Low-carbon electricity generation through demand and carbon neutrality prediction 2035 of a transitional economy: decomposition and LEAP learning approaches

1 Introduction

The ever-increasing global climate change evolving worldwide is compelling societies to transition away from fossil fuels and plan pathways to a zero-emission future. To be efficient, the low-carbon transition will need to emerge at an unparalleled speed and scale across all regions and economic sectors (Vrontisi et al., 2020). Obviously, the transition to climate neutrality will pressure nations to undergo a transformation. Several countries around the world have become centers of heavy industry and resource extraction, adding significantly to the carbon footprint (Lu S. et al., 2024). Amid such huge environmental pressures, the low-carbon transition has arisen as a critical means to attain green and sustainable growth in developing countries (Balta-Ozkan et al., 2015). Nevertheless, despite the benefits to various countries, structural change will be marked by short- to long-run growth in energy prices and huge fuel imports (i.e., oil, coal, and gas).

As an emerging country, Pakistan is heavily dependent on resource-based industries, which are responsible for a large share of the country's CO₂ emissions (Lin and Raza, 2020). In addition, Pakistan accounted for 0.6% of global CO₂ emissions in 2018, ranking 27th among the world's CO₂ emitters (International Energy Agency, 2024), highlighting its significant share of the world's emissions. Recently, CO₂ emissions of different sectors (i.e., electricity, industry, transport, residential, commercial, and agriculture) in Pakistan have contributed 53.25, 36.03, 43.61, 17.66, 2.89, and 0.03 MtCO₂ (International Energy Agency, 2024). As a developing country, Pakistan uses more fuels every year by importing fossil fuels (88.87%) from the Gulf (Bakhsh et al., 2017), which is contributor to growing CO₂ emissions. CO₂ emissions from the electricity sector reached 155.55 MtCO₂, where oil, coal, and natural gas contributed 65, 41.9, and 48.75 MtCO₂ in 2024, respectively. Research on a low-carbon electricity transition in Pakistan has excelled at detecting internal factors of energy production, such as low-carbon electricity structure, generation intensity, energy security (ES), energy intensity (EI), carbon productivity, and CO₂ emissions.

The development in low-carbon electricity exemplifies the country's transition from industrial might to sustainable growth, contributing lessons and hope for an eco-friendly future. Heavily dependent on oil, coal, and gas, which has resulted in huge CO₂ emissions, Pakistan has turned toward expanding its low-carbon industry by exploiting low-carbon electricity sources since 2013 (Hydrocarbon Development Institute of Pakistan, 2024). As shown in Figure 1, Pakistan's electricity generation is based on four major sources (i.e., hydro, nuclear, renewable, and thermal electricity), showing that the generation trend is stable and growing steadily, which is evidence of the strong contribution of electricity generation.¹ Recently, it was noted that annual reductions of CO₂ emissions of 36.9-Gt are possible by 2050 using renewable energy, including the previously stated sources (International Renewable Energy Agency, 2021), where solar energy is the most plentiful energy source since it does not depend on direct sunlight. It is the most capable clean energy source, which is supposed to spur countries' social, economic, and environmental growth by replacing traditional energy sources (Hao et al., 2022; Xu et al., 2026). Thus, these sources are leading technologically and are significantly contributing to mitigating CO₂ emissions.

Comparative work at this level is rare, and analysis is naturally limited to a single sector or multiple countries. For example, Lin and Zhou (2022) used the entropy-weighted TOPSIS method to estimate the regional population, economic growth, urbanization, industrial structure, and environmental factors in China during the period 2000–2017 and found that all the factors had a significant impact on green development. Csereklyei et al. (2023) analyzed the coal revenues and growth in renewable energy production in Australia using wind, solar, and coal energy from 2017 to 2021. They estimated that a one-megawatt-hour rise in wind and solar energy generation would cause a reduction in coal generation. Sovacool and Valentine (2010) used the predictive framework under economic, policy, and nuclear power factors to evaluate the causality relationship between China and India. They concluded that these countries would fully develop nuclear energy capabilities to the point that it would eventually impact carbon reduction policies and sustainable development. Wang and Yan (2022) used decomposition analysis to determine the effects of various factors in China, such as fossil fuels, ESs, EI, economic growth, and CO₂ emissions between 2000 and 2016. They found that economic growth and population increased CO₂ emissions; however, energy structures have not significantly changed in CO₂ emissions, which will require substitution policies to reduce carbon emissions. Amponsah et al. (2014) examined heat coming from electricity generation, renewable energy technologies, and global emissions using the life cycle assessment method and found that fossil fuel electricity emissions were higher than those of renewable energy sources, with the exception of nuclear-based electricity generation. Li et al. (2024) used the Logarithmic Mean Divisia Index (LMDI) method to analyze the effects of electricity consumption, economic growth, energy efficiency, electricity intensity, fuel emission, and population factors in China in the period 1991–2020. They estimated that economic and fossil fuel factors were spurring CO₂ emissions; however, industrial and energy structures can mitigate CO₂ emissions and sustain economic growth.

However, this study discusses low-carbon electricity generation (LCEG) using six major driving factors and demand scenarios. To this end, the present research looks at which low-carbon electricity sources are linked with positive or negative designs. Under the country's economic planning framework, the country can forecast and support the relevant transition and decarbonize the electricity system. Thus, the study's core research questions are as follows: What are the low-carbon electricity factors that meet carbon neutrality objectives? What is the LCEG structure? What is the fuel demand and its mitigation projections? Presently, no research has analyzed these questions in a large sample of Pakistan. In an empirical assessment, several studies have concentrated on the analysis of electricity consumption and compared several renewable and social factors, for instance, Raza and Lin (2024), Shahbaz et al. (2024), Khan and Abbas (2016), and Valasai et al. (2017) employed traditional methods to analyze the relationships between energy, economics, and CO₂ emissions at national and sector levels. However, they failed to discuss low-carbon structure (LCS), electricity generation, ES, EI, carbon productivity, and carbon emissions in the country's policy analysis. Therefore, the significance is broadly recognized, and gaps in the literature leave room for further research.

