THE DEMAND FOR BANK LOANS IN SINGAPORE

Abstract

Bank loans constitute 70 per cent of the bank's local asset and a large part of bank profit comes from the interest on the loans portfolio. The demand for bank loans is an economic indicator which reflects the health of the economy. It affects the money supply through the creation of demand deposits. An econometric model for the demand for bank loans in Singapore will provide insight on the variables influencing decision on borrowing. This will help both the bankers and the authorities to implement policies concerning the demand of bank loans. The objective of this academic exercise is to formulate an econometric model for the demand for bank loans in Singapore. This is done by using Ordinary Least Squares (OLS) method on quarterly data covering from 1977 first quarter to 1989 fourth quarter. This study is, however, limited by the non-availability of the quarterly data for private fixed capital formation. Taking into account the limitation of data, this study has shown that the main determinants for the demand for real bank loans are the income variable as proxied by GDP, private fixed investment and the lagged dependent variable. The real prime rate and the real inventory do not have significant effect on the demand for real bank loans. This is in contrast with the usual expectation that the cost of borrowing (real prime rate), is an important factor when making decision on borrowing. The results also indicate that the demand for real bank loans responds more sharply to an increase in real fixed investment than a rise in real inventory. The coefficient on the lagged dependent variable suggests that the rate of adjustment to the desired level of demand is very slow. The estimation of sectoral demand for real bank loans further supports the above finding. Due to time constraint, this study does not use Zellner’s Seemingly Unrelated Regression (SUR) to estimate the sectoral demand equations. This method provides more efficient estimators if the disturbances in the set equations are correlated. In addition, further research may test the validity of aggregating the sectoral equation into aggregate equation.