Gerenciamento de resultados em períodos de crises econômicas: uma análise dos mercados acionários brasileiro e norte-americano

Abstract

This dissertation investigated earnings management (EM) by means of discretionary accruals (DA) in companies listed in Brazilian and US capital markets in periods of economic crises. Taking into account that accounting practices both affect and are affected by the environment in which they are found, more specifically by cultural, legal, political and economic factors, it was expected that, based on a conducted review of the literature, that at times characterized as economic distortions companies would identify proper occasions for obtaining private gains by means of EM practices. In order to achieve that purpose, two samplings were composed, one by companies listed in the São Paulo Stock Exchange (BM&F Bovespa) with quarterly data for the period from 1999 to 2009, amounting to 8,600 observations, and another of companies listed in Nyse and Nasdaq, with 99,931 observations, for the period from 1999 to 2010. It should be mentioned that the non-inclusion of 2010 for Brazilian companies occurred due to the convergence of the International Financial Reporting System (IFRS) in that year, which might lead to the creation of undesired biases in the research results. The approach used was that of Aggregate Discretionary Accruals (ADA) with regression based on the panel data technique, and initially obtaining the discretionary accruals by estimation in the Modified Jones Model (1995) adapted with the variables Return on Assets (ROA) and quarter. Subsequently, the DA were regressed against the variable of interest to the study, the dummy crisis variable, together with the other control terms: sectors, Book-to-Market ratio, provision for income and social contribution taxes, and tax burden, the last two ones applicable only to the Brazilian data sample. The statistical tests showed that in both equity markets significant relationships were evidenced between the crisis variable and the discretionary accruals, indicating that, at such times, companies manage their earnings more than in other periods. More specifically, the estimated coefficients had negative values, which shows that the companies have reduced their DA on those dates, possibly aiming at maintaining their profitability standards, which typically characterizes the practice of target earnings. Also regarding the identification of variables that might complement the aggregate accrual models, and improving the explanatory capacity of such tools in the search for discretionary accruals as proxies for earnings management, one can verify that items such as (i) sectors, (ii) taxation, (iii) market values compared to book values and (iv) quarters, representing the adjustment of temporary trends that escape the study's purpose have shown to be extremely important for the gauging of estimators that could reduce regression biases