approved by parliament; and through their performance audits, they assess value for money -efficiency, effectiveness and economy -of public spending, and more broadly, the performance of government in implementing public policies.
The literature on budget institutions has traditionally focused on analysis the impact of legislative oversight and external auditing on fiscal outcomes, with mixed findings however.More recently, with the emergence of results-based management and performance-based auditing, the focus has shifted to asserting their impact on government performance.Empirical evidence from latin America and the Caribbean in Figures 1 and 2 suggests that legislative budget oversight and audit institutions have a greater impact on government effectiveness than fiscal outcomes.Increasingly, audit agencies are moving from a focus on government compliance with financial legislation and budget appropriations, to an emphasis on value for money and government performance in managing public resources and implementing public policies.This shift entails moving from an oversight role to assuming more advisory functions so as to recommend practical improvements in the implementation of public policies.There is, indeed, an important "implementation gap" between public policies enacted by government and their effective implementation by the bureaucracy, which audit agencies can help bridge.Why, then, budget accountability fails?Why has increased oversight of the budget not led to greater accountability of government?This article assesses the quality of the links between parliaments and audit agencies in the oversight of public spending and the enforcement of government accountability.It gauges the critical factors enabling or hindering the effectiveness of those links and analyses how this relation should work, how it actually works, and how it could be improved.The article focuses on key processes as proxies to evaluate the quality of those linkages, in particular the discharge of government and the follow-up of audit findings.
Audit agencies are critical partners and advisers of parliaments in the oversight of the budget and the enforcement of accountability on government.Nevertheless, the linkages between them are not as effective as they could be due to a combination of technical capacity constraints and political economy disincentives.As a result, as the article reveals, there are important coordination failures in the architecture of financial scrutiny and budget oversight.We suggest that the effective functioning of the system of checks and balances in public budgeting depends inasmuch on the agility of the functional links between accountability institutions, as on the efficacy of individual accountability institutions acting in isolation.In turn, these dysfunctions are due to the interaction of institutional rules and individual incentives of key actors in the budget process.ultimately, it is argued, the ex-post control of budget execution is inherently political process, inasmuch as the other stages of the budget process.

PARlIAMENTS AND AuDIT AGENCIES IN THE EX-POST CONTROl OF THE BuDGET
The literature on the political economy of the budget process provides useful insights to conceptualize the role and functions of independent audit agencies as auxiliaries to parliaments in the oversight of government finances (Santiso, 2009).The literature and practice of external auditing underscore two critical factors affecting their effectiveness: their degree of independence (political independence) and the intensity of their accountability functions (enforcement powers).This independence requires the existence of an appropriate constitutional, statutory and legal framework guaranteeing the credibility and impartiality in the appointment of external auditors, security of tenure and legal immunity in the discharge of their duties, as well as financial, managerial and administrative autonomy of audit agencies, including unrestricted access to government information, and their autonomy in reporting and publicizing audit results.
There are two important controversies in the theory and practice of external auditing of government finances: (i) whether audit agencies are independent or autonomous with enforcement capabilities, and (ii) whether they are oversight agencies or accountability institutions with direct sanctioning powers (Santiso, 2009).Different approaches emphasize different definitions, degrees and features of independence in terms of political, constitutional, financial, and administrative independent.Audit agencies tend to emphasize the importance of their absolute independence.However, absolute independence is hard to attain and not necessarily advisable.The concept of "embedded autonomy" (Evans, 1995) drawn from the institutional literature is useful to conceptualize the place of audit agencies in the institutional architecture of the budget process.undoubtedly, audit agencies should be as autonomous as fiscal councils (Hemming and Joyce, 2013), tax authorities (Taliercio, 2004), and central banks (Boylan, 2001).However, they are "embedded" in the broader system of checks and balances and thus the question is not whether audit agencies ought to be totally independent, but rather how much independence is enough and how much independence is too much.Independence is a means for effectiveness, not an end in itself.From this perspective, the most critical considerations are those related to the guarantees for political impartiality, technical competence, and protection from government interference.
In the system of checks and balances in public budgeting, there is a symbiotic relation between parliaments and audit agencies, the latter often acting as advisers of the former.In the political economy literature, the institutional and functional linkages between audit agencies and parliaments are often couched in a principal-agent framework.Audit agencies are conceptualized as "agents" or "auxiliary institutions" of parliaments, assisting their "principals" in mitigating asymmetries in fiscal information.Their effectiveness is contingent on the quality of its linkages with its "principal".This why we refer to supreme audit institutions as autonomous audit agencies: they ought to be impartial, capable and competent, and are inserted in the broader system of checks and balances.If audit agencies are to be effective, they can rarely function in isolation.As O'Donnell (1999, p. 39) underscores: "effective horizontal accountability is not the product of isolated agencies but of networks of agencies.In particular, their ultimate effectiveness depends on decisions by courts and legislatures to enforce government accountability".From this perspective, there is or should be a symbiotic relation between audit agencies and parliaments.These two institutions are more effective acting in concert than acting in isolation and part of the "accountability gap" in public budgeting is due to coordination failures between them.
The second debate in the accountability literature relates to the core functions of audit agencies and the extent of their statutory powers to hold government to account.Accountability institutions have autonomous enforcement powers in their own right; while oversight agencies require a "principal" to act upon their recommendations and hold government to account.A distinct feature of an accountability institution is its ability to hold answerable and enforce sanctions on another state institution.In a democratic system of institutionalized checks and balances, only a state power can enforce accountability on another state power.The three powers of the state are the executive, the legislature and the judiciary.In this democratic triangle of powers, audit agencies need the relay of a power of the state to effectively restrain government and hold it to account.This is why audit agencies are most often institutionally embedded in either the judiciary or the legislative branches.Their main powers, contributions and functions center on the oversight of government, on behalf of another state power with the authority to enforce accountability on government 4 .
These two dimensions highlight the centrality of the institutional and functional linkages between parliaments and audit agencies and parliaments to bridge the "accountability gap" in the oversight of public spending.In democratic systems of government, there is a distinction between the financial scrutiny undertaken by audit agencies and the political accountability enforced by parliaments.As such, audit agencies are best conceptualized as oversight agencies, with sufficient autonomy to perform their responsibilities in an effective and impartial manner; but enforcing accountability on government ultimately rests with the other two branches of government, in particular parliament.
Indeed, in the various institutional models for organizing the external audit function there is either an institutional or a functional linkage between parliaments and audit agencies.The institutional set-up of audit agencies varies reflecting the constitutional, legal, financial and political traditions of countries' public financial management systems.Standard typologies of audit agencies classify them according to their institutional features and functions and include three ideal types: (i) the monocratic model, (ii) the court model and (iii) the board or collegiate model (Santiso, 2009;Stapenhurst and Titsworth, 2001).
• The monocratic model or the Westminster model is that of a uninominal audit agency headed by a single auditor-general or president, and generally acting as an auxiliary institution to parliament, albeit with ample autonomy.
4 Another important distinction is between managerial accountability and political accountability: "political accountability involves holding those in public office responsible for performance and decisions, while managerial accountability involves the more technical aspects of fiscal and administrative responsibility" (Newell and Bellour, 2002, p. 6).While managerial accountability is part of the process of bureaucratic delegation, political accountability is embedded in the process of legislative delegation.However, managerial accountability is often inoperative in the absence of political accountability.
Revista de Economia Política 35 (3), 2015 • pp.601-621 under this model, audit agencies focus on ex-post auditing, rather than ex--ante control, and privilege financial and performance auditing over compliance control.The controls they perform seek to correct rather than penalize.
• The court model or Napoleonic model is that of a collegiate court of auditors or tribunals of accounts, endowed with quasi-judicial powers in administrative matters, often acting as an administrative tribunal.The quasi-judicial features and functions of the court model privilege legal and financial compliance over performance auditing.The links with the legislature are weaker than in the monocratic model; yet those with the judiciary are also ambiguous.
• The board model or collegiate model is an institutional hybrid.It is an agency with collegial decision-making similar to that found in tribunals, headed by a board of auditors, but without jurisdictional authority or quasi-judicial powers.under this model, audit agencies emit an audit opinion on the reliability and probity of government accounts, usually for the legislature to consider.
All three models share a central common feature: their institutional independence from government and a connection to parliament.In most parliaments a standing parliamentary committee or subcommittee is in charge of relations with audit agencies, guiding its mandate and setting its budget.The parliamentary committee responsible for controlling the budget before and during its implementation, the budget and finance committee, is generally distinct from the committee responsible for controlling it after its implementation, the public accounts committee5 .
Nevertheless, in some parliaments a single parliamentary committee is responsible for entire budget cycle through various subcommittees.This model is, by design, potentially more effective, "the same members who scrutinize audit findings in the subcommittee subsequently make decisions on upcoming budgets in the budget committee.This link is weaker in parliaments with separate audit and budget committees, or, worse, without any audit committee at all" (Wehner, 2013) 6 .In presidential systems of government and in those countries following the court model, a subcommittee of the finance and budget committee is in charge of the relations with the audit agency, considers audit reports and discharges government7 .
Furthermore, external audit agencies are gradually shifting their emphasis from checking compliance with budgetary legislation to improving government performance.
The emergence of performance auditing emphasizes the advisory role of audit agencies as partners of parliaments in the oversight of government finances.This synergy is all the more important as parliaments seek to reassert their influence in the budget process (lienert, 2013;Shick, 2009;Santiso, 2008Santiso, , 2006;;Stapenhurst, 2008), reasserting their constitutional prerogatives and strengthening their technical capacities.
The changing role of parliament in public budgeting is part of broader reforms in the management of public finances that seek to redress the "accountability gap".Nevertheless, and despite parliaments' renewed assertiveness, the "downstream" part of the budget process remains generally weaker than its "upstream" part (Andrews, 2013).Several indicators underscore weaknesses in legislative oversight and external auditing of the budget.The Public Expenditure and Accountability Framework (PEFA) highlights structural weaknesses in the scope, nature and follow-up of external audit; parliamentary scrutiny of the annual budget law; and parliamentary scrutiny of external audit reports8 .Figure 3 shows the general weakness of the downstream part of the budget process around the world and in particular external scrutiny and auditing.Figure 4 highlights the shortcomings in parliamentary scrutiny of the budget proposal (measured by PEFA Performance Indicator 27) and more severely of external audit reports (measured by PEFA Performance Indicator 28) in latin America and the Caribbean.Measures of parliamentary scrutiny of the budget in latin America and the Caribbean suggest parliamentary influence on the budget process is not so much hindered by limited budgetary powers, but rather by the effective use of those budgetary powers and their limited technical capacities (Santiso and Varea, 2013).There is a critical synergy between the oversight of the budget performed by audit agencies and the accountability functions of parliaments.However, those linkages are not as effective as they could be.The quality of those linkages can be gauged using four proxies: (i) the role of parliament in providing an enabling environment for independent external auditing and the autonomy of the audit agency; (ii) parliament's consideration of audit reports and follow-up to audit findings; (iii) parliament's certification of public accounts and the discharge of government; and (iv) parliament's examination of the draft budget law.
enabling external auditing.Parliaments are critical to provide the enabling environment for independent external auditing and for guaranteeing the autonomy of audit agencies.First, parliaments enact the enabling legislation providing audit agencies with effective powers and autonomy, through the audit law, the organic budget law, and the public financial management law.In the recent past, many countries have updated their legislation for public financial management and external auditing, strengthening the role of parliament and upgrading the responsibilities of audit agencies.The organic budget laws and fiscal responsibility laws adopted by several emerging economies, such as Brazil in 2000, have enhanced the autonomy of audit agencies and parliamentary oversight functions.Second, parliaments participate in the selection of the auditor general, approve the agency's budget appropriation, which is often ring-fenced, guide the work of the audit agencies, and request special audit reports as part of their investigatory powers.Third, parliaments are one of the main clients and beneficiaries of the audit agency's work.
Following-up on audit findings.An important shortcoming in the oversight of the budget is the lack of follow-up of audit reports and findings.Admittedly, the prime responsibility for acting upon audit findings rests with audited government agencies themselves.Nevertheless, audit agencies and parliaments have a role in ensuring that their recommendations are acted upon.Often, yet not systematically, audit agencies provide regular follow-up to the audit recommendations of past audit reports.Parliaments, through their public accounts committees, do so less systematically, reducing the effectiveness of the ex-post scrutiny of budget execution.Good practice recommends formal arrangements in parliament for systematically following up on audit report findings and that the public accounts committee should report on the extent to which the executive has addressed the recommendations of the external auditor.The annual report of the audit agency and the annual audit of public accounts provide parliament with another opportunity to discuss the outcome of the previous year's budget, to assess government performance in managing public spending, and, if needed, to take corrective actions.However, reporting lags do not often allows the audit of the previous year's budget to feed into the consideration of next year's budget.
Discharging government.Many parliaments impose ex-post reporting requirements and carry out some form of ex-post scrutiny of budget execution.They generally examine and approve the implementation of the budget through the discharge procedure and, most often, a legislative act.In some countries, parliament is legally required to adopt a budget execution law, through which it decided on whether or not to discharge government and officially close the budget cycle.Parliamentary decisions tend not to have legal consequences per se, but can have serious political consequences through a motion to censure or impeach government.They may also influence budget allocations in future budgets.The cooperation between audit agencies and parliamentary public accounts committees is critical in the discharge of government.Audit agencies certify government public accounts and emit an audit opinion on those accounts.Parliamentary public accounts committees consider the audit opinion and decide whether or not to discharge government.These critical linkages are often dysfunctional.
For example, in Argentina, Berensztein et al. (2000) argue that the external audit agency's audit are sparse, usually untimely, do not go beyond legal compliance and are generally ineffective in improving budget efficiency or punishing corruption practices.The audit agency has abstained from emitting an audit opinion on government accounts, arguing that they did not provide sufficiently reliable information on the true nature of budget execution.The problem is further compounded by the lack of legislative action on audit findings.Between 1994 and 2006 parliament did not approve or disapprove government accounts.According to Abuelafia et al. (2009, p. 47), "the main problem affecting the way control mechanisms work relates to their being highly politicized; that is they are subject to political interference".Oddly enough, in Argentina the opposition chairs the audit agency, while the majority chairs the public accounts committee.In addition, the discharge of government is a legislative act that could thus be vetoed by the ruling party's majority in the commission.In Brazil, legislative scrutiny of the consolidated financial statements of the federal government using the audit findings is subject to signifi-Revista de Economia Política 35 (3), 2015 • pp.601-621 cant delays at times reaching almost a decade 9 .Approval of the government's accounts after the end of the president's term has not been unusual 10 .Interestingly, the audit of the consolidated financial statements of the federal government tend to be completed in a timely manner, but parliamentary scrutiny is subject to long delays despite the parliament having a permanent public accounts committee and a strong parliamentary budget office.According to the OECD (2012, p. 24), during the past decade the public accounts committee has taken, on average, two and a half years to initiate its review of the consolidated financial statements of the federal government after receiving the preliminary audit opinion.The consideration by the federal parliament of the public accounts committee's recommendation is marked by even more delays.By 2012, the federal parliament had only concluded deliberations on the government accounts of 2001 (OECD, 2012, p. 25).
examining the budget proposal.The interaction between the audit agency and parliament is also important to close the feedback loop in the budget process.Ideally, parliament ought to have examined the audit report of the previous fiscal year when considering the government's budget proposal for the following fiscal year.Thus, to be relevant, the budget execution law ought to be examined a few months after the end of the fiscal year and coincide with the pre-budget parliamentary debate.However, in many countries, time lags and delays do not allow for the audit report of the previous year to feed back into the consideration of next year's budget proposal.
WHY OVERSIGHT FAIlS TO BECOME ACCOuNTABIlITY ultimately, the effectiveness of audit agencies depends on the agility of their functional linkages with the other components of the national systems of integrity and the architecture of fiscal accountability.The literature on "delegative democracy" (O'Donnell, 1994) and "horizontal accountability" (O'Donnell, 1999(O'Donnell, , 1998) ) emphasizes the dynamic interdependence between accountability institutions inserted in the interwoven system of checks and balances.From this perspective, audit agencies are part of a broader system of accountability and their effectiveness 9 Brazilian presidents are required to render accounts to parliament every year since 1934.The audit agency prepares a report on the public accounts of the federation within three months of receiving them and issues a "preliminary opinion" or "prior judgment" ("parecer prévio às contas") to guide discussions in parliament on whether to approve the government's statements or not.Based on the audit agency's opinion, the parliamentary public accounts committee emits its own recommendation to parliament by unanimity vote.The process then follows normal legislative procedures, being examined and voted on by both chambers successively.However, the audit agency's preliminary opinion is not a formal audit ruling and seldom includes any reference to corrective measures or a follow-up of the irregularities detected in previous years. 10The audit agency has usually recommended approval, except in 1992, though it typically includes reservations and recommendations.On several occasions, to clear the backlog, the legislature approved the public accounts of several years simultaneously, often in the same day, as in 2002.
Brazilian Journal of Political Economy 35 (3), 2015 • pp.601-621 is conditioned by the effectiveness of their interaction with other budgetary actors, in particular parliament.The rationale is that, for accountability to bite, the "horizontal accountability" exercised by audit agencies needs to translate into "vertical accountability" on government enforced by parliament (Moreno et al., 2003).The accountability powers of audit agencies are delegated to them by parliaments, and, as such, audit agencies act as oversight agents on behalf of parliaments."Horizontal accountability" by oversight agencies does not generally involve self-enforceable sanctions.As such, the absence of direct enforcement and sanctioning powers of audit agencies requires them to develop effective relations with parliaments to act upon audit recommendations and follow-up audit findings.As Melo et al., (2009Melo et al., ( , p. 1220) argue, "horizontal accountability is thus a function of the quality of the vertical accountability between elected officials and voters" mediated by parliaments and "if this relationship is flawed, then the horizontal exchange will be deficient".
A reason for these disjunctions lies in that parliaments and audit agencies have different "bounded rationalities" (Rubinstein, 1998;Simon, 1991).Audit agencies tend to be dominated by technical rationality and parliaments by political rationality.The incentives motivating parliamentary oversight are primarily political, determined by the balance of political power, electoral rules, and committee politics framing parliamentarians' incentive structures.Another reasons is the inconsistency in the time horizons of politicians and auditors, between the longer-term perspective of audit work and the shorter-term time horizons of parliamentarians determined by the electoral cycles.In some countries, committee politics, including the partisan nature of key committees and the rotation of their members, have an important influence on their capacities to engage in public budgeting.In Colombia, for example, members of parliamentary committees rotate every year, although a constitutional amendment is under consideration to change this.In addition, at any time, parliaments are overseeing three different budgets and parliamentarians have more incentives to engage in the debate on the preparation of the budget of the next fiscal year, than scrutinizing the execution the previous year's budget.
Strengthening the partnership between parliaments and audit agencies, while maintaining the autonomy of audit agencies, can significantly enhance budget oversight and government accountability.Efforts to improve this relationship have nevertheless focused on strengthening institutional capabilities, rather than modifying the underlying political incentives framing the interaction between parliaments and audit agencies.Admittedly, strengthening institutional capacities and function linkages has proved challenging and remains incipient.
First, external audit agencies could improve their effectiveness by increasing their reliance on internal control systems tasked with ensuring compliance with legal and financial regulations, in particular through ex-ante control (Wescott, 2008).This would also allow external audit agencies to concentrate more on expost performance auditing and selective auditing of high-risk areas or sectors, including through comprehensive sector reviews.Similarly, audit agencies could also strengthen their functional linkages with key integrity institutions such as anti-corruption commissions, financial intelligence units, financial crime commissions, and anti-money laundering agencies.
Second, audit agencies and parliaments could engage more actively with the agencies responsible for evaluating public policies.There are indeed unexplored and underutilized synergies between independent external auditing (especially performance auditing), government monitoring mechanisms, and independent evaluation of public policies.The evaluation of public policies remains, however, a relatively weak area of public sector management (lopez-Acevedo, 2012), partly because of the lack of political incentives of governments to evaluate themselves.Several countries are seeking to strengthen the independent evaluation of government policies by creating autonomous evaluation agencies, such as Mexico, Colombia and Chile.
Third, and beyond improving the functional linkages among horizontal accountability institutions within government, several countries have sought to strengthen the technical budget capabilities of government and the technical interface with audit agencies.This has helped bridge the information asymmetries between government and parliaments.While parliaments are mainly driven by political incentives of its members, the technical rationality of parliaments can be enhanced through the strengthening of their own technical capacities to undertake analysis of fiscal policies, audit reports, and budget proposals.This has resulted in an uneven strengthening of legislative budget offices in several countries, which act as an institutional devise to enhance the linkages between parliaments (and its committees) and audit agencies.legislative budget offices improve parliamentarians' ability to analyze, review and make better decisions on the budget (Johnson and Stapenhurst, 2007;Santiso and Varea, 2013).These have emerged in many countries in recent years, modeled after the powerful united States' Congressional Budget Office created in 1974 11 .However, parliamentary budget offices tend to focus primarily on the analysis of fiscal policies and budget formulation, advising the finance and budget committees of parliament mainly during the approval stage of the budget.They tend to be less engaged with the downstream stages of the budget process and the ex-post scrutiny of budget execution.Resource asymmetries subsist, however.Most central budget offices in latin America and the Caribbean have more than forty employees while parliaments seldom have specialize legislative budget offices and when they do these are usually staffed with less than ten employees.
11 These include: united Kingdom in 2010, Canada in 2006, Korea in 2003, Mexico in 1998, Philippines in 1990.They have emerged in developed and developing countries alike, including in fragile states.In uganda, the Budget Act of 2001 established a Parliamentary Budget Committee to scrutinize macroeconomic and fiscal targets and a Parliamentary Budget Office to analyze budget proposals.liberia created a Legislative Budget Office in 2010, which conducts fiscal impact analysis and reviews draft budget reports.In 2013, South Africa is setting up its parliamentary budget office (Verwey, 2009;Fölscher and Cole, 2006) and Colombia is considering it, as part of a broader modernization of parliament.Economy 35 (3), 2015 • pp. 601-621 Nevertheless, the effectiveness of the architecture of budget accountability are primarily ultimately hinges upon and is embedded in the political economy dynamics of executive-legislative budget relations.The political incentives to use existing audit information to hold government to account are often missing.In majoritarian political systems where parliament is dominated by the president's political party and party loyalty is adhered to strictly, there is very little incentive for parliamentarians to deviate from the party line during the examination of budget execution.Parliamentarians from the ruling party have limited interest in finding fault with the government and opposition parliamentarians lack incentives to debate an often technically arcane audit report.As a result, as ladipo (2009, p. 47) notes "parliaments usually conduct a perfunctory vote to accept the audit agency's audit on the consolidated public accounts.In general, few detailed questions are asked about the contents of the annual report or the audit findings, and there is no followup to audit recommendations to correct any deficiencies identified".Moreover, parliamentarians have few incentives to invest in financial scrutiny and in strengthening parliaments' own technical capacities for budget oversight.

Brazilian Journal of Political
There are a number of critical political economy factors affecting the effectiveness of the linkages between parliaments and audit agencies in the oversight of government finances and the enforcement of government accountability.Melo et al., (2009) highlight two critical factors: institutional design and political competition.First, the choice of institutional design and the distribution of budget powers defined by the constitutions and organic budget laws define the budget powers of parliaments and the institutional configuration of the audit agency.There have indeed been very few reforms in the original institutional design of audit agencies and their relationship to parliaments.Budget rules shape the interactions between the executive and legislative during the budget cycle.Moreno et al., (2003) acknowledge that institutional design and budget rules governing the accountability institutions affect the autonomy and the performance of those institutions.The rules of appointment of the auditor general (or the collegial board of the audit agency) affect the effectiveness of the audit agencies and the impartiality of its work.In Argentina, for example, while the head of the collegiate audit agency is a member of the opposition, its collective board is dominated by the ruling party, therefore hindering its autonomy.In fact, the ruling party has tended to dominate both the collegiate governing body of the audit agency and the parliament's public accounts committee, which gives it de facto veto power over the certification of public accounts and the discharge of government.The result is that the audit agency risks to be hierarchically depending on those very authorities it is meant to control.In addition, parliamentarians' careers depend more on their political standing in the provinces and with the federal government, than their standing in the legislature.They thus have few incentives to specialize and invest in strengthening parliament's oversight capacities.In sum, flaws in institutional design are compounded by the broader political economy of executive-legislative budget relations.
Budget rules, such as the president's veto power on legislative amendments, have been particularly effective at disciplining the upstream part of the budget process and in particular its adoption after its consideration by parliament.In most latin American countries, presidents have veto power (including line-item veto power) on the budget amendments proposed by parliamentarians.In Brazil, it has also discretion on the execution of those amendments.Pereira and Mueller (2004) and Alston and Mueller (2005) show that the strategic management of amendment appropriations (in effect, pork barrel) has been an important policy tool of Brazilian presidents to advance their legislative agenda under coalition governments and discipline a fragmented parliament as a result of the open-list proportional representation electoral system with large district magnitude.In addition, the open-list proportional representation systems encourages voters to support candidates based on personal qualities and their ability to augment federal transfers to states.Pereira and Mueller (2004, p. 715) provide evidence that the "Brazilian president routinely uses his or her powers to reward and punish legislators for supporting or opposing his or her interests in congress.Although legislators have a limited role in the budgetary process, that role can have important electoral consequences for their political careers".It is logical to believe that the same logic applies to the sanctioning of government in the ex-post control of budget execution through the discharge process.Parliamentarians have little incentives to risk loosing the pork included in their budget amendments in the budget proposal for the following year, as "there is a direct link between the voting behavior of the members of congress and their individual budget amendments" (Alston and Mueller, 2005, p. 29).
Second, the nature of the political system and electoral rules are central to parliamentary politics and critical to shape the political incentives of parliamentarians to effectively use audit findings to hold government to account.The balance of political power and the degree of political competition is a critical factor to incentivize a more effective relation between parliaments and audit agencies.The existence of a strong cohesive opposition and programmatic political parties tend to augment the incentives of parliaments to oversee government and hold it to account.Majoritarian electoral systems with large electoral districts tend to generate stronger oppositions and more cohesive political parties.As such, the legislative committee system tends to be more disciplined and with greater incentives to use external audit findings more effectively.In addition, parliaments tend to have more incentives in strengthening their technical budget capacities, have greater access to fiscal information and rely more effectively on audit agencies.As Khagran et al., (2013, p. 12) note, "single-party governments will most likely face weak demand by the legislature for information taken by the executive" and "the higher the level of political party competition and therefore the probability of losing power in the next election, the more a government will have an incentive to promote transparency and reduce discretion, in order to tie the hands of its competitors in the case of electoral defeat".However, party fragmentation augments the collective action dilemmas of parliaments in the oversight of government.Proportional representation systems (especially with open lists) with large district magnitude or majoritarian electoral systems with small electoral districts tend to lead to multiparty representation, party fragmentation and weak coalition government.
The degree of political competition and the likelihood of alternation in power have proved to be important triggers to reform audit arrangements, strengthen legislative budget capacities, and effectively use audit findings.As Melo et al.(2009Melo et al.( , p. 1219) note, "When elites face greater risk of being replaced in office by their rivals, they have more incentives to bind their hands and delegate more power to autonomous accountability institutions.Incumbents who face no credible threats have fewer incentives to bind their own hands because they would prefer to keep their discretionary powers unchecked".Whener and De Renzio (2013b) also show a correlation between the degree of political competition and budget transparency.
The interplay between institutional design and political incentives play out in different ways, though.In Mexico, the alternation in power in 2000 triggered an important reform of the federal audit agency and a significant strengthening of the technical budget capacities of parliament.The partisan use of these accountability resources has nevertheless hindered their effectiveness and sustainability.In Chile, the existence of two strong political parties and recurrent alternation in power has strengthened the checks and balances in the budget process.In Brazil, Melo et al., (2009) show that the degree of political competition and the likelihood of power alternation are critical determinants of the effectiveness of accountability arrangements.Analyzing the effectiveness of 33 state audit agencies in Brazil, Melo el al. (2009Melo el al. ( , p. 1239) demonstrate that "elite turnover at the gubernatorial level strengthens accountability, because it creates incentives for delegation of power to independent institutions" [...] "political competition and alternation in power among elites groups influences the likelihood that a governance appoints a senior auditor for the board (of the collegial audit agency).When auditors are members of the boards, (state audit agencies) are more active and initiate more audits on the basis of suspicions by auditors".
Third, in an effort to strengthen the vertical accountability relationship in the budget process, audit agencies have often sought to strengthen their direct relations with voters through the publicity of their audit reports and a more effective use of the media.Both parliaments and audit agencies can enhance their impact through a more agile and effective relationship with civil society organizations that are specializing in budget analysis and fiscal transparency.The specialized media plays a critical role and both audit agencies and parliaments should adopt pro-active communication strategies.Beyond the publicity of audit reports, audit agencies and parliaments should make their reports more accessible to the public in terms of contents, avoid overly technical and legalistic terminology, and work with the specialized media to disseminate them 12 .Civil society organizations and policy think tanks specialized in budget analysis have emerged in both developed and developing countries, advocating for greater budget transparency.Such organizations provide a further source of independent budget analysis.In addition, the publicity of audit reports has proved an effective way for audit agencies to press for greater transparency and accountability in government.

CONCluSION
Most practical budgeting may take place in a twilight zone between politics and efficiency Aaron Wildavsky, 1961, p. 187 The interactions between audit agencies and parliaments are a critical juncture in the cycle of accountability and the system of integrity.However, these linkages are often ineffective, diminishing the overall effectiveness of the system of checks and balance in the management of public finances.This explains why "oversight potential does not always translate into effective oversight" (Pelizzo andStapenhurst, 2007, p. 391). As Hallerberg et al. (2009, p. 2) note, "institutional arrangements can either exacerbate or contribute to resolving a number of potential problems" and, as such, institutional arrangements and agency coordination have a significant impact on the efficacy of the overall system of fiscal control and budget policymaking.
The article contains four important findings First, the strengthening budget accountability will not necessarily emerge from increasing the formal budgetary powers of parliaments and audit agencies, but rather through a more effective use of the budgetary powers they already have.This would require, in turn, enhancing thee incentives, motivations and capacities of individual parliamentarians to do so.Second, the analysis of the interactions between parliament and audit agencies reveals important gaps in the system of accountability in the budget process.It shows that those dysfunctions are systemic rather than agency-specific.As Hallerberg et al. (2009, pp. 4-6) underscore, "the budget process should not be examined as part of an isolated and technical discussion […] (it) is part of a broader policymaking process".Third, those dysfunctions are primarily coordination failures, rather than agency failures.The agility of the linkages among the different components of the system of budget oversight is often more important than the effectiveness of each individual oversight agency and accountability institution taken in isolation.Fourth, the analysis shows that the linkages between parliaments and audit agencies are marked by principal-agent challenges.However, it suggests that modifying the framework of incentives of the principal, that is parliament, can strengthen the effectiveness of its agent, the audit agency.In other words, changes in the incentives of the principal improve the effectiveness of the agent.
ultimately, the budget is an inherently political process framed by the interaction between institutional rules and political incentives.Of particular relevance are the choice of institutional arrangements and budget rules framing executive-legislative relations in the budget process, as well as the degree of competition in the political system, the political incentives provided by electoral rules, and the dynam-Brazilian Journal of Political Economy 35 (3), 2015 • pp.601-621 ics of parliamentary politics.In the end, budgeting is governing, "budgeting is a subsystem of politics, not vice versa -because of the current tendency to overload budgeting" (Wildavsky, 1992, p. 439).
The effectiveness of the linkages between parliaments and audit agencies in the ex-post control of budget execution cannot be dissociated from the broader political economy of the budget process and the dynamics of executive-legislative relations.Strong audit institutions cannot in themselves prevent the abuse of power by those who are holding it.As Johnson et al. (2012, pp. 21-22) underscore, "even the most detailed and robust audit reports are unlikely to have much effect without effective budgetary oversight and scrutiny by parliament, and the ability to hold spending ministries to account".More research is undoubtedly needed on the political economy of budget accountability institutions and the institutional architecture of checks and balances in the oversight of the budget.

Figure 1 :
Figure 1: Government effectiveness and legislative oversight in Latin America and the Caribbean 2

Figure 1 :
Figure 1: Government effectiveness and legislative oversight.Graph shows the bivariate association between the World Bank Institute's government effectiveness index (2011) and the International Budget Partnership's strength of legislative oversight index (2010).Countries in Latin America and the Caribbean are depicted by large circles.Countries in other regions are depicted by smaller circles.The estimated regression line for a regression of the effectiveness index on the oversight measure is shown as a thick black line.Ninety-five percent confidence intervals are shown as dashed lines.

Figure 2 :
Figure 2: Government effectiveness and external auditing in Latin America and the Caribbean 3

Figure 2 :
Figure 2: Government effectiveness and auditing institutions.Graph shows the bivariate association between the World Bank Institute's government effectiveness index (2011) and the International Budget Partnership's strength of the supreme auditing institution index (2010).Countries in Latin America and the Caribbean are depicted by large circles.Countries in other regions are depicted by smaller circles.The estimated regression line for a regression of the effectiveness index on the strength of auditing measure is shown as a thick black line.Ninety--five percent confidence intervals are shown as dashed lines.

Figure 3 :
Figure 3: Weakness in external scrutiny and independent auditing

Figure 4 :
Figure 4: Shortcomings in parliamentary scrutiny of the budget proposal and audit reports