Dynamic impacts of energy use, agricultural land expansion, and deforestation on CO₂ emissions in Malaysia

Abstract

This study empirically investigates the nexus among energy use, agricultural land expansion, deforestation, and carbon dioxide (CO₂) emissions in Malaysia. Time series data from 1990 to 2019 were utilized using the bounds testing (ARDL) approach followed by the Dynamic Ordinary Least Squares (DOLS) method. The DOLS estimate findings show that the energy usage coefficient is positive and significant with CO₂ emissions, indicating a 1% increase in energy consumption is related to a 0.91% rise in CO₂ emissions. In addition, the coefficient of agricultural land is positive, which indicates that agricultural land expansion by 1% is associated with an increase in CO₂ emissions by 0.84% in the long run. Furthermore, the forested area coefficient is negative, which means that decreasing 1% of the wooded area (i.e., deforestation) has a long-term effect of 5.41% increased CO₂ emissions. Moreover, the pairwise Granger causality test results show bidirectional causality between deforestation and energy use; and unidirectional causality from energy use to CO₂ emissions, agricultural land expansion to CO₂ emissions, deforestation to CO₂ emissions, agricultural land expansion to energy use, and deforestation to agricultural land expansion in Malaysia. The empirical findings reveal that increased energy use, agricultural land expansion, and deforestation have a negative impact on environmental quality in Malaysia. Thus, the effective implementation of policy measures to promote renewable energy, climate-smart agriculture, and sustainable management of forest ecosystems could be useful for reducing environmental degradation in Malaysia.

1 Introduction

Global climate change is a burning issue due to the atmospheric concentrations of greenhouse gases (GHGs) dominated by CO₂ which is largely emitted from human induced activities like the combustion of fossil fuels and deforestation (IPCC 2014; Begum et al. 2020; Uddin 2021; Çıtak et al. 2021). The continuous increase in CO₂ emissions is projected to have tremendous effects on the global climate system resulting in catastrophic consequences that will impact all aspects of society (Seriño 2018). Therefore, reducing CO₂ emissions and increasing the environmental quality became a global concern to secure sustainable development as well as to minimize the climate change detrimental effects (Begum et al. 2015). Within this framework, determining the factors affecting CO₂ emissions is important for selecting appropriate strategies to increase environmental quality (Inekwe et al. 2020). If the functional connection between natural resources and contemporary development processes cannot be avoided, environmental damage is unavoidable. Such issues are greater in nations like Malaysia, where energy security and environmental sustainability are concurrently essential (Begum et al. 2015; Zhang et al. 2021). In Malaysia, the government established a voluntary goal to decrease carbon emissions by 45% by 2030 (MESTECC 2018). Overall knowledge of Malaysia’s vulnerability to climate change is increasingly essential for policymakers to strike a balance between policies aimed at mitigating climate change and achieving sustainable development and executing both. The biggest difficulty facing this dual objective is the trade-off between pollution and development. Thus, a significant concern is whether the objectives of sustainable growth and better environmental quality (emission reduction) are mutually incompatible. A major issue emerging is how Malaysia can reduce CO₂ emissions, and this issue may be handled by examining the country’s primary CO₂ sources.

However, energy, ecology, and environmental problems are strongly linked in the networks of human and natural systems, usually deconstructed into three distinct elements of reductionist thinking, despite the transdisciplinary inherent in basic and applied research (Chen 2016). Efforts in one subject, without understanding broader connections and effects, are frequently invalid in improving sustainability. For example, increasing energy consumption may lead to possible environmental contamination and ecological deterioration. Such trade-offs require comprehensive analysis and balanced solutions (Chen 2016). Energy demand in Southeast Asia grew quickly owing to rapid population expansion and economic development. Energy is important for meeting domestic requirements, enabling industry and commerce (Pratiwi and Juerges 2020). However, the Malaysian economy subsidizes petroleum prices (Abdullah et al. 2009) which alternatively encourages additional energy use by accelerating economic growth. Zhang et al. (2021) reported that energy-using CO₂ emissions in Malaysia increased by 2% in 2018, the fastest rise in seven years. In 2018, Malaysia’s CO₂ emissions were 250.3 million tons, up from 241.6 million tons in 2017. Gan and Li (2008) predicted a 4.3% rise in energy consumption in Malaysia in 2030. Nevertheless, the transport and industrial sectors in Malaysia are the major energy users that have combined used 80% of the total energy consumption (MNRE, 2018). Malaysia has been facing an increasing demand for energy consumption over the past years due to urbanization and industrial development (Begum et al. 2017). In line with energy demand, Malaysia has also experienced an increasing trend of CO₂ emissions. Therefore, the country is highly concerned about the growing emission intensity, especially from the energy sector. Therefore, it is crucial to analyze the nexus between energy demand and CO₂ emission that would indicate policy options.

Moreover, CO₂ emissions and climate change were connected to agriculture, one of the main drivers of environmental deterioration, and deemed ultrasensitive to climate change (Seo 2016; Naseem et al. 2020). The Agriculture, Forestry and Other Land Use (AFOLU) sector are responsible for about one-fifth of the global annual CO₂ emissions, which is the second-largest CO₂ emission source and is a major contributor to global climate change (IPCC 2014). Agricultural GHG emissions are mainly generated by deforestation, livestock emissions, soil, nutrition management of fossil-fuel-based fertilizer, fossil-fuel-driven agricultural equipment, and bush-burning operations as well as the burning of biomass fuel (Balsalobre-Lorente et al. 2019). Agricultural GHG emissions are a particularly difficult issue in emerging nations like Malaysia because agricultural change is linked with industrialization. Agriculture in Malaysia is dominated by oil palm and rubber production, which plays a major role in economic development. Malaysia is the world’s largest exporter and second-largest oil palm producer, contributing 37.7% to agricultural GDP (DOSM 2020). The oil palm business is also a major source of employment, with a total workforce of 1.16 million people, representing 40.45% of the agricultural jobs in Malaysia (MPOC 2014; Sarkar et al. 2020). In addition, Malaysia is the world’s fifth-biggest rubber producer (Yusoff et al. 2019). However, the oil palm and rubber sectors have been linked with major environmental problems, such as deforestation, peatland exploitation, and biomass burning that contribute to climate change by releasing CO₂ into the atmosphere (Uning et al. 2020).

The increasing population and growing need for food, transportation, and other infrastructure impose a strain on forest ecosystems in Malaysia (MNRE 2018). Since the agricultural industry revolution, CO₂ emissions from deforestation and forest degradation have grown fast. Global forest cover has changed quickly in recent decades, with the tropical forest cover substantially decreased by an annual loss of 2101 square kilometers (Hansen et al. 2013). Among the many types of environmental degradation, tropical deforestation is said to be one of the main problems to be handled seriously, especially to ensure the sustainability of the global ecosystem (Murshed et al. 2021). Due to high CO₂ emissions from deforestation, significant effects from climate change have occurred in Malaysia, which is also one of South-East Asia’s top CO₂ emitters (Begum et al. 2020). Peninsular Malaysia has lost more than half of its natural forests due to industrialization, urbanization, settlements, mining, farming, agricultural fires, clearing forests for oil-palm plantation, and other forms of agriculture (Raihan et al. 2018; Jaafar et al. 2020; Begum et al. 2020). However, forest ecosystems function as both sources and sinks of carbon through which they have a crucial influence on the global climate system (Matthew et al. 2018). Forests play a significant role in climate change mitigation by absorbing the atmospheric CO₂ and storing it in tree biomass, which is called carbon sequestration (Raihan et al. 2019). About 300 billion tons of CO₂ emissions from the atmosphere are yearly absorbed by the global forest ecosystems, and about three billion tons of it is anticipated to leak into the environment annually due to deforestation (Aziz et al. 2020). Deforestation causes long-term ecological problems by accelerating the effects of climate change, thus attributing to desertification, flooding, soil erosion, and loss of natural habitats (Murshed et al. 2021). With temperatures anticipated to rise 1.5ºC above pre-industrial levels between 2030 and 2052 under the current global warming and climate change scenario, the role of forest ecosystems in absorbing carbon from the atmosphere is becoming increasingly important (IPCC 2018). Therefore, it is crucial to investigate the impact of forest ecosystems on CO₂ emissions in Malaysia.