Thus, the current study's contributions are as follows. First, it uses the six major driving factors to investigate LCEG in Pakistan and reveals the driving mechanism from the viewpoint of energy systems over four decades as per the Five-Year National Economic Planning and annual periods from 1985 to 2024. This efficiently fills a gap in current research that has largely depended on econometric models, giving valuable direction for Pakistan and related developing countries experiencing fast industrialization and urbanization. Surprisingly, despite the significance of Pakistan's low-carbon transition, there has been no published empirical study analyzing the association between LCEG factors in the recent period. Second, the study emphasizes the significant national impact of the electricity sector, representing substantial variations in the effects across various factors' contexts and related models. The finding highlights the need for decarbonization policies to be tailored to the specific characteristics of Pakistan, which is of great value for other countries with similar regional differences in economic growth, population and environmental challenges. Third, the study applies two methodologies [i.e., LMDI and Long-range Energy Alternatives Planning system (LEAP)] (Stockholm Environment Institute, 2008). These methods are valuable in estimating decomposing energy with the relative contributions of different factors and develop some changes in energy utilization, its awareness, and in-depth analysis of future energy. Thus, to quantify electricity demand projections, economic policies, and carbon emissions, it is necessary to predict energy demand, for which we employed the LEAP model in conjunction with the LMDI decomposition method, which has produced good outcomes (Wang et al., 2024). The integration between methods will provide an in-depth analysis of CO₂ emission leading factors and reveal threshold effects. Moreover, to forecast electricity demand and CO₂ emissions in the period 2025–2035, first, a time-series method is employed to check consumption patterns, which are identified as the additional mitigation potentials under five electricity consumption scenarios. Trends in a dataset over time will be estimated, and the estimate will be appropriate for predicting future values of data involving electricity consumption from various sources. Compared with past studies on driving factors of electricity-related factors and CO₂ emissions in Pakistan, these factors can provide a more comprehensive and detailed description, especially in the mitigation of CO₂ emissions under Energy Vision 2035 (2014), which focuses on more key issues of energy systems.

The remainder of this study is organized as follows. Section 2 describes the literature review. Section 3 presents the data, and Section 4 outlines the methods used. Empirical results and a discussion are presented in Section 5, while conclusions and policy suggestions are provided in Section 6.

2 Literature review

In the available literature, we found several studies related to energy and the environment at national, regional, and sector levels. Since Grossman and Krueger (1995) ground breaking work, several studies have been done on the interrelationship between economic and environmental factors. For instance, Gong and Zhang (2024) analyzed the role and mechanism of digital finance in China's sustainable growth in the period 2011–2020 and found that digital finance enhanced rural economic growth by changing the energy transition. Raza et al. (2024) employed the translog production model to check energy substitution and nonenergy factors for India and found that capital and energy were good substitutes in a country's economic growth. Boulanouar et al. (2024) used the Environmental Kuznets Curve (EKC) and Autoregressive Distributed Lag (ARDL) methods to examine energy, economic, and CO₂ emission mitigation and found that each factor's coefficient was higher in the short term vs. the long run. Tachega et al. (2024) employed the SMB-DEA and PMG-ARDL methods for Africa from 2010 to 2018 using energy, economics, and environmental quality and found that traditional and renewable energy were efficient at growth rates of 89% and 58%, respectively. In addition, the causes of CO₂ emissions have received substantial attention. For example, Acheampong (2018) used the panel autoregression method to check the causal relationship between 116 countries' economic growth, emissions, and energy consumption over the period 1990–2024. However, they found that CO₂ emissions caused economic growth. Saidi and Hammami (2015) employed the GMM panel model to analyze the impact of economic growth and CO₂ emissions on energy consumption for 58 countries for the period 1990–2012 and found that economic growth had a positive impact on energy use and CO₂ emissions. A similar finding of ASEAN countries was described by Saboori and Sulaiman (2013) using the ARDL method for the period 1971–2009, who determined that there was a significant relationship between energy and CO₂ emissions in all countries. Consequently, based on several findings at various national and regional levels, the energy revolution should be taken to enhance the sustainable use of energy, which will stabilize the estimated models and future framework to achieve sustainable development.

In the empirical literature, several studies (e.g., Lu S. et al., 2024; Hammond and Norman, 2012; Chong et al., 2019; Ozdemir, 2023; Raza and Lin, 2024; Huang, 2020) explored the relationship between various factors, such as EI, CO₂ emissions, gross domestic product (GDP), population, total energy, industrial energy consumption, and residential energy between 1978 and 2020. They suggested a dynamic relationship between energy and social indicators using econometric methods (see Table 1). Furthermore, several studies performed scenario analyses of future energy-related CO₂ emissions in numerous countries. Gan et al. (2013) conducted a quantitative analysis on a low-carbon society, Malaysia, to 2035. They estimated that alternative scenarios would enhance technologies and energy efficiency across various sectors while mitigating CO₂ emissions.

Table 1. Literature on decomposition analysis and its driving factors at local and international levels.

