Modeling the effects of environmental degradation, renewable energy, and technological advancement on food security in Sierra Leone: Does institutional quality matter?

doi:10.21203/rs.3.rs-8231003/v1

Download PDF

Research Article

Modeling the effects of environmental degradation, renewable energy, and technological advancement on food security in Sierra Leone: Does institutional quality matter?

https://doi.org/10.21203/rs.3.rs-8231003/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Background

Achieving food security in Sierra Leone is increasingly constrained by environmental degradation, climate variability, and weak institutional capacity. Despite growing challenges, empirical evidence on the combined influence of environmental stressors, renewable energy use, technological adoption, and institutional quality on food security remains scarce. Food security in this study is assessed through a multidimensional lens, focusing on the production of key staple and export crops.

Methods

The study analyzes both direct and mediated effects using time series data from 1990 to 2023. Dynamic Autoregressive Distributed Lag (DYNARDL) simulations and Fully Modified Ordinary Least Squares (FMOLS) are employed to capture short- and long-term dynamics, nonlinearities, and asymmetric responses to shocks.

Results

Renewable energy adoption and technological progress significantly enhance agricultural output, with institutional quality serving as a partial mediator. Conversely, air pollution and temperature variability consistently reduce crop yields, underscoring agriculture’s vulnerability to ecological and climatic stress. Nonlinear estimates indicate diminishing returns when energy and technology inputs exceed optimal levels, while dynamic simulations reveal asymmetric effects between positive and negative shocks. Robustness checks confirm that clean energy and modern inputs bolster production of rice, maize, cassava, cocoa, vegetables, and fruits, whereas environmental degradation uniformly depresses output.

Conclusion

The findings highlight the urgent need for integrated strategies that combine clean energy transition, sustainable technological innovation, environmental stewardship, and institutional reform. The study offers policy-relevant insights for building climate-resilient food systems in Sierra Leone and other low-income, environmentally vulnerable regions.

Food security

Environmental degradation

Renewable energy

Technological advancement

Institutional quality

Sierra Leone

Food security remains one of the most urgent developmental and humanitarian challenges of the twenty-first century, particularly in low income, climate vulnerable, and institutionally fragile countries such as Sierra Leone. Globally, over 2.4 billion people lack regular access to safe, nutritious, and sufficient food, and approximately 600 million are projected to experience hunger by 2030 [1]. Sub-Saharan Africa is disproportionately affected, with nearly 20% of its population facing severe food insecurity, compared to 8.5% in Asia, 7% in Oceania, and 6.5% in Latin America and the Caribbean [1]. Alarmingly, one in every five Africans goes to bed hungry [2]. Food insecurity varies significantly among countries in the Global South. In Sierra Leone, over half of the population is food insecure due to persistent poverty, limited agricultural capacity, and elevated exposure to environmental and climate risks [3].

In response to the global food crisis, the United Nations adopted “Zero Hunger” as the second Sustainable Development Goal¹. This goal aims to eliminate hunger, ensure food access, improve nutrition, and promote sustainable agriculture. However, progress toward this objective in Sierra Leone is increasingly constrained by the combined effects of environmental degradation, climate variability, energy scarcity, and limited technological development. Agriculture in Sierra Leone is primarily dependent on rainfall and is characterized by low productivity, weak value chains, and inadequate irrigation systems [4]. Additionally, land degradation, irregular rainfall patterns, deforestation, and increasing greenhouse gas emissions have intensified food insecurity and contributed to a persistent cycle of poverty, malnutrition, and vulnerability [5].

Climate change is widely acknowledged as a key driver of hunger, although its effects vary considerably across different regions and agricultural systems [6, 7]. In Africa, where rain-fed agriculture remains dominant, the impacts of climate change are particularly severe [8]. Projections suggest that average temperatures across the continent could increase by 2 to 4 degrees Celsius by the year 2100, nearly 1.5 times greater than the global average [9]. Empirical studies indicate that changes in temperature, precipitation, and atmospheric carbon dioxide concentrations negatively affect crop yields, reduce food quality, raise food prices, and increase the likelihood of famine [5, 10–13]. Conversely, other studies suggest that certain crops may benefit from specific climate conditions, thereby improving food availability, affordability, and access in localized contexts [14–16]. These mixed findings highlight the complexity and context dependence of the relationship between climate change and food security.

Sierra Leone is already undergoing significant climatic changes. The dry season typically spans from November to April, with temperatures ranging from 25°C to 35°C and occasional extreme heat events [4]. Under the most adverse climate scenario, the country’s average annual temperature is projected to increase from a baseline of 26.5°C to approximately 28°C by 2050 [4]. Rising temperatures, coupled with changes in precipitation patterns and sea level rise, elevate the risk of floods, droughts, and land degradation. Although future rainfall trends remain uncertain, Sierra Leone currently experiences among the highest levels of precipitation worldwide, averaging over 2,600 millimeters annually². Projections suggest an increase in both prolonged dry spells and extreme rainfall events, which pose serious threats to agricultural productivity and food security. If left unaddressed, these changes will likely exacerbate existing vulnerabilities.

Sierra Leone’s dependence on rainfall and its limited adaptive capacity make its agricultural system particularly sensitive to climatic and environmental shocks. Disruptions in domestic food production also have international implications, given the country’s role in exporting crops such as cocoa, coffee, and oil palm. Any decline in these sectors may adversely affect both the global food supply chain and the national economy. Moreover, pollution from fossil fuel consumption and other unsustainable energy practices has contributed to increased greenhouse gas emissions³. In recognition of these interconnected challenges, global institutions such as the Intergovernmental Panel on Climate Change and the Sustainable Development Goals framework have emphasized the need for a transition to clean energy systems.

In recent years, scholars and policymakers have increasingly focused on the potential of renewable energy and technological innovation to enhance food security through climate resilient and energy efficient agricultural practices [5, 13, 17, 18]. Renewable sources such as solar, hydro, and wind energy can reduce dependence on fossil fuels, support the development of irrigation and storage infrastructure, and help mitigate environmental degradation [18]. In parallel, technological advancements including improved seed varieties, digital extension services, fertilizer and pesticide application, and precision farming have the potential to increase agricultural productivity and market efficiency [17]. However, the extent to which these innovations improve food security depends on a country’s institutional capacity, particularly in terms of policy coordination, regulation, and investment [19].

Despite the growing body of research on food security, limited empirical work has explored the combined effects of renewable energy, environmental degradation, and technological progress within a unified analytical framework, especially in fragile contexts such as Sierra Leone. Most existing studies examine these drivers separately or focus on related outcomes such as carbon emissions [20, 21], economic growth [22], or climate adaptation [23]. Furthermore, food security is seldom the central variable of analysis and is often represented by a single indicator. Additionally, while institutional quality is frequently identified as an important contextual factor influencing development outcomes [19, 24], its potential role as a mediator or moderator in the relationship between environmental and technological drivers and food security remains largely unexplored. In low-income countries with weak governance structures and high exposure to climate risks, understanding how institutional quality shapes this relationship is vital for designing effective policy interventions.

This study addresses these gaps by investigating the combined effects of renewable energy use, environmental degradation, and technological advancement on food security in Sierra Leone. A key focus is placed on examining whether institutional quality serves as a mediating or moderating factor in these relationships. The study seeks to provide an integrated and context specific contribution to the literature on food security, environmental economics, and development governance. The research is guided by the following questions: (1) What are the effects of renewable energy adoption, environmental degradation, and technological advancement on food security in Sierra Leone? (2) Are there nonlinear or threshold effects that influence the direction or intensity of these relationships? (3) To what extent does institutional quality moderate the adverse effects of environmental degradation on food security? (4) Does institutional quality mediate the influence of renewable energy and technological advancement on food security outcomes? (5) What policy recommendations emerge for enhancing food security in resource constrained and environmentally vulnerable settings?

This study makes four main contributions to the existing literature. First, it develops a novel integrated framework that jointly examines renewable energy, environmental degradation, and technological advancement as interrelated drivers of food security over both the short term and the long term. This approach contrasts with earlier studies that have analyzed these factors in isolation [5, 12]. The study also uses a multidimensional measure of food security, drawing on the Food Production Index and data on the output of major crops such as rice, maize, cassava, cocoa, vegetables, and fruits. Second, the research introduces a composite index of institutional quality, developed using Principal Component Analysis. This index is employed both as a mediating and moderating variable to assess how governance influences the relationships between environmental and technological factors and food security. While institutional quality is often mentioned in theoretical discussions, few empirical studies rigorously evaluate its role using statistical interaction or mediation models [19]. Third, the study focuses on Sierra Leone, a post conflict, low-income country that is underrepresented in empirical food security research. Most studies in the region focus on larger or more politically stable economies [14, 25, 26]. This research therefore enhances the geographical and contextual relevance of the literature. Fourth, the study adopts a multi-method econometric strategy, applying Dynamic Nonlinear Autoregressive Distributed Lag, Kernel Regularized Least Squares, and Fully Modified Ordinary Least Squares models. These methods allow for the estimation of dynamic and nonlinear relationships and address potential issues of endogeneity and serial correlation [27, 28]. The inclusion of nonlinear and asymmetric specifications permits the identification of threshold effects that conventional models often overlook. These methodological innovations also support dynamic simulations and counterfactual analysis under various policy scenarios, thereby enhancing the practical policy relevance of the findings.

The structure of this study is as follows: Section One outlines the background and research rationale; Section Two reviews the literature and theoretical framework; Section Three details the data and methodology; Section Four presents and interprets the results; and Section Five concludes with key findings and policy implications.

2.1. Food security in Sierra Leone: A contextual overview

Food security remains a central development concern in Sierra Leone, influenced by complex structural, environmental, economic, and institutional dynamics. According to the Food and Agriculture Organization (FAO), food security exists when all individuals, at all times, have physical, social, and economic access to sufficient, safe, and nutritious food that meets their dietary needs and food preferences for an active and healthy life [29]. This comprehensive definition highlights four interrelated dimensions: availability, access, utilization, and stability. Sierra Leone faces persistent challenges in ensuring food security across these dimensions. The country is classified as a low-income and fragile state and ranked 185th out of 193 countries in the 2024 Human Development Index⁴.. Agriculture is the backbone of the economy, employing over 75% of the labor force and contributing approximately 64% to the gross domestic product [30]. Despite its centrality to national livelihoods, the agricultural sector is predominantly subsistence-based and marked by low productivity, poor access to quality inputs, limited mechanization, inadequate infrastructure, and minimal technological adoption.

Environmental degradation and climatic variability further threaten agricultural productivity. Practices such as slash-and-burn cultivation, coupled with widespread deforestation and unsustainable mining, have led to severe land degradation, soil erosion, and reduced arable land quality [30, 31]. These issues are exacerbated by increasingly frequent and intense climatic events, floods, erratic rainfall, and prolonged dry spells, that undermine production cycles and food availability. Recent data from the 2023 Comprehensive Food Security and Vulnerability Analysis (CFSVA) indicate that more than 50% of the population, approximately 4.7 million individuals, are food insecure, with over 1.2 million classified as severely food insecure [3]. These statistics underscore the urgent need for integrated and resilient strategies to address both immediate and structural drivers of food insecurity.

In response to these challenges, the government of Sierra Leone has articulated several policy frameworks aimed at promoting agricultural transformation and food system resilience. The Medium-Term National Development Plan (2019–2023) identified food and nutrition security as a national priority. Building on this agenda, the National Agricultural Transformation Programme (2019–2025) emphasizes sustainable agricultural intensification, climate-smart practices, and improved access to markets and inputs [32]. Furthermore, a key policy initiative relevant to the current study is the Renewable Energy Policy (2020), which aims to expand access to clean and affordable energy in rural and underserved regions. This policy encourages the deployment of renewable energy technologies, such as solar-powered irrigation systems, cold storage facilities, and decentralized mini-grids, to enhance agricultural productivity and reduce post-harvest losses. By aligning with SDGs 2 (Zero Hunger) and 7 (Affordable and Clean Energy), the policy envisions a synergistic role for renewable energy in supporting agricultural resilience and food security. Despite these policy efforts, major challenges remain. Institutional weaknesses, inadequate financing, limited technical capacity, and fragmented coordination among stakeholders continue to hinder effective implementation. Moreover, empirical research on how renewable energy, environmental degradation, and technological advancement jointly affect food security in Sierra Leone is limited. In particular, the mediating role of institutional quality in shaping these interactions has received insufficient scholarly attention.

