Exploring the role of geothermal energy consumption in achieving carbon neutrality and environmental sustainability Authors: Ghalieb Mutig Idroes, Mohd Afjal, Mohsin Khan, Mohammad Haseeb, Irsan Hardi, Teuku Rizky Noviandy, Rinaldi Idroes Published: 2024 ABSTRACT While there is extensive research on the environmental impacts of renewable energy sources, a notable gap remains in understanding the specific environmental effects of geothermal energy. This knowledge shortfall is particularly critical for Indonesia, which holds the world's second-largest geothermal potential but has yet to effectively harness these valuable resources. This study aims to address this gap by providing a preliminary evaluation of the dynamic impact of geothermal energy consumption and various macroeconomic variables on carbon dioxide (CO2) emissions in Indonesia from 1995 to 2020. The study employs the Fully Modified Ordinary Least Squares (FMOLS) method and conducts robustness checks using the Dynamic Ordinary Least Squares (DOLS) method. Our findings reveal that geothermal energy consumption, military expenditure, and technology innovation contribute to a reduction in CO2 emissions, while non-renewable energy consumption increases CO2 emissions. Additionally, the pairwise Granger causality test results indicate a unidirectional causality running from geothermal energy consumption to CO2 emissions, suggesting that trends in geothermal energy use can reliably predict changes in CO2 levels. This study provides valuable insights for policymakers and stakeholders in formulating sustainable policies for carbon reduction in Indonesia. Abbreviations: Augmented Dickey-Fuller Autoregressive Distributed Lag Carbon Dioxide Emissions Cumulative Sum of Recursive Residuals Cumulative Sum of Squares of Recursive Residuals Dynamic Ordinary Least Squares Environmental Kuznets Curve Fully Modified Ordinary Least Squares Gross Domestic Product Geothermal Energy Consumption Capital Formation Globalization Green Solow Model International Energy Agency Labor Force Manufacturing Value Added Military Expenditure Non-renewable Energy Our World in Data Phillips-Perron Technology Innovation World Development Indicator Introduction The growing threat of climate change, evidenced by rising temperatures, sea levels, and more frequent extreme weather events, demands a swift transition to cleaner energy sources [[1], [2], [3], [4], [5]]. Energy consumption is a key driver of global warming and climate change, with two primary categories: non-renewable and renewable energy [[6], [7], [8], [9]]. Non-renewable energy has been proven to have detrimental environmental impacts, while renewable energy is recognized as a more sustainable alternative [10,11]. Consequently, reducing the reliance on non-renewable energy and prioritizing renewable energy sources has become a central goal for countries worldwide [[12], [13], [14], [15]]. Renewable energy sources encompass solar, hydro, wind, tidal, geothermal, and biomass [10,[16], [17], [18]]. Among these, geothermal energy stands out as one of the most sustainable due to its minimal carbon footprint and consistent energy output. Geothermal energy's relevance in combating climate challenges is increasingly acknowledged, with empirical evidence supporting its role as a critical solution [[19], [20], [21], [22]]. Studies by Raihan et al. [23] and Terrapon-Pfaff et al. [24] advocate for a rapid transition to renewable energy sources, emphasizing geothermal energy's contribution to reducing CO2 emissions in leading geothermal-consuming countries [25]. Geothermal energy offers a distinctive advantage over other renewable sources due to its reliability and independence from external meteorological conditions [22,26,27]. Unlike intermittent renewables that depend on sunlight or wind, geothermal energy harnesses the Earth's internal heat, providing a stable and sustainable power source [28]. This reliability is derived from the Earth's inexhaustible internal heat, making geothermal energy a cornerstone of sustainable energy strategies [19]. Research by Umar et al. [25] and Bilgili et al. [29] reinforces the premise that geothermal technology offers a resilient energy solution with significantly lower environmental impacts compared to traditional fossil fuels. Given these attributes, geothermal energy is increasingly recognized as a viable alternative for achieving environmental sustainability and mitigating climate change. The integration of geothermal with other renewable energy sources is particularly strategic in regions like Indonesia, where geographical and environmental conditions favor such synergies [30,31]. Research by Olabi et al. [32] underscores the potential for hybrid renewable systems that combine geothermal with solar power to optimize energy reliability and reduce dependency on imported oil and gas. In Indonesia, these hybrid systems could play a transformative role, enhancing energy independence and supporting the country's shift away from its current reliance on coal. Explorations into using geothermal as a primary power source, particularly during peak loads supplemented by other renewables, could significantly contribute to national energy stability and security [33,34]. Indonesia possesses the world's second-largest geothermal reserves, positioning it advantageously to capitalize on this clean energy resource [22,35,36]. The nation's rich volcanic terrain harbors extensive biodiversity and offers vast potential for sustainable energy production [22,36]. Despite this, Indonesia's current energy infrastructure is predominantly coal-dependent, significantly contributing to its rising CO2 emissions. Embracing geothermal energy could be transformative, positioning Indonesia as a leader in the pursuit of carbon neutrality and enhancing environmental sustainability. The collaborative research by Atiz et al. [37] illustrates the potential benefits of integrating geothermal with solar energy, showcasing how such innovative approaches can amplify energy production while curtailing emissions. Further exploration into geothermal energy exploitation, as proposed by Gupta & Vashistha [38] and Liu et al. [39], underlines the significant potential for geothermal solutions to enhance Indonesia's energy landscape and meet its environmental targets. These innovative approaches could stabilize the national grid and play a pivotal role in transitioning Indonesia towards a more sustainable and economically viable energy future. Although the advantages of renewable energy are widely acknowledged, the specific impact of geothermal energy on Indonesia's CO2 emissions remains insufficiently explored. Studies by Deng et al. [40] and Gong et al. [41] have underscored the operational efficiencies of geothermal systems; however, the wider economic and environmental ramifications of expanding geothermal energy within the Indonesian context are not well documented. This knowledge gap is a critical barrier to formulating effective policies and strategies that could maximize the environmental benefits of Indonesia's natural geothermal assets. The theoretical foundation of this study is anchored in environmental economics and the principles of sustainable development. Furthermore, this study employs the Green Solow Model (GSM), a theoretical framework developed by Brock & Taylor [42], which expands upon previous models to illustrate the interdependent relationship between economic