Recently, the rise in extreme weather events coming from globe has brought renewed interest in the need for CO₂ emission mitigation. In focusing on the topic, researchers have concentrated on forecasting CO₂ emissions to measure the influence of socioeconomic development on climate change. Ouedraogo (2017) employed the LEAP model it to investigate the energy demand and the carbon emissions under alternative strategies from 2010 to 2040. Song et al. (2007) analyzed chemical absorption in Korea under the economic and environmental policies at the time and found that these energy-related policies were useful in climate change and helped reduce CO₂ emissions. McPherson and Karney (2014) investigated long-term scenario alternatives for Panama's electricity sectors and found that the electricity composition, marginal cost, global warming, and resource diversity were forecasted under various development pathways. Nieves et al. (2019) analyzed Colombia's energy demand and greenhouse gas (GHG) emissions until 2030 and 2050 and estimated that economic growth and technological substitution of various sectors would be more efficient with greater CO₂ emission reductions in positive scenarios. Huang et al. (2023) employed LEAP and LMDI methods to determine the relationship between China's CO₂ emissions and energy consumption form 2000 to 2060 and found that CO₂ emissions were predicted to peak in 2028 and then fall to achieve carbon neutrality by 2060. Finally, Wang et al. (2024) predicted China's CO₂ emissions in Gansu province before 2030 and achieved carbon neutrality by 2060 using both LMDI and LEAP methods. They proposed that achieving carbon peak neutrality by 2030 would require a 29.53% reduction in EI from 2020, with 46.37 and 64% rises in electricity utilization and green energy generation, respectively. Thus, a huge mitigation of 123.84 Mt can be seen by 2060. Liya and Jianfeng (2018) employed a LEAP analysis for China's environmental and economic effects based on various electric resources during 2012–2050. They found that the natural gas scenario was better than the carbon capture scenario. Wakiyama and Kuramochi (2017) employed an ARMA model to predict CO₂ emissions from electricity in the residential sector of Japan. They determined that Japan could reduce electricity utilization and CO₂ emissions in 2030, which is greater than the Japanese post-2020 mitigation target.

In addition, Fang and Yang (2025) analyzed medium- and long-term energy demand in the Yangtze River Delta using the LEAP model from 2023 to 2060 and found that low-carbon energy would replace the current energy source after 2029. Lu L. et al. (2024) used the LEAP model to determine that GHG emissions for an industrial park would drop from 2020 to 2040 by 17%−30%. Shahid et al. (2025) forecasted fuel demand up to 2041 in Pakistan and found that growing dependence on cleaner energy and domestic resources rather than imported fuels would guarantee ES under sustainable development goals (SDGs). El-Sayed et al. (2023) estimated the Egyptian energy mix case from 2020 to 2050 and determined that low-carbon scenarios have minimum long-term production cost compared to natural gas, nuclear, and other fuels. Finally, Huang et al. (2023) employed LEAP and decomposition analysis for China's CO₂ commitments from 2030 to 2060. According to their investigation, emissions are predicted to rise in 2028, and China will need to cut about 2000 Mt, while other scenarios projected that China would be unbale to attain their carbo peak neutrality goals. Thus, it can be concluded that most earlier studies focused on energy, environment, and economic growth factors, with limited predictions for a specific industry or country. Some studies discussed the energy efficiency of renewable or low-carbon electricity using limited factors, for example, CO₂ emissions, GDP, total energy consumption, total electricity consumption (TEC), EI, and energy generation; however, low-carbon electricity structure, electricity intensity, and carbon production were seldom employed as driving factors in Pakistan, even though these factors may contribute significantly to CO₂ emissions. The argument on the factor relationship and CO₂ emissions has no data presented in the literature; hence, this study aims to fill this gap by analyzing the effects of LCS, electricity generation intensity, ES, EI, carbon productivity, and CO₂ emission factors on LCEG in Pakistan.

3 Data collection

The current study employed annual time-series data of four decades (1985–2024). Given the recent availability of data, we focused on six major variables, including LCEG, total electricity generation (TEG), total energy production, TEC, GDP, and CO₂ emissions. The study employs data for actual scenarios for electricity demand and mitigation based on renewable generation, starting from 2013 to 2024 and then predicted toward Renewable Energy Vision 2035. The data relating to electricity generation, energy production, electricity consumption, and energy production were collected from different annual books of the Hydrocarbon Development Institute of Pakistan (2024). All the energy-related data are organized in million tons of oil equivalent (Mtoe). The data relating to GDP were collected from the World Development Indicators (2025) and arranged in billions of US$. The data relating to CO₂ emissions were collected from the BP Statistical Review of World Energy and are organized in metric tons (Mt) of CO₂. Table 2 presents a description of all the variables involving mean, median, maximum, minimum, standard deviation, skewness, kurtosis, observation, and units. The analysis considers the changes in major variables in which GDP, CO₂, and TEP rank highest in all variables, signifying intense industrialization. The skewness of all variables is positive, showing a right tail. LCEG and TEG ranks lowest in all estimation, reflecting its lower intensity. Positive kurtosis means factors have a sharper peak and heavier tails compared to normal distribution. These differences emphasize the necessity of differentiated policy and decarbonization objectives.

Table 2. Description of variables.

LCEG includes the electricity generated through hydro, nuclear, wind, biofuels, and PV solar. TEG is the total electricity generated through thermal, hydro, nuclear, wind, biofuels, and PV solar. TEP is the sum of coal, oil, gas, LPG, and electricity including imported energy. TEC is the sum of thermal, hydro, nuclear, renewable, and auxiliary consumption. GDP is the overall gross domestic product, and CO₂ emissions are carbon dioxide emissions using the balances based on IPCC guidelines.

In addition, LCEG and carbon mitigation potential in Pakistan have been changing recently. The study framework provides factor decompositions and their estimations, which are essential for discussing the impact of each factor. The research framework is described in Figure 2.

4 Methodology and scenarios

4.1 The LMDI decomposition method

In the context of LCEG and CO₂ emissions, the key driving factors are measured by employing the LMDI method. To comprehend the effects of electricity generation, we developed a new approach to theoretical decomposition analysis following Ang et al. (1998). In the current study, we employed LMDI and additive methods to investigate low-carbon electricity and its factors' effects at various levels and dimensions. These estimations can be used to approximate the impacts of various indicators that provide useful information to governments and policymakers. Thus, this study analyzes Pakistan's national electricity generation and mitigation potential under the 5-year economic planning period of Pakistan. Technological progress, on the other hand, is another facet that is increasing the efficiency of LCEG to mitigate CO₂ emissions and enhance energy efficiency. The LMDI method has been widely employed to investigate the factors that impact energy utilization, social factors, and CO₂ emissions (Ang, 2004). Electricity is an important source, and there is some degree of substitution effect between electricity generation and socioeconomic factors, for instance, economic growth and the environment. Therefore, the internal structure and electricity sustainability impacts cannot be ignored when analyzing electricity generation factors. This method has several advantages; for example, it can be used to evaluate several issues without residual problems, estimate a zero value, be easily interpreted, and be convenient to apply. For this reason, several studies used this method in different ways, for instance, Zhang et al. (2013) and Li et al. (2024) for China's electricity and CO₂ emissions; Moutinho et al. (2015) for Eastern, Western, Northern, and Southern Europe; and Tang et al. (2024) for Pakistan's coal and economic development. In light of the preceding discussion, the decomposition method is suitable for developing a novel framework for measuring mitigation strategies to develop low-carbon electricity systems. Thus, this method includes decomposing the impacting factors into six major influencing factors: LCS, electricity generation intensity, ES, EI, carbon productivity, and CO₂ emissions. Employing Kaya's (1990) approach, the factors adding to low-carbon electricity and carbon mitigation are analytically estimated as follows in Equation 1:

LCEGt=∑iLCEGt=∑itLCEGtTEGt×TEGtTEPt×TEPtTECt×TECtGDPt×GDPtCO2t×CO2t(1)

where LCEG^*t*, TEG^*t*, TEP^*t*, TEC^*t*, GDP^*t*, CO2t, and t represent the low-carbon electricity generation, total electricity generation, total energy production, total electricity consumption, gross domestic product, carbon dioxide emissions, and time at the national level. Based on Equation 1, the influencing factors can be described as

=∑itLCSt×EGIt×ESt×EIt×CO2Pt×Ct(2)

where LCS^*t*, EGI^*t*, ES^*t*, EI^*t*, CO2Pt, and C^*t*represent the low-carbon structure, electricity generation intensity, energy security, EI, carbon productivity, and carbon emissions. Moreover, each factor in the model includes assumptions that are considered for policy measures. The first factor, LCSt=LCEGtTEGt, shows that if the numerical value is higher, the low-carbon electricity structure will be enhanced. The second factor, EGIt=TEGtTEPt, higher renewable electricity generations intensity will be useful. The third factor, ESt=TEPtTECt, is called energy security. If the numerical value of ES^*t* is higher, then the effect of LCEG^*t*will be at its maximum. Also, the contribution of ES^*t* to renewable energy highlights the country's energy self-sufficiency, which shows that the positive values reveal an energy production surplus, strengthening ES^*t* and renewable initiatives. Thus, evaluating renewable energy efficiency is integral to achieving carbon neutrality. The fourth factor, EIt=TECtGDPt, is called energy intensity. A lower EI^*t*value is preferred. The fifth factor, CO2Pt=GDPtCO2t, is called carbon production. If the numerical value of CO2Pt is higher, then the maximum economic growth will be seen with minimum pollution. The sixth factor, C^*t*, is called the carbon emission factor. If the numerical value is lower, then green development is progressing, which will also show that fuel efficiency and technical enhancements are applicable in mitigating CO₂ emissions. Thus, employing the LMDI method, the change in LCEG from the base year (1985) to the target year (2024) is estimated as

ΔLCEGt=LCEGt-LCEG0=ΔLCEGLCS+ΔLCEGEGI +ΔLCEGES+ΔLCEGEI+ΔLCEGCO2P+ΔLCEGC(3)

where Δ shows the change in overall growth in LCEG. However, ΔLCEG_*LCS*, ΔLCEG_*EGI*, ΔLCEG_*ES*, ΔLCEG_*EI*, ΔLCEG_{C_*O*2P}, and ΔLCEG_*C* indicate impacts of low-carbon structure, electricity generation intensity, ES, EI, carbon productivity, and carbon emissions, respectively. Thus, the effects of each factor's decomposition are expressed as in Equations 4–10:

ΔLCEGLCS=∑(LCEGt-LCEG0)(ln LCEGt-ln LCEG0)×ln (LCStLCS0)(4)

ΔLCEGEGI=∑(LCEGt-LCEG0)(ln LCEGt-ln LCEG0)×ln (EGItEGI0)(5)

ΔLCEGES=∑(LCEGt-LCEG0)(ln LCEGt-ln LCEG0)×ln (EStES0)(6)

ΔLCEGEI=∑(LCEGt-LCEG0)(ln LCEGt-ln LCEG0)×ln (EItEI0)(7)

ΔLCEGCO2P=∑(LCEGt-LCEG0)(ln LCEGt-ln LCEG0)×ln (CO2PtCO2P0)(8)

ΔLCEGC=∑(LCEGt-LCEG0)(ln LCEGt-ln LCEG0)×ln (CtC0)(9)

However, the aggregated LCEG^*t*represents the amount of changeability, as follows in Equation 10:

=(LCEGt−LCEG0)(lnLCEGt−lnLCEG0)(ln(LCStLCS0)+ln(EGItEGI0)+ln(EStES0)+ln(EItEI0)+ln(CO2PtCO2P0)+ln(CtC0))(10)

4.2 The LEAP model estimation

We integrate the LEAP and LMDI methods to conduct an in-depth analysis of the factors of these emissions. The research applies the LEAP model to measure electricity demand and CO₂ emission mitigation potential of Pakistan's electricity sector. The LEAP-Pakistan model employed in the current study takes into account electricity demand under various fuels (i.e., TEC, hydro, nuclear, thermal, and renewable energy) and CO₂ emission mitigation potential scenarios from each fuel type in Pakistan. The analysis covers the period from 2013 to 2035. The major components of the analysis are various electricity-related fuels and CO₂ emission mitigation under each fuel over the forecasted period. Each level includes the process of energy substitution and efficiency. For this, we employed three scenarios, i.e., electricity forecast [also called business as usual (BAU)], low-demand, and high-demand scenarios. The LEAP model is an energy-environment tool that was developed at the Stockholm Environment Institute (Stockholm Environment Institute, 2008) to measure the impacts of physical, economic, environmental, and technological initiatives. The model is useful and gives wide information on technical characteristics, prices, present practices, and energy production (Jun et al., 2010; Özer et al., 2013). Being a full energy accounting framework, it considers both demand- and supply-side technologies and considers overall system effects. Thus, the LEAP model explains how energy is utilized, substituted, and generated under the limitations of selected energy variables, such as hydro, renewable, nuclear, and thermal. The method is based on a bottom-up electricity generation simulation process, involving production capacity with the capacity factor of various fuels. As per the CO₂ mitigation scenario in the target year 2035, it can be estimated based on low-carbon energy generation. Thus, annual gross electricity production (G) analysis decomposes each fuel, which can be estimated as follows in Equation 11:

Gyear=∑ingi(11)

where i is the fuel type, n represents the energy employed for the electricity generation, and g_*i* is the gross electricity generation from fuel i. Moreover, the primary electricity demand for g_*i* is estimated through the method with the electricity production per unit of electricity consumption (Mtoe), as shown in Equation 12:

(EC)i=∑i(gi*0.0000010)/φi(12)

where EC is the electricity consumption in Mtoe of fuel i. (EC)_*h*, (EC)_*t*, (EC)_*n*, and (EC)_*r*denote the electricity consumption of hydro, thermal, nuclear, and renewable. φ_*i*is the fuel efficiency (%) of fuel i, and 0.0000010 is used for unit conversion as 1Toe = 0.0000010 Mtoe. Thus, various production and consumption fuel types in different scenarios have varying results respecting electricity demand. The logical structure in the LEAP model is presented in Figure 3. The framework is based on baseline, high-demand, and low-demand scenarios. The process can be summarized in four main steps: electricity production as per the demand projection, electricity types, fuel-related CO₂ emissions, and carbon mitigation potential.

4.3 Scenario criteria for electricity demand and CO₂ mitigation pattern to 2035

Scenarios for electricity demand patterns up to 2035 are fixed under the assumption of the reduction potential of electricity demand in Pakistan, as presented in Table 3. Table 3 shows four types of energy-related scenarios (i.e., TEC scenario 1, TEC-BAU, TEC-low demand, and TEC-high demand scenarios) from the base year 2013 to the current year 2035. Aside from the BAU scenario, three different scenarios, based on common access to modern energy services consistent with the low-carbon energy transition initiatives, are developed. The CO₂ emissions mitigation scenarios are developed to assess pollution reduction under the initiatives of Renewable Energy Vision 2035 of Pakistan. The objective is to see the reduction potential; hence, the low-mitigation scenario is taken up. These scenarios reveal future consumption patterns and how further actions can maintain or reduce the level of electricity combustion by establishing five electricity-related scenarios under BAU, low demand, and high demand from 2013 to 2035 using the LEAP model. A more in-depth description is given in Table 3.

Table 3. Electricity demand and CO₂ mitigation scenarios.

5 Empirical results and discussion

5.1 Factor decomposition of LCEG under economic planning period

This section gives econometric outcomes and a discussion. To obtain relevant outcomes, decomposition and LEAP methods are applied to check the factor effects and predict the mitigation potential of CO₂ emissions. LCEG in Pakistan grew by 4.15% and CO₂ emissions increased by 3.13% from 1985 to 2024. Thus, it is necessary to analyze electricity generation and related factors to investigate the country's sustainability. Furthermore, we employed the decomposition method to estimate the major driving factors, which are directly associated with electricity generation, economic growth, and pollution. These factors are low-carbon structure (ΔLCEG_*LCS*), electricity generation intensity (ΔLCEG_*EGI*), ES (ΔLCEG_*ES*), EI (ΔLCEG_*EI*), carbon productivity (ΔLCEG_{C_*O*2P}), carbon emissions (ΔLCEG_*C*), and total factor effects (ΔLCEG_*total*) within the economic planning period (i.e., 1985–1989, 1990–1994, 1995–1999, 2000–2004, 2005–2009, 2010–2014, 2015–2019, 2020–2024, and 1985–2024). Following Equations 4–9, the decomposition results are described in Figure 4.

Figure 4 describes the various driving factors that add to the impact of LCEG. It is obvious that the total effects of ΔLCEG_*total* during 1985–2024 show an optimistic impact of 415.95 Mtoe. Also, we found that there were positive and significant growing impacts of ΔLCEG_*total*during the planning period excluding 1995–1999. The reduction in TEG during this period was estimated due to the consumption boom, spurring the use of electrical appliances and energy imports (Consortium for Development Policy Research, 2018). The maximum impact was seen in the current planning period (2020–2024), with a growth rate of 167.68 Mtoe, contributing to significant development, while the third planning phase (1995–1999) experienced negative growth of −4.53 Mtoe.

The ΔLCEG_*LCS* effect shows variations, with the most significant growth rate of 157.20 Mtoe occurring in 2020–2024, making an improved contribution of low-carbon electricity. Recently, the percentage contributions of hydro, nuclear, renewable, and thermal electricity were 25.4%, 8.4%, 6.8%, and 59.4%, respectively (Pakistan Economic Survey, 2023–2024). These are explained by the start of the China-Pakistan Economic Corridor (CPEC) power projects, which have addressed historical gaps in electricity production and enhanced sustainability. However, before 2013, renewable energy was not part of the electricity mix, and even thermal power, a dominant electricity supply source, has declined over the past few decades. For example, the energy sector in Pakistan is the biggest carbon emitter, which has recently increased its renewable energy production capacity from all kinds of renewable energy sources (e.g., solar, wind, biomass, bagasse) (Hydrocarbon Development Institute of Pakistan, 2024). It will try to meet low-carbon electricity demand and sustain the economic development of Pakistan. In addition, the country's dependence on imported and costly fossil fuels for electricity generation emphasizes the need for a substitution in the fuel mix. For this, Pakistan has set a target to mitigate carbon emissions by 50% by 2030, and clean energy resources will play a dominant role in achieving this objective.