2.2. Renewable energy and food security

The relationship between renewable energy and food security has attracted increasing scholarly attention, especially in regions where agricultural productivity is constrained by limited access to affordable and reliable energy [33]. Renewable energy sources such as solar, hydro, wind, and biomass offer alternative energy options that may support agricultural systems in achieving greater efficiency and resilience [18, 34, 35]. These technologies are used to improve irrigation systems, enhance water pumping, support postharvest processing, and enable cold storage facilities [36, 37]. Through these applications, renewable energy has the potential to reduce postharvest losses and improve food availability, which may influence food security outcomes. Studies from SSA countries suggest that the integration of renewable energy into agricultural practices could yield positive results. For instance, Burney, Woltering [34] observed that solar-powered irrigation systems were associated with improved crop yields and enhanced household food access in West Africa. Similarly, Majeed, Khan [18] reported that the use of renewable energy in postharvest processing was linked to reduced spoilage and improved food quality in rural areas. These findings reflect the growing interest in renewable energy as a tool for improving food system performance. However, the extent to which such outcomes can be replicated in other contexts remains an open empirical question.

In Sierra Leone, the deployment of renewable energy in the agricultural sector is still at a relatively early stage. Although the country possesses considerable solar and hydro potential, several structural and institutional challenges may limit the scale and impact of renewable energy adoption. These challenges include high initial investment costs, limited technical capacity, the absence of strong policy incentives, and regulatory constraints. As a result, agriculture continues to rely heavily on traditional and often inefficient energy sources. The extent to which renewable energy can be expanded and integrated into agriculture, and its potential effects on food security, requires further investigation [35]. Nevertheless, a number of small-scale renewable energy initiatives have been implemented in rural communities, including solar-powered rice mills and local mini-grid projects⁵. While some of these projects have reported positive operational outcomes, their scalability and long-term impact remain uncertain. These initiatives suggest that renewable energy may offer opportunities for supporting agricultural modernization in Sierra Leone, particularly if appropriate institutional and policy conditions are established to facilitate wider adoption. Building on this discussion, the following hypothesis is developed:

H₁

Renewable energy adoption has a positive effect on food security in Sierra Leone in the long run.

2.3. Environmental degradation and agricultural viability

Environmental degradation poses a significant threat to sustainable agricultural development, particularly in countries that depend heavily on natural resources for food production. As ecosystems deteriorate, their ability to support essential ecological functions, such as soil fertility maintenance, water regulation, and biodiversity preservation, is greatly diminished. According to Rockström [38], the breakdown of ecological systems undermines the core foundations upon which agriculture depends, leading to reduced productivity and increased exposure to environmental shocks. In Sierra Leone, the agricultural sector is increasingly affected by widespread deforestation, poor land use practices, unregulated mining, and the expansion of farming activities into environmentally sensitive areas [31]. These activities have contributed to severe land degradation, particularly in regions that are critical for staple food production. Soil erosion, the loss of vegetative cover, and declining water resources have been widely observed, contributing to stagnant crop yields and a growing risk of food insecurity. Crops such as rice, cassava, and maize, which form the dietary staple for much of the population, are particularly vulnerable to these environmental pressures.

The United Nations Environment Programme (UNEP, 2020) emphasizes that the degradation of land and natural ecosystems diminishes a country’s natural capital and impairs the delivery of vital ecosystem services. In many rural areas, where environmental degradation is most severe, food production has become increasingly unpredictable and, in some cases, insufficient to meet household and community needs. These challenges suggest an urgent need for sustainable land management strategies, reforestation programs, and the integration of climate resilient farming practices. However, the effectiveness of these interventions in restoring ecological balance and improving food security in Sierra Leone remains a critical area for further empirical investigation. Based on this discussion, the following hypothesis is developed:

H₂

Environmental degradation (proxied by air pollution, and temperature change) reduces food crop production in Sierra Leone.

2.4. Technological advancement: Inputs and productivity

Technological innovation is widely recognized as a major driver of agricultural productivity and a critical component in addressing food security challenges [26, 34]. Historical evidence from the Green Revolution illustrates how the adoption of improved seed varieties, chemical fertilizers, irrigation systems, and crop protection methods contributed to significant increases in agricultural output across many regions. In more recent contexts, emerging technologies such as digital agriculture, precision farming, and mobile-based advisory services have opened new opportunities for improving farming efficiency and sustainability [5, 17]. In the present study, technological advancement is proxied by the use of chemical fertilizers and pesticides. These inputs are central to modern agricultural practices, as they help improve crop yields by correcting soil nutrient deficiencies and managing pest and disease pressures. Ongoma, Brouziyne [39] emphasize that closing the yield gap in SSA requires not only increased input use but also complementary investments in infrastructure, research, and agricultural extension systems to support adoption.

In Sierra Leone, however, the use of such modern inputs remains relatively limited. Farmers often face barriers such as high input costs, inadequate access to credit, underdeveloped input markets, and weak extension services. Moreover, low levels of education and limited technical knowledge among smallholder farmers constrain the effective and efficient use of available technologies. While some government and donor-led programs have aimed to increase access to agricultural inputs, their impact has frequently been curtailed by poor implementation, lack of continuity, and weak institutional coordination. Improving the availability, affordability, and appropriate use of fertilizers and pesticides may contribute to enhanced agricultural productivity and increased food availability. However, realizing these potential benefits will likely require a coordinated and systemic approach that addresses supply-side constraints, builds institutional capacity, and promotes farmer education. Based on this discussion, the following hypothesis is developed:

H₃

Technological advancement, proxied by fertilizer use and pesticide application, has a positive and significant impact on food security in Sierra Leone.

2.5. Theoretical foundation

The theoretical foundation of this study suggests that institutional quality may serve as a mediating variable through which renewable energy adoption and technological advancement influence food security outcomes. Grounded in institutional theory and development economics, this perspective is based on the premise that structural progress in the energy and technology sectors is unlikely to result in sustained social or economic development without effective institutional frameworks [40, 41]. Institutions, encompassing formal rules, governance capacity, regulatory effectiveness, and accountability mechanisms, are critical in enabling the transformation of technological and energy innovations into inclusive and sustainable development outcomes [42]. Figure 1 shows the study’s conceptual framework, which outlines the hypothesized relationships among the key variables,

Within this framework, renewable energy and technological advancement are not only associated with improvements in production efficiency and environmental sustainability, but may also contribute to institutional development. The implementation of renewable energy technologies often requires coherent policy design, legal infrastructure, stakeholder coordination, and administrative capacity [21, 22, 43], all of which can enhance institutional performance. Likewise, the adoption of advanced technologies in agriculture and environmental management typically depends on responsive public administration, adequate investment in human capital, and transparent regulatory oversight [44, 45] These dynamics indicate a potential pathway through which progress in the energy and technology domains may support the strengthening of institutional quality.

Institutional quality, in turn, may influence the extent to which such advancements improve food security. Effective institutions help reduce uncertainty, protect property rights, allocate resources efficiently, and ensure the equitable delivery of services [46, 47]. These functions are essential for establishing the conditions necessary to enhance food availability, accessibility, utilization, and stability. For instance, effective governance can promote the equitable distribution of technological and energy resources, reinforce the resilience of food systems, and support the capacity of households and communities to adapt to environmental and economic shocks. Moreover, institutional quality may shape the relationship between environmental degradation and food security. In contexts where institutions are capable of enforcing environmental regulations, managing land and water resources, and promoting sustainable agricultural practices, the adverse effects of environmental degradation may be mitigated. Even in ecologically vulnerable settings, institutional capacity may play a role in maintaining agricultural productivity and protecting food distribution systems [48]. This framework outlines a mechanism in which renewable energy and technological advancement may lead to improvements in institutional quality, which subsequently influences governance, policy implementation, and the management of resources critical to food security. Institutional quality is therefore conceptualized not only as a background condition but also as a potential pathway through which development interventions may impact food system outcomes. Based on this theoretical foundation, the following hypotheses are developed for empirical testing in this study:

H₄

Renewable energy adoption has a positive effect on institutional quality.

H₅

Technological advancement positively influences institutional quality.

H₆

Institutional quality positively impacts food security outcomes.

H₇

Institutional quality mediates the relationship between renewable energy adoption and food security.

H₈

Institutional quality mediates the relationship between technological advancement and food security.

3.1. Study area and data sources

This study focuses on Sierra Leone, a low-income country in West Africa characterized by persistent food insecurity, environmental degradation, limited access to renewable energy, and weak institutional frameworks. These conditions make Sierra Leone a relevant case for examining the effects of renewable energy, environmental degradation, and technological advancement on food security, with institutional quality as a moderating factor. The period from 1990 to 2023 is selected due to data availability and its coverage of key historical, political, and environmental developments, including the civil conflict (1991–2002), post-war recovery, and recent shifts toward renewable energy and agricultural modernization. The study utilizes secondary annual time series data from reputable international sources, including the World Bank, FAO, and the Worldwide Governance Indicators. Variables include indicators of food security, renewable energy consumption, environmental degradation, technological advancement, and institutional quality.

3.2. Dependent variables

The primary dependent variable in this study is the Food Production Index, which reflects the aggregate volume of food output in Sierra Leone. While this index serves as a useful proxy for food availability, it does not fully capture the multidimensional nature of food security. To address this limitation, the study incorporates additional indicators in the robustness analysis, including the production of rice, vegetables, maize, cocoa, cassava, and fruits. These variables are selected to reflect broader dimensions of food security, particularly staple crop availability and dietary diversity. Rice and cassava are emphasized due to their status as the principal staple foods in Sierra Leone, while maize plays a vital role in household consumption and contributes to food variety. The inclusion of these crop-specific variables enables a more comprehensive and nuanced assessment of food security outcomes.

3.3. Independent variables

This study incorporates six key independent variables to assess the drivers of food security in Sierra Leone. These variables reflect critical environmental and technological dimensions that influence agricultural production and food availability. The variables are organized and described as follows: First, renewable energy refers to energy derived from non-fossil sources such as hydro, solar, wind, and biomass. Improved access to these sources can support agricultural activities through enhanced mechanization, irrigation, storage, and processing, leading to greater efficiency in food production and distribution. Second, climate mitigation involves efforts to reduce greenhouse gas emissions and promote sustainable agricultural practices. This variable is proxied through investment in clean technologies and environmental policy initiatives, enabling an assessment of the relationship between mitigation efforts and food system performance. Third, environmental degradation is represented by two variables: air pollution and temperature change. Air pollution, commonly measured by carbon dioxide emissions or particulate matter, may negatively impact crop health, soil quality, and ecosystem balance. Temperature change reflects climate variability, which can affect growing conditions, crop cycles, and yields, particularly relevant for Sierra Leone’s rain-fed agriculture. Fourth, technological advancement is proxied by fertilizer and pesticide use. Fertilizers enhance soil fertility and crop productivity, while pesticides are used to control pests, diseases, and weeds. Although these inputs contribute to yield improvements, their long-term environmental and health implications require careful consideration. Furthermore, a technological advancement index is developed using additional variables to capture broader innovation trends. This index is constructed using Principal Component Analysis (PCA), which reduces dimensionality and captures the shared variance among technological indicators. Table 1 presents the PCA results, and Fig. 2 displays the scree plot of eigenvalues. A Kaiser-Meyer-Olkin (KMO) value of 0.6481 confirms strong sampling adequacy and the robustness of the index. These variables provide a multidimensional framework for analyzing how environmental conditions and technological practices influence food security dynamics in Sierra Leone over time.