growth, environmental degradation, and renewable energy usage. By transitioning to renewable energy, reliance on fossil fuels decreases, leading to a reduction in environmental degradation. The GSM builds on Solow's classic economic growth theory by integrating environmental considerations [43]. This theoretical model has been applied in previous research by Kwakwa & Aboagye [44] and Idroes et al. [21], investigating the impact of various macroeconomic variables on the environment. The analysis includes key control variables to strengthen the results: GDP, non-renewable energy use, globalization, manufacturing, military expenditure, technology innovation, capital formation, and labor. Including these factors helps ensure more reliable and comprehensive findings. Through an in-depth analysis of geothermal energy's role in Indonesia's energy and environmental landscape, this study investigates the impact of geothermal energy consumption on CO2 emissions from 1995 to 2020, employing the Fully Modified Ordinary Least Squares (FMOLS) method, supported by the Dynamic Ordinary Least Squares (DOLS) method. It provides valuable insights into leveraging geothermal potential to achieve sustainability and climate goals. By exploring the relationship between geothermal energy and emissions, this study offers a critical resource for policymakers, stakeholders, and the global community to develop effective strategies for a sustainable energy transition. Addressing the lack of literature on geothermal energy's impact on emissions using an econometric approach, the study enriches understanding within the Indonesian context and paves the way for future research and policies to achieve carbon neutrality and environmental sustainability globally. The paper is structured as follows: Section 2 reviews the relevant literature, Section 3 describes the research gap, Section 4 presents the materials and methods, Section 5 presents the results and discussion, and Section 6 concludes with policy recommendations, limitations, and suggestions for future research. Review of literature The relationship between geothermal energy utilization and CO2 emissions reduction is critical in global efforts to combat climate change. Geothermal energy's capacity to provide a clean, renewable energy source has been documented to significantly reduce greenhouse gas emissions. The study by Umar et al. [25] specifically highlights the efficacy of geothermal energy in lowering the carbon footprint of nations with substantial geothermal consumption. This evidence supports the premise that transitioning to geothermal energy can be a strategic component of national and global strategies to achieve carbon neutrality. The hypothesis posits that environmental degradation increases with economic growth up to a certain point, after which it begins to decline as societies prioritize and have the means to invest in cleaner technologies and practices [21]. With its low operational costs and minimal environmental footprint, geothermal energy embodies the transition towards more sustainable economic development models [29]. Further research in this area could explore specific policy measures and investments required to harness geothermal energy's full potential for sustainable economic development. The advancement of technology plays a pivotal role in unlocking the full potential of geothermal resources. Innovative approaches, such as those proposed by Huang et al. [45], demonstrate the feasibility of accessing previously untapped geothermal reserves, thereby expanding the scope of geothermal energy production. However, the path to widespread adoption of these technologies is fraught with challenges, including high initial capital costs, technical complexity, and the need for specialized expertise. Overcoming these barriers is essential for countries like Indonesia to exploit their geothermal resources fully. Ahmed et al. [26] provide an in-depth analysis of the efficiency of shallow geothermal systems in both heating and cooling applications, a theme that finds resonance in the work of Christodoulides et al. [46], who dive into the modeling system, thereby bolstering the case for shallow geothermal energy's pivotal role in sustainable energy strategies. This discourse is further expanded by Kumar et al. [47], who navigate through the technological evolution within the geothermal sector, acknowledging both the significant strides made and the persistent hurdles that challenge further advancements. Parallel to these discussions, the application of geothermal energy in sustainable water desalination has garnered notable attention, particularly from Shelare et al. [48], their research underscores the critical utility of geothermal energy in devising sustainable desalination processes, a crucial development given the escalating global demand for potable water and the imperative for environmentally benign desalination methods. Al-Qadami et al. [27] explore geothermal energy's potential in pavement energy harvesting, demonstrating its versatility and contribution to a renewable energy mix while showcasing its capacity to drive innovative sustainable solutions. The efficiency and environmental benefits of geothermal energy are comprehensively illustrated by Cadelano et al. [49], who demonstrate how integrating geothermal energy with adept energy management policies can significantly enhance building energy efficiency while concurrently reducing CO2 emissions. Their focused study on a museum setting exemplifies the wider applicability of geothermal solutions across diverse infrastructures, further advocating for geothermal energy's role in cost reduction and sustainability enhancement. Agemar et al. [50] offer a detailed overview of deep geothermal energy production in Germany, highlighting the complex potential and the accompanying technical and regulatory challenges inherent in deep geothermal system exploration. This examination highlights Germany's commitment to developing geothermal resources while revealing the complex challenges of deploying deep geothermal systems. In an evolving energy landscape, the dynamic interactions between various energy sources, industrial development, and carbon emissions have become a focal point of research, especially within emerging economies. Shahzad et al. [51] employ the Autoregressive Distributed Lags (ARDL) methodology to dissect the interplay between coal, geothermal energies, industrial advancement, and carbon emissions within the Philippines, unveiling the nuanced relationship these factors share. Their findings illuminate geothermal energy's capacity to reduce carbon emissions while fostering industrial growth, a narrative gaining traction amidst the global pivot towards renewable energy sources. In Indonesia, renowned for its significant geothermal reserves, Idroes et al. [22] explore the environmental footprint of biomass versus geothermal energy, elucidating the superior environmental profile of geothermal energy. This comparison is pivotal for advocating geothermal energy's broader adoption to diminish carbon emissions. The temporal relationship between geothermal energy usage and CO2 emissions is meticulously analyzed by Bilgili et al. [29], who investigate these dynamics across various frequency cycles. Their insights affirm that escalating geothermal energy use is synonymous with substantial CO2 emission reductions, aligning with global sustainability