Over the last decade, the nexus approach to energy, ecology, and the environment has gained traction, and it is projected to discover more information on the widening challenges of climate change, energy consumption, agricultural land expansion, and deforestation (Chen 2016; Ahmad et al. 2020). The IPCC (2014) reported that population size, economic activity, lifestyle, energy use, land use patterns, technology, and inadequate climate policy are the primary drivers of GHG emissions. As a result, several studies examined the relationship between environmental pollution indicators such as CO₂ emissions and economic growth using time series data for Malaysia (Azlina and Mustapha 2012; Saboori et al. 2012; Begum et al. 2015; Begum et al. 2020; Zhang et al. 2021). However, there is a research gap in investigating the dynamic impacts between CO₂ emission and other variables by utilizing econometric approaches. A limited study explores the dynamic impacts of energy use, agricultural land expansion, and deforestation on CO₂ emissions in Malaysia. Therefore, the present study aims to fill this literature gap by using the Dynamic Ordinary Least Squares (DOLS) approach to investigate the dynamic interaction of energy use, agricultural land expansion, and deforestation on CO₂ emission in Malaysia. This study is expected to contribute to the recent literature and policymaking in Malaysia in five ways. Firstly, this investigation establishes a relationship between agricultural land expansion and CO₂ emissions, which is a pioneering attempt to reveal the impact of agricultural land expansion on CO₂ emissions in Malaysia. Secondly, this research portrays the special role of the forest ecosystems, which is commonly neglected when explaining the determinants of CO₂ emissions. Thirdly, this study creates a link between deforestation and agricultural land expansion in Malaysia which is a ground-breaking effort to reveal the nexus between deforestation, agricultural land expansion, and related CO₂ emissions in Malaysia. Next, several unit root tests, cointegration tests, and diagnostic tests are employed to verify the precision of the results. Finally, the study’s outcomes would provide further comprehensive and valuable insights to policymakers for designing effective policies related to cleaner energy, climate-smart agriculture, sustainable forest management, and emission reduction in Malaysia.

The rest of the article is structured as follows. The Introduction is followed by Sect. 2 with a brief exposure on the current status of CO₂ emissions, energy use, agricultural land, and forested area in Malaysia. Section 3 exhibits the Literature Review, where relevant research studies have been discussed. The fourth section is the Methodology section, followed by the Empirical Findings in Sect. 5 and Discussion in Sect. 6. Subsequently, Sect. 7 presents the conclusion, policy implications, recommendations, limitations, and future research opportunities.

2 Current status of CO₂ emissions, energy use, agricultural land, and forested area in Malaysia

Malaysia is one of the few developing nations in the world that has transitioned from a resource-based economy in the 1970s to a multi-sector economy in the 1990s, with a focus on manufacturing and services (Begum et al. 2015). Positive economic development, driven mostly by the export of manufactured goods, fostered this change. As a result of the misuse or misallocation of non-renewable energy sources, quicker economic expansion leads to higher urbanization and energy usage, which in turn promotes CO₂ emissions. Malaysia has been facing an increasing demand for energy consumption over the past years (Begum et al. 2017). In line with energy demand, Malaysia has also experienced an increasing trend of CO₂ emissions. Figure 1 presents the annual trend of CO₂ emissions in Malaysia. Total CO₂ emissions in 1960 were about 3,568 kt which increased gradually to 56,190 kt in 1990. Due to rapid industrialization, total CO₂ emissions in Malaysia increased drastically to 239,620 kt in 2018. From 1990 to 2018, Malaysia’s average yearly increase rate of CO₂ emissions was approximately 11.66%. Consequently, lowering CO₂ emissions has become a key priority in Malaysia to preserve environmental sustainability and mitigate the harmful effects of climate change.

Fig. 1

Annual trend of CO₂ emissions in Malaysia. Source: World Bank (2021)

Fig. 2

The major sources of CO₂ emissions in Malaysia for 2014. Source: MNRE (2018)

Malaysia is the world’s 26th greatest emitter of energy-related CO₂; its contribution amounts to 0.66% of the world’s total emissions (Chik and Rahim 2014). Malaysia’s total GHG emissions in 2014 were 248,195 Gg CO₂eq, with the energy sector accounting for 80% of those emissions (MNRE 2018). Therefore, the country is highly concerned about the growing emission intensity, especially from the energy sector. Figure 2 depicts the major sources of CO₂ emissions in Malaysia for 2014. Energy industries are the main culprit of CO₂ emissions in Malaysia, followed by transport, manufacturing industries, and cement production. From 1990 to 2014, the energy sector’s emissions rose at an annual rate of 5.8% on average. Between 2005 and 2014, the energy sector’s emissions increased by 28% (MNRE 2018). The annual trend of energy use in Malaysia is presented in Fig. 3. Energy use per capita in 1971 was 547 kg which increased to 3,003 kg per capita in 2014. In Malaysia, primary energy supply and energy demand have increased in lockstep with population and economic growth, and this pattern is anticipated to continue. As a result, it is necessary to ensure that the economy has access to reliable and affordable energy while avoiding negative environmental impacts and ensuring long-term energy security. (MNRE 2018).

Fig. 3

Annual trend of energy use in Malaysia. Source: World Bank (2021)

Moreover, agricultural production is responsible for approximately 20% of global GHG emissions (FAO 2020) and 8.22% of Malaysia’s GHG emissions in 2016 (Ridzuan et al. 2020). Agriculture activities in Malaysia that contribute to the CO₂ emission are mainly oil palm plantation, rubber plantation, and other agricultural crops (Omar et al. 2016). Malaysian agriculture has been linked to various unsustainable practices and land clearing, both of which contribute to the country’s significant CO₂ emissions. Furthermore, increasing agricultural production to ensure food security necessitates increased energy consumption, which leads to increased emissions. (Zhang et al. 2019). However, Malaysia’s agriculture industry has remained one of the top three contributors to the country’s gross domestic product (GDP), accounting for 8.2% of GDP in 2020 (World Bank 2021). Malaysia’s agriculture is mainly dominated by oil palm cultivation, which accounts for 37.7% of the country’s GDP. While oil palm cultivation is important to Malaysia’s economic development, it is also responsible for the extinction of several species-rich forest ecosystems, and ongoing unsustainable farming methods are endangering these delicate and fragile ecosystems. Malaysia’s oil palm plantations cover about 5.2 million hectares that occupy almost 16% of the total land area and nearly 61% of the total agricultural land (Yamada et al. 2016). Figure 4 presents the annual trend of agricultural land in Malaysia. Total agricultural land in Malaysia increased drastically from 30,847 square km in 1961 to 67,217 square km in 1989. Nevertheless, agricultural land increased gradually after 1990, and it turned into 85,710 square km in 2018.