Table 1
Principal component results for technological advancement
	Technological advancement index
Variables	Component	Eigenvalue	Proportion	Cumulative	KMO-MSA
Internet subscriptions (total)	PC1	3.26153	0.6523	0.6523	0.1096
Mobile phone subscriptions (total)	PC2	1.10034	0.2201	0.8724	0.7521
Individual using the internet (% of population)	PC3	0.51712	0.1034	0.9758	0.7117
Mobile cellular subscriptions (per 100 people)	PC4	0.0724307	0.0145	0.9903	0.6472
Fixed broadband subscriptions (per 100 people)	PC5	0.0485791	0.0097	1.0000	0.7295
Overall Total					0.6481
Note: KMO-MSA = Kaiser-Meyer-Olkin Measure of Sampling Adequacy

3.4. Control variables

This study includes three control variables: arable land, rural population, and GDP per capita to account for additional factors that may influence food security outcomes in Sierra Leone. Arable land refers to the area of land suitable for crop production and is an important factor in determining agricultural capacity, as the availability of arable land affects the potential for food production. Given Sierra Leone’s reliance on agriculture, changes in arable land may have implications for food availability. Rural population is included because a significant portion of the population depends directly on agriculture for their livelihoods. Variations in the size and distribution of the rural population can influence the agricultural labor supply, demand for food, and pressure on land and natural resources. GDP per capita is used as an indicator of overall economic development and income levels. Changes in income may affect access to food through purchasing power, infrastructure development, and investment in agricultural technologies. Including these variables allows the study to control for socioeconomic and resource-related factors that could affect food production and availability.

3.5. Mediation variable

This study examines institutional quality as a mediating variable linking renewable energy adoption and technological advancement to food security in Sierra Leone. Mediation occurs when the effect of an independent variable on a dependent variable is transmitted through a third variable, in this case, institutional quality. Effective institutions, defined by regulatory strength, rule of law, accountability, and policy implementation capacity, are essential for translating energy and technology interventions into development outcomes. Without strong institutions, these efforts may underperform. Institutional frameworks support coordination, oversight, and service delivery, which in turn enhance agricultural productivity and improve food access. To assess this mediation effect, a composite Institutional Quality Index is constructed using PCA, with a Kaiser Meyer Olkin value of 0.8179 indicating strong sampling adequacy as shown in Table 2 and Fig. 3. The analysis tests whether institutional quality channels the impact of energy and technology on food security. If validated, this would underscore governance as a key mechanism for strengthening food system resilience in resource constrained settings.

Table 2
Principal component results for institutional quality
	Institutional quality index
Variables	Component	Eigenvalue	Proportion	Cumulative	KMO-MSA
Control of corruption	PC1	3.68876	0.6148	0.6148	0.3007
Governance effectiveness	PC2	1.17693	0.1962	0.8109	0.8843
Political stability and absence of violence or terrorism	PC3	0.504492	0.0841	0.8950	0.7869
Regulatory quality	PC4	0.264831	0.0441	0.9392	0.7685
Rule of law	PC5	0.236639	0.0394	0.9786	0.8985
Voice and accountability	PC6	0.128347	0.0214	1.0000	0.9235
Overall Total					0.8179
Note: KMO-MSA = Kaiser-Meyer-Olkin Measure of Sampling Adequacy.

3.6. Model specification

The empirical strategy of this study is based on the dynamic autoregressive distributed lag (DYNARDL) framework developed by Jordan and Philips [28], used to examine the effects of renewable energy, air pollution, temperature change, climate mitigation, fertilizer use, pesticide use, arable land, rural population, and farmers' income on food production in Sierra Leone. DYNARDL simulations estimate the impact of a counterfactual shock to one independent variable while holding others constant, allowing for a clear assessment of dynamic effects over time. This approach helps determine whether changes in a covariate lead to positive or negative variations in the dependent variable, accounting for the evolving nature of time-series data [49]. Though based on the ARDL model, DYNARDL enhances interpretation by addressing issues related to data limitations and model instability. It also allows for the evaluation of how changes in regressors influence food production dynamically. The general model specification is presented in Eq. (1):

$\:FOP=f(RE,\:AP,\:CT,\:CM,\:FERT,\:PEST,\:AL,\:RP,\:GDP$)(1)

In the model, FOP denotes food production, RE represents renewable energy, AP stands for air pollution, CT captures change in temperature, CM refers to climate mitigation, FERT indicates fertilizer use, PEST represents pesticide use, AL denotes arable land, RP refers to rural population, and GDP captures economic growth. To facilitate elasticity interpretation and stabilize variance, Eq. (2) is transformed into its logarithmic form. This log-linear specification allows the estimated coefficients to be directly interpreted as elasticities, indicating the percentage change in food production resulting from a one percent change in the respective independent variable. The transformed equation is expressed as follows:

while $\:{\epsilon\:}_{t}$ is an error term

3.7. Empirical methodology

3.7.1. Unit root test

The first step of the analysis is to test the stationarity of the variables to avoid spurious regression and ensure the reliability of the results. Stationarity confirms that the relationships among variables are consistent over time. The DYNARDL approach requires the dependent variable to be integrated of order one, I(1), while allowing the explanatory variables to be either I(0), I(1), or mutually cointegrated. To determine the order of integration, unit root tests were conducted using the Augmented Dickey-Fuller (ADF) test by Dickey and Fuller [50], the Phillips-Perron (PP) test by Phillips and Perron [51], and the Zivot-Andrew’s test by Zivot and Andrews [52]. In the three tests, the null hypothesis assumes the presence of a unit root (non-stationarity), while the alternative hypothesis indicates stationarity. These tests help verify the suitability of the data for analysis under the DYNARDL framework.

3.7.2. PSS bound test

To assess the existence of a long-run relationship among the study variables, the bounds testing procedure developed by Pesaran, Shin [53], commonly referred to as the PSS bounds test, was employed. The selection of the optimal lag length for the model was guided by standard information criteria, including the Likelihood Ratio (LR), Final Prediction Error (FPE), Akaike Information Criterion (AIC), Hannan-Quinn Criterion (HQC), and Schwarz Bayesian Information Criterion (SBIC). In the bounds testing framework, if the calculated F-statistic exceeds the upper critical bound, the null hypothesis of no cointegration is rejected, indicating a long-run association among the variables. If the F-statistic lies between the lower and upper bounds, the result is inconclusive. Conversely, if the F-statistic is below the lower bound, the null hypothesis cannot be rejected, suggesting no evidence of cointegration. When a long-run relationship is confirmed, both the ARDL and DYNARDL models can be estimated to explore the dynamic interactions among the variables over time.

3.7.3. Estimation of autoregressive distributed lag model

The Autoregressive Distributed Lag (ARDL) model applied in this study is based on the framework introduced by Pesaran, Shin [53]. Compared to other time series methods, the ARDL approach offers several important advantages. As noted by Haug [54], the ARDL model is suitable for small sample sizes and can accommodate variables that are either stationary at level, I(0), or first difference, I(1), without requiring pre-testing for unit roots under strict assumptions. Additionally, the model allows for the simultaneous estimation of both long-run and short-run dynamics within a unified framework. This flexibility makes it particularly useful for examining the temporal relationships between economic variables. Odhiambo [55] further supports the use of ARDL in studies involving mixed orders of integration and modest sample sizes, provided that none of the variables are integrated of order two or higher. Given these properties, the ARDL model is well-suited for estimating both long-run and short-run effects in the context of this study. The general ARDL specification used for estimation is presented in Eq. (3):

where, $\:{\beta\:}_{1},\:{\beta\:}_{1}\dots\:\dots\:..{\beta\:}_{9}$ are the long-run relation and $\:{\phi\:}_{1},\:{\phi\:}_{1}\dots\:\dots\:\dots\:{\phi\:}_{9}$ are the short-run relation, $\:{\epsilon\:}_{t}$ is the error correction term.

3.7.4. Stability test

The stability and robustness of the ARDL model are evaluated through a series of diagnostic and stability tests. To detect serial correlation in the residuals, the Breusch-Godfrey Lagrange Multiplier (LM) test is applied. Heteroskedasticity is assessed using Cameron and Trivedi’s decomposition of the Information Matrix (IM) test. The normality of residuals is examined using the Skewness/Kurtosis test, complemented by visual inspection through the standardized normal probability plot, as well as comparisons between the quantiles of residuals and those of a theoretical normal distribution. To evaluate the structural stability of the estimated model over time, the Cumulative Sum (CUSUM) test based on recursive residuals is employed. These diagnostic procedures help ensure that the model’s estimates are statistically sound and not affected by specification errors or instability in the underlying data.

3.7.5. Dynamic simulated ARDL model

To address certain limitations in the traditional ARDL framework, Jordan and Philips [28] introduced the DYNARDL model. This approach enhances the standard ARDL methodology by enabling dynamic simulation, visualization, and forecasting of the effect of changes in a specific explanatory variable while holding other variables constant. The DYNARDL model is particularly useful for generating impulse response-like plots that depict how the dependent variable reacts over time to sustained changes in a selected regressor. The model employs simulation techniques using 5,000 draws from the estimated multivariate normal distribution of the parameters, thereby providing robust confidence intervals for the dynamic effects. This allows for a more intuitive and policy-relevant interpretation of the estimated relationships. The general specification of the DYNARDL model is presented in Eq. (4):

3.7.6. Kernel-based regularized least square (KRLS)

The KRLS model provides a flexible, nonparametric approach to estimating relationships between variables while preserving interpretability. For each observation, KRLS generates closed-form estimates of pointwise marginal derivatives, effectively capturing the marginal effects of covariates. These effects can be interpreted in a manner similar to coefficients in traditional linear regression models, offering intuitive insights into how changes in explanatory variables are associated with variations in the outcome variable [56]. In addition to estimating marginal effects, KRLS allows for the exploration of interaction dynamics by examining the correlations between the pointwise marginal derivatives of one variable and those of others. This feature enables a nuanced understanding of how the influence of one covariate may vary depending on the levels or behavior of other variables in the model.

3.8. Fully Modified Ordinary Least Squares for robustness analysis

While the ARDL and DYNARDL models are effective for estimating short- and long-run relationships, they have limitations when applied to robustness checks involving alternative dependent variables. ARDL assumes a valid long-run equilibrium and is best suited for models with a single dependent variable and regressors integrated at I(0) or I(1), but may suffer from small-sample bias and sensitivity to lag selection. DYNARDL, though useful for dynamic simulations, inherits these structural limitations. In contrast, the Fully Modified Ordinary Least Squares (FMOLS) estimator addresses issues of endogeneity and serial correlation by adjusting the OLS estimator through non-parametric corrections [57]. FMOLS is particularly suitable for long-run estimation in cointegrated systems and offers more robust results when testing alternative indicators of food security. Therefore, FMOLS is employed to assess the robustness of the main findings using additional dependent variables.

4.1. Descriptive Statistics

The descriptive statistics in Table 3, along with the correlation matrix in Table 4, provide insights into the central tendencies and variability of the study’s core constructs, highlighting their interconnected influence on food security outcomes in Sierra Leone. Among the dependent variables, the Food Production Index (FPI) records a mean of 70.49 and a standard deviation (SD) of 30.87, indicating significant variation in food production performance across the observed period. Disaggregated crop-level outputs also exhibit substantial variability; rice, the principal staple, has a mean of 766,752.9 tons and an SD of 408,104.6, reflecting high inter-annual fluctuations. Similarly, fruit, vegetable, maize, cassava, and cocoa outputs display wide dispersions, with cassava showing the greatest variability (SD = 125,698.5), suggesting its heightened sensitivity to both environmental and agronomic factors.