goals. Furthering this discourse, Alsaleh et al. [19] evaluate geothermal power's impact on carbon dioxide emissions, reinforcing the premise of geothermal energy as a feasible solution for emission reduction and sustainability pursuit. This body of work contributes to the expanding evidence supporting renewable energy's role in climate change mitigation. The economic and technological aspects of geothermal energy consumption and its environmental impact in newly industrialized countries have been thoroughly examined [52]. They underscore the significance of economic complexity and technological innovation in amplifying geothermal energy's environmental benefits, positing geothermal consumption as a crucial determinant in ecological sustainability achievement. Alsaleh & Wang [53] probe into the impact of Information and Communication Technology (ICT) on geothermal energy sustainability, suggesting that ICT could significantly enhance geothermal resource management and utilization efficiency. This perspective opens new pathways for leveraging technology to bolster geothermal energy sustainability. It is forecasted geothermal energy development's impact on CO2 emission reductions by 2030, highlighting its substantial potential to contribute to global climate change mitigation efforts. Their projections offer a strategic framework for integrating geothermal energy into comprehensive climate action plans. Finally, Aissa & Hartono [54] evaluate the impact of the geothermal energy sector's development on Indonesia's electricity sector, delineating both the economic benefits and challenges associated with geothermal energy exploitation. These studies provide a rich tapestry of insights into geothermal energy's multifarious impacts on carbon emissions, economic development, and sustainability. Research gap Current state of research Renewable energy research has greatly expanded our understanding of how energy sources interact with economic and environmental sustainability. However, there remains a significant gap in our knowledge regarding the impact of geothermal energy within Indonesia's unique context. Despite being a global leader in geothermal production, detailed insights into how geothermal energy interacts with various economic variables to affect CO2 emissions are lacking. This is particularly critical as Indonesia seeks to leverage its extensive geothermal resources to achieve carbon neutrality and improve environmental sustainability. Identified gap This study identifies a specific research gap: the insufficient analysis of geothermal energy's role within Indonesia's broader economic framework, especially in relation to CO2 emissions. Existing research has not comprehensively explored how geothermal energy consumption correlates with critical economic factors such as economic growth, non-renewable energy consumption, globalization, manufacturing value added, military expenditure, technology innovation, capital formation, and labor force. Study objectives and contributions The objective of this research is to conduct a thorough examination of the effects of geothermal energy consumption on CO2 emissions in Indonesia, analyzing data from 1995 to 2020. By incorporating an extensive range of economic variables—economic growth, non-renewable energy consumption, globalization, manufacturing value-added, military expenditure, technology innovation, capital formation, and labor—this study will provide a nuanced analysis of the interactions between geothermal energy and these factors. The methodology section will elaborate on the methodological details, including the conversion of all data to natural logarithms to address heteroscedasticity. This approach allows us to focus on the current discussion's identified research gap and objectives. Significance of addressing the research gap A significant gap in research has been identified regarding the impact of geothermal energy consumption on CO2 emissions in Indonesia. The findings from this investigation will be utilized to inform emission reduction policies and sustainable development initiatives. The insights gained will be available to policymakers, stakeholders, and researchers concerned with optimizing Indonesia's geothermal resources. The contribution to existing literature on the environmental impacts of renewable energy is expected to be substantial, and Indonesia's geothermal energy sector is positioned to be recognized as essential to achieving both environmental and economic targets. Materials and methods Variable description and data sources This study utilizes annual data from 1995 to 2020 in Indonesia. The selection of this timeframe was driven by the availability of geothermal energy data, with these years providing the most comprehensive dataset. The decision to conclude the dataset in 2020 was specifically made to avoid the impact of the COVID-19 pandemic, as the economic disruptions during this period would introduce significant shocks to the data, potentially skewing the regression results. In this research, we employ proxies for several variables to better capture the underlying concepts. For example, geothermal electricity generation is used as a proxy for geothermal energy consumption, CO2 emissions represent environmental degradation, and patent applications are utilized as a proxy for technological innovation. Table 1 offers a detailed overview of the variables. To address potential heteroscedasticity, all data were transformed into their natural logarithms.Table 1The summary of the variable.Table 1Types of VariableVariableUnit of MeasurementDescription & Abbreviations of the VariablesSourceIndependentGeothermal Energy ConsumptionElectricity Generation in Gigawatt-hours (GWh)Geothermal energy (GEO) utilizes Earth's heat for sustainable power, offering a low-emission alternative that minimizes environmental impact compared to traditional sources like fossil fuels.IEADependentCO2 emissionsKilotonnes (kt)Carbon dioxide (CO2) emissions, arising from human activities like the burning of fossil fuels, serve as a proxy measure for environmental degradation.WDIControlEconomic GrowthConstant 2015 (US$)Economic growth is the sustained increase in a nation's production and consumption of goods and services, measured by rising Gross Domestic Product (GDP). It contributes to higher standards of living, increased employment, and overall prosperity.WDINon-renewable EnergyTerawatt-hour (TWh)Non-renewable energy (NRE) comes from limited and cannot be quickly replaced by sources like coal, oil, natural gas, or nuclear power; using these sources can lead to environmental issues such as pollution and climate change.OWIDGlobalization IndexIndex (scale 1–100)Globalization (GLB) involves eliminating barriers to trade and encouraging trade, affecting economic, social, and political activities worldwide.KOFManufacturing Value addedConstant 2015 (US$)Manufacturing value added (MANU) measures the value created during production, reflecting the economic contribution and increased worth of manufactured goods.WDIMilitary ExpenditureCurrent (US$)Military expenditure (MIL) represents the total government spending on its military forces, encompassing defense-related investments, personnel salaries, and the procurement of equipment and services to maintain and enhance national security.WDITechnology InnovationTotal patent applications (residents & nonresidents)Technology innovation (TECH) refers to creating