Abstract:
Despite the rapid growth in the banking sector in Sri Lanka, there
is an increasing trend of NPLs in most of the Sri Lankan state
banks including ABC bank. Therefore, the main purpose of this
study is to develop a time series model to forecast the NPLs in ABC
bank, while bridging the gap of developing novel model to forecast
future NPLs. Based on the nature of the research objective it has
been applied quantitative research methodology for the study.
Secondary data on gross NPLs of ABC bank from fourth quarter of
2008 to fourth quarter of 2018 has been taken for the analysis and
applied the technique of ARIMA as the main analytical tool.
Moreover, MAPE and MAD have been used to check the accuracy of
the fitted model. ARIMA (0,2,1) model was identified as the best
fitted model to forecast the future NPLs in ABC bank after checking
the accuracy through MAPE and MAD by using both training and
validation datasets. According to the findings, the increasing trend
of gross NPLs in ABC bank will continue in future and there will be
nearly Rs. 64929.6 million of gross NPLs in this bank at the end of
2019. This is the only study which has been touched the area of
forecasting NPLs in Sri Lankan context. Hence, it would provide
insights to the rare literature, while enhancing the knowledge of
financial forecasting. The study would be favorable to the
management of ABC bank along with the Central Bank of Sri Lanka
in order to mitigate the risk attached with NPLs.