The effect of asset liability management on profitability of Commercial Banks in Kenya
To effectively compete in the market place, banks manage their assets and liabilities taking into consideration the risk level, earnings, liquidity, profit, solvency, the level of loans and deposits to mitigate losses and thus improve profitability. Asset liability management is comprehensive and dynamic framework used to measure, monitor and manage the market risk of a bank. Considering that the Kenyan banking sector has been competitive and ALM is critical for success of financial institutions, this descriptive study set out to determine the effect of asset liability management on profitability of commercial banks in Kenya. The study collected secondary data from published financial statements of 44 commercial banks in Kenya for the period 2010 to 2014. The regression analysis establish that 47.7 percent of variations in financial performance proxied by ROA are explained by variations in the study independemt variables namely; Size, Capital structure and asset liability management position of the bank. The findings show that there is a statistically significant positive relationship between bank size and financial performance and a statistically significant negative relationship between capital structure and financial performance. A unit increase in ALM position caused by increase in advances and decrease in deposits causes a decline in average financial performance of the banks. The study therefore recommends that bank managers should put in place mechanisms to attract deposits and low cost funding so as to manage any potential liquidity mismatches that may force the banks to resort to expensive debt capital. Given that bigger banks are seen to perform better than smaller banks, policy should be geared towards making the smaller banks equally competitive so that they can contribute in financial inclusion in the country. Future studies should revisit the presumed linear relationships between the study variables and choice should be made on use of fixed or random effects with the panel data.