new and innovative technological solutions, including developing novel products, systems, or methods that contribute to progress and advancements in various fields.WDICapital FormationConstant 2015 (US$)Gross fixed capital formation (GFCF) refers to the total value of investments made in a country's physical assets, such as buildings, machinery, and infrastructure, during a specific time period.WDILaborTotal labor (Person)The labor force (LF) encompasses the total number of individuals who are either employed or actively seeking employment within a specific region or country. It serves as a key demographic for assessing workforce participation and economic productivity.WDINote: IEA – International Energy Agency; WDI – World Development Indicators; OWID – Our World in Data; KOF – KOF Globalisation Index. Table 1The summary of the variable.Table 1Types of VariableVariableUnit of MeasurementDescription & Abbreviations of the VariablesSourceIndependentGeothermal Energy ConsumptionElectricity Generation in Gigawatt-hours (GWh)Geothermal energy (GEO) utilizes Earth's heat for sustainable power, offering a low-emission alternative that minimizes environmental impact compared to traditional sources like fossil fuels.IEADependentCO2 emissionsKilotonnes (kt)Carbon dioxide (CO2) emissions, arising from human activities like the burning of fossil fuels, serve as a proxy measure for environmental degradation.WDIControlEconomic GrowthConstant 2015 (US$)Economic growth is the sustained increase in a nation's production and consumption of goods and services, measured by rising Gross Domestic Product (GDP). It contributes to higher standards of living, increased employment, and overall prosperity.WDINon-renewable EnergyTerawatt-hour (TWh)Non-renewable energy (NRE) comes from limited and cannot be quickly replaced by sources like coal, oil, natural gas, or nuclear power; using these sources can lead to environmental issues such as pollution and climate change.OWIDGlobalization IndexIndex (scale 1–100)Globalization (GLB) involves eliminating barriers to trade and encouraging trade, affecting economic, social, and political activities worldwide.KOFManufacturing Value addedConstant 2015 (US$)Manufacturing value added (MANU) measures the value created during production, reflecting the economic contribution and increased worth of manufactured goods.WDIMilitary ExpenditureCurrent (US$)Military expenditure (MIL) represents the total government spending on its military forces, encompassing defense-related investments, personnel salaries, and the procurement of equipment and services to maintain and enhance national security.WDITechnology InnovationTotal patent applications (residents & nonresidents)Technology innovation (TECH) refers to creating new and innovative technological solutions, including developing novel products, systems, or methods that contribute to progress and advancements in various fields.WDICapital FormationConstant 2015 (US$)Gross fixed capital formation (GFCF) refers to the total value of investments made in a country's physical assets, such as buildings, machinery, and infrastructure, during a specific time period.WDILaborTotal labor (Person)The labor force (LF) encompasses the total number of individuals who are either employed or actively seeking employment within a specific region or country. It serves as a key demographic for assessing workforce participation and economic productivity.WDINote: IEA – International Energy Agency; WDI – World Development Indicators; OWID – Our World in Data; KOF – KOF Globalisation Index. The summary of the variable. Note: IEA – International Energy Agency; WDI – World Development Indicators; OWID – Our World in Data; KOF – KOF Globalisation Index. Theoretical framework and empirical model This study employs the Green Solow Model (GSM), a theoretical framework developed by Brock & Taylor [42], which expands upon previous models to illustrate the interdependent relationship between economic growth, environmental degradation, and renewable energy usage. By transitioning to renewable energy, reliance on fossil fuels decreases, leading to a reduction in environmental degradation. GSM builds on Solow's classic economic growth theory by integrating environmental considerations [43]. It posits that while production generates positive economic outputs [55], it also produces negative environmental externalities. Utilizing a Cobb-Douglas production function, Solow's model allows for substituting capital and labor [56]. This research further enhances Solow's model by incorporating a resource variable (R) to explicitly account for environmental factors, as demonstrated in Equation (1).(1)Q=f(K,L,R)orQ=KαLβRγWhere α, β, and γ denote positive parameters, specifically, α > 0, β > 0, and γ > 0. Additionally, Q symbolizes both economic growth and environmental degradation, while K, L, and R stand for the quantities of capital, labor, and natural resources (specifically, geothermal energy). This study explored the linkage between geothermal energy and carbon dioxide emissions (CO2), a proxy measure of environmental degradation. In line with recent literature on the determinants and sources of CO2 emissions, the empirical model shown in Equation (2) is formulated to achieve this objective.(2)CO2t=fGEOt,GDPt,NREt,GLBt,MANUt,MILt,TECHt,GFCFt,LFtWhere CO2 represents carbon dioxide emissions, GEO indicates geothermal energy, GDP stands for economic growth, NRE refers to non-renewable energy, GLB represents globalization, MANU denotes manufacturing value-added, MIL indicates military expenditure, TECH stands for technology innovation, GFCF represents capital formation, and LF refers to the labor force. Before analysis, all variables undergo a transformation into logarithmic form to simplify complexity and improve the clarity of the findings. Equation (2) can be expressed in log-linear econometric model formats, as Equation (3) demonstrated.(3)lnCO2t=β0+β1lnGEOt+β2lnGDPt+β3lnNREt+β4lnGLBt+β5lnMANUt+β6lnMILt+β7lnTECHt+β8lnGFCFt+β9lnLFt+εtWhere the term β0 denotes the intercept, while εt signifies the error term, the coefficients β1 through β9 are indicative of estimated parameters. It is expected that β1 through β9 will exhibit positive and negative signs, respectively. Moreover, Johansen [57] cointegration tests are a common method to assess the presence of cointegration among the datasets, while this study employs the ARDL bounds test for cointegration. Establishing a cointegration relationship between variables is essential for conducting subsequent analyses. Data stationary method with ADF and P-P In assessing the stationarity of time series data, we utilize two common tests: the Augmented Dickey-Fuller (ADF) and Phillips-Perron (P-P) tests. The ADF test examines the presence of a unit root, indicating non-stationarity, while the P-P test accounts for autocorrelation and heteroscedasticity. Both tests entail comparing the test statistic to critical values to determine stationarity [58,59]. This methodology provides a robust framework for verifying the stationarity of time series data, which is a crucial step in conducting accurate analyses. Cointegration with ARDL bounds test This study employed the Autoregressive Distributed Lag (ARDL) bounds testing approach developed by Pesaran et al. [60] to detect cointegration between variables. The ARDL method offers distinct advantages, particularly its ability to handle series with different orders of integration without requiring all variables to be integrated in the same order [61,62]. It also performs well with small sample sizes, providing reliable estimates of long-run relationships [17,63,64]. Therefore, the ARDL bounds test is effective whether the variables are integrated at I(0), I(1), or are fractionally integrated [65]. The ARDL bounds test is represented in Equation (4).