Fig. 4

Annual trend of agricultural land in Malaysia. Source: World Bank (2021)

Furthermore, forest ecosystems have a crucial role in sustaining the equilibrium of the global climate system by absorbing and storing atmospheric carbon, managing hydrological systems, protecting biodiversity, and providing habitats for wildlife (Begum et al. 2020). Malaysia has forest land, approximately 67% of the total land area (World Bank 2021), which helps mitigate climate change by absorbing atmospheric CO₂ and improving the national carbon sink. However, Malaysia has faced environmental degradation, biodiversity loss, and the destruction of wildlife habitat because of the forest cover changes during the previous few decades due to extensive deforestation. Deforestation and forest degradation have led to global anthropogenic CO₂ emissions, as well as severely affecting tropical forest ecosystems and natural carbon sequestration (Begum et al. 2020). The annual trend of forested area in Malaysia is presented in Fig. 5. Between 1990 and 2010, Malaysia experienced a high rate of deforestation, losing approximately 8.6% of its forest cover. This could be due to a variety of factors, including industrialization, urbanization, settlements, mining, farming, agricultural fires, forest clearing for oil palm and rubber plantations, and other types of agriculture (Begum et al. 2020). In 1990, the total forested area was nearly 206,185 square km which reduced to approximately 189,477 square km in 2010. However, the forested area has been on an upward trend since 2010 due to afforestation, reforestation, forest protection, and conservation. The forested area in Malaysia was reported as about 192,143 square km in 2018.

Fig. 5

Annual trend of forested area in Malaysia. Source: World Bank (2021)

3 Literature Review

Several studies around the world have established the positive relationship between energy use and CO₂ emissions by using numerous econometric approaches. Adebayo and Kalmaz (2021) used ARDL, FMOLS, and DOLS estimators to uncover a positive and significant interaction between energy usage and CO₂ emissions in Egypt by utilizing the data covering the years from 1971 to 2014. Kirikkaleli and Kalmaz (2020) found a long-run positive effect of energy consumption on CO₂ emissions in Turkey throughout 1960–2016 by utilizing FMOLS and DOLS techniques. Odugbesan and Adebayo (2020) investigated the symmetric and asymmetric effects of energy consumption on CO₂ emissions towards environmental sustainability in Nigeria by utilizing the yearly data spanning from 1981 to 2016 employing ARDL, FMOLS, and DOLS techniques. The findings from Odugbesan and Adebayo (2020) showed that energy consumption has a long-run linear relationship with CO₂ in Nigeria. Moreover, by applying ARDL, FMOLS, and DOLS estimators to examine the long-run and causal effects of energy usage on CO₂ emissions in Mexico using the yearly data spanning between 1971 and 2016, Adebayo (2020) found that energy usage impacts CO₂ emissions positively. In addition, he revealed unidirectional causality from CO₂ emissions to energy usage. Furthermore, by employing an ARDL estimator using yearly data between 1980 and 2016, Adebayo (2021) revealed that energy usage positively triggers CO₂ emissions in Indonesia. Using the ARDL estimator for Kazakhstan spanning 1980–2011, Akbota and Baek (2018) found that energy consumption increases CO₂ emissions. By employing the ARDL approach, Nondo and Kahsai (2020) revealed a positive relationship between energy intensity and CO₂ emissions in South Africa from 1970 to 2016.

Furthermore, a number of studies reported the positive association between energy use and CO₂ emissions from a group of countries. By utilizing FMOLS and DOLS estimators using the time series data from 1971 to 2014, Vo et al. (2019) revealed that the level of CO₂ emissions is positively associated with energy consumption in five ASEAN countries (Indonesia, Myanmar, Malaysia, the Philippines, and Thailand). In addition, Vo et al. (2019) found unidirectional causality from energy consumption to CO₂ emissions. Irfan and Shaw (2017) investigated the relationship between energy consumption and CO₂ emissions in South Asian countries over 1978–2011 by employing a nonparametric additive model. Irfan and Shaw (2017) discovered that energy consumption positively impacts CO₂ emissions in the panel of countries. Adebayo et al. (2020) found a positive interconnection between CO₂ emissions and energy usage by utilizing the ARDL model for MINT countries using the time coverage from 1980 to 2018. Zmami and Ben-Salha (2020) investigated the effects of energy consumption on CO₂ emissions in Gulf Cooperation Council (GCC) countries between 1980 and 2017 using the STIRPAT model and the ARDL approach. Zmami and Ben-Salha (2020) reported that energy consumption has a positive and significant impact on CO₂ emissions in the long run.

In Malaysia, Ang (2008) exposed that energy consumption and environmental pollution are positively associated in the long run. By applying the cointegration and vector error correction model for Malaysia over the period 1970–2010, Azlina and Mustapha (2012) reported unidirectional causality running from CO₂ emissions to energy consumption. Using the time series data for the newly industrialized countries spanning 1971–2007, Hossain (2011) found that higher energy consumption increases CO₂ emissions over time. Begum et al. (2015) investigated the dynamic impacts of energy consumption on CO₂ emissions using the ARDL bounds testing approach and DOLS technique for Malaysia by using the data over 1970–2009. Begum et al. (2015) demonstrated that per capita energy consumption has a long-term positive impact on carbon emissions. Using time-series data from 1980 to 2016, Sarkar et al. (2019) revealed unidirectional causality of energy demand with carbon emissions in Malaysia.

However, the agriculture sector is the primary cause of environmental degradation due to its extensive use of fossil fuels and nitrogen-rich fertilizers (Aziz et al. 2020). According to Reynolds and Wenzlau (2012), the agriculture sector uses a lot of fossil fuel for pumping water, irrigation, and nitrogen-rich fertilizer, which accounts for 14–30% of total GHG emissions. Holly (2015) reported that one of the major sources of CO₂ emissions in agriculture is nitrous oxide and methane from livestock practices and soil management. Thus, reducing emissions from the agriculture sector has become a critical issue for sustainable development. Over the last few years, the impact of agriculture on environmental pollution has been a recurring topic of discussion. Recent studies have been performed with various econometric methods to investigate the effects of agriculture on environmental degradation. Doğan (2018) utilized ARDL, FMOLS, DOLS, and CCR techniques to explore the relationship between agricultural production and CO₂ emissions in China by using annual data covering1971-2010. The findings from Doğan (2018) revealed that agriculture increases a country’s long term CO₂ emissions. By performing the ARDL approach and Bayer-Hanck cointegration test for Nigeria using annual time series data from 1981 to 2014, Agboola and Bekun (2019) found that agricultural production has a positive but insignificant impact on CO₂ emissions. Using the ARDL method for the period 1990–2016, Burakov (2019) found that the agricultural sector is a statistically significant determinant of carbon emission in Russia. In addition, the pairwise Granger causality test by Burakov (2019) revealed unidirectional causality running from agriculture to CO₂ emissions.

Gokmenoglu et al. (2019) investigated the long-run equilibrium relationship between CO₂ emissions and agriculture in the case of China for the period of 1971–2014. Using the ARDL approach, Gokmenoglu et al. (2019) reported that agricultural development is positively associated with CO₂ emissions. Using the Johansen-Juselius cointegration test and VECM model for Tunisia spanning 1980–2011, Jebli and Youssef (2017a) showed that an increase in agricultural value-added increases CO₂ emissions. By applying the ARDL model for data from 1961 to 2012, Sarkodie and Owusu (2016) found that agricultural production leads to increased CO₂ emissions in Ghana. By performing FMOLS, Maki cointegration, and Toda-Yamamoto causality tests for Pakistan from 1971 to 2014, Gokmenoglu and Taspinar (2018) reported that agriculture deteriorates environmental quality, and there is bidirectional causality between agriculture and CO₂ emissions. Naseem et al. (2020) examined the asymmetrical impact of agriculture on CO₂ emissions in Pakistan from 1969 to 2018 by utilizing the NARDL approach. The findings by Naseem et al. (2020) revealed that agriculture value-added has a significant positive impact on the CO₂ emissions in Pakistan. In addition, Naseem et al. (2020) found unidirectional causality running from agriculture to CO₂ emissions. Ullah et al. (2018) performed ARDL, Johansen-cointegration test, and Granger causality test by using the data from 1972 to 2014. Ullah et al. (2018) found that six different types of agricultural activities increase CO₂ emissions in Pakistan. They also revealed a bidirectional causality between agriculture and CO₂ emissions.