The ΔLCEG_*EGI*effect had positive and negative fluctuations. The optimistic figures for 1985–1989, 1990–1994, and 2015–2019 highlighted the enhancements in electricity generation efficiency by 8.17, 3.04, and 54.13 Mtoe, respectively. However, we found that the maximum period shows a negative share due to technological development and reliance on imported fuel, which is not only affecting the country's economy but also spurring CO₂ emissions. The renewable electricity era from 2013 onward is a positive sign of growing renewable electricity, which is consistent with the findings of Raza and Lin (2022), who showed that Pakistan could engage in energy substitution, economic development, and inter-fuel substitutions; however, the huge technological costs associated with renewable energy will impact actual progress. The ΔLCEG_*ES* effect shows that electricity generation is maximally positively significant in most phases, excluding 1985–1989, 1995–1999, and 2015–2019. The positive growth in the 1990–1994, 2000–2004, 2005–2009, 2010–2014, and 2020–2024 planning periods show that there is an energy production excess, strengthening ES and renewable energy production. As shown in Figure 4, the ΔLCEG_*ES* during the phases 1985–1989, 1995–1999, and 2015–2019 reflects the negative effect by 0.08, 17.77, and 57.91 Mtoe, showing a reduction in EI linked with economic progress. However, the maximum reduction in EI was estimated during the COVID-19 epidemic. It is clear that the EI (ΔLCEG_*EI*) shows a positive effect when ES is negative or show low impact, which shows a negative impact on energy development. Moreover, a positive changeover seemed to occur during 1985–2024, with a positive growth of 32.27 Mtoe/billion US dollars, which suggests that Pakistan still relies on imported energy with an efficient and productive relationship between energy inputs and desired outputs. Thus, the outcomes suggest that achieving the goal with respect to CO₂ emissions mitigation in Pakistan is necessary for a basic energy infrastructure.

Concerning the ΔLCEG_{C_*O*2P}effect, a mixed trend has been observed. The positive trend during the first three phases (1995-1999; 2000–2004 and 2005–2009) shows that the economy is efficiently progressing with a reduction in CO₂ emissions. It is evident that the country's economy and pollution reduction have been significant since the start of the renewable energy period (2013) with negative effects, which is also consistent with the study of Qudrat-Ullah (2022), who proposed that Pakistan can significantly achieve CO₂ emissions targets and electricity consumers can save 23% annually on their electricity bills. The CO₂ emission (ΔLCEG_*C*) effect on LCEG shows a maximum aggregated emissions trend by 462.90 Mt during 1985–2024, respectively. As shown in Figure 4, all the economic periods show a positive fluctuation with increasing or decreasing trends. However, it can be seen that during the most recent phase, 2020–2024, there was a reduction in CO₂ emissions of 0.49% to 35.94 Mt. Overall, it is noted that CO₂ emissions remained 462.90 Mt and positively affected LCEG during the period. Finally, it can be seen that the investigated factors are of utmost importance in providing a comprehensive model for understanding and developing low-carbon electricity productivity.

5.2 LCEG and factor decomposition based on annual effects during the period 1985–2024

To further estimate the annual variation in the driving factors, we checked the effects of each factor during the estimated period, as shown in Figure 5. The analysis of each factor provides a mixed direction for the substitution of renewable energy. It is obvious that most of the factors show a wide fluctuation in the current decade because of the renewable contribution and indigenous production. According to the Pakistan Economic Survey (2023–2024), Pakistan is taking initiatives to meet its energy demand and reduce GHG emissions. For instance, several scholars have shown that Pakistan's electricity is the biggest sector in demand and supply, growing energy-related import bills, and growing environmental issues (Lin and Raza, 2020; Tang et al., 2024). For this reason, the government is actively pursuing large-scale renewable energy projects under CPEC and Pakistan's Energy Vision 2035 (Government of Pakistan, 2014) to achieve its global (i.e., Paris Agreement) and local clean energy goals. Despite the positive trends in EI and CO₂ emission mitigation, there has been a possible reduction in the considered factors, suggesting a slow transition toward more sustainable electricity consumption, especially after 2013. For example, there have been variations in the share of low-carbon electricity since 2000, which leads to mitigating CO₂ emissions in the future. The addition of renewable energy efficiency, a sectoral low-carbon energy portfolio, and significant industrial productivity will yield insights into sustainability and efficiency. This analysis will permit policymakers, researchers, and industrialists to make future decisions regarding multiple investments in different aspects.

5.3 Electricity demand and potential scenario

Electricity consumption is the major cause of demand. For this, we employed electricity-related indicators of the considered scenarios, including total electricity demand, hydro, thermal, nuclear, and renewable electricity demand in the period 2013–2035. The scenario analysis takes 2013 as the base year and 2035 as the final year. These inputs are needed to investigate the level of electricity activities as a whole. These demand projection were selected from the base period of 2013 because renewable electricity (i.e., biofuel, solar, and wind) first started in this period (Hydrocarbon Development Institute of Pakistan, 2024), which will provide insights for sound and progressive government policies. All the data are projected up to 2035 in order to predict Pakistan's Energy Vision 2035 (2014). Furthermore, the additional values after 2023–2035 are estimated under Equations 11, 12 with their prediction results under a significance level of 5%. Three scenarios were generated in this model and used in the analysis: baseline scenario, also called BAU, low-demand scenario, and high-demand scenario. However, the low-demand scenario, also called the mitigation scenario, assumes the implementation of more determined energy conservation and emission mitigation actions. Thus, the outcomes of each fuel scenario are presented in Figures –. These scenarios will benefit the energy sector and policymakers and give direction for technological and sustainable enhancement.

