Modelling and Forecasting Energy Intensity, Energy Efficiency and CO₂ Emissions for Pakistan

The aim of this thesis is to examine the significant environmental issues, especially, Green House Gases (GHGs) emissions and specifically Carbon Dioxide (CO₂) emissions which are mainly caused by energy use. This thesis consists of three core chapters. Chapters 2 and 4 discuss how to stabilize and forecast CO₂ emissions for Pakistan while chapter 3 discusses the energy efficiency of Asian developing countries. Exogenous Technical Change (TC) and endogenous TC models are considered in the chapter 2 for the stabilization of CO₂ emissions. Specifically, the estimated results show that endogenous TC model (which is estimated by following the Kalman Filter (KF) technique) does a better job in comparison. The results also point out the existence of a trade-off between GDP growth and fuel prices. Inter-fuel substitutions are estimated using the Almost Ideal Demand System (AID). Results suggest that stabilization can be achieved just in short run but it needs too much time for the implementation in the long run plans. In chapter 3, a parametric Stochastic Frontier model Approach (SFA) is used for a panel of 19 countries including Pakistan over the period of 1980 to 2013. The individual and relative energy efficiency over time of all counties is estimated. The focus is to find either energy intensity a good indicator of energy efficiency or not. According to the estimated results, energy intensity is not a good indicator of energy efficiency but the energy efficiency estimated using SFA after controlling for some of the economic factors (fuel prices, population, income, etc.) it is. In chapter 4, the relationship between CO₂ emissions and income, and energy consumption and income are found to support the Environmental Kuznets Curve (EKC) hypothesis. Univariate (Grey Prediction Model (GM), Exponential Smoothing (ES), Holt-Winter (H-W)) and multivariate model solving techniques are used to predict CO₂ emissions and their forecasting abilities are compared. A new technique, Out Of Sample Grey Prediction (OOSGP), is introduced after providing a critique of the GP model to get better forecast results. The findings of this study provide a valuable reference with which Pakistan’s government could formulate measures to reduce CO₂ emissions by curbing the unnecessary consumption of energy.