Furthermore, several studies have uncovered a relationship between agriculture and CO₂ emissions in a set of countries. By utilizing panel cointegration tests, FMOLS, and DOLS techniques in a 1992–2013 sample of BRICS countries, Liu et al. (2017) testified that agricultural production generates environmental pollution. Using the Pedroni cointegration test, FMOLS, DOLS, and VECM for a panel of E7 countries spanning 1990–2014, Aydoğan and Vardar (2020) reported that agriculture has a positive relationship with environmental degradation. Balsalobre-Lorente et al. (2019) employed DOLS and FMOLS methods to investigate the long-run effects of agricultural activities on carbon emissions for Brazil, Russia, India, China, and South Africa (BRICS) over the period 1990–2014. Empirical evidence confirmed that agriculture exerts a negative impact on the environment in BRICS countries. By employing the FMOLS and VECM approach using the data from 1990 to 2014, Qiao et al. (2019) found that agriculture significantly increases CO₂ emissions in G20 countries. Applying the Emirmahmutoglu-Kose panel causality test and augmented, pooled, and mean group estimators for eleven developing countries in the Central and West African regions, Olanipekun et al. (2019) reported that agriculture has a positive impact on environmental pressure. By utilizing the ADL cointegration test for BRIC countries using annual time series data from 1971 to 2016, Pata (2021) found that the impact of agricultural value-added on environmental pollution is positive but statistically insignificant. In addition, the bidirectional causality relationship showed that agriculture is an important factor in environmental pollution and that agricultural activities are affected by environmental problems.

The influence of agriculture on environmental degradation has been a frequent issue of controversy throughout the previous decade. Several studies reported the negative impact of agriculture on environmental deterioration. By utilizing the ARDL approach using data from 1968 to 2010, Dogan (2016) reported that CO₂ emissions in Turkey are negatively associated with agricultural production. Prastiyo et al. (2020) found negative impacts of agriculture value-added on CO₂ emissions utilizing the ARDL technique for Indonesia using the data over 1970–2015. Liu et al. (2017) employed the ARDL model to uncover a negative influence of agricultural value added on CO₂ emissions in ASEAN countries using data over the span time of 1970–2013. Jebli and Youssef (2017b) reported that an increase in agricultural value added decreases CO₂ emissions in North African countries throughout 1980–2011 by utilizing OLS, FMOLS, and DOLS techniques. By employing the ARDL model using the data from 1996 to 2017, Wang et al. (2020) found that agriculture value-added decreases CO₂ emissions in G7 countries. By using panel data from 1982 to 2015, Anwar et al. (2019) found that higher agricultural value added has an inverse association with CO₂ emissions in low- and lower-middle-income nations. Rafiq et al. (2016) found that agricultural value added has a key impact in decreasing emissions using the STIRPAT model and yearly data from 1980 to 2010 for a set of 53 nations.

Moreover, using country-specific panel data from 1990 to 2014 for 86 different countries, Parajuli et al. (2019) established a negative association between forested area and CO2 emissions and a positive relationship between agricultural land and CO₂ emissions. Aziz et al. (2020) investigated the role of forestry and agriculture in testing EKC in Pakistan by employing the Quantile ARDL approach using the data from 1990 to 2018. Aziz et al. (2020) found a positive link between agricultural production and CO₂ emissions, whereas a negative relationship between forested area and CO₂ emissions. Using the ARDL estimator for Pakistan spanning 1990–2014, Waheed et al. (2018) reported that agriculture increases environmental degradation, whereas CO₂ emission can be reduced by increasing forested area. By employing econometrics approaches using time series data from 1991 to 2010, Islam et al. (2017) found a negative relationship between forested area and CO₂ emissions in Malaysia, Indonesia, and Thailand. Shittu et al. (2018) utilized the ARDL technique on secondary data from 1981 to 2014 and found that deforestation positively influences environmental degradation in Malaysia. Begum et al. (2020) examined the dynamic impacts of forested area on CO₂ emissions in Malaysia by utilizing the DOLS approach using time series data from 1990 to 2016. Begum et al. (2020) revealed a negative and significant coefficient of forested area, which implied that declining one hectare of forested area (i.e., deforestation) has an impact of three kilo tons of CO₂ emissions rise in Malaysia. However, most environmental studies have concentrated on GHG emissions into the atmosphere, leaving agricultural land expansion and deforestation as key determinants of environmental quality, especially in Malaysia. Therefore, this study attempts to explore the dynamic impacts of energy use, agricultural land expansion, and deforestation on CO₂ emissions in Malaysia.

4 Methodology

4.1 Data

This study provides an empirical analysis of the dynamic impacts of energy use, agricultural land expansion, and deforestation on CO₂ emissions in Malaysia by using the dynamic ordinary least squared (DOLS) approach of cointegration by Pesaran and Shin (1999) and Pesaran et al. (2001). Time series data from 1990 to 2019 for Malaysia, were obtained from the World Development Indicator (WDI) dataset. The variables are transmuted into a logarithm to ensure that data are normally distributed. The variables with their logarithmic forms, measurement units, and data sources are presented in Table 1. This research considers CO₂ emissions as the dependent variable while energy use, agricultural land, and forested area as explanatory variables.

Table 1 Variables with their logarithmic forms, units, and data sources

4.2 Econometric Model

To examine the long-term effects of energy use, agricultural land expansion, and deforestation on CO₂ emissions in Malaysia, we have derived the following equation (Begum et al. 2020; Adebayo 2021):

$$ \text{CO}_{2t} =f (\text{EU}_{t}; \text{AGL}_t; \text{FA}_t)$$

(1)

The following equation depicts the econometric model:

$$ \text{CO}_{2t} ={{\uptau }}_{0}+{{\uptau }}_{1}{\text{EU}}_{\text{t}}+{{\uptau }}_{2}{\text{AGL}}_{\text{t}}+{{\uptau }}_{3}{\text{FA}}_{\text{t}}$$

(2)

Further Eq. (2) can be expended as an econometric model in the following form:

$$ \text{CO}_{2t} ={{\uptau }}_{0}+{{\uptau }}_{1}{\text{EU}}_{\text{t}}+{{\uptau }}_{2}{\text{AGL}}_{\text{t}}+{{\uptau }}_{3}{\text{FA}}_{\text{t}}+{{\upepsilon }}_{\text{t}}$$

(3)

where τ₀ and ${\epsilon }_{t}$ stand for intercept and error term, respectively. In addition, τ₁, τ₂, and τ₃ denote the coefficients.

Moreover, the logarithmic arrangement of Eq. (3) can be as follows:

$$ \text{LCO2}_{t}={{\uptau }}_{0}+{{\uptau }}_{1}{\text{LEU}}_{\text{t}}+{{\uptau }}_{2L}{\text{AGL}}_{\text{t}}+{{\uptau }}_{3L}{\text{FA}}_{\text{t}}+{{\upepsilon }}_{\text{t}}$$

(4)

4.3 Stationarity techniques for data

Conducting a unit root test is essential in avoiding spurious regression. It verifies that variables included in regression are stationary by differencing them and estimating the equation of interest using the stationary processes (Mahadeva and Robinson 2004). The obligation to determine the order of integration before investigating cointegration amongst the variables is recognized in the empirical literature. Several researchers have suggested that owing to the power difference of unit root tests regarding the size of the sample, it is vital to utilize more than one unit root test to evaluate the integration order of the series (Saboori et al. 2017; Adebayo et al. 2021). Therefore, we applied the Augmented Dickey-Fuller (ADF) test introduced by Dickey and Fuller (1979), and Dickey-Fuller generalized least squares (DF-GLS) test proposed by Elliott et al. (1996) to detect the autoregressive unit root. This study did the unit root test to confirm that no variable exceeded the order of integration.