(4)ΔYt=α0+∑i=1pαiΔYt−i+∑j=0qφjΔXt−j+δ1Yt−1+δ2Xt−1+εt The ARDL bounds test equation, presented as Equation (4), includes ΔYt, representing the first difference of the dependent variable Y at time t, which captures short-term fluctuations in Y. The intercept term α0 provides a baseline level independent of other variables. Coefficients αi reflect the short-run dynamics of Y through lagged differences, while φj capture similar dynamics for the independent variable X. Furthermore, ‘p’ and ‘q’ denote the optimum lag lengths for each variable. Coefficients δ1 and δ2 measure the long-run impacts of the lagged levels of Yt−1 and Xt−1, respectively, highlighting the sustained effects of both variables over time. Furthermore, The ARDL bounds test for our model is represented in Equation (5).(5)ΔlnCO2t=β0+∑i=0pβ1ΔlnCO2t−i+∑i=0qβ2ΔlnGEOt−i+∑i=0qβ3ΔlnGDPt−i+∑i=0qβ4ΔlnNREt−i+∑i=0qβ5ΔlnGLBt−i+∑i=0qβ6ΔlnMANUt−i+∑i=0qβ7ΔlnMILt−i+∑i=0qβ8ΔlnTECHt−i+∑i=0qβ9ΔlnGFCFt−i+∑i=0qβ10ΔlnLFt−i+δ1lnCO2t−1+δ2lnGEOt−1+δ3lnGDPt−1+δ4lnNREt−1+δ5lnGLBt−1+δ6lnMANUt−1+δ7lnMILt−1+δ8lnTECHt−1+δ9lnGFCFt−1+δ10lnLFt−1+εt Furthermore, the null and alternative hypotheses for no cointegration among the variables in Equation (5) are.H0δ1=δ2=δ3=δ4=δ5=δ6=δ7=δ8=δ9=δ10=0Haδ1≠δ2≠δ3≠δ4≠δ5≠δ6≠δ7≠δ8≠δ9≠δ10≠0 δ1=δ2=δ3=δ4=δ5=δ6=δ7=δ8=δ9=δ10=0 δ1≠δ2≠δ3≠δ4≠δ5≠δ6≠δ7≠δ8≠δ9≠δ10≠0 The ARDL bounds test uses the F-statistic and critical values by Pesaran & Timmermann [66] to assess cointegration. The test involves estimating an ARDL model via ordinary least squares to explore long-run equilibrium relationships among variables [62]. The null hypothesis posits no cointegration. If the F-statistic exceeds the upper critical value, cointegration is indicated, and the null hypothesis is rejected. The null hypothesis cannot be rejected if it is below the lower critical value. When the F-statistic falls between the critical values, the result is inconclusive. This test helps determine if long-run relationships exist between variables. The robustness check of FMOLS with DOLS Fully Modified Ordinary Least Squares (FMOLS) is a technique devised by Phillips & Hansen [67] to estimate cointegrating regressions while addressing issues such as endogeneity and serial correlation commonly encountered in time series analysis. By incorporating lagged values of both the dependent and independent variables into the regression equation, FMOLS corrects for endogeneity, ensuring consistent estimation of long-run parameters. This method efficiently estimates long-run relationships in cointegrated time series data, making it a valuable tool in econometrics. Dynamic Ordinary Least Squares (DOLS), developed by Stock & Watson [68], is another method used to estimate cointegrating regressions in time series analysis, particularly in serial correlation and heteroscedasticity. DOLS extends the standard OLS regression by including lagged first differences of the variables, capturing the dynamic adjustment process in the long-run relationship. This approach ensures robust estimation of the long-run parameters while allowing for flexibility in modeling the relationship between variables. DOLS is widely employed in econometrics for its ability to handle dynamic specifications and address issues of serial correlation and heteroscedasticity effectively. Additionally, this research employs DOLS as a robustness check for the FMOLS regression. This approach has been used by several researchers, including Raihan & Tuspekova [62], Pattak et al. [69], Hardi et al. [70], Idroes et al. [10], Hardi et al. [71] and Kwakwa & Aboagye [44], to strengthen the initial findings. By using DOLS as a complementary technique to FMOLS, we ensure the reliability and robustness of our findings [62,71,72]. This approach allows us to corroborate the results obtained from FMOLS estimation, further strengthening the validity of our analysis. Through the integration of both FMOLS and DOLS methodologies, this research endeavors to provide a comprehensive and rigorous examination of the long-run relationships within the analyzed time series data [69,73]. The pairwise granger causality test This study extends its analysis by incorporating pairwise Granger causality tests [74]. Granger causality examines whether one variable can forecast another, with past values of the former enhancing predictions of the latter. It's crucial to understand that Granger causality implies predictive rather than true causality. However, identifying Granger causality can illuminate potential causal relationships between variables. The pairwise Granger causality approach involves testing every possible variable pair, revealing dynamic interdependencies and influence mechanisms not evident in standard analyses. By integrating these tests, we gain deeper insights into temporal precedence and potential influence pathways, enriching our understanding of complex interactions within the time series data. The summary of the stages of the analysis process The workflow of the study, shown in Fig. 1, presents a step-by-step analytical process that begins with selecting variables, followed by choosing the data period, conducting descriptive statistics, performing unit root tests with ADF and P-P, checking for cointegration with the ARDL Bound Test, applying the FMOLS method, assessing robustness with DOLS, analyzing causality with pairwise Granger causality tests, engaging in a discussion of the results, and concludes with summarizing findings and offering policy recommendations.Fig. 1Workflow of the study.Fig. 1 Workflow of the study. Results and discussion Descriptive statistics Table 2 summarizes descriptive statistics for key variables across 26 observations. CO2 emissions have a mean of 392,772, ranging from a minimum of 223,679 to a maximum of 605,291, with a standard deviation of 110,048. GEO averages 7894 with a wide range from 2210 to 15,563 and a standard deviation of 3771. GDP shows a high mean of 6.31E+11, varying between 3.74E+11 and 1.05E+12, with considerable dispersion (standard deviation of 2.28E+11). NRE, non-renewable energy, has a mean of 1,478, a range from 839 to 2,093, and a standard deviation of 364. GLB has a mean of 60.10, spanning from 50.37 to 64.44 with minimal variation. MANU, manufacturing, averages 1.36E+11, ranging from 7.96E+10 to 2.13E+11, with substantial variability. MIL has a mean of 4.40E+09, ranging from 9.19E+08 to 9.39E+09, with a standard deviation of 2.82E+09. TECH shows a mean of 5,733, ranging from 1846 to 11,481, with moderate variability. GFCF has an average of 1.98E+11, ranging from 8.82E+10 to 3.49E+11. Finally, LF has a mean of 1.12E+08, spanning from 8.87E+07 to 1.36E+08, with a standard deviation of 1.45E+07.Table 2Descriptive statistics.Table 2VariableCO2GEOGDPNREGLBMANUMILTECHGFCFLFMean39277278946.31E+111477.9660.101.36E+114.40E+0957331.98E+111.12E+08Median37804776655.75E+111499.9760.861.30E+113.27E+0948731.74E+111.12E+08Max.605291155631.05E+122093.4664.442.13E+119.39E+09114813.49E+111.36E+08Min.22367922103.74E+11839.4850.377.96E+109.19E+0818468.82E+1088673402Std. Dev.11004837712.28E+11363.823.664.33E+102.82E+0925838.26E+1014502788Obs.26262626262626262626 