Forecasting the determinants of environmental degradation: a gray modeling approach

Abstract

ABSTRACT The primal contribution of present research work is to analyze and predict the environmental degradation indicators for Pakistan; by the implications of the marginal impact of economic growth, energy consumption, urbanization, and agriculture land on carbon dioxide (CO2) emissions and nitrous oxide (N2O) emissions for the period 1971–2014. To quantify the marginal impact, the autoregressive-distributed lag (ARDL) bounds testing co-integration approach is implemented concerning CO2 and N2O emissions in the long-run and short-run. For the prediction of environmental degradation indicators, the gray prediction approach GM(1,1), a first-order single variable prediction model, is utilized by fitting the real data of leading factors of environmental degradation. The findings indicate that GDP and energy consumption induce CO2 emissions, whereas urbanization and agriculture land induce N2O emissions in the long-run. Further, the emissions from designated variables up to the year 2030 are forecasted using GM(1,1) model. Mean absolute percentage error (MAPE) is employed to examine forecasting accuracy and resulting MAPE values lie in the appropriate range. GM(1,1) model establishes forecasting results that are worthwhile and can contribute to environmental policymaking to reduce the polluted environment in Pakistan.