First, the total electricity demand of Pakistan from 2013 to 2024 is applied (Hydrocarbon Development Institute of Pakistan, 2024) as a criterion to regulate the share of overall consumption and way of growth in forecasting electricity in Pakistan for the study, as shown in Figure 6. As shown in Figure 6, the electricity demand share from 2025 to 2035 shows substantial growth. It is estimated that the share of Pakistan's electricity consumption was 6.25 Mtoe in 2013 and 9.60 Mtoe in 2024, with an average growth rate of 0.53%. During the period between 2025 and 2035, the share of electricity consumption is estimated to add about 0.21%, which means that the share of the electricity sector is consistently growing. The outcomes are consistent and in line with Raza and Lin (2025), who proposed that fuel substitution and structure are useful in growing low-carbon energy. This shows that electricity-related fuel can be further adjusted at the maximum suitability under low-carbon demand.

Second, total hydroelectricity demand shows a zigzag growth pattern from 2013 to 2024, with a growing trend of 0.31%, as shown in Figure 7. Also, Jamil (2020) used the ARIMA model for hydroelectricity consumption prediction for 2030 and determined that no dams will be constructed due to financial issues, which has created a margin between the installed and available potential of hydroelectricity in Pakistan. The hydroelectricity shares from 2025 to 2035 contributed an average growth rate of 0.14%, which means that hydroelectricity contributes significantly over time. The historical measures show that precipitation variations and decreasing water availability have damaged riverine ecology, impaired water security, and influenced hydropower production (Pakistan Economic Survey, 2023–2024). Growth in renewable energy and hydroelectricity from 33 to 62% by 2031 will come through various projects, such as CPEC, and the promotion of local resources, inviting private companies, and building small dams (Integrated Generation Capacity Expansion Plan, 2022).

Third, thermal electricity is similar to hydroelectricity, which on average decreased by 0.58% during 2013–2024, as shown in Figure 8. The thermal electricity share will on average decline by 0.85% between 2025 and 2035, which shows that thermal electricity will decrease due to renewable energy substitution. If we consider the low-demand scenario, thermal electricity will decline by 3% in 2035. For example, the government of Pakistan has decided to add about 6,000 MW of PV solar capacity to quickly reduce pollution from energy generation. The results are consistent with a study by Tahir and Asim (2018), who proposed that this capacity will impact supply and demand of electricity, leading to several hours of power cuts. Consequently, the government should activate its natural resources, including natural gas, solar systems, biomass, and wind, for example, for potential development with the help of international companies.

Fourth, in 1972, Pakistan was the world's 15th nuclear energy producer under Pakistan Atomic Energy. Due to project life cycles, nuclear energy is linked to multiple projects and shows continuous growth. As shown in Figure 9, nuclear electricity demand grew by 4.85% from 2013 to 2024, while demand between 2025 and 2035 will arrive by 0.77%, respectively. This shows that nuclear electricity has the maximum possibility of growing into a future energy source. For instance, nuclear power plants in Pakistan are a reliable source of electricity. These plants can run for a longer time and store enough energy on-site, which can protect against short- and long-run variations and capacity factors. According to Raza and Tang (2024), nuclear energy is clean and has zero GHG emissions. Thus, the low-demand and high-demand scenarios show an increasing and consistent trend over the period. Furthermore, nuclear electricity is the best energy source, which cut about 10 Mt of GHG during 2023; thus, the high-demand scenario is the best choice (Pakistan Economic Survey, 2023–2024), which can be enhanced by several nuclear power plants.

Finally, renewable energy (i.e., solar, wind and biomass) has shown an increasing trend over the period. As presented in Figure 10, Pakistan's renewable energy grew by 7.62% during 2013–2024, because Pakistan's energy sector is paving the way toward transitioning from imported fossil fuels to renewable energy sources, for instance, through substantial investment in solar and wind energy. This is consistent with Shakeel et al. (2016), who found that renewable power generation in 2013 started with a determined goal of removing the gap between generation and consumption. Based on huge investments in renewables, the predictions were estimated at an average growth rate of 0.75% between 2025 and 2035. For example, the government of Pakistan has been concentrating on strengthening the regulatory framework and incentivizing private sector investment in renewable energy that will guarantee ES and climate change mitigation. Moreover, the private power infrastructure board is easing 24 power generation projects, of which 22 are based on renewable energy.

5.4 Fuel-based CO₂ emission mitigation scenario from 2013 to 2035

Based on decomposition results and scenario analysis, it is necessary to adopt a carbon mitigation scenario. The modeling framework is applied to investigate and project the energy demand and CO₂ emission mitigation under the planning for the period 2013–2023. The long-term prediction of energy demand and carbon mitigation is of key importance in Pakistan due to the steady growth in energy needs, the fuel import dependence (i.e., oil, energy technologies, and gas), and the global concerns over energy-induced climate issues. In this scenario, we divide the scenario into five fuels, including TEC, hydro, thermal, nuclear, and renewable electricity. This scenario is adopted to examine CO₂ mitigation from the existing electricity resources in Pakistan. This scenario is imperative because it reduces CO₂ emissions under the condition that the installed power offers a maximum advantage. No power plants are operating under capacity, and all low-carbon electricity consumption sources except thermal are used. In this scenario, we have considered only the mitigation scenario (low-demand scenario), and the carbon reduction of each energy source is described from 2023 to 2035, as shown in Table 4. However, it is obvious that the installed electricity from 2013 to 2022 will not change, and hence carbon emissions from various fuels are expected to decrease due to a rising capacity of low-carbon electricity except thermal electricity (Hydrocarbon Development Institute of Pakistan-HDIP, 2024). Thus, the main outcomes related to CO₂ emission mitigation by fuel type are presented in Table 4. The electricity-related CO₂ emissions were reduced by 6.45% yearly between 2013 and 2035. However, electricity output increases at an average rate of 4.34% per year during the estimated period. For instance, a study by Özer et al. (2013) analyzed Turkey's electricity sector using fossil fuel consumption and investigated the prediction analysis until 2030. According to their analysis, CO₂ emissions will grow due to fossil fuel consumption to meet electricity demand, and mitigation potential is possible in growing renewable energy. Hence, it can be said that energy models can be developed based on available renewable energy sources to promote energy access, permit a transition to cleaner energy, and reduce pollution for a developing economy.