Table 2Descriptive statistics.Table 2VariableCO2GEOGDPNREGLBMANUMILTECHGFCFLFMean39277278946.31E+111477.9660.101.36E+114.40E+0957331.98E+111.12E+08Median37804776655.75E+111499.9760.861.30E+113.27E+0948731.74E+111.12E+08Max.605291155631.05E+122093.4664.442.13E+119.39E+09114813.49E+111.36E+08Min.22367922103.74E+11839.4850.377.96E+109.19E+0818468.82E+1088673402Std. Dev.11004837712.28E+11363.823.664.33E+102.82E+0925838.26E+1014502788Obs.26262626262626262626 Descriptive statistics. Unit root test The unit root test results in Table 3 show that all variables are stationary in the first difference. Both the ADF and P-P tests significantly reject the null hypothesis of a unit root at the 1 % level, uniformly supporting this finding. Both of these unit root tests reliably indicate that the variables are integrated into order one. Determining the integration order sets the stage to accurately specify the cointegration framework used later in the analysis.Table 3The results of ADF and P-P unit root test.Table 3VariablesADFP-Pt-stat.Prob.t-stat.Prob.At LevellnCO2−1.32880.5998−2.88230.0617lnGEO−1.66740.4337−1.55100.4919lnGDP0.46680.98190.43680.9805lnNRE−2.43300.1434−4.4062∗0.0020lnGLB−3.5102∗∗0.0162−3.5085∗∗0.0163lnMANU−0.49930.8756−0.48090.8793lnMIL−0.62510.8478−0.59630.8545lnTECH−1.43280.5500−1.24730.6372lnGFCF−0.13150.9353−0.14340.9339lnLF−1.26970.6254−1.53260.5010At First DifferencelnCO2−5.2238∗0.0003−6.8562∗0.0000lnGEO−3.7201∗0.0108−3.6782∗∗0.0114lnGDP−3.7619∗0.0095−3.7619∗0.0095lnNRE−4.3460∗0.0025−4.2366∗0.0032lnGLB−4.1379∗0.0050−3.4775∗∗0.0179lnMANU−4.6904∗0.0011−4.6882∗0.0011lnMIL−5.0684∗0.0004−5.0696∗0.0004lnTECH−5.8089∗0.0001−7.3007∗0.0000lnGFCF−3.6595∗∗0.0119−3.6062∗∗0.0134lnLF−4.1467∗0.0041−4.0748∗0.0046Note: ∗∗ and ∗ indicate significance levels at 5 % and 1 %, respectively. Table 3The results of ADF and P-P unit root test.Table 3VariablesADFP-Pt-stat.Prob.t-stat.Prob.At LevellnCO2−1.32880.5998−2.88230.0617lnGEO−1.66740.4337−1.55100.4919lnGDP0.46680.98190.43680.9805lnNRE−2.43300.1434−4.4062∗0.0020lnGLB−3.5102∗∗0.0162−3.5085∗∗0.0163lnMANU−0.49930.8756−0.48090.8793lnMIL−0.62510.8478−0.59630.8545lnTECH−1.43280.5500−1.24730.6372lnGFCF−0.13150.9353−0.14340.9339lnLF−1.26970.6254−1.53260.5010At First DifferencelnCO2−5.2238∗0.0003−6.8562∗0.0000lnGEO−3.7201∗0.0108−3.6782∗∗0.0114lnGDP−3.7619∗0.0095−3.7619∗0.0095lnNRE−4.3460∗0.0025−4.2366∗0.0032lnGLB−4.1379∗0.0050−3.4775∗∗0.0179lnMANU−4.6904∗0.0011−4.6882∗0.0011lnMIL−5.0684∗0.0004−5.0696∗0.0004lnTECH−5.8089∗0.0001−7.3007∗0.0000lnGFCF−3.6595∗∗0.0119−3.6062∗∗0.0134lnLF−4.1467∗0.0041−4.0748∗0.0046Note: ∗∗ and ∗ indicate significance levels at 5 % and 1 %, respectively. The results of ADF and P-P unit root test. Note: ∗∗ and ∗ indicate significance levels at 5 % and 1 %, respectively. Cointegration with ARDL bound test results The ARDL bound test results in Table 4 show the F-statistic of 7.4072 exceeds the upper and lower critical values at the 5 % significance level. This rejects the null hypothesis of no cointegration, indicating the variables are cointegrated. The significant F-statistic points to a long-term relationship between the variables. With confirmed cointegration, the analysis can proceed using the ARDL approach to estimate both long-term and short-term effects.Table 4The ARDL bounds test results.Table 4F-bounds testNull Hypothesis:No levels relationshipTest StatisticValueSignif.I(0)I(1)F-statistic7.4072At 10 %1.82.8K9At 5 %2.042.08At 2.5 %2.243.35At 1 %2.53.68 Table 4The ARDL bounds test results.Table 4F-bounds testNull Hypothesis:No levels relationshipTest StatisticValueSignif.I(0)I(1)F-statistic7.4072At 10 %1.82.8K9At 5 %2.042.08At 2.5 %2.243.35At 1 %2.53.68 The ARDL bounds test results. FMOLS and robustness check with DOLS The dynamic impact of the independent variable on CO2 emissions is presented in Table 5 through the FMOLS results. In the FMOLS model, GEO has a negative impact on CO2 emissions; a 1 % increase in GEO is associated with a decrease in CO2 emissions of 0.1462 %. Furthermore, NRE exhibits a positive relationship with CO2 emissions, where a 1 % increase in NRE leads to a 0.8035 % increase in CO2 emissions. Additionally, GLB has a negative effect on CO2 emissions, indicating that a 1 % rise in GLB corresponds to a 0.5654 % decrease in CO2 emissions. Furthermore, MIL also shows a negative impact on CO2 emissions, with a 1 % increase in MIL leading to a decrease of 0.0795 % in CO2 emissions. Then, TECH has a negative impact on CO2 emissions, where a 1 % increase in TECH is correlated with a 0.0821 % reduction in CO2 emissions.Table 5Results of the FMOLS test.Table 5VariableCoeff.t-Stat.Prob.lnGEO−0.1462∗∗−2.80360.0134lnGDP0.13090.23100.8204lnNRE0.8035∗5.33240.0001lnGLB−0.5654∗∗∗−2.09660.0534lnMANU0.54101.24890.2309lnMIL−0.0795∗∗−2.58830.0206lnTECH−0.0821∗∗−2.59710.0202lnGFCF−0.0476−0.35790.7254lnLF0.60950.91440.3750C−14.3948−1.89240.0779R20.9938R2 Adjusted0.9900Note: ∗∗∗, ∗∗, and∗ indicate the significance level at 10 %, 5 %, and 1 %, respectively. Table 5Results of the FMOLS test.Table 5VariableCoeff.t-Stat.Prob.lnGEO−0.1462∗∗−2.80360.0134lnGDP0.13090.23100.8204lnNRE0.8035∗5.33240.0001lnGLB−0.5654∗∗∗−2.09660.0534lnMANU0.54101.24890.2309lnMIL−0.0795∗∗−2.58830.0206lnTECH−0.0821∗∗−2.59710.0202lnGFCF−0.0476−0.35790.7254lnLF0.60950.91440.3750C−14.3948−1.89240.0779R20.9938R2 Adjusted0.9900Note: ∗∗∗, ∗∗, and∗ indicate the significance level at 10 %, 5 %, and 1 %, respectively. Results of the FMOLS test. Note: ∗∗∗, ∗∗, and∗ indicate the significance level at 10 %, 5 %, and 1 %, respectively. Furthermore, the DOLS robustness check presented in Table 6 confirms the significant negative impact of GEO on CO2 emissions, implying that a 1 % increase in GEO reduces CO2 emissions by 0.1254 %. It also confirms the positive impact of NRE on CO2 emissions, with a 1 % increase in NRE resulting in a 0.8261 % increase in CO2 emissions. Moreover, MIL shows a significant negative effect on CO2 emissions; a 1 % increase in MIL is associated with a 0.0692 % decrease in CO2 emissions. Additionally, TECH continues to display a significant negative impact on CO2 emissions, where a 1 % increase in TECH leads to a 0.0717 % decrease in CO2 emissions. These results suggest that increases in geothermal energy usage and technology innovation are associated with reductions in CO2 emissions. In contrast, increases in non-renewable energy usage are linked to increases in emissions. Furthermore, higher military spending is consistently associated with lower CO2 emissions in both models.Table 6Robustness results using the DOLS test.Table 6VariableCoeff.t-Stat.Prob.lnGEO−0.1254∗∗−2.28900.0360lnGDP0.27640.49720.6258lnNRE0.8261∗4.80830.0002lnGLB−0.3872−1.25960.2259lnMANU0.33640.80450.4329lnMIL−0.0692∗∗∗−2.04680.0575lnTECH−0.0717∗∗∗−1.98520.0645lnGFCF−0.0414−0.29930.7685lnLF0.38230.53270.6016C−10.4438−1.31690.2064R20.9950R2 Adjusted0.9921Note: ∗∗∗, ∗∗, and∗ indicate the significance level at 10 %, 5 %, and 1 %, respectively. Table 6Robustness results using the DOLS test.Table 6VariableCoeff.t-Stat.Prob.lnGEO−0.1254∗∗−2.28900.0360lnGDP0.27640.49720.6258lnNRE0.8261∗4.80830.0002lnGLB−0.3872−1.25960.2259lnMANU0.33640.80450.4329lnMIL−0.0692∗∗∗−2.04680.0575lnTECH−0.0717∗∗∗−1.98520.0645lnGFCF−0.0414−0.29930.7685lnLF0.38230.53270.6016C−10.4438−1.31690.2064R20.9950R2 Adjusted0.9921Note: ∗∗∗, ∗∗, and∗ indicate the significance