4.4 ARDL bounds test for cointegration

We applied the ARDL bounds test proposed by Pesaran et al. (2001) to capture the cointegration amongst the series. The ARDL bounds test for cointegration valuation has many advantages over the other one-time integer methods (Sam et al. 2019; Sarker and Khan 2020; Tong et al. 2020; Adebayo and Akinsola 2021). Firstly, it can be utilized when series have a mixed order of integration as the ARDL bounds test does not have obligatory assumptions. Secondly, it is significantly more reliable, particularly for a small sample size. Thirdly, it offers an accurate estimation of the long-term model (Adebayo and Akinsola 2021). Therefore, the ARDL bounds testing approach can be used irrespective of whether the fundamental returning system is in sequence to apart in the I(2), and the cointegration order happens at I(0) or I(1). The ARDL bounds test is depicted as follows in Eq. (5):

$$ {{\Delta }\text{LCO}2}_{\text{t}}={{\uptau }}_{0}+{{\uptau }}_{1}{\text{LCO}2}_{\text{t}-1}+{{\uptau }}_{2}{\text{LEU}}_{\text{t}-1}+{{\uptau }}_{3}{\text{LAGL}}_{\text{t}-1}+{{\uptau }}_{4}{\text{LFA}}_{\text{t}-1}+\sum _{\text{i}=1}^{\text{q}}{{\upgamma }}_{1}{{\Delta LCO}2}_{\text{t}-\text{i}}+\sum _{\text{i}=1}^{\text{q}}{{\upgamma }}_{2}{{\Delta LEU}}_{\text{t}-\text{i}}+\sum _{\text{i}=1}^{\text{q}}{{\upgamma }}_{3}{{\Delta LAGL}}_{\text{t}-\text{i}}+{\sum _{\text{i}=1}^{\text{q}}{{\upgamma }}_{4}{{\Delta LFA}}_{\text{t}-\text{i}}}+{\upepsilon }_{\text{t}}$$

(5)

where Δ is the first difference operator and q is the optimum lag length in the above Eq. (5).

The ARDL bounds test follows the F-distribution, and its critical values were proposed by (Pesaran and Timmermann 2005). The estimation procedure begins with Eq. (5) and uses OLS to enable the F-test to determine the joint significance of the coefficient of the lagged variables. The essence of this procedure is to examine the likelihood of any possible long-run relationship among the respective variables. In this regard, the null hypothesis (H₀) states that cointegrating relationships do not exist among the regressors. The F-statistics can be compared against the critical values of the upper and lower bounds, as in Pesaran et al. (2001). If the F-statistics are higher than the upper critical value, the null hypothesis is rejected, which means the existence of a long-run relationship among the respective variables. On the other hand, if the F-statistics are less than the lower critical value, the null hypothesis is accepted. Alternatively, if the F-statistics are observed within the lower and upper critical values, the test is inconclusive.

4.5 DOLS cointegration test

This study employed DOLS, an extended equation of ordinary least squares estimation to analyze the time series data. DOLS cointegration test contains explanatory factors as well as leads and lags of their initial difference terms to regulate endogeneity and to calculate the standard deviations using a covariance matrix of errors which is resistant to serial correlation. The inclusion of the different terms’ leads and lags confirms that the error term is orthogonalized. The standard deviations of the DOLS estimators provide a valid test for the statistical significance of the variables because they have a normal asymptotic distribution (Wang 2012). When a mixed order of integration occurs, the DOLS approach is effective at allowing individual variables in the cointegrated outline to be integrated by estimating the dependent variable on explanatory variables in levels, leads, and lags.

The DOLS estimation’s main advantage, however, is the presence of mixed order integration of individual variables in the cointegrated outline. For example, in DOLS estimation, one of I(1) variables was regressed against other variables, some of which were I(1) variables with leads (p) and lags (-p) of the first difference, while others were I(0) variables with a constant term (Alcántara and Padilla, 2009). As a result of aggregating the leads and lags among explanatory variables, this estimate solves small sample bias, endogeneity, and auto correlation issues (Stock and Watson, 1993). However, after confirming cointegration among the variables, the study proceeds with the DOLS estimation of the long-run coefficient by using Eq. (5).

4.6 Cointegration regression to check the robustness of DOLS estimation

This study utilized the fully modified OLS (FMOLS) and Canonical Cointegrating Regression (CCR) to check the robustness of DOLS outcomes. The FMOLS regression was developed by Hansen and Phillips (1990) to incorporate optimum estimates of cointegrating regression. The FMOLS method alters least squares to account for the serial correlation consequences and the endogeneity in the independent variables that arise from the interaction of cointegrating. It facilitated polynomial regression of deterministic variables, stationary error, and integrated procedures. The errors can serially be associated, and the regressors can be endogenous (Hong and Wagner 2011). The FMOLS method is capable of estimating nonstationary I(1) data where it can use standard regression techniques (OLS) of the nonstationary (unit root) data to add to the problem of spurious regressions. Moreover, Park (1992) first introduced the CCR procedure that involves data transformation that uses only the stationary component of a cointegrating model. A cointegrating relationship supported by the cointegrating model would remain unchanged after such data transformation. The CCR transformation makes the error term in a cointegrating model uncorrelated at the zero frequency with regressors. Consequently, the CCR procedure yields asymptotically efficient estimators and provides asymptotic chi-square tests that are free from nuisance parameters. FMOLS and CCR techniques allow asymptotic coherence to be acquired by evaluating the influence of serial correlation. Therefore, long-term elasticity is evaluated in this research by utilizing FMOLS and CCR estimators using Eq. (5).

4.7 Pairwise Granger causality test

The present research intends to capture the causal effects between the variables. Therefore, we utilize the pairwise linear Granger-causality test proposed by Granger (1969) to examine if there is a causal association between the variables. Granger causality is a ‘statistical concept of causality based on prediction’ which has a number of advantages over other time-series analysis methods (Winterhalder et al. 2005) and hence is adopted in the present study. If one-time series Y can help predict the future of another time series X, then Y “Granger-causes” X. The time series of these two variables have data length T, denoting their values at time t by X_t and Y_t (t = 1,2, …,T), respectively. However, X_t and Y_t can be modeled by a bivariate autoregressive model:

$$ \text{X}_{t}\sum _{\text{l}=1}^{\text{p}}\left({\text{a}}_{11,\text{l}}{\text{X}}_{\text{t}-\text{l}}+{\text{a}}_{12,\text{l}}{\text{Y}}_{\text{t}-\text{l}}\right)+{{\upepsilon }}_{\text{t}}$$

(6)

$$ \text{Y}_{t}\sum _{\text{l}=1}^{\text{p}}\left({\text{a}}_{21,\text{l}}{\text{X}}_{\text{t}-\text{l}}+{\text{a}}_{22,\text{l}}{\text{Y}}_{\text{t}-\text{l}}\right)+{{\upxi }}_{\text{t}}$$

(7)

where p is the model order, a_ij,l (i,j = 1,2) are coefficients of the model, and ${{\upepsilon }}_{\text{t}}$ and ${{\upxi }}_{\text{t}}$ represent residuals. The coefficients can be estimated by ordinary least squares, and the Granger causality between X and Y can be detected by F tests (Tam et al. 2013).

