Earnings Management Around Zero: A motivation to local politician signalling competence

Abstract

Literature earnings management aims to determine what causes/motivates managers to disclose earnings close to zero and to use this as an instrument to influence users' decisions. However, limited research has been carried out on this subject in the public sector. Therefore, the main purpose of this paper is to evaluate whether local politicians (in Portuguese municipalities), aiming to demonstrate their high level of competence and skills, engage in earnings management in such a way as to ensure that earnings are positive but close to zero. We examined whether political competition is a determining factor of earnings management close to zero and if managers use discretionary accruals in order to do this. Results indicate that, in order to report positive net earnings close to zero, discretionary accruals are used. This study identified the overriding tendency to avoid the reporting of losses in those municipalities where political competition is greatest.

Key words:

• Public accounting
• agency theory
• public choice theory
• earnings management
• municipalities
• political competition

Notes

This refers to taxes that are municipal revenues and to taxes that are government revenues and assigned to municipalities as funds.

This centre collects information relating to the reporting of municipal financial statements to prepare the ‘Anuários Financeiros dos Munícipios Portugueses’ (2003, 2004, 2005, 2006, 2007 and 2008), compiled by the following authors: João Carvalho (University of Minho), Maria José Fernandes (Polytechnic Institute of Cávado and Ave), Pedro Cam[otilde]es (University of Minho) and Susana Jorge (University of Coimbra).

The test has used two other variables: (variations in sales, services and taxes) and (variations in sales, services, taxes, and transferences and subsidies). Reliability of the estimates is also high and the study results are not sensitive to the use of one, or any other, variable.

The interval amplitude was calculated using the same methodology as Silverman (1986): 0.79 × σ × n ^− 1/5, where n is the number of observations in the sample and σ is the standard deviation. Although space precludes the demonstration of a graph, the analysis using larger and smaller intervals was also carried out. However, the results were more or less the same.

Subdivision according to variable PComp, calculated for 2001 election results.

Subdivision according to variable PComp, calculated using election results of 2005.