Determinants of SME’s failure in Malaysia and Nigeria

Small and medium-sized enterprises (SMEs) play an important role toward economic development worldwide. Predicting bankruptcy among SMEs can have a significant impact on the economy as an effective early warning signal. This study develops bankruptcy prediction models for Malaysian and Nigerian SMEs by combining financial, non-financial, corporate governance and macroeconomic variables using the logistic regression and artificial neural network (ANN) methods. The accuracy rates obtained by those two methods were then compared. In developing the estimated model, 1,556 (632) SMEs from the manufacturing sector in Malaysia (Nigeria) are selected. Three sub-samples are created representing 3-years, 2-years and 1-year prior to bankruptcy, with total observations of 666 (344), 470 (172) and 420 (116) respectively. Each sub-sample comprises 50 percent nonbankrupt and 50 percent bankruptcy firms, from years 2000 to 2014. The findings show that four of the financial variables, namely, lower profitability,
high leverage, insufficient liquidity and high operating expenses are associated with bankruptcy among SMEs in Malaysia and Nigeria. As for the non-financial variables, the results indicate that young SMEs and those located in less industrialised states are more likely to go bankrupt. In addition, the corporate governance variables, such as number of directors, independent director, managing director duality, controlling shareholder, ethnicity and gender of managing director are found significant. Moreover, high unemployment rate is associated with bankruptcy among SMEs in Malaysia and Nigeria, while high inflation rate as well as lending rate are associated with SMEs bankruptcy only in Nigeria. The result shows that the ANN model leads to a higher predictive accuracy rate compared to the logistic regression model for Malaysia and Nigeria. The study reveals that SMEs should increase the number of independent directors, discourage CEO duality and reduce ownership concentration. Financial institutions
could use this study as a reference model to manage credit risk of SMEs while the government agencies may use it to improve their existing policies.