POPULATION AGEING AND ECONOMIC GROWTH IN MALAYSIA: EVIDENCE FROM AN ARDL BOUNDS TESTING APPROACH

Abstract

Malaysia’s transition towards an ageing society may influence economic performance through fertility, longevity, dependency and labour force participation. This study investigates these relationships using annual data for 1994–2024 and an autoregressive distributed lag bounds testing framework. The preferred theoretical specification relates the natural logarithm of real GDP to the logarithms of fertility, life expectancy, the old age dependency ratio and labour force participation. Unit-root testing shows, however, that the logged old age dependency ratio and logged labour force participation exhibit second-order persistence. Their first log differences are therefore employed to ensure that every variable included in the ARDL is integrated of order zero or one. BIC selects ARDL(1,1,1,1,1), with controls for the 1998 Asian financial crisis and the 2020 pandemic. The bounds-test F-statistic is 1.449, below the 5% lower bound of 2.86. Hence, the null hypothesis of no level relationship cannot be rejected and long-run coefficients or an error correction model are not estimated. A stationary short-run model finds no statistically significant demographic coefficient at the 5% level. The crisis controls are negative and significant, indicating that exceptional macroeconomic shocks dominate annual changes in real GDP during the sample. Residual and stability tests also identify non-normality, serial correlation in the conditional ARDL and CUSUMSQ instability. The findings do not imply that demographic change is economically unimportant; rather, they show that this 30 years sample does not provide sufficiently robust evidence of a stable long-run ARDL relationship.