Among the independent variables, renewable energy use shows a mean of 48.23 and a relatively high SD of 25.12, pointing to considerable shifts in energy access or adoption over time. Air pollution, measured by carbon emissions per capita, averages 41.02 with moderate variability (SD = 3.49), while temperature change has a mean of 0.82 degrees Celsius and a low SD of 0.29, indicating gradual but consistent climatic variations throughout the period. Climate mitigation efforts, proxied by environmental expenditure or clean technology investment, reveal a mean of 27.86 and an SD of 13.84, reflecting inconsistent policy engagement. Technological inputs also demonstrate considerable heterogeneity, with fertilizer use averaging 6.11 kilograms per hectare (SD = 4.11), indicating low but uneven application, and pesticide use showing a higher mean of 227.65 and SD of 100.25, suggesting irregular usage. The Technological Advancement Index, constructed via principal component analysis, records a normalized mean close to zero and an SD of 1.81, capturing aggregate innovation levels. Furthermore, the mediation variable, institutional quality, has a mean of approximately 7.14e-09 and an SD of 1.92, reflecting variability in governance capacity and institutional effectiveness. These descriptive metrics underscore the complex and interrelated dynamics of technological change, environmental stressors, institutional conditions, and renewable energy adoption in shaping food production outcomes in a fragile context.

Table 3
Variables units of measurement, data sources, summary statistics results
Variables	Data source	Obs.	Mean	Std. Dev.	Min	Max
Dependent variables
Food production index	WDI	34	70.494	30.867	29.200	130.560
Rice production in tons	FAOSTAT	34	766752.9	408104.6	199134	1978905
Fruit production in tons	FAOSTAT	34	216144.2	49476.32	152985	282814.1
Vegetable production in tons	FAOSTAT	34	30420.89	16534.7	8100	80000
Maize production in tons	FAOSTAT	34	26687.28	11649.21	3636	44000
Cassava production in tons	FAOSTAT	34	1596921	1256985	182400	3810418
Cocoa production in tons	FAOSTAT	34	17159.85	10858.95	5400	50150
Independent variables
Renewable energy (% of total final energy consumption)	WDI	34	48.23426	25.16116	6.090347	87.39964
Air pollution (micrograms per cubic meter)	WDI	34	41.02414	3.487	34.42033	51.80617
Change in temperature (°c)	FOASTAT	34	0.8238235	0.2954723	0.184	1.343
Climate mitigation	EPI	34	27.85829	11.85368	3.215277	40.84271
Fertilizer use (kilograms per hectare of arable land)	WDI	34	6.109105	4.113657	0.1859504	16.44104
Pesticides use in tons	FAOSTAT	34	227.6462	100.2515	59.230	405.88
Technological advancement index	Authors construction	34	8.82e-09	1.805971	1.687671	4.247467
Control Variables
Arable land (% of land area)	WDI	34	15.8974	6.988035	6.705459	23.41203
Rural population	WDI	34	3563001	684642.5	2740329	4712505
GDP per capita (current US$)	WDI	34	546.336	316.5026	143.7437	1146.416
Mediation Variable
Institutional quality index	Authors construction	28	7.14e-09	1.920614	4.727294	1.983487
Note: WDI represents the World Development Indicators, FAOSTAT stands for the Food and Agriculture Organization Statistical Database, and EPI denotes the Environmental Performance Index.

Table 4
Correlation matrix
Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)
(1) FP	1.000
(2) IQ	0.879	1.000
(3) RE	0.740	0.569	1.000
(4) CT	0.476	0.384	0.257	1.000
(5) AP	-0.211	0.101	0.296	0.213	1.000
(6) CM	0.568	0.728	0.214	0.528	0.013	1.000
(7) FERT	0.455	0.540	0.354	0.206	0.183	0.599	1.000
(8) PEST	0.160	0.176	0.584	0.048	0.270	0.284	-0.008	1.000
(9) GDP	0.930	0.830	0.712	0.464	0.223	0.560	0.504	0.175	1.000
(10) AL	0.894	0.901	0.553	0.666	0.297	0.724	0.525	0.033	0.890	1.000
(11) RP	0.899	0.816	0.733	0.594	0.380	-0.503	0.383	0.347	0.916	0.904	1.000

4.2. Preliminary analysis findings

As a preliminary step, unit root tests were conducted to ensure that all variables are stationary and suitable for ARDL and DYNARDL estimation. Table 5 reports the results of the Augmented Dickey-Fuller (ADF), Phillips-Perron (PP), and Zivot-Andrews (ZA) tests. These confirm that the dependent variable, food production, is stationary at first difference (I(1)), while none of the variables are integrated at the second difference (I(2)). Air pollution is the only variable found to be stationary at level (I(0)) across all three tests. In contrast, renewable energy, temperature change, climate mitigation, fertilizer use, pesticide use, arable land, rural population, and GDP per capita are stationary at first difference. Table 6 presents the lag length selection criteria, indicating lag 2 as optimal for both the ARDL and DYNARDL models. To assess long-run relationships among the variables, the Pesaran, Shin, and Smith (PSS) bounds testing procedure is applied [53]. As shown in Table 7, the computed F-statistic (5.273) and t-statistic (–4.383) exceed the upper critical bounds at the 5% significance level (3.772–4.176) and also surpass the critical values at the 1% and 10% levels. These findings are further supported by Kripfganz and Schneider [58] approximate p-values (p < 0.01), leading to the rejection of the null hypothesis of no cointegration. Overall, both the bounds test and approximate p-values confirm the existence of a long-run relationship between food production and the explanatory variables.

Table 5
Synopsis of stationarity test for unit root.
Variable	Dickey–Fuller test	Phillips–Perron test	Zivot-Andrew’s test	Break point
Level
Food production	-0.702	-0.793	-2.376	2012
Renewable energy	0.564	0.291	-3.520	1998
Air pollution	-3.745***	-3.634***	-4.819***	2011
Change in temperature	-2.728	-2.507	-5.369	2003
Climate mitigation	-1.712	-1.959	-2.762	2010
Fertilizer use	-2.913	-2.992	-3.221	2017
Pesticides use	-0.822	-1.371	-4.278	2003
Arable land	-0.525	-0.743	-3.739	2010
Rural population	0.465	0.121	-4.935	2012
GDP per capita	-0.893	-0.841	-2.878	2014
1st Difference
Food production	-5.619***	-5.616***	-5.169***	2004
Renewable energy	-5.360***	-5.425***	-5.133***	2013
Air pollution	-6.695***	-7.353***	-6.900***	1998
Change in temperature	-9.206***	-10.917***	-9.458***	1996
Climate mitigation	-5.572***	-5.576***	-5.014***	2006
Fertilizer use	-6.444***	-6.488***	-4.624***	1997
Pesticides use	-4.169***	-4.279***	-3.961***	2017
Arable land	-4.262***	-4.364***	-4.488***	2002
Rural population	-2.640**	-2.763**	-4.392**	2002
GDP per capita	-6.023***	-6.009***	-5.529***	2002

Table 6
Result of the Lag-order selection criteria
Lag-order	LL	LR	df	P-values	FPE	AIC	HQIC	SBIC
0	54.5964				4.5e-14	-2.34718	-2.19385	-1.91624
1	360.656	612.12	100	0.000	1.0e-18	-13.1924	-11.5058	-8.45204
2	566.113	410.91*	100	0.000	1.4e-20*	-18.7428*	-15.5229*	-9.69297*
Number of obs. = 34
Note: that * represent the lag order selection criteria

Table 7
Pesaran, Shin, and Smith (2001) PSS bounds test
K	10%		5%		1%		P-value
	I (0)	I (1)	I (0)	I (1)	I (0)	I (1)	I (0)	I (1)
F = 5.273	1.896	3.216	2.273	3.772	3.194	5.120	0.009***	0.003***
t = -4.383	-1.609	-4.113	-1.979	-4.176	-2.722	-3.526	0.000***	0.006***

4.3. Autoregressive distribution lag and Dynamic Autoregressive distribution lag model estimation

The long run and short run estimation results from both the ARDL and dynamic ARDL (DYNARDL) simulation models reveal several important relationships between key variables and food production in Sierra Leone, as summarized in Table 8. Renewable energy consumption has a positive and statistically significant effect on food output. Specifically, a 1% increase in renewable energy use is associated with a 0.381% increase in food production in the ARDL model and a 0.108% increase in the DYNARDL simulation, supporting H1. Although the magnitude differs between models, likely due to their distinct treatment of dynamic adjustments, the consistent significance and direction underscore the role of renewable energy in enhancing agricultural productivity. Access to clean energy supports irrigation, mechanization, postharvest storage, and agro-processing, particularly in rural areas with limited infrastructure. This reduces reliance on fossil fuels and promotes sustainable agricultural systems. These findings are consistent with He, Osabohien [35], who found that renewable energy enhances agricultural output by lowering production costs and increasing input efficiency, and Burney, Woltering [34], who emphasized its role in improving rural productivity.

Air pollution, in contrast, has a negative and statistically significant long run effect on food production. A 1% rise in pollution results in a 0.625% and 0.215% decrease in food output in the ARDL and DYNARDL models, respectively, supporting H2. This finding aligns with studies that link air pollution, through carbon emissions and particulate matter, to soil degradation, impaired photosynthesis, and climatic stress [59]. The increased vulnerability of traditional and low input farming systems to these stressors underscores the urgent need for clean technologies and stronger environmental regulation to protect food security.

Similarly, temperature variability negatively affects food production. A 1% increase in temperature volatility reduces output by 0.053% in the ARDL model and 0.055% in the DYNARDL model. These findings support H2 and are consistent with Subedi, Poudel [60] and Alimagham, van Loon [6], who show that rising temperatures reduce yields by shortening growing seasons, intensifying drought risk, and promoting the spread of pests and diseases. This impact is especially severe in rainfed agricultural systems like those in Sierra Leone. The results emphasize the importance of climate resilient strategies, such as the adoption of drought tolerant crop varieties, precision agriculture, and early warning systems.

Climate mitigation efforts, measured through indicators such as afforestation and adoption of clean technologies, have a positive and statistically significant effect on food production. A 1% increase in climate mitigation is associated with a 0.068% and 0.041% rise in food output in the ARDL and DYNARDL models, respectively. This finding supports the view that environmental sustainability enhances agricultural resilience and productivity [13, 61, 62]. Mitigation activities improve ecological stability, reduce climate related shocks, and promote soil health, thereby creating favorable conditions for long term productivity.

Technological advancement, measured through fertilizer and pesticide use, is generally positively associated with food production, supporting H3. A 1% increase in fertilizer application raises output by 0.041% in the ARDL model and 0.012% in the DYNARDL model, both statistically significant. This aligns with findings by Oyetunji, Bolan [63] and Chandio, Gokmenoglu [17], which highlight the role of nutrient management in improving yields. Pesticide use, however, shows a positive but statistically insignificant effect, possibly due to variability in usage, improper application, or resistance development. Ahmad, Ahmad [64] note that while pesticides can protect yields, their benefits depend heavily on proper application and integrated pest management.

Arable land availability also has a strong positive and significant effect on food production. A 1% increase in arable land corresponds to a 0.706% and 0.349% rise in output in the ARDL and DYNARDL models, respectively. This confirms the central role of land in agrarian economies like Sierra Leone. However, expanding cultivation must be managed carefully to avoid deforestation and biodiversity loss. Zhou, Chen [65] warn of the risks posed by unregulated land expansion. Therefore, land tenure reform, sustainable intensification, and soil conservation practices are critical to achieving sustainable productivity gains.

Rural population is also positively associated with food production. A 1% increase in rural population results in a 0.136% and 0.064% increase in output in the ARDL and DYNARDL models, respectively. This reflects the labor-intensive nature of smallholder agriculture in Sierra Leone. Li, Wang [66] argue that rural population growth, if supported by investments in infrastructure, training, and services, can boost agricultural productivity. However, this must be managed carefully to avoid putting pressure on limited natural resources.

Economic growth, measured by GDP per capita, contributes positively to food production. A 1% increase in GDP per capita leads to a 0.074% and 0.094% rise in food output in the ARDL and DYNARDL models, respectively. This supports findings by Dethier and Effenberger [67], who show that income growth facilitates access to better technology, education, and credit. However, the relatively moderate effect indicates that growth alone is not enough. Complementary investments in agriculture and rural development are essential for turning macroeconomic gains into improved food security.

