LIFE EXPECTANCY - POVERTY DYNAMICS: AN AUTOREGRESSIVE DISTRIBUTED- LAG (ARDL) APPROACH

Abstract

This study examines the relationship between males' life expectancy at birth (LEM) and poverty
(POV), particularly the unilateral relationship from the predictor POV to the dependent LEM. To
enhance the robustness of parameter estimations, the Naira-Dollar exchange rate (EXR) was
included as a control variable alongside POV. Yearly time series data collected on LEM, POV,
and EXR spanning 1981 to 2023 was used in our study. Moreover, to determine the possible
presence of short and long-run relationships among these three series, we used the
Autoregressive Distributed-Lag (ARDL) model for examining these series. Basic pre-test results
of ARDL such as first difference stationary conditions (I(1)s) and lag selection criteria jointly
selected the ARDL(1, 3, 2) model as the optimal model for examining the series. Moreover,
diagnostic checking on the model’s residual showed that it is non-autocorrelated and non-
spurious (R 2 (0.999367=) < Durbin-Watson (=2.404865)). The findings established that changes
in EXR and POV have an immediate lag on life expectancy, with EXR fluctuations having
complex short and long-term effects and POV having significant delayed negative effects.
Further findings revealed that the Error Correction Term (ECT) has the correct sign (-0.061242)
which indicates a 6.1% adjustment rate back to the long-run equilibrium per period.