**3. Methodology and Scenario Assumptions**

This study aims to address the research questions on what are the projected impacts on greenhouse gas emissions of different potential energy mixes for ASEAN countries, and which energy mix should the region pursue? Thus, this study employs energy models of ASEAN countries using the Long-range Energy Alternatives Planning (LEAP) system software, an accounting system used to develop projections of energy balance tables based on final energy consumption and energy input/output in the transformation sector. LEAP is an energy modelling software that can conduct a variety of analyses of energy systems including Demand Analysis, Transformation Analysis, Resource Analysis, and Environmental Analysis. The data structures in a LEAP are organized using a hierarchical tree. The types of data entered at each branch depend on the type of branch, its position in the tree (for example whether it is a Demand or Transformation branch). At the heart of LEAP is the concept of scenario analysis. Scenarios are self-consistent story-lines of how a future energy system might evolve over time in a particular demographic and socioeconomic setting and under a particular set of policy conditions. Using LEAP, scenarios can be built and then compared to assess their energy requirements, social costs and benefits and environmental impacts. All scenarios start from a common base year. We can use scenarios to ask an unlimited number of "what if" questions, such as: What if more efficient appliances are introduced, what if different electric generation capacity expansion plans are pursued, what if indigenous reserves of oil and gas are discovered, what if renewable energy technologies are introduced, etc.

In LEAP, the energy consumption is calculated as the product of an activity level and an annual energy intensity (energy use per unit of activity). Overall activities are defined as the products of the individual activities entered along a complete branch of the Demand tree. Total energy consumption is thus calculated by the equation:

$$\text{Energy Consumption} = \text{(Activity level)} \times \text{(Energy intensity)}.\tag{1}$$

In general, Equation (1) can apply only when energy intensity data is available. However, in most case, energy intensity by activity is not available. Thus, regressions are used to predict the future energy consumption by activity and sector.

In this study, data availability varies in the 10 ASEAN Member States (AMS). It is very challenging to collect long-term historical data in countries such as Cambodia, the Lao PDR, and Myanmar. Further, there are many missing data points in the historical data that need to be estimated. The LEAP application is very useful in dealing with such minimal data, and it allows expert judgement on how the future growth of demand in each fuel should be estimated. If good historical data are available, linear forecasting is used to forecast future values based on a time series of historical data. The new values are predicted using linear regression, assuming a linear trend (y = mx + c) where the Y term corresponds to the variable to be forecast and the X term is years. Multiple regressions are used to predict the future growth of energy demand by sector, such as transport, industry, and the commercial and residential sectors. Whenever the energy intensity data is not available, energy demand equations are applied and could be summarized in the equations below:

$$\sum\_{i=1}^{10} TFEC\_{ijkt} = TFEC\_{1jkt} + TFEC\_{1jkt} + \dots + TFEC\_{10jkt} \tag{2}$$

where index (*i*) represents the country in ASEAN; index (*j*) is the energy consumption by sector; index (*k*) is the energy type; and index (*t*) is time. Thus, the Total Final Energy Consumption in ASEAN is the aggregated summation of all energy consumption in 10 countries of ASEAN. The TFEC of each country is derived from regression estimates of energy consumption by sector (*j*) and by energy type (*k*) at time (*t*) as follows:

$$TFEC\_{i\bar{j}kt} = \infty\_0 + \beta\_1 X\_{i\bar{j}kt} + \beta\_2 TFEC\_{i\bar{j}k(t-1)} + \varepsilon\_{i\bar{j}kt} \tag{3}$$

where:

*TFECijkt* is the Total Final Energy Consumption of country (*i*), by sector (*j*), by energy type (k), at time (t). *TFECijk*(*t*−1) is the lag variable. The variables *β*1*Xijt* are the independent variables including per capita gross domestic product, energy relative price, population growth, car ownership, and floor area. The summation of all energy demand by sector forms the Total Final Energy Consumption (TFEC). From here, the LEAP will further generate the Total Primary Energy Supply (TPES) by certain assumption of the losses in the transformation sector. Due to many equations run by each country in ASEAN, this study omits the results of the equation, and it shows only the results of the demand generated by LEAP.

In this modelling work using the LEAP application, the baseline for the 10 AMS was 2017—the latest available baseline data. For future energy demand, the projected demand growth is based on government policies, population and economic growth, and other key variable such as energy prices, using the International Energy Agency (IEA) world energy model (IEA, 2019) [22]. The BAU case is future predicted energy demand based on the government's current energy policies. However, the APSs are somewhat different to the BAU case in terms of policy changes and targets, as they have a greater share of renewables, including possible nuclear uptake if the government's alternative policies include nuclear as an energy option and more efficient power generation and energy efficiency in the final energy consumption.

Key variables and assumptions used in the model include the average annual growth rate of the population and the GDP, and energy efficiency and renewable targets (Figure 1 and Table 1).

**Figure 1.** Average Annual Growth Rate of GDP (%) and Population in AMS, 2017–2050. AMS = ASEAN Member States, ASEAN = Association of Southeast Asian Nations, GDP-AAGR = average annual growth rate of GDP, POP-AAGR = average annual growth rate of the population. Source: Authors' calculations.



AMS = ASEAN Member State, APS = alternative policy scenario, ASEAN = Association of Southeast Asian Nations, Lao PDR = Lao People's Democratic Republic. Source: Kimura and Phoumin (2019) [23].

This study generates and compares four possible energy mixes and their effects on greenhouse gas emissions over time. We break these pathways into four types, each with their own assumptions, including the APS, APS\_RE, APS\_EI, and APS\_EmT. The APS refers to Alternative Policy Scenario and assumes that states will increase more efficient final energy consumption, more efficient power generation, a higher share of renewables, and the introduction of nuclear power plants, based on each AMS government policy. The assumptions used in the APS are described in the table below. The APS\_RE is the APS with a higher share of renewable targets at the ASEAN level. In the APS\_RE, the targets are

increases of 23%, 30%, and 50% in the share of renewables in the primary energy supply by 2025, 2030, and 2050, respectively, from 2005 levels. The increase in the renewable share is expected from solar, wind, geothermal, and hydro. As hydro and geothermal energy are limited by resources, the maximum share is set based on the resource endowment. The APS\_EI is the APS using energy intensity reduction targets of 30%, 40%, and 50% from 2005 levels by 2025, 2030, and 2050, respectively. A greater reduction in energy intensity means that the energy consumption per unit of GDP becomes more efficient as a result of the application of energy efficiency, technological development, or any economic structural transformation of the economies shifting from energy-intensive sectors such as industry to less energy-intensive sectors such as services. The APS\_EmT is the APS using emission reduction targets of 40% and 80% from the BAU scenario by 2030 and 2050, respectively. This is the top-down policy target in which the energy mix composition needs to be changed towards cleaner energy to meet such targets. This will have many policy implications if the AMS wish to reduce emissions by as much as half from the BAU scenario by 2050.