The error correction term is negative and statistically significant at the 5% level, confirming the presence of a stable long run relationship among the variables. The coefficient indicates that approximately 25% of deviations from the long run equilibrium are corrected each year, suggesting that full adjustment occurs over four years. The R-squared values of 0.7390 and 0.7865 in the ARDL and DYNARDL models, respectively, show that the models explain 73.9% and 78.7% of the variation in food production. These high explanatory powers confirm the robustness of the models and the importance of environmental, technological, and socioeconomic factors in determining food production outcomes in Sierra Leone.

Table 8
Regression results from the ARDL and DYNARDL simulation.
Variable	Autoregressive distribution lag (ARDL)				Dynamic Autoregressive distribution lag (DYNARDL)
	Coeff.	Std. E.	t > z	P > z	Coeff.	Std. E.	t > z	P > z
$\:\text{E}\text{C}\text{T}-1$	-0.525	0.119	-4.38	0.000	-0.345	0.161	-2.14	0.041
Long-run
Renewable energy	0.381	0.051	1.08	0.001	0.108	0.034	0.42	0.000
Air pollution	-0.625	0.016	-1.50	0.008	-0.215	0.066	-0.81	0.006
Change in temperature	-0.053	0.012	-0.50	0.035	-0.055	0.016	-0.46	0.016
Climate mitigation	0.068	0.008	0.16	0.009	0.041	0.007	0.36	0.008
Fertilizer use	0.041	0.007	1.06	0.017	0.012	0.001	0.59	0.033
Pesticides use	0.030	0.109	0.28	0.783	0.098	0.072	1.37	0.182
Arable land	0.706	0.229	3.06	0.005	0.349	0.183	1.91	0.067
Rural population	0.136	0.042	0.95	0.017	0.064	0.036	0.84	0.049
GDP per capita	0.074	0.006	0.49	0.003	0.094	0.012	0.12	0.000
Short-run
Renewable energy	0.200	0.094	1.03	0.043	0.371	0.089	1.76	0.000
Air pollution	-0.328	0.113	-1.43	0.065	-0.456	0.121	-1.45	0.000
Change in temperature	-0.028	0.056	-0.49	0.627	-0.052	0.016	-0.21	0.005
Climate mitigation	0.063	0.034	0.16	0.033	0.146	0.006	0.51	0.004
Fertilizer use	0.022	0.009	1.07	0.015	0.045	0.018	0.32	0.000
Pesticides use	0.016	0.058	0.27	0.788	0.264	0.161	1.63	0.114
Arable land	0.370	0.157	2.35	0.026	0.703	0.245	2.88	0.008
Rural population	0.072	0.015	0.96	0.008	0.884	0.044	0.45	0.000
GDP per capita	0.038	0.011	0.50	0.007	0.041	0.013	0.49	0.026
Number of obs. = 34					Number of obs. = 34
R-squared = 0.7390					R-squared = 0.7865
Adj R-squared = 0.6743					Adj R-squared = 0.6414
Root MSE = 0.1023					Root MSE = 0.0993
Log likelihood = 38.530252

4.4. Diagnostic and stability tests

The diagnostic results of the ARDL model, as presented in Table 9, provide strong support for the validity and robustness of the estimated parameters. To assess the presence of serial correlation in the residuals, the Breusch-Godfrey LM test was applied. The test fails to reject the null hypothesis of no serial correlation at the 5% significance level, indicating that the residuals are free from autocorrelation and that the model does not suffer from misspecification in this regard. To examine the presence of heteroscedasticity, Cameron and Trivedi’s decomposition of the Information Matrix (IM) test was employed. The test results indicate that the null hypothesis of homoscedasticity cannot be rejected, as the p-value exceeds the 5% threshold. This suggests that the variance of the residuals is constant, satisfying the assumption of homoscedasticity and affirming the reliability of the estimated standard errors. Normality of the residuals was tested using the Skewness/Kurtosis test for normality. The test fails to reject the null hypothesis of normal distribution at the 5% significance level, implying that the residuals are approximately normally distributed. This is further validated visually through the standardized normal probability plot (Fig. 4) and the quantile-quantile (Q-Q) plot comparing the residuals’ quantiles to those of a normal distribution (Fig. 5). Both graphical analyses suggest that the residuals closely follow a normal distribution, reinforcing the findings of the formal statistical test.

Additionally, the stability of the model parameters over time was examined using the cumulative sum (CUSUM) of recursive residuals (Fig. 6). The CUSUM test statistic remains within the 95% confidence bands throughout the sample period, indicating parameter stability and the absence of structural breaks. This is a critical validation for time series models, particularly when the analysis extends over a long historical period or is subject to potential regime changes. These diagnostic tests confirm that the ARDL model meets key classical assumptions, including no autocorrelation, homoscedasticity, normality of residuals, and parameter stability. These findings enhance the credibility of the estimated relationships and ensure that the inference drawn from the model is statistically sound. The use of these standard diagnostic procedures aligns with best practices in empirical research and has been widely adopted in similar studies for validating model reliability and robustness.

Table 9
Model diagnostic test
A. Breusch–Godfrey LM test for autocorrelation
lags(p)			F	Df		Prob > F
1			12.750	(1,33)		0.1311
2			6.682	(2,32)		0.0738
3			4.532	(3,31)		0.1095
4			3.728	(4,30)		0.1514
B. Cameron & Trivedi's decomposition of IM-test
Source			Chi2	df		P-value
Heteroskedasticity			44.00	43		0.4290
Skewness			2.90	9		0.9679
Kurtosis			1.37	1		0.2418
Total			48.28	53		0.6584
C. Skewness and kurtosis tests for normality					----- Joint test -----
Variable	Obs.	Pr(skewness)		Pr(kurtosis)	Adj chi2(2	Prob > chi2
res1	44	0.4071		0.1930	2.53	0.2819

4.5. Novel dynamic ARDL simulations

To examine the dynamic and asymmetric effects of key variables on food production, this study employs the DYNARDL simulation model. This technique models the response of food output to 10% positive and negative shocks in variables such as renewable energy use, air pollution, temperature variations, climate mitigation efforts, and the application of fertilizers and pesticides. Predicted responses are illustrated using dark blue dots, while confidence intervals are represented by green and light blue bands. Each simulation isolates a single explanatory variable, holding all others at their mean values, thereby providing a clear interpretation of its independent effect. The framework captures both short- and long-term adjustments, offering insights into the persistence and magnitude of impacts that inform data-driven policy formulation.

Figure 7 presents the projected response of food production to changes in renewable energy consumption over a 20-period horizon. The left panel simulates a 10% increase, while the right reflects a corresponding decrease. A positive shock leads to a gradual and sustained increase in food output, plateauing at approximately a 40% cumulative gain by the tenth period. This trajectory underscores the role of renewable energy in supporting agricultural systems through improved mechanization, irrigation, and post-harvest processing. Conversely, reduced access to renewable energy induces a sharp and persistent decline of similar magnitude, highlighting the sector’s vulnerability to energy shortfalls. The contrasting outcomes reinforce the critical need for sustained investments in clean energy to enhance food system resilience, especially in energy-insecure contexts such as Sierra Leone.

Figure 8 shifts focus to air pollution and its detrimental effects on agricultural output. A 10% increase in pollution results in a rapid and sustained contraction in food production, stabilizing near a 6% decline. This likely reflects the compounded effects of environmental degradation, including diminished soil quality and increased toxicity. In contrast, a 10% decrease in pollution produces a more modest but consistent improvement in output, with gains exceeding 4%. The asymmetry between these responses suggests that while pollution reduction yields benefit, the damage from increased emissions is more severe and enduring. These findings support the need for stronger environmental governance to ensure agricultural sustainability.

Figure 9 explores the effects of temperature variation. A positive deviation causes a continual decline in food production, stabilizing after eight periods. This reflects the adverse impact of excessive heat on crop performance, including heightened evapotranspiration and reduced soil moisture. In contrast, cooler conditions lead to a rise in output, peaking midway through the projection horizon and remaining above baseline levels. The asymmetry between warming and cooling responses highlights that the losses from heat stress outweigh the benefits of moderate cooling. This underscores the need for climate adaptation strategies such as drought-resistant crop varieties and efficient water management to mitigate the sector’s vulnerability to temperature fluctuations.

Figure 10 assesses the implications of climate mitigation policies. Enhanced mitigation efforts are associated with sustained improvements in food production, with effects stabilizing after ten periods. The narrowing confidence intervals over time suggest growing certainty in this upward trend. In contrast, a reduction in mitigation efforts leads to a significant and prolonged decline in output. These findings reveal a clear asymmetry: while proactive climate action produces lasting benefits, policy reversals impose enduring costs. This highlights the strategic importance of maintaining long-term commitments to climate mitigation in order to support agricultural productivity.

Figure 11 evaluates the role of fertilizer application. A 10% increase results in a steady rise in food production, reflecting improved soil fertility and crop yields. The response stabilizes over time, with the narrowing confidence intervals indicating increasing reliability of the projections. Conversely, a decrease in fertilizer use leads to only a modest and short-lived decline, which flattens quickly. This weaker response may reflect already low baseline input levels, limiting the marginal impact of further reductions. The findings emphasize the potential for expanded input use to enhance productivity, particularly in under-resourced agricultural systems.

Lastly, Fig. 12 examines the effects of pesticide use. Increased application yields a strong and sustained rise in food output, reaching a steady state after several periods. This suggests that effective pest control is crucial to maintaining yields, especially in high pest-pressure environments. In contrast, reduced pesticide uses results in a rapid and prolonged decline in food production, stabilizing at a substantially lower level. The pronounced asymmetry underscores the importance of consistent pest management. In regions like Sierra Leone, ensuring access to safe and affordable pesticide solutions is essential for yield protection and food security. These simulations reveal the distinct and often asymmetric effects of key policy and environmental variables on agricultural performance. They underscore the importance of targeted interventions in energy access, environmental protection, climate adaptation, and input availability to enhance resilience and productivity within the food sector.

4.6. Kernel-based regularized least squares

To assess the causal impact of the explanatory variables on food production, pointwise derivatives were calculated using Kernel Regularized Least Squares (KRLS), as presented in Table 10. The model demonstrates a high level of predictive accuracy, with an R-squared value of 0.9913. This indicates that the selected variables collectively explain approximately 99.13% of the variation in food production, reflecting strong model performance and the relevance of the chosen predictors. The mean marginal effects highlight the average influence of each variable on food production. Specifically, renewable energy use (0.846%), pesticide application (0.084%), arable land (0.127%), rural population (0.363%), and fertilizer use (0.040%) exhibit positive and relatively strong marginal effects, underscoring their importance in enhancing agricultural output. In contrast, air pollution (–0.043%) and temperature change (–0.034%) display negative marginal impacts, suggesting that environmental stressors adversely affect productivity. Climate change mitigation has a modest but positive effect (0.029%), while GDP per capita shows a negligible marginal impact (0.016%). At the 1% significance level, all variables except GDP per capita are statistically significant, indicating that the remaining predictors have meaningful effects on food production within the context of the model. These findings emphasize the importance of environmental quality, energy access, and agricultural inputs in shaping food production outcomes, while suggesting that broader economic indicators like GDP per capita may have limited direct influence in this specific context.

Additionally, Fig. 13 presents a series of LOWESS smoothed plots that illustrate the nonlinear relationships between food production and six key explanatory variables: renewable energy use, air pollution, temperature change, climate change mitigation, fertilizer application, and pesticide application. These visualizations provide valuable insights into the direction and shape of each relationship across the observed range of food production, revealing the complexity of interactions between environmental and agricultural factors. For instance, the relationship between renewable energy use and food production follows an inverted U-shape. This indicates that while initial increases in renewable energy adoption are associated with higher agricultural output, further expansion beyond a certain threshold may result in diminishing or even negative returns, possibly due to inefficiencies or resource constraints arising at higher levels of energy integration.