Table 4. CO₂ emissions mitigation scenario by electricity fuel type, 2013–2035 (unit: Mt).

Figure 11 presents the variation in yearly CO₂ emissions through the prospect of the study for BAU and CO₂ mitigation scenarios. If the maximum share of low-carbon electricity becomes accessible from 2025 onward, the reduction in CO₂ emissions until 2035 is expected to be 68.551%. For instance, a study by Sadiqa et al. (2018), who discussed the electricity transition roadmap for Pakistan's overall energy demand by 2050, found that CO₂ emissions can be significantly reduced, by 100%. Thus, the cumulative CO₂ emissions would be reduced by an average of 0.026 Mt from 2025 to 2035. As shown in Figure 11, the BAU scenario is established first as a benchmark, and a mitigation scenario is established because of the gradual growth share of low-carbon electricity consumption. As shown in Figure 1, the electricity generation of various fuels is increasing as a result of the growing electricity demand and the mitigation potential of CO₂ emissions. Therefore, Pakistan should implement strong renewable energy policies to sustain economic development and mitigate pollution and ES while meeting local and international demand.

6 Conclusion and policy recommendations

6.1 Conclusion

The present research analyzes the major electricity generation and CO₂ emissions in Pakistan, spanning the data from 1985 to 2024, employing LMDI and LEAP models. Based on empirical analysis, the main outcomes are as follows.

First, using the decomposition analysis, the six major factors (i.e., low-carbon structure, electricity generation intensity, ES, EI, carbon productivity, and CO₂ emissions) are divided into the Five-Year Economic Planning period, and then all the factors' effects are analyzed annually. The results prove that aggregated LCEG has a rising impact; however, a positive and growing impact has been observed in the current decade, particularly in the current planning period. Similarly, the low-carbon structure factor provided amazing results during 2020–2024, showing a significant impact on the country's economic and environmental development. The intensity of electricity generation and ES demonstrate technological progress; however, the negative impact of EI has a significant influence on energy technology. This proves that Pakistan can adopt technologies in the energy sector to fight ES, CO₂ emission reduction, and economic sustainability. Carbon production effects show that the economy of Pakistan is improving due to significant advances in LCEG, which is lowering CO₂ emissions. Consequently, a clear reduction, of 23.77%, has been observed in the current phase, showing that Pakistan is committed to renewable agreements, such as the Paris Agreement and Vision 2035 (2014).

Second, electricity demand-based scenarios show that total electricity demand, nuclear, and renewable electricity are increasing in the baseline and high-demand scenarios. This guarantees that the low-carbon electricity will continue to grow and can be a factor in the country's security.

Finally, various electricity sources provide a significant contribution to mitigating CO₂ emissions. As Pakistan is an emerging country and relies heavily on fossil fuels, CO₂ emissions cannot be fully stopped from the baseline scenario. However, we estimated that the mitigation scenario between 2013 and 2035 would reduce emissions by 6.45% annually accompanied by an increasing rate of electricity of 4.34%. Thus, the current study provides relative outcomes for policymakers to put forth national and international-level policies for mitigating long-run CO₂ emissions in the electricity sector of Pakistan, which can be discussed in environmental discussions.

6.2 Policy suggestions

First, based on current results, we noticed that a low-carbon electricity structure was useful for energy efficiency. Thus, the government should pay attention to and encourage long-term investments, renewable energy technologies, and dam reconstruction for hydroelectricity, which will ultimately increase ES and reduce pollution. For example, 250 engineering structures, including dams, ponds, spillways, and drainage systems, are under consideration and can be utilized to increase electricity efficiency.

Second, comparison of scenario analyses suggested that Pakistan could cut CO₂ emissions by 6.45% using low-carbon electricity up to 2035. To do this, ES efforts can be increased by various electricity-consuming sectors or by substituting fossil fuel electricity for low-carbon electricity-generating resources, i.e., hydro, nuclear, wind, biofuels, and photovoltaic (PV) solar by following the long-term goals of the Paris Agreement to achieve net zero emissions in the second half of the century.

Finally, we could not find low-carbon electricity costs or sectoral information. Thus, further study and scenarios can be investigated in comparing costs and sectoral progress. Energy models support monitoring energy indicators, measuring the time effects of energy consumption, and evaluating progress on SDGs. Moreover, energy prediction models use machine learning with a focus on renewable energy, load, and cost forecasting. These prediction models will be employed to formulate policies with extrinsic inputs (i.e., population, income, growth factors, and technologies) to measure energy utilization patterns. Also, to review the additional mitigation potential of the current scenario under Energy Vision 2035, further studies should be conducted on the mitigation potential of electricity consumption in other sectors and overall energy demand. Also, electricity demand forecasts are based on economic growth and population; hence, these factors might provide efficient results.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Xiaonan Wei, China University of Geosciences, China

Reviewed by: Izzet Ari, Social Sciences University of Ankara, Türkiye

Lu Gao, University of South China, China

Footnotes

LEAP, Long-range Energy Alternatives Planning System; LMDI, Logarithmic Mean Divisia Index; TEP, total energy production; TEG, total electricity generation; TEC, total electricity consumption; GDP, gross domestic product; CO₂, carbon dioxide; LCS, low-carbon structure; EGI, electricity generation intensity; ES, energy security; CO₂P, carbon productivity; (EC)_*h*, electricity consumption of hydro; (EC)_*t*, electricity consumption of thermal; (EC)_*n*, electricity consumption of nuclear, and renewable; (EC)_*r*, electricity consumption of renewable.

Footnotes

Note that low-carbon electricity generation includes hydro, nuclear, wind, biofuels, and photovoltaic (PV) solar during the estimated period. However, renewable energy is the aggregate of wind, biofuels, and PV solar.