5 Empirical Findings

5.1 Summary statistic of data

The outcomes of the summary measures amid variables are shown in Table 2 with the statistical values of different normality tests (skewness, probability, kurtosis, and Jarque-Bera) used. Each variable includes 30 observations of time series data from 1990 to 2019 for Malaysia. The negative values of skewness by the variables imply that all the variables adhere to normality. Furthermore, the research employed kurtosis to evaluate if the series is light-tailed or heavy-tailed relative to normal distribution. The empirical findings indicate that all the series are platykurtic as their values are less than 3. In addition, the results of the Jarque-Bera probability reveal that all of the parameters are normal. The unit root test for stationarity of the variables, as well as further analysis of the DOLS estimation, are based on these statistics.

Table 2 Summary statistics of the variables

5.2 Correlation between the variables

Correlation analysis to test for linear relationships between the variables is presented in Table 3. The findings reveal that all the variables are correlated to one another. LCO2, LEU, and LAGL indicate a very strong and positive correlation with each other which implies that when the value of one variable increases, the value of the other variable also tends to increase and vice versa. Nevertheless, LFA shows a negative correlation with all the other variables, which reveals that when the value of the forested area increases, the value of the other variable tends to decrease and vice versa. The correlation analysis led us to proceed with the unit root tests for the stationarity of the variables.

Table 3 The results of the correlation analysis

5.3 Results of unit root tests

We used unit root tests to check for integration orders. To find the autoregressive unit root, we used the augmented Dickey-Fuller (ADF), and Dickey-Fuller generalized least squares (DF-GLS) approaches based on trends and constants. Table 4 presents the findings of unit root testing using ADF and DF-GLS. The ADF test result reveals that LCO2, LEU, and LFA were stationary at the level and stayed stationary after taking the first difference, whereas LAGL was non-stationary at the level but became stationary at the first difference. In addition, the DF-GLS test revealed that LAGL and LFA were stationary at the level and stayed stationary after taking the first difference, whereas LCO2 and LEU were found non-stationary at the level but became stationary at the first difference. Hence, the occurrence of mixed orders integration for variables estimated by the ADF and DF-GLS justifies using the ARDL bounds test and DOLS cointegration test.

Table 4 Results of unit root tests by Augmented Dickey-Fuller (ADF), and Dickey-Fuller generalized least squares (DF-GLS)

5.4 Results of ARDL bounds test

After the stationarity characteristics of the series are affirmed, we proceed to conduct the ARDL bounds test for cointegration valuation. This analysis chose a reasonable lag period to measure the F-statistic constructed on the minimum values of the Akaike Information Criterion (AIC) proposed by Akaike (1987). Table 5 depicts the ARDL bounds test results to explore the cointegration link among the variables. The outcomes are presented in such a manner that the existence of a long-run association between the parameters is verified if the estimated value of the F-test is greater than the values of both limits (lower and upper bound). The findings reveal that the estimated F-statistic value 5.243503 is higher than 10%, 5%, 2.5%, and 1% of the crucial upper limit in the order zero and one, which rejects the null hypothesis by indicating that long-run relationship exists among the respective variables.

Table 5 Findings from cointegration with bounds testing

5.5 DOLS outcomes

The outcomes of the DOLS estimated by using Eq. (5) are presented in Table 6. The estimated coefficient of LEU is positive and significant at a 1% level, which implies that a 1% increase in energy use will lead to a 0.91% increase in CO₂ emissions when other indicators are held constant. The finding reveals that energy use triggers environmental degradation in the long run. Furthermore, the estimated coefficient of agricultural land is positive and significant at a 1% level, which indicates that agricultural land expansion by 1% is associated with an increase in CO₂ emissions by 0.84% in the long run. This reveals that agricultural land expansion deteriorates environmental quality. Finally, the estimated coefficient of forested area is negative and significant at a 1% level, which implies that a 1% reduction of the forested area due to deforestation has an impact of rising 5.41% of CO₂ emissions in Malaysia. The empirical finding suggests that deforestation prompts environmental degradation while increasing forested area improves the environmental quality as the forest ecosystems absorb the atmospheric CO₂ and store it in tree biomass. The result demonstrates that the forest ecosystems can be utilized as a tool to keep the environment clean by avoiding deforestation and increasing forest protection and conservation in Malaysia.

Moreover, it is notable that the indications of the calculated coefficients are compatible both from the theoretical point of view and practical expectations. In addition, we evaluated our estimated model’s goodness of fit using various diagnostic tests. First, the values of R² and adjusted R² are 0.9747 and 0.9718, respectively, indicating a very excellent fitting of the calculated regression model. This indicates 97% of the variance in the change of dependent variable can be explained by the independent factors. Second, the F- statistic indicates that the calculated DOLS regression is supported by its dependent and independent variables. The p-value of F- statistic is 0.0000, suggesting that the linear relation of the model is statistically significant. Third, the root mean square error (RMSE) offers an accurate estimate of model predictions for multiple periods. The value of RMSE is 0.0677 (near to 0) and non-negative, suggesting the results of the DOLS model was an almost perfect fit to the data

Table 6 The outcomes of DOLS: dependent variable LCO2

5.6 Robustness check

We utilized the FMOLS and CCR estimators to verify the consistency of DOLS estimation. The FMOLS and CCR estimation results for the model are presented in Tables 7 and 8. The outcomes of FMOLS and CCR provide evidence of the robustness of the DOLS estimation. The FMOLS and CCR results confirmed the coefficient of energy use and agricultural land is positive and significant at a 1% level. The results further validated the inverse relationship between CO₂ emissions and deforestation at a 1% level of significance. Hence, it can be stated that energy use, agricultural land expansion, and deforestation increase CO₂ emissions in Malaysia while increasing forested area helps to reduce CO₂ emissions. The results of the FMOLS and CCR are duly in line with the findings from DOLS outcomes. The R² and adjusted R² values from FMOLS and CCR estimation reflect the model’s goodness of fit, indicating that the independent variables can explain 97% of the variation in the dependent variable’s change.

Table 7 The results of FMOLS: dependent variable LCO2

Table 8 The results of CCR: dependent variable LCO2

5.7 Diagnostic inspection

This study performed normality, heteroscedasticity, and serial correlation analysis to validate the intensity of the cointegration valuation. The diagnostic test results are presented in Table 9. The model indicates normality and nonexistence of autocorrelation and heteroscedasticity. In addition, we employed the cumulative sum of squares of recursive residuals (CUSUM of Squares) test to check the stability of the model. The CUSUM of Squares plot at 5% significance level is presented in Fig. 6. The CUSUM of Squares plot from the recursive estimation of the model indicates that the model was stable as the residuals were within the critical bounds of the 5% significance level, and there is nonexistence of structural break.

Table 9 The results of diagnostic tests

Fig. 6

Plot of CUSUM of Squares (critical bounds at 5% significance level)

5.8 Results of Pairwise Granger causality test

The relationship between the variables indicates that there is the existence of Granger-causality, which the F-statistic determines. The summary of pairwise Granger causality is presented in Table 10, including the causality direction between the variables, such as left to right (→), right to left (←), and bidirectional causality (↔) when both variables cause each other. The pairwise Granger causality test results indicate that LEU and LCO2, LAGL and LCO2, LFA and LCO2, LAGL and LEU, LFA and LAGL show unidirectional causality due to statistically significance leading to the rejection of the null hypothesis. This indicates that energy use causes CO₂ emissions, agricultural land expansion causes CO₂ emissions, deforestation causes CO₂ emissions, agricultural land expansion causes energy use, and agricultural land expansion causes deforestation in Malaysia. Furthermore, the pairwise Granger causality test shows bidirectional causality between LFA and LEU, which implies that deforestation causes energy use and energy use causes deforestation.

Table 10 The results of pairwise Granger causality test

6 Discussion

The present study investigates the interconnection between energy use and environmental pollution for the case of Malaysia, and the empirical outcomes from the DOLS model illustrate that energy use exerts a positive and significant impact on CO₂ emissions in the long run. The result indicates that an increase in energy usage deteriorates the quality of the environment in Malaysia. Our finding is supported by several studies, such as, Adebayo (2021), Akbota and Baek (2018), Adebayo and Kalmaz (2021), Nondo and Kahsai (2020), Kirikkaleli and Kalmaz (2020), Adebayo (2020), Odugbesan and Adebayo (2020), Vo et al. (2019), Irfan and Shaw (2017), Zmami and Ben-Salha (2020), and Adebayo et al. (2020). Furthermore, our result about the positive association between energy use and CO₂ emissions by the present study is in line with other Malaysian studies by Ang (2008); Azlina and Mustapha (2012); Hossain (2011); Begum et al. (2015); and Sarkar et al. (2019). In addition, we found unidirectional causality from energy use to CO₂ emissions, which is supported by Ang (2008), Hossain (2011), Sarkar et al. (2019), Vo et al. (2019), Adebayo (2020), and Adebayo (2021).

However, energy consumption has a substantial and positive influence on CO₂ emissions, as the electricity, industrial, and transportation sectors account for a large percentage of CO₂ emissions in Malaysia. Begum et al. (2015) reported that technological advancements that increase energy efficiency (e.g., solar, wind, nuclear) are favorable to lowering CO₂ emissions while maintaining economic development. Nevertheless, Southeast Asia’s renewable energy development is accelerating due to fast population expansion and limited fossil fuel supply. (Pratiwi and Juerges 2020). With the impending threat of climate change, renewable energy is viewed as a viable alternative option for sustainable development as well as climate change mitigation (Seriño 2018; Pratiwi and Juerges 2020). Renewable energy delivers significant economic advantages in addition to decreasing carbon emissions, such as increased energy availability, improved energy security, and the use of local renewable resources (Seriño 2018). As a result of the growing global environmental consciousness, it is important to shift Malaysia’s energy balance to renewables in order to enable the use of sustainable energy sources and build an environmentally sustainable ecosystem.

Moreover, this research uncovers a positive link between agriculture and CO₂ emissions, implying that agricultural land expansion leads to environmental degradation. The outcome indicates that the agriculture sector is a major source of CO₂ emissions in Malaysia. However, numerous studies established the positive link between agriculture and CO₂ emissions, which supports our finding. For example, Sarkodie and Owusu (2016), Liu et al. (2017), Jebli and Youssef (2017a), Doğan (2018), Gokmenoglu and Taspinar (2018), Waheed et al. (2018), Ullah et al. (2018), Agboola and Bekun (2019), Balsalobre-Lorente et al. (2019), Burakov (2019), Qiao et al. (2019), Gokmenoglu et al. (2019), Aydoğan and Vardar (2020), Olanipekun et al. (2019), Aziz et al. (2020), Naseem et al. (2020) and Pata (2021). In addition, the present study reveals unidirectional causality from agriculture to CO₂ emissions, which is in line with the findings by Burakov (2019) and Naseem et al. (2020).

The outcome of our study suggests that traditional agricultural methods should be replaced by modern technologies, which would increase agricultural production and reduce emissions by reducing the necessity for agricultural land expansion to satisfy the demand from the growing population. Bayrakcı and Koçar (2012) reported that diversified renewable energy sources could be introduced in agricultural activities to reduce emissions from the agriculture sector. For example, (i) solar energy can be used for lighting, product drying, and irrigation; (ii) biofuels such as bioethanol and biogas, as well as different agricultural wastes, can be utilized as energy sources; (iii) wind energy can be utilized to power generators, irrigate farms, and grind certain crops; (iv) hydropower can be used to generate energy, irrigate crops, and provide drinking water, as well as facilitating the equitable distribution of water among farmers.

Furthermore, the IPCC (2014) reported that achieving the goal of reducing GHGs emissions from the agriculture sector would not only result in a cleaner environment, but it will also provide new sources of income as more farming operations may be carried out. Various international organizations have recently developed a climate-smart agriculture (CSA) method to change agricultural growth in order to mitigate environmental damage (FAO 2019). These projects have a long-term impact on global climate change reduction and mitigation. Ridzuan et al. (2020) reported that the agriculture industry might help to reduce GHGs emissions by using correct farming techniques. Carbon release may be stored by agricultural operations when proper management and technology are used, resulting in a reduction in carbon footprint.

However, our investigation reveals that forested area has a negative impact on CO₂ emissions in Malaysia. Thus, it confirms that reducing the forested area through deforestation increases CO₂ emissions and contributes to global climate change. Instead, forest ecosystems improve the quality of the environment as the forests absorb the atmospheric CO₂ and store it in tree biomass. Our empirical finding indicates that enhancing forest carbon sink by increasing forested area mitigates environmental degradation in the long run. However, Islam et al. (2017), Waheed et al. (2018), Shittu et al. (2018), Parajuli et al. (2019), Aziz et al. (2020), and Begum et al. (2020) established a negative relationship between forested area and CO₂ emissions which validate our result. As the second-biggest source of anthropogenic CO₂ emissions into the atmosphere, forest loss has been considered a driver of environmental deterioration (IPCC 2014). Nordhaus (2006) reported that controlling deforestation is the simplest approach to reduce CO₂ emissions. Raihan et al. (2019) reported that enhancing forest carbon sequestration is the most cost-effective way to reduce environmental degradation and mitigate global climate change.

Reversing forest losses via restoration, enhancement, and conservation is a crucial aim for climate change mitigation, and it is a hot topic in today’s climate debate (Matthew et al. 2018). Implementing cost-effective mitigation strategies in the forestry sector to prevent deforestation and forest degradation can reduce global carbon emissions and avert climate change at the lowest cost (Raihan et al. 2018). Furthermore, forestry-based mitigation strategies (forest protection, afforestation, natural regeneration) would serve a multifunctional purpose, including carbon sequestration, biodiversity conservation, ecosystem enhancement, and community outputs of goods and services (Raihan and Said 2021). As the forest carbon sequestration rate in Malaysia is relatively high due to the rapid growth of plants (Raihan et al. 2021), Malaysian forests have a huge potential to mitigate global climate change by reducing CO₂ emissions and increasing forest biomass by enhancing the national carbon sink through the widespread implementation of forestry-based mitigation measures.

Moreover, the pairwise Granger causality test shows bidirectional causality between deforestation and energy use which implies that deforestation causes energy use and energy use causes deforestation. However, deforestation requires machinery for timber harvesting and transportation of harvested timber to wood industries which increases energy consumption. Conversely, increased energy use from urbanization, industrialization, and agricultural production requires more land for settlement, agricultural land, building industries, road, and highways, which lead to deforestation. Furthermore, an interesting result from the pairwise Granger causality test is the significant unidirectional causality from forested area to agricultural land, implying that agricultural land expansion causes deforestation in Malaysia. However, one of the most prevalent causes of deforestation in developing nations is land clearing and conversion for agricultural use (Galinato and Galinato 2016). Agricultural growth can account for two-thirds of the change in tropical forest cover. In tropical areas, encroachment of large-scale agricultural production, small-scale agricultural production, and shifting cultivation account for 32%, 26%, and 15% of forest cover change, respectively (FAO 2010). López (2000) estimated that a one-hectare increase in the cultivated area results in a 4.4-hectare reduction in forest cover because the conversion of forest cover to agricultural land necessitates extra land clearing for human settlement and infrastructure supporting agricultural production.

Nevertheless, the environmental degradation generated by agriculture in Malaysia is linked to CO₂ emissions from deforestation due to agricultural land expansion for oil palm plantations. Uning et al. (2020) reported that oil palm production had transformed large areas of Malaysian forest ecosystems over the last three decades, and it is anticipated to be one of the primary sources of GHG emissions connected to land use. However, oil palm production is the mainstay of Malaysian agriculture that contributes the lion’s share in Malaysia’s economic growth. Thus, high profit from oil palm production triggers agricultural land expansion for oil palm plantations, leading to deforestation due to the scarcity of available plantation areas. Consequently, shrinkage of forested area deteriorates the quality of the environment by releasing a substantial amount of CO₂ back into the atmosphere, which was once stored in forests. Furthermore, enacting fire-related practices to remove vegetation and forest structure is part of the conversion process from forest to oil palm plantations. Controlled burning has played a key role in CO₂ emissions from oil palm production, as well as the long-term deterioration of tropical forest ecosystems. Hence, preventing forest ecosystems from being deforested for oil palm production has been identified as a major challenge in Malaysia’s effort against climate change and environmental degradation.

7 Conclusion and policy implications

This study investigates the dynamic impacts of energy use, agricultural land expansion, and deforestation on CO₂ emissions by using time series data from 1990 to 2019. The current study utilized ADF and DF-GLS unit root tests to capture the integration order of the series. We applied the ARDL bounds test to capture the cointegration amongst the series. The DOLS estimator was employed to capture the long-run impacts of energy use, agricultural land expansion, and deforestation on CO₂ emissions in Malaysia. In addition, We employed the FMOLS and CCR test as a robustness check to the DOLS estimation. Furthermore, the pairwise Granger causality test was utilized to check the causal relation amid study variables. The empirical findings indicate that increased energy use, agricultural land expansion, and deforestation in Malaysia have an adverse effect on environmental degradation in the long run. The policy recommendation is drawn following the study outcomes to promote green energy, climate-smart agriculture, and sustainable management of forest ecosystems that will ensure emission reduction in Malaysia.

This research’s outcomes have been recommended in adopting the promotion of energy intensity diversification of Malaysia. Developing and adopting sound policies to moderate Malaysia’s energy and manufacturing sector practices will improve the country’s sustainable growth. If the government imposes CO₂ emission limitations on businesses and industries, this will continue to manage CO₂ pollution levels. The fear of severe action or hefty taxes for those who violate this regulation will help to reduce pollution. Furthermore, Malaysia should introduce clearer policies toward enhancing energy efficiency and energy usage programs to minimize unnecessary energy waste. To enhance energy efficiency and therefore reduce CO₂ emissions, the government should expand investment in fossil fuel energy cleansing technology. Furthermore, carbon intensity should be reduced by using more renewable energy sources for energy generation, such as hydropower, ocean power, geothermal, wind power, and solar. Government and policymakers should pay more attention to promoting renewable energy in Malaysia. Moreover, government spending on renewable energy infrastructure and scientific advancements should be increased. Malaysia must develop measures to lower the cost of renewable energy and discourage the use of fossil fuels in companies and households since renewable energy use can help to reduce emissions. Research and development of green technology should be undertaken to boost domestic investment and reduce pollution. However, because the ongoing COVID-19 epidemic has had an impact on energy consumption patterns, the government should carefully create policies. In Malaysia, for example, residential energy use has increased while transportation energy consumption has decreased. Hence, authorities should focus on boosting energy-efficient resident electric appliances and more cheap renewable options for the household sector.

However, this study recommends that the Malaysian government exercise caution when designing policies to improve agricultural production, particularly through agricultural land expansion for oil palm and rubber plantations, as this might harm the environmental quality. Increased efforts are needed to boost agricultural productivity by implementing modern agro-based technologies such as high-yield and disease-resistant crop varieties and land management, as well as encouraging farmers to abandon traditional farming practices in pursuit of more advanced agrarian techniques. Moreover, agricultural productivity and value-added component may be improved at a greater level with the help of contemporary agricultural technology and the availability of good seeds and other agricultural inputs. Sustainable agriculture by developing organic and low carbon agriculture systems can reduce emissions and enhance carbon sequestration. To achieve long-term agricultural productivity, the government should encourage more efficient energy infrastructure and support the switch to cleaner, more efficient energy sources in agriculture. Governments should support the use of renewable energy, particularly clean renewable energy like solar and wind, because it boosts agricultural productivity while also helping to battle global warming and climate change. Subsidies for renewable energy use in agriculture, according to this study, would help the industry become more competitive on worldwide markets while emitting less pollution. For a carbon-neutral environment, irrigation methods can be switched from non-renewable to renewable energy sources. Other important agricultural changes include encouraging farming communities to use solar tube wells for irrigation, organic farming, tunnel farming, changing traditional tillage to no-till, and reducing fertilizer use to decrease environmental impact. These contemporary agriculture technologies can help large farms cut personnel, improve productivity, and cut emissions. For the sake of sustainable agriculture and pollution reduction, excessive use of fertilizers and pesticides must be avoided, and green production must be prioritized. The agriculture industry may have a significant positive impact on the environment by using an organic framework. Furthermore, boosting agricultural investment in Malaysia through improved international collaboration will aid in the reduction of emissions from Malaysia’s agriculture sector.

The outcomes of our investigation recommend the policymakers in Malaysia implement effective environmental and climate-resilient policies with more focus on reducing CO₂ emissions by enhancing forest ecosystems. To reduce CO₂ emissions from deforestation, Malaysia’s government should boost financial investment and implement robust forest laws and policies. Strong forest conservation policies are essential to protect biodiversity by avoiding deforestation in Malaysian forests, which are rich in species diversity. Forest conservation with the participation of the local community is a novel approach to preserve forest conservation while ensuring the local community’s livelihood. Furthermore, Malaysia’s forestry sector has enormous tourist potential, and encouraging forest protection through ecotourism might boost the country’s economy. Moreover, through the formation of private forest plantation areas, the government should promote private investment in forest development. Malaysia, on the other hand, may improve its climate change mitigation potential by implementing many forestry-based mitigation measures, such as forest conservation, afforestation, reforestation, sustainable forest management, enhanced natural regeneration, agroforestry, urban forestry, and wood-based bioenergy. Finally, the actual implementation of forestry policy could help Malaysia to be an emission-free country by enhancing the national carbon sink while maintaining national green growth and sustainable management of the forest ecosystems.

Moreover, the results also suggest that a higher degree of energy use, agricultural land expansion, and deforestation for oil-palm plantations trigger environmental degradation in the long run, which could be reduced by implementing more structured environmental policy, agriculture policy, renewable energy policy, green technology policy, and forestry policy. Although the current study has produced substantial empirical findings in the instance of Malaysia, our analysis has numerous flaws that might be addressed in future research. One of the critical drawbacks of our analysis is the unavailability of the data related to the forested area beyond the period of study, which limits the power of the econometric techniques used. However, this study has examined the dynamic impacts of energy use, agricultural land expansion, and deforestation on CO₂ emissions in Malaysia. Further studies can explore the other determinants of CO₂ emissions, such as urbanization, industrialization, tourism, etc. Furthermore, this study utilized CO₂ as an indicator for environmental pollution. Future research should investigate more environmental pollution indicators, such as water pollution and land pollution.