level at 10 %, 5 %, and 1 %, respectively. Robustness results using the DOLS test. Note: ∗∗∗, ∗∗, and∗ indicate the significance level at 10 %, 5 %, and 1 %, respectively. In summary, the robustness check with DOLS largely supports the FMOLS findings, particularly regarding the impact of geothermal energy, non-renewable energy, military expenditures, and technology on CO2 emissions. Differences in significance levels and changes in significance for globalization suggest that some relationships may be model-specific and warrant further investigation. Variables showing consistent significance across both models can be considered more robust findings. The graphical findings are displayed in Fig. 2.Fig. 2The graphical results of FMOLS with robustness testing using DOLS.Fig. 2 The graphical results of FMOLS with robustness testing using DOLS. The results of diagnostic tests and parameters stability test The diagnostic tests presented in Table 7 indicate a robust statistical model. The high R-squared values (R2 > 0.9938 and R2 Adjusted >0.9900) suggest an excellent fit, explaining a significant proportion of the variance in the dependent variable. The Jarque-Bera statistic confirms that the residuals are normally distributed, with a p-value of 0.7101, implying that the assumption of normality is not violated. The stability of the model is supported by the CUSUM and CUSUM of Squares tests, with the former showing significance below the 0.05 threshold, indicating that the model parameters are stable over time (Fig. 3). Serial correlation and heteroscedasticity are not concerns, as indicated by the Breusch-Godfrey LM test (p = 0.5117). Furthermore, the Breusch-Pagan-Godfrey, Harvey, and Glejser tests have p-values well above the conventional significance levels. The ARCH test (p = 0.5008) further supports the absence of autoregressive conditional heteroscedasticity. Lastly, the Ramsey test concludes with a p-value of 0.0675, suggesting that the model's specification is appropriate, with no omitted variables that could bias the results. Overall, these diagnostic tests provide confidence in the model's validity and the reliability of the inferences drawn from it.Table 7The results of the diagnostic test.Table 7Diagnostic testsCoeff.Prob.DecisionR2>0.9938–The model is well-fittedR2 Adjusted>0.9900–The model is well-fittedJarque-Bera0.68450.7101Residuals are normally distributedCUSUM & CUSUMQ–<0.05The model is stableBreusch-Godfrey LM test0.70310.5117No serial correlation existsBreusch-Pagan-Godfrey test0.76720.6476No heteroscedasticity existsHarvey test1.57500.2050No heteroscedasticity existsGlejser test0.97970.4914No heteroscedasticity existsARCH test0.46790.5008No heteroscedasticity existsRamsey test2.11400.0675The model is properly specifiedFig. 3Results of CUSUM (a) and CUSUMQ (b) test.Fig. 3 Table 7The results of the diagnostic test.Table 7Diagnostic testsCoeff.Prob.DecisionR2>0.9938–The model is well-fittedR2 Adjusted>0.9900–The model is well-fittedJarque-Bera0.68450.7101Residuals are normally distributedCUSUM & CUSUMQ–<0.05The model is stableBreusch-Godfrey LM test0.70310.5117No serial correlation existsBreusch-Pagan-Godfrey test0.76720.6476No heteroscedasticity existsHarvey test1.57500.2050No heteroscedasticity existsGlejser test0.97970.4914No heteroscedasticity existsARCH test0.46790.5008No heteroscedasticity existsRamsey test2.11400.0675The model is properly specified The results of the diagnostic test. Results of CUSUM (a) and CUSUMQ (b) test. The pairwise granger causality test Based on the results of the pairwise Granger causality test presented in Table 8, unidirectional causality exists, running from geothermal energy consumption, manufacturing value-added, and labor force to CO2 emissions. Furthermore, unidirectional causality is observed running from CO2 emissions to military expenditures, technology innovation, and capital formation.Table 8Results of the pairwise Granger causality test.Table 8Null Hypothesis:F-StatisticProb.GEO ≠ CO23.0623∗∗∗0.0941CO2 ≠ GEO1.35000.2577GDP ≠ CO22.45810.1312CO2 ≠ GDP1.47720.2371NRE ≠ CO20.73410.4008CO2 ≠ NRE0.08840.7690GLB ≠ CO20.08410.7745CO2 ≠ GLB2.02390.1689MANU ≠ CO23.7696∗∗∗0.0651CO2 ≠ MANU0.49090.4909MIL ≠ CO20.388150.5397CO2 ≠ MIL5.90188∗∗0.0237TECH ≠ CO20.426840.5203CO2 ≠ TECH6.52376∗∗0.0181GFCF ≠ CO20.95580.3389CO2 ≠ GFCF3.82362∗∗∗0.0634LF ≠ CO25.59804∗∗0.0272CO2 ≠ LF1.185630.2880Note: ∗∗∗, ∗∗, and∗ indicate the significance level at 10 %, 5 %, and 1 %, respectively. Table 8Results of the pairwise Granger causality test.Table 8Null Hypothesis:F-StatisticProb.GEO ≠ CO23.0623∗∗∗0.0941CO2 ≠ GEO1.35000.2577GDP ≠ CO22.45810.1312CO2 ≠ GDP1.47720.2371NRE ≠ CO20.73410.4008CO2 ≠ NRE0.08840.7690GLB ≠ CO20.08410.7745CO2 ≠ GLB2.02390.1689MANU ≠ CO23.7696∗∗∗0.0651CO2 ≠ MANU0.49090.4909MIL ≠ CO20.388150.5397CO2 ≠ MIL5.90188∗∗0.0237TECH ≠ CO20.426840.5203CO2 ≠ TECH6.52376∗∗0.0181GFCF ≠ CO20.95580.3389CO2 ≠ GFCF3.82362∗∗∗0.0634LF ≠ CO25.59804∗∗0.0272CO2 ≠ LF1.185630.2880Note: ∗∗∗, ∗∗, and∗ indicate the significance level at 10 %, 5 %, and 1 %, respectively. Results of the pairwise Granger causality test. Note: ∗∗∗, ∗∗, and∗ indicate the significance level at 10 %, 5 %, and 1 %, respectively. Discussion The consistent findings from both FMOLS and DOLS estimations indicate a significant impact of geothermal energy consumption on reducing CO2 emissions, which are further reinforced by the Granger causality analysis, suggesting a unidirectional causality from geothermal energy consumption to CO2 emissions. This alignment strengthens the argument for the effectiveness of geothermal energy in mitigating carbon emissions and underscores its potential as a key component in sustainable energy transition strategies in Indonesia. The theoretical underpinning for this lies in geothermal energy - as a renewable energy source that harnesses heat from the Earth's interior, it provides a carbon-neutral alternative to fossil fuels. These results align with Umar et al. [25], Lund & Boyd [75] but are contrary to Idroes et al. [22] and Shahzad et al. [51], who found that geothermal energy will increase CO2 emissions. These results are consistent with prior research emphasizing the role of renewable energy in decoupling economic growth from carbon emissions, highlighting the importance of investing in sustainable energy infrastructure [22,25]. Moreover, adopting a dynamic approach in this study offers valuable insights into the long-term linkage between geothermal energy consumption and CO2 emissions. Demonstrating that changes in geothermal energy consumption precede changes in CO2 emissions provides empirical support for the notion that increasing the adoption of geothermal energy technologies can lead to reductions in carbon emissions over time in Indonesia. This finding is particularly significant for policymakers and stakeholders seeking evidence-based strategies to transition toward cleaner energy systems. The significant positive impact of non-renewable energy consumption on CO2 emissions corroborates prior research highlighting the adverse environmental consequences of reliance on fossil fuels. These results align with Kartal et al. [76], Voumik et al. [77], Ansari [78], and Hanif et al. [79]. This reinforces the urgency of reducing dependence on non-renewable energy sources and accelerating the adoption of renewable alternatives like geothermal energy. By doing so, Indonesia can mitigate the negative impacts of climate