This paper tries to examine the performance of VECM model to forecast the amount of loanable funds in the banking system of Nepal. For this, monthly data of 14 years starting from July 2007 to June 2021 have been used in the systematic process of modeling and forecasting practices. The VECM model was estimated with the training dataset covering from 2007 to 2015 followed by examination and validation with the testing dataset covering 2015 to 2020. The empirical results reveal that the supply side factors (government expenditure and BoP) of loanable funds have got dominant power in comparison to the demand side factors (industrial investment and consumption) while determining the amount of loanable funds in the banking system. The forecast performance indicators confirm that the selected VECM model is capable enough in explaining the variations of the determinants that bring changes in the monthly amount of loanable funds of the banking system. As suggested by the results, the VECM modeling approach could be used for forecasting of loanable funds at the BFIs’ level in the banking system of Nepal.

Forecasting Nepal’s Gross Domestic Product (GDP) holds paramount importance for effective resource planning and allocation. In this research, Artificial Neural Networks (ANNs) have been introduced to predict the GDP time series, wherein the data have been dissected into linear and nonlinear components. The linear aspects have been handled by the ARIMA model, while the ANNs managed the nonlinear elements. Additionally, the study has delved into hybrid models, resulting in additive and multiplicative combinations of ARIMA and ANN. These hybrid models have aimed to enhance forecasting performance, minimize errors, and improve accuracy compared to standalone models. The findings revealed that both ANN and hybrid models surpassed other approaches in terms of prediction accuracy.