X hits on this document





44 / 70

and the culture of public administration. Five research findings and five policy recommendations can be drawn from this paper.

Research findings

First, the research uncovers a positive correlation between the credibility of external auditing and the quality of financial governance, in terms of budget transparency, bureaucratic effectiveness and corruption control. However, institutional design only partially explains the agency effectiveness. For example, evidence from Argentina, Brazil and Chile suggest that there is no strong, direct and unambiguous relationship between the probity of public administration and the existence of ex-ante compliance control. Causality relations between external auditing and fiscal performance remain uncertain, however.

Second, AAAs are part of the broader system of financial governance and fiscal control in which they are embedded. Dysfunctions are systemic, not agency-specific. Therefore, the efficacy of the linkages between the various components of the ‘national integrity systems’ is critical. The nature and quality of the AAAs’ relations with the executive (for enforcing administrative accountability), the judiciary (for enforcing judicial accountability) and the legislature (for enforcing political accountability) are critical. The case of Argentina illustrates that, while the AAA is technically competent, it fails to have a deterrent impact because of audit findings are not effectively acted upon by the legislature. Similarly, the dysfunctions in the links between the AAA and the judicial system in Brazil hamper the effective enforcement of audit recommendations.

Clearly, the prevalence of corruption in many countries requires strong preventive, corrective and punitive systems of fiscal control. However, as this article posits, AAAs are essentially oversight agencies with indirect enforcement powers, rather than accountability institutions with direct sanctioning powers. Their contribution to fiscal control resides in the support they provide, as auxiliary institutions, to those state powers with the mandate with enforcing accountability on government as part of the system of checks and balances and the separation of powers, namely the legislature and the judiciary. As Moreno et al. (2003:117) stress, ‘agents cannot hold other agents accountable, only their principals can.’ What matters most is not whether AAAs are endowed with direct sanctioning powers, but rather that enforcement actually occurs.

The ‘functional linkages’ hypothesis also draws attention to the balance between technical capacities and political incentives to improve the effectiveness of oversight agencies. Technical improvements in auditing techniques can help enhance agency performance. However, the impact of AAAs ultimately hinges upon the incentives of political actors to effectively use the financial information they provide. This, in turns, depends on the configuration of political power and the incentives provided by electoral rules and party systems. There thus exists unexplored potential to improve government accountability by tackling the dysfunctions in the political incentives for exercising oversight. This requires understanding the synergies between the institutions of horizontal and vertical accountability (Persson and Tabellini 2001, 1999).

Third, institutional independence is a critical guarantee of impartiality and credibility, but should not be an end in itself. Securing the political independence of AAAs is critical to their credibility and effectiveness, but what matters most is shielding them from partisan interference and political capture. In Argentina, the AGN is undermined by the ‘political filtering’ of its audit rulings and the resulting fiction of control. Brazil also shows that while the TCU is largely independent from executive interference, it is not free from partisan influence.

However, excessive independence and institutional insulation can also hamper effectiveness, as in Chile. There exists a paradox of independence: while AAAs should be sufficiently independent to


Document info
Document views256
Page views256
Page last viewedSun Jan 22 02:01:04 UTC 2017