Limits on government spending: evidence of manipulation from Brazilian municipalities

ABSTRACT

This paper adopts a data-driven approach to detect manipulation in the statements of personnel expenses of Brazilian municipalities used by the Courts of Accounts to verify compliance with the limits determined by the Fiscal Responsibility Law. First, it tests the null hypotheses that the empirical density is continuous at the three legal cut-offs, and shows a large discontinuity at the threshold that brings about the most severe penalties. Next, it investigates whether sorting is attained through fiscal adjustment programmes or by changing accounting practices. The histogram obtained from an alternative data source, which is not so prone to manipulation, is undoubtedly smoother. There is also evidence that public officials overstate deductions from gross spending to avoid the legal sanctions.

KEYWORDS:

• Numerical fiscal rule
• personnel expenses
• manipulation
• Brazilian municipalities

JEL CLASSIFICATION:

• H72
• H83

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ http://www1.worldbank.org/publicsector/pe/BudgetLaws/BRLRFEnglish.pdfhttp://www1.worldbank.org/publicsector/pe/BudgetLaws/BRLRFEnglish.pdf.

² The spending ceiling for municipal governments is 60%, but 6% is the share of the Legislative branch. The alert and prudential limits are equal to 90% and 95% of the ceiling, respectively.

³ Tax autonomy may be more effective than numerical rules to stimulate fiscal discipline at the lower levels of government according to the empirical evidence (Von Hagen and Eichengree 1996; Foremny 2014).