Bank Income Smoothing and Loan Loss Provisioning Practices in Africa

The primary objective of the thesis is to investigate whether African banks use loans loss provisions estimates to smooth reported earnings, and to determine the factors that influence the extent of earnings smoothing among African banks. Earnings smoothing via loan loss provision has been examined in several regions, but the case of Africa remain unexplored in the literature. In the thesis, earnings smoothing is viewed as an earnings management practice while loan loss provisions estimate is considered to be the tool used by African banks to smooth reported earnings. Using African bank data obtained from Bankscope database, I test the earnings smoothing hypothesis for 370 African banks during the 2002 to 2014 period using the specific-accrual approach. The specific-accrual approach estimates a specific discretionary accrual as a function of its non-discretionary determinants and other factors that influence the manipulation of the specific accrual. The model specification expresses discretionary loan loss provisions as a function of earnings before provisions and tax, its non-discretionary determinants and other factors that influence the decision regarding the level of bank provisions for each period. The findings indicate that African banks manipulate loan loss provisions estimates to smooth reported earnings and this behaviour is influenced by bank differences, accounting disclosure differences and institutional differences across African countries. The primary contribution to knowledge of the thesis is its extension of our understanding of the role of discretionary accruals in the bank financial reporting, focusing on African banks - a context that has not been extensively examined in the literature. Also, the thesis extends the bank earnings smoothing debate to the African context and the findings of this study are useful to bank regulators in Africa in their evaluation of whether bank loan loss provisions solely reflect credit risk considerations or whether bank loan loss provisions estimates reflect opportunistic considerations of African bank managers. Finally, the findings are useful to local accounting standard setters in the region in their evaluation of several accounting numbers that bank managers might use to manipulate reported earnings.