change and reduce its vulnerability to energy-related geopolitical risks and fluctuations in fossil fuel prices. As also observed in both FMOLS and DOLS estimations, the significant negative impact of technology on CO2 emissions in the long term adds another dimension to the discussion on environmental sustainability and energy transition. This finding emphasizes the importance of technological innovation and advancement in driving emissions reductions and fostering Indonesia's sustainable development. Advanced technologies enable industries to optimize resource use, improve energy efficiency, and deploy cleaner production processes, thereby mitigating the environmental footprint associated with economic activities [[80], [81], [82], [83], [84]]. The negative relationship between technology and CO2 emissions aligns with the theory of the Environmental Kuznets Curve (EKC), suggesting that as economies progress and technological advancements are made, ecological degradation can eventually be reversed [85,86]. In this context, leveraging technology alongside renewable energy sources like geothermal power becomes crucial for achieving long-term sustainability goals. Alongside energy consumption patterns and technologies, this study also provides empirical evidence of the significant adverse impacts of globalization and military expenditure on CO2 emissions in the long term, offering additional perspectives into the complex dynamics shaping environmental sustainability. While globalization may initially result in higher emissions due to increased economic activity, it can also facilitate the transfer of cleaner technologies and best practices, ultimately contributing to emissions reductions in the long term [[87], [88], [89]]. Similarly, military operations often entail significant energy consumption and emissions, suggesting that reducing military expenditure and reallocating resources towards sustainable development initiatives could yield environmental benefits by curbing emissions from defense-related activities [[90], [91], [92]]. Initiatives such as investments in peacebuilding, conflict resolution, and diplomacy can contribute to stability and resilience, addressing underlying drivers of environmental degradation and promoting sustainable peace. Conclusions and policy recommendations The primary goal of this study is to investigate the impact of geothermal energy consumption on Indonesia's CO2 emissions from 1995 to 2020 using the FMOLS method, with additional validation through the DOLS method. The findings reveal a negative coefficient for geothermal energy consumption on CO2 emissions, highlighting the significant potential of geothermal energy in reducing CO2 emissions in Indonesia. In contrast, the positive impact of non-renewable energy consumption on CO2 emissions indicates that an increase in non-renewable energy use will lead to higher CO2 emissions, underscoring the urgency of transitioning to renewable alternatives like geothermal energy to mitigate climate change and reduce energy-related risks. Additionally, the study finds that both technological innovation and military expenditure have a negative impact on CO2 emissions, further emphasizing the critical role of technological advancements and strategic military spending in driving emissions reductions and fostering sustainable development. Harnessing the combined power of technological innovation and renewable energy sources is vital for achieving Indonesia's long-term ecological sustainability goals. Given that Indonesia holds 40 % of the world's geothermal potential, the government should prioritize creating an investment-friendly environment for geothermal energy. This includes offering tax breaks, subsidies, and streamlined permitting processes to attract both domestic and international investors. Additionally, funding research and development in geothermal technologies, fostering international collaborations, and investing in workforce training will drive innovation and sector growth. Strengthening regulatory frameworks with clear licensing guidelines and feed-in tariffs will enhance market certainty and stimulate investment. Public awareness campaigns and community engagement are vital to ensure social acceptance and alignment with sustainability goals. Moreover, robust monitoring and reporting systems should be established to track progress in emissions reductions and demonstrate geothermal energy's impact on climate change mitigation and sustainable development. To manage emissions effectively, the government should reduce reliance on fossil fuels by implementing stricter regulations and incentives for renewable energy adoption, promoting clean technologies, and enhancing fuel efficiency standards. Investment in emission-reducing technological innovations and collaboration with international research organizations is crucial. Additionally, optimizing military expenditure by integrating sustainable practices and energy-efficient technologies can further reduce emissions. Establishing comprehensive monitoring and reporting systems to track emissions from fossil fuels, technological advancements, and military operations will ensure policy effectiveness and progress toward Indonesia's climate goals. Integrating these strategies will help Indonesia address emissions challenges and move toward a more sustainable future. This study has limitations that suggest areas for future research. The 1995 to 2020 dataset may not fully capture recent policy changes, technological advancements, or economic shocks like the COVID-19 pandemic, indicating a need for extending the analysis with updated data. Additionally, the study focuses solely on Indonesia, overlooking regional variations in geothermal potential and policies, which future research could address by examining provincial disparities. While using the FMOLS and DOLS methods provides robust insights, exploring alternative econometric techniques like panel data analysis could enhance understanding broader causal relationships. The socio-economic impacts of geothermal energy adoption, such as job creation, social acceptance, and local community effects, are also not examined, warranting further investigation to offer a more comprehensive view of the transition to renewable energy. Moreover, factors such as international trade, financial development, and urbanization, which could influence CO2 emissions, are not included in the current analysis. Considering these additional determinants and addressing regional and socio-economic aspects would enhance understanding of sustainable energy transitions and support more effective policy development for Indonesia's climate goals. CRediT authorship contribution statement Ghalieb Mutig Idroes: Formal analysis, Data curation, Conceptualization. Mohd Afjal: Resources, Project administration. Mohsin Khan: Resources. Mohammad Haseeb: Methodology. Irsan Hardi: Writing – review & editing. Teuku Rizky Noviandy: Writing – original draft, Conceptualization. Rinaldi Idroes: Visualization, Investigation. Data availability The data used in this study are sourced from publicly archived databases. Ethics declaration Review and/or approval by an ethics committee as well as informed consent was not required for this study because this article did not involve any direct experimentation/studies on living beings. Funding No Funding received for this study. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.