Similarly, air pollution shows a generally negative association with food production, suggesting that elevated pollution levels undermine crop productivity. This observation is consistent with existing research on the harmful effects of air pollutants on soil quality and plant health. In contrast, the impact of temperature change appears relatively flat and nonlinear, indicating that its influence on food production is weak or inconsistent across the dataset. Climate change mitigation, on the other hand, displays a positive association with food production, especially at higher levels of implementation. This suggests that investments in environmental protection and sustainable practices can bolster agricultural resilience and productivity. Fertilizer application exhibits a concave pattern, with food production increasing at moderate application levels but plateauing thereafter, implying diminishing returns from excessive input use. A similar trend is evident for pesticide application: while moderate usage enhances productivity, overuse corresponds with stagnation or decline, likely due to ecological degradation or pest resistance. These findings underscore the nonlinear and context-dependent nature of the relationships between key environmental variables and food production. They highlight the necessity for carefully calibrated, evidence-based policy interventions that promote agricultural efficiency while safeguarding environmental integrity.

Table 10
Kernel Regularized Least Squares results
Variables	Avg.	SE	t	p>\|t\|	P25	P50	P75
Renewable energy	0.846	0.123	6.880	0.000	1.348	0.616	0.446
Air pollution	-0.043	0.012	-0.344	0.003	-0.621	0.141	0.550
Change in temperature	-0.034	0.020	-0.697	0.041	-0.048	0.017	0.078
Climate mitigation	0.029	0.010	2.805	0.008	-0.005	0.002	0.059
Fertilizer use	0.040	0.009	4.381	0.000	-0.002	0.049	0.088
Pesticides use	0.084	0.018	4.431	0.000	-0.033	0.157	0.257
Arable land	0.127	0.015	8.210	0.000	0.044	0.059	0.163
Rural population	0.363	0.054	6.741	0.000	0.158	0.477	0.563
GDP per capita	0.016	0.018	0.867	0.392	-0.045	0.009	0.061
Diagnostics
Lambda	0.05979	Sigma	9	R²	0.9913	Obs.	34
Tolerance	0.042	Eff. df	24.91	Looloss	1.319

4.7. Robustness and moderating effects analysis

To enhance the robustness of the analysis, this study incorporates six additional dependent variables that reflect the multidimensional nature of food security beyond aggregate food production. Specifically, rice, vegetable, maize, cocoa, cassava, and fruit production are included due to their economic and nutritional significance in Sierra Leone. The Fully Modified Ordinary Least Squares (FMOLS) estimator is employed to address endogeneity and serial correlation, ensuring efficient and unbiased parameter estimates. As presented in Table 11, the results indicate that renewable energy use and climate mitigation have positive and statistically significant effects on overall food production, as well as on rice, vegetable, cassava, and fruit production. For maize and cocoa, the effects are positive but statistically insignificant. In contrast, air pollution and temperature variability exhibit negative and significant effects across all food categories, suggesting that environmental degradation hampers agricultural output through soil deterioration and climatic stress.

Technological inputs produce mixed results. Fertilizer use has a positive and significant effect on food, rice, vegetable, maize, and cocoa production but shows a negative and significant effect on fruit production, and a negative yet insignificant effect on cassava. Pesticide use positively and significantly influences vegetable, cocoa, and fruit production, while its effect is positive but insignificant for overall food and maize production, and negative and insignificant for rice and cassava. These results are consistent with the ARDL and DYNARDL simulation outcomes in Table 8, reinforcing the observed relationships. The findings highlight the importance of renewable energy and climate mitigation strategies in enhancing food security, while emphasizing the need for more targeted and sustainable application of agricultural inputs across different crop systems.

Table 12 presents the moderating effects of institutional quality on the production of major staple foods. The interaction terms between air pollution and institutional quality, as well as between climate mitigation and institutional quality, are positive and statistically significant for food, rice, and cassava production. These results suggest that stronger institutional frameworks enhance the effectiveness of environmental and climate-related interventions in improving agricultural productivity. Conversely, the interaction between temperature variability and institutional quality is negative and significant across the same food categories, indicating that even strong institutions may not fully offset the adverse effects of climatic fluctuations. These findings underscore the critical role of institutional capacity in shaping agricultural outcomes and reveal the continued vulnerability of food systems to environmental stressors.

Table 11
Robustness check results
Variables	Food production	Rice production	Vegetable production	Maize production	Cocoa production	Cassava production	Fruit production
Renewable energy	0.559***	0.344***	0.405***	0.408^NS	0.481^NS	1.281***	0.052**
	(0.124)	(0.047)	(0.060)	(0.498)	(0.555)	(0.453)	(0.022)
Air pollution	-0.623***	-1.317**	-0.193***	-0.784	-1.882**	0.032^NS	-0.031^NS
	(0.204)	(0.554)	(0.054)	(0.678)	(0.757)	(0.616)	(0.012)
Change in temperature	-0.129**	-0.180^NS	-0.049^NS	-0.109^NS	-0.421**	-0.174***	-0.028***
	(0.062)	(0.126)	(0.081)	(0.155)	(0.173)	(0.041)	(0.009)
Climate mitigation	0.057***	0.029^NS	0.172***	0.058^NS	0.011^NS	0.163***	0.022***
	(0.021)	(0.092)	(0.059)	(0.113)	(0.126)	(0.034)	(0.007)
Fertilizer use	0.044**	0.128**	0.178***	0.359***	0.057***	-0.023^NS	-0.009**
	(0.010)	(0.055)	(0.035)	(0.066)	(0.015)	(0.061)	(0.004)
Pesticides use	0.055^NS	-0.064^NS	0.364*	0.185^NS	0.657*	-0.388^NS	0.044*
	(0.153)	(0.279)	(0.178)	(0.341)	(0.381)	(0.310)	(0.022)
Control variables	Yes	Yes	Yes	Yes	Yes	Yes	Yes
R-squared	0.937	0.853	0.893	0.789	0.552	0.925	0.993
Adjusted R-squared	0.912	0.796	0.851	0.708	0.377	0.895	0.989
Observations	34	34	34	34	34	34	34
Notes: NS refer to not significant, *, , * represent 1%, 5% and 10% significance level

Table 12
Moderating effect of institutional quality and environmental degradation on staple food in Sierra Leone
Food Production
Variables	Model-1	Model-2	Model-3	Model-4	Model-5	Model-6
Air pollution*IQ	0.065***			0.021***
	(0.011)			(0.006)
Change in temperature*IQ		-0.607***			-0.131**
		(0.184)			(0.057)
Climate mitigation*IQ			0.069***			0.022***
			(0.013)			(0.006)
Control variables	No	No	No	Yes	Yes	Yes
Constant	4.166***	4.267***	4.206***	2.860***	2.285***	2.007***
	(0.075)	(0.118)	(0.077)	(0.174)	(0.041)	(0.082)
R-squared	0.667	0.272	0.649	0.937	0.910	0.937
Adjusted R-squared	0.656	0.248	0.638	0.928	0.898	0.928
Rice Production
	Model-1	Model-2	Model-3	Model-4	Model-5	Model-6
Air pollution*IQ	0.079***			0.038***
	(0.011)			(0.009)
Change in temperature*IQ		-0.743***			-0.296***
		(0.207)			(0.083)
Climate mitigation*IQ			0.086***			0.041***
			(0.012)			(0,009)
Control variables	No	No	No	Yes	Yes	Yes
Constant	13.413***	13.537***	13.463***	11.625***	10.251**	9.170***
	(0.074)	(0.133)	(0.072)	(0.419)	(6.358)	(0.804)
R-squared	0.685	0.291	0.689	0.866	0.823	0.874
Adjusted R-squared	0.675	0.268	0.679	0.847	0.798	0.856
Cassava Production
	Model-1	Model-2	Model-3	Model-4	Model-5	Model-6
Air pollution*IQ	0.125***			0.038***
	(0.034)			(0.014)
Change in temperature*IQ		-1.214***			0.037
		(0.441)			(0.131)
Climate mitigation*IQ			0.131***			0.018**
			(0.037)			(0.005)
Control variables	No	No	No	Yes	Yes	Yes
Constant	13.854***	14.061***	13.929***	41.480***	38.301***	40.133***
	(0.216)	(0.283)	(0.224)	(13.774)	(14.833)	(14.129)
R-squared	0.471	0.215	0.436	0.908	0.909	0.908
Adjusted R-squared	0.454	0.190	0.417	0.895	0.897	0.895
Notes: IQ denotes Institutional Quality, *, , * represent 1%, 5% and 10% significance level

4.8. Mechanism analysis

To better understand how renewable energy adoption and technological advancement influence food security in Sierra Leone, a mediation analysis was conducted using the FMOLS estimation technique. FMOLS is appropriate in this context as it addresses endogeneity and serial correlation, producing reliable long run estimates in cointegrated systems. Following the frameworks of previous studies [68, 69], the analysis was conducted in two stages. The first stage examines whether renewable energy use and technological advancement significantly affect institutional quality, proposed as the mediating channel. Institutional quality, defined by regulatory effectiveness, government accountability, and the rule of law, plays a key role in enabling effective policy implementation, resource management, and public service delivery. These institutional capabilities are critical for transforming energy and technological investments into improved agricultural productivity and food access.

If renewable energy and technological advancement significantly enhance institutional quality, this supports the hypothesis that institutional improvements help transmit the benefits of these interventions to food security outcomes. The second stage evaluates whether institutional quality significantly influences food security when energy and technology are controlled. If institutional quality remains significant and simultaneously reduces the direct effects of the independent variables, this provides evidence of mediation. This approach clarifies how institutional dynamics contribute to the broader impacts of development interventions. In Sierra Leone, where institutions are often weak and resources constrained, identifying this mediation mechanism is essential for formulating effective and integrated food security strategies.

The results of the initial mediation analysis, presented in Table 13, indicate that both renewable energy adoption and technological advancement have a positive and statistically significant effect on institutional quality, supporting the hypothesized mediating relationship. These effects hold with and without the inclusion of control variables, suggesting the robustness of the results. Specifically, a one unit increase in renewable energy adoption is associated with a 0.542% increase in institutional quality without control variables, and a 0.233% increase when controls are included, confirming H4. Similarly, technological advancement leads to a 0.308% increase in institutional quality without controls, and 0.075% with controls, supporting H5. These results suggest that investments in renewable energy and technological innovation contribute to institutional development by strengthening governance, regulatory capacity, and service delivery systems. Although the effect sizes decline when controls are included, their continued significance highlights the role of energy and technology as drivers of institutional improvement. This evidence forms a strong basis for investigating the mediating role of institutional quality in shaping food security outcomes.

To further explore this mediating effect, institutional quality was added as an additional covariate in the FMOLS models. The findings, reported in Tables 14 and 15, offer further support for the mediating role of institutional quality and underscore the importance of renewable energy and technological advancement in enhancing food security in Sierra Leone with and without control variables, lending support to H6. Table 14 shows that, without control variables, renewable energy exerts a positive and statistically significant influence on food production (0.483%), rice (0.608%), cassava (0.055%), and fruit production (0.345%). Institutional quality also demonstrates a positive and significant effect on food (0.215%), rice (0.269%), cassava (0.437%), and fruit production (0.091%). These results indicate that institutional quality partially mediates the effects of renewable energy, particularly in enhancing the production of rice, cassava, and fruit (supporting H7).

Table 15 presents the outcomes for technological advancement. Without control variables, technological advancement positively and significantly affects food (0.036%), rice (0.089%), and fruit production (0.047%), while the effect on cassava production (0.036%) is positive but statistically insignificant. Institutional quality again shows a consistently positive and significant impact on food (0.206%), rice (0.234%), cassava (0.446%), and fruit production (0.073%) (supporting H8). These findings emphasize that institutional mechanisms, including policy enforcement, administrative efficiency, and governance quality, are essential channels through which technological improvements enhance food security. The consistent significance of institutional quality across models confirms its mediating role and highlights the importance of integrating technological and institutional reforms to improve food security in vulnerable settings such as Sierra Leone.

Table 13
Indirect effect of renewable energy and technological advancement on the mediating variable
	Institutional quality		Institutional quality
Variables	Model-1	Model-2	Model-3	Model-4
Renewable energy	0.542***	0.233***
	(0.149)	(0.045)
Technological advancement			0.308***	0.075***
			(0.083)	(0.012)
Control variables	No	Yes	No	Yes
Constants	3.124***	2.760***	4.479***	2.897***
	(0.246)	(0.483)	(0.537)	(0.627)
R-squared	0.733	0.795	0.647	0.813
Adjusted R-squared	0.709	0.759	0.615	0.781
Notes: *, , * represent 1%, 5% and 10% significance level

Table 14
Mediation effect of renewable energy and institutional quality on food security
	Food production		Rice production		Cassava production		Fruit production
Variables	Model-1	Model-2	Model-3	Model-4	Model-5	Model-6	Model-7	Model-8
Renewable energy	0.483***	0.426***	0.608***	0.230***	0.055***	0.775***	0.345**	0.064*
	(0.047)	(0.031)	(0.094)	(0.023)	(0.017)	(0.096)	(0.143)	(0.035)
Institutional quality	0.215***	0.096***	0.269***	0.097***	0.437***	0.119	0.091***	0.015***
	(0.021)	(0.032)	(0.036)	(0.036)	(0.064)	(0.070)	(0.013)	(0.005)
Control variables	No	Yes	No	Yes	No	Yes	No	Yes
Constants	1.902***	2.516***	10.515***	9.512***	13.876***	48.289***	10.648***	3.297***
	(0.158)	(0.189)	(1.926)	(1.284)	(3.470)	(13.241)	(0.700)	(0.631)
R-squared	0.868	0.936	0.781	0.869	0.745	0.879	0.796	0.993
Adjusted R-squared	0.857	0.921	0.762	0.839	0.724	0.851	0.779	0.992
Notes: *, , * represent 1%, 5% and 10% significance level

Table 15
Mediation effect of technological advancement and institutional quality on food security
	Food production		Rice production		Cassava production		Fruit production
Variables	Model-1	Model-2	Model-3	Model-4	Model-5	Model-6	Model-7	Model-8
Technological advancement	0.036**	0.124***	0.089***	0.350***	0.036^NS	0.088	0.047***	0.039***
	(0.016)	(0.042)	(0.029)	(0.117)	(0.063)	(0.183)	(0.008)	(0.012)
Institutional quality	0.206***	0.149***	0.234***	0.191***	0.446***	0.153*	0.073***	0.014**
	(0.023)	(0.037)	(0.031)	(0.054)	(0.065)	(0.085)	(0.008)	(0.005)
Control variables	No	Yes	No	Yes	No	Yes	No	Yes
Constants	4.243***	40.195***	13.453***	20.238***	14.134***	17.521***	12.313***	8.052*
	(0.038)	(10.075)	(0.051)	(9.514)	(0.108)	(8.521)	(0.014)	(3.981)
R-squared	0.876	0.934	0.831	0.903	0.748	0.867	0.924	0.993
Adjusted R-squared	0.866	0.918	0.817	0.880	0.727	0.835	0.918	0.991
Notes: NS refer to not significant, *, , * represent 1%, 5% and 10% significance level

This study provides robust empirical evidence on the dynamic and structural relationships between renewable energy adoption, technological advancement, environmental conditions, institutional quality, and food security in Sierra Leone. By employing a comprehensive methodological framework including Autoregressive Distributed Lag (ARDL) models, dynamic ARDL simulations, Kernel Regularized Least Squares (KRLS), and Fully Modified Ordinary Least Squares (FMOLS), the research advances the literature on sustainable energy transitions, environmental resilience, and food system stability in low income and climate vulnerable economies.

The findings demonstrate that renewable energy adoption has a consistently positive and statistically significant effect on food security. Access to clean energy facilitates mechanization, irrigation, post-harvest processing, and cold storage, particularly in rural areas where infrastructure deficits constrain agricultural productivity. These results align with SDG 7 (Affordable and Clean Energy) and SDG 2 (Zero Hunger) by emphasizing the role of renewable energy in improving agricultural production and reducing food insecurity. Technological progress, proxied by the use of fertilizers and pesticides, also enhances productivity. However, the nonlinear effects observed suggest diminishing returns at higher levels of input application. This reflects the Environmental Kuznets Curve hypothesis and highlights the need for efficiency and sustainability in agricultural intensification.

Environmental stressors, particularly air pollution and temperature variability, exert consistently negative impacts on crop yields, showing the sector’s vulnerability to ecological degradation and climate volatility. These findings highlight the urgency of SDG 13 (Climate Action) and SDG 15 (Life on Land), as climate and environmental pressures threaten food security. The application of resilience theory suggests that renewable energy and technological innovation act as adaptive capacities that strengthen agricultural systems against environmental shocks.

Mediation analysis confirms that institutional quality partially transmits the effects of energy and technological interventions to food security. This finding underscores the importance of governance, regulatory coherence, and administrative effectiveness in achieving productive outcomes, consistent with institutional theory and the aims of SDG 16 (Peace, Justice, and Strong Institutions). Institutional quality therefore emerges as a crucial enabling condition that enhances the effectiveness of clean energy and technology in building food system resilience.

Robustness checks confirm that renewable energy and technological inputs consistently increase the production of rice, maize, cassava, vegetables, cocoa, and fruits, while air pollution and temperature variability reduce yields across these crops. Overall, the study demonstrates that achieving sustainable food security in Sierra Leone requires integrated strategies that combine clean energy transitions, sustainable technological innovation, environmental stewardship, and institutional reform. These insights align with the 2030 Agenda for Sustainable Development and contribute to theoretical debates on the connections between energy, environment, institutions, and food system resilience in vulnerable economies.

Based on these findings, several policy implications emerge. First, expanding renewable energy access in agriculture should be a strategic priority. Investments in decentralized energy infrastructure, such as solar-powered irrigation systems and off-grid agro-processing facilities, can significantly enhance productivity and resilience, particularly in underserved rural communities. Second, institutional capacity and governance must be strengthened. Given the mediating role of institutional quality, reforms aimed at improving regulatory frameworks, enhancing policy coordination, and increasing administrative efficiency are essential to maximize the returns on energy and technological investments. Third, environmental sustainability should be integrated into agricultural development planning. The observed negative effects of air pollution and temperature variability call for stringent emissions regulation, promotion of climate-smart agricultural practices, and investment in ecosystem restoration to improve long-term resilience.

Fourth, the efficient use of agricultural inputs should be encouraged. While fertilizers and pesticides contribute positively to food security, their nonlinear effects indicate the need for targeted policies promoting optimal input use. This includes farmer training, agricultural extension services, and the adoption of integrated nutrient and pest management systems to enhance input efficiency and minimize ecological harm. Fifth, investment in rural development and land tenure security is critical. The positive association between rural population and food output reflects the labor-intensive nature of agriculture in Sierra Leone. Policymakers should therefore invest in rural infrastructure, education, healthcare, and legal mechanisms to secure land rights, enabling smallholder farmers to engage in long-term planning and resource conservation. Sixth, economic growth strategies should be aligned with agricultural transformation. The limited figures and statistically significant effect of gross domestic product per capita on food output suggests that broader economic expansion does not necessarily translate into improvements in the agricultural sector. Therefore, growth must be complemented by targeted investments in agricultural value chains, rural finance, and research and development.

This study advances the understanding of how energy access, technological change, environmental stressors, and institutional quality interact to influence food security in fragile agrarian settings. The use of multiple analytical techniques reinforces the robustness of the findings and enhances their relevance for policymaking. For Sierra Leone and similar economies, the results underscore the need for coordinated, multisectoral strategies that simultaneously address energy poverty, environmental degradation, weak institutions, and rural underdevelopment. Future research should consider spatial variation, long-term equity outcomes, and the role of climate adaptation strategies in shaping sustainable food security. Achieving meaningful agricultural transformation will require policy frameworks that are context-specific, evidence-based, and institutionally embedded.

Acknowledgments

The first author, Abdul Salami Bah, is affiliated with the College of Economics and Management, Northwest A&F University, China. I am profoundly grateful to the China Scholarship Council (CSC) for its generous financial support, which has been instrumental in facilitating my doctoral research and academic development. The views and findings presented in this work are solely those of the authors and do not necessarily reflect the official positions of the CSC or the Government of the People's Republic of China. We also extend our sincere appreciation to the editors and anonymous reviewers for their insightful feedback and constructive suggestions, which have greatly enhanced the quality of this manuscript.

Author’s Contribution

Abdul Salami Bah: Writing–original draft, Formal analysis, Data curation, Investigation, Conceptualization, Methodology, Software development, Visualization, Supervision, Project administration, Writing–review & editing, Resources. Nazir Muhammad Abdullahi: Data validation, Conceptualization, Methodology, Writing–review & editing. Saffa Mohamed Massaquoi: Data validation, Writing–review & editing, Resources. Aklok Getnet: Data curation, Writing–review & editing. Abdulai Jusufu: Formal analysis, Data validation, Visualization, Writing–review & editing. Bello Nasiru Abdullahi: Formal analysis, Data curation, Visualization, Writing–review & editing. Chernor A.U Bah: Formal analysis, Data curation, Visualization, Writing–review & editing. Kenneth Lansana Mangoh: Formal analysis, Data curation, Visualization, Supervision, Project administration, Writing–review & editing.

Funding

This study did not receive any financial support from any organization.

Data availability statement

The replication code and data used in the analysis of this study are available from the authors upon request.

Ethical Considerations:

This study involves no human ethical concerns, as it relies exclusively on secondary data obtained from publicly accessible sources, including the World Bank database and FAOSTAT. Because no human participants or primary data collection were involved, a clinical trial registration number is not applicable.

Consent for publication

Not applicable.

Conflict of Interest

The authors declare that they have no conflict of interest.

FAO, Africa - Regional Overview of Food Security and Nutrition 2023. 2023.
Oduoye, M.O., et al., The outlook of food security and food safety in Africa: correspondence. Ann Med Surg (Lond), 2023. 85(4): p. 1314-1315.
WFP, World Food Programme (WFP) State of food security in Sierra Leone 2020 comprehensive food security and vulnerability analysis. 2021.
IMF, International Monetary Fund (IMF). Climate Change and Food Insecurity: Options for Building Resilience in Sierra Leone International Monetary, 2024. 2024(322): p. 98.
Abdullahi, N.M., et al., Modeling the dynamic nexus between climate change, ICT, and agriculture in sub-Saharan Africa: a panel quantile ARDL. GeoJournal, 2024. 89(5).
Alimagham, S., et al., Climate change impact and adaptation of rainfed cereal crops in sub-Saharan Africa. European Journal of Agronomy, 2024. 155.
Pickson, R.B., et al., The impacts of climatic conditions on cereal production: implications for food security in Africa. Environ Dev Sustain, 2023: p. 1-28.
Pickson, R.B., et al., Examining the impacts of climate change and political instability on rice production: empirical evidence from Nigeria. Environ Sci Pollut Res Int, 2023. 30(23): p. 64617-64636.
Umeh, S.C. and L.A. Gil-Alana, Trends in temperatures in Sub-Saharan Africa. Evidence of global warming. Journal of African Earth Sciences, 2024. 213.
Ahmed, M., J. shuai, and H. Ali, The effects of climate change on food production in India: evidence from the ARDL model. Environment, Development and Sustainability, 2023. 26(6): p. 14601-14619.
Xu, X., et al., Quantifying the impacts of climate variability and human interventions on crop production and food security in the Yangtze River Basin, China, 1990-2015. Sci Total Environ, 2019. 665: p. 379-389.
Murray-Tortarolo, G.N., V.J. Jaramillo, and J. Larsen, Food security and climate change: the case of rainfed maize production in Mexico. Agricultural and Forest Meteorology, 2018. 253-254: p. 124-131.
Brenya, R., et al., Food security in sub-Sahara Africa: Exploring the nexus between nutrition, innovation, circular economy, and climate change. Journal of Cleaner Production, 2024. 438.
Haque, M.I. and M.R. Khan, Impact of climate change on food security in Saudi Arabia: a roadmap to agriculture-water sustainability. Journal of Agribusiness in Developing and Emerging Economies, 2020. 12(1): p. 1-18.
Atiah, W.A., et al., Climate variability and impacts on maize (Zea mays) yield in Ghana, West Africa. Quarterly Journal of the Royal Meteorological Society, 2021. 148(742): p. 185-198.
Samarasinghe, B.K.D.J.R., et al., Climate change and rice production in Sri Lanka: short-run vs. long-run symmetric and asymmetric effects. Frontiers in Sustainable Food Systems, 2025. 9.
Chandio, A.A., et al., Does internet technology usage improve food production? Recent evidence from major rice-producing provinces of China. Computers and Electronics in Agriculture, 2023. 211.
Majeed, Y., et al., Renewable energy as an alternative source for energy management in agriculture. Energy Reports, 2023. 10: p. 344-359.
Akbar, A., et al., An institutional framework for enhanced food security amidst the COVID-19 pandemic: Strategic implementation and outcomes. Journal of Agriculture and Food Research, 2025. 21.
Justice, G., G. Nyantakyi, and S.H. Isaac, The effect of renewable energy on carbon emissions through globalization. Heliyon, 2024. 10(5): p. e26894.
Yang, Z., et al., Can renewable energy investment reduce carbon dioxide emissions? Evidence from scale and structure. Energy Economics, 2022. 112.
Chen, C., M. Pinar, and T. Stengos, Renewable energy consumption and economic growth nexus: Evidence from a threshold model. Energy Policy, 2020. 139.
Girgibo, N., et al., Risks of climate change effects on renewable energy resources and the effects of their utilisation on the environment. Energy Reports, 2024. 11: p. 1517-1534.
Uddin, I., et al., Enhancing institutional quality to boost economic development in developing nations: New insights from CS-ARDL approach. Research in Globalization, 2023. 7.
Giller, K.E., The Food Security Conundrum of sub-Saharan Africa. Global Food Security, 2020. 26.
Choruma, D.J., et al., Digitalisation in agriculture: A scoping review of technologies in practice, challenges, and opportunities for smallholder farmers in sub-saharan africa. Journal of Agriculture and Food Research, 2024. 18.
Hossain, M.R., et al., Green energy transition in OECD region through the lens of economic complexity and environmental technology: A method of moments quantile regression perspective. Applied Energy, 2024. 365.
Jordan, S. and A.Q. Philips, Cointegration Testing and Dynamic Simulations of Autoregressive Distributed Lag Models. The Stata Journal: Promoting communications on statistics and Stata, 2018. 18(4): p. 902-923.
FAO, An Introduction to the Basic Concepts of Food Security. 2008.
Massaquoi, S.M., et al., Does Financial Credit Obtained From Financial Institutions Influence Agricultural Productivity While Balancing Economic Growth and Sustainability? Empirical Evidence From Sierra Leone Using the VAR Approach. SAGE Open, 2025. 15(2).
Musa, M.F., et al., The Effects of Bushfires on Farming Industry Among Small-Scale Farmers in Kori Chiefdom, Moyamba District, Sierra Leone. International Journal of Agricultural Economics, 2023.
IFAD, Republic of Sierra Leone Country Strategic Opportunities Programme 2020-2025. 2020.
Hanif, M.H., Dynamic ARDL simulations of Pakistan’s energy intensity: the role of non-fossil energy, innovation, and financial development. Environment, Development and Sustainability, 2025.
Burney, J., et al., Solar-powered drip irrigation enhances food security in the Sudano-Sahel. Proc Natl Acad Sci U S A, 2010. 107(5): p. 1848-53.
He, J., et al., Green economic growth, renewable energy and food security in Sub-Saharan Africa. Energy Strategy Reviews, 2024. 55.
Gupta, E., The impact of solar water pumps on energy-water-food nexus: Evidence from Rajasthan, India. Energy Policy, 2019. 129: p. 598-609.
Kyire, S.K.C., et al., Climate change and food security nexus in Ghana: The role of renewable energy. Sci Total Environ, 2025. 963: p. 178454.
Rockström, J., Steffen, W., Noone, K., Persson, Å., Chapin, F. S., Lambers, J. H., ... & Foley, J. A., A safe operating space for humanity. Nature, 2009. 461(7263),: p. 472-475.
Ongoma, V., et al., Closing yield gap for sustainable food security in sub-Saharan Africa – progress, challenges, and opportunities. Frontiers in Agronomy, 2025. 7.
Glover, J.L., et al., An Institutional Theory perspective on sustainable practices across the dairy supply chain. International Journal of Production Economics, 2014. 152: p. 102-111.
Fuenfschilling, L. and B. Truffer, The structuration of socio-technical regimes—Conceptual foundations from institutional theory. Research Policy, 2014. 43(4): p. 772-791.
Faundez, J., Douglass North’s Theory of Institutions: Lessons for Law and Development. Hague Journal on the Rule of Law, 2016. 8(2): p. 373-419.
Akono, E.B. and V.C. Kemezang, Balancing environmental sustainability and international trade: a dynamic analysis of factors influencing renewable energy generation in Cameroon. Environment, Development and Sustainability, 2025.
Hidrobo, M., et al., Making complementary agricultural resources, technologies, and services more gender-responsive. Global Food Security, 2024. 42.
Cook, B.R., P. Satizábal, and J. Curnow, Humanising agricultural extension: A review. World Development, 2021. 140.
Lio, M. and M.-C. Liu, Governance and agricultural productivity: A cross-national analysis. Food Policy, 2008. 33(6): p. 504-512.
Oppong, C., Leveraging Governance and Accountability for Advancing Sustainable Development Goals in Africa. Business Strategy & Development, 2025. 8(2).
Stringer, L.C., et al., Adaptation and development pathways for different types of farmers. Environmental Science & Policy, 2020. 104: p. 174-189.
Sarkodie, S.A. and P.A. Owusu, How to apply the novel dynamic ARDL simulations (dynardl) and Kernel-based regularized least squares (krls). MethodsX, 2020. 7: p. 101160.
Dickey, D.A. and W.A. Fuller, Distribution of the Estimators for Autoregressive Time Series with a Unit Root. Journal of the American Statistical Association, 1979. 74: p. 427-431.
Phillips, P.C.B. and P. Perron, Testing for a unit root in time series regression. Biometrika, 1988. 75(2): p. 335-346.
Zivot, E. and D.W.K. Andrews, Further Evidence on the Great Crash, the Oil-Price Shock, and the Unit-Root Hypothesis. Journal of Business & Economic Statistics, 2002. 20(1): p. 25-44.
Pesaran, M.H., Y. Shin, and R.J. Smith, Bounds testing approaches to the analysis of level relationships. Journal of Applied Econometrics, 2001. 16: p. 289-326.
Haug, A.A., Temporal Aggregation and the Power of Cointegration Tests: a Monte Carlo Study. 2002.
Odhiambo, N.M., Finance-growth-poverty nexus in South Africa: A dynamic causality linkage. The Journal of Socio-Economics, 2009. 38(2): p. 320-325.
Hainmueller, J. and C. Hazlett, Kernel Regularized Least Squares: Reducing Misspecification Bias with a Flexible and Interpretable Machine Learning Approach. Political Analysis, 2017. 22(2): p. 143-168.
Wagner, M., P. Grabarczyk, and S.H. Hong, Fully modified OLS estimation and inference for seemingly unrelated cointegrating polynomial regressions and the environmental Kuznets curve for carbon dioxide emissions. Journal of Econometrics, 2020. 214(1): p. 216-255.
Kripfganz, S. and D.C. Schneider, Response Surface Regressions for Critical Value Bounds and Approximate p‐values in Equilibrium Correction Models1. Oxford Bulletin of Economics and Statistics, 2020. 82(6): p. 1456-1481.
Sun, F., Y. Dai, and X. Yu, Air pollution, food production and food security: A review from the perspective of food system. Journal of Integrative Agriculture, 2017. 16(12): p. 2945-2962.
Subedi, B., A. Poudel, and S. Aryal, The impact of climate change on insect pest biology and ecology: Implications for pest management strategies, crop production, and food security. Journal of Agriculture and Food Research, 2023. 14.
IPCC, Climate Change 2021: The Physical Science Basis. Intergovernmental Panel on Climate Change. 2021.
Erdogan, S., U.K. Pata, and M.T. Kartal, Does economic transition enhance environmental sustainability? The case of Russia. Environment, Development and Sustainability, 2025.
Oyetunji, O., N. Bolan, and G. Hancock, A comprehensive review on enhancing nutrient use efficiency and productivity of broadacre (arable) crops with the combined utilization of compost and fertilizers. Journal of Environmental Management, 2022. 317.
Ahmad, M.F., et al., Pesticides impacts on human health and the environment with their mechanisms of action and possible countermeasures. Heliyon, 2024. 10(7).
Zhou, Y., et al., Identifying the contradiction between the cultivated land fragmentation and the construction land expansion from the perspective of urban-rural differences. Ecological Informatics, 2022. 71.
Li, K., L. Wang, and L. Wang, Consumption as the catalyst: Analyzing rural power infrastructure and agricultural growth through panel threshold regression and data-driven prediction. Applied Energy, 2024. 365.
Dethier, J.-J. and A. Effenberger, Agriculture and development: A brief review of the literature. Economic Systems, 2012. 36(2): p. 175-205.
Ketu, I. and S.M. Nguea, Globalization and the “zero hunger” goal in Africa: Starving in an open world? Journal of International Development, 2024. 36(7): p. 2769-2789.
Cao, L. and S.M. Nguea, Vulnerable energy, vulnerable food: Assessing the effects of energy vulnerability on food security in Africa. Energy, 2025. 320.

United Nations Sustainable Development Goals https://sdgs.un.org/goals
World Bank (2025). Sierra Leone: Country Climate and Development Report https://reliefweb.int/report/sierra-leone/sierra-leone-country-climate-and-development-report
EPA (2020) report. Environment Protection Agency (EPA) Sierra Leone. https://unfccc.int/sites/default/files/NDC/2022-06/SIERRA%20LEONE%20INDC.pdf
HDI (2024) report, United Nations Human Development Index https://www.bmz.de/en/countries/sierra-leone
UNOPS (2022) report, Rural Renewable Energy Project (RREP), Sierra Leone end-line Impact Evaluation Report. https://www.theigc.org/sites/default/files/2022/06/Levine-et-al.-Final-Report-2022.pdf

No competing interests reported.

Download PDF

Editorial decision: Revision requested
15 Dec, 2025
Reviews received at journal
14 Dec, 2025
Reviews received at journal
08 Dec, 2025
Reviewers agreed at journal
08 Dec, 2025
Reviews received at journal
04 Dec, 2025
Reviewers agreed at journal
04 Dec, 2025
Reviewers agreed at journal
04 Dec, 2025
Reviewers invited by journal
03 Dec, 2025
Editor invited by journal
01 Dec, 2025
Editor assigned by journal
30 Nov, 2025
Submission checks completed at journal
30 Nov, 2025
First submitted to journal
28 Nov, 2025

You are reading this latest preprint version

Modeling the effects of environmental degradation, renewable energy, and technological advancement on food security in Sierra Leone: Does institutional quality matter?

Status:

Version 1

Abstract

Background

Methods

Results

Conclusion

Figures

1. Introduction

2. Literature Review

2.1. Food security in Sierra Leone: A contextual overview

2.2. Renewable energy and food security

2.3. Environmental degradation and agricultural viability

2.4. Technological advancement: Inputs and productivity

2.5. Theoretical foundation

3. Materials and Methods

3.1. Study area and data sources

3.2. Dependent variables

3.3. Independent variables

3.4. Control variables

3.5. Mediation variable

3.6. Model specification

3.7. Empirical methodology

3.7.1. Unit root test

3.7.2. PSS bound test

3.7.3. Estimation of autoregressive distributed lag model

3.7.4. Stability test

3.7.5. Dynamic simulated ARDL model

3.7.6. Kernel-based regularized least square (KRLS)

3.8. Fully Modified Ordinary Least Squares for robustness analysis

4. Results and Discussion

4.1. Descriptive Statistics

4.2. Preliminary analysis findings

4.3. Autoregressive distribution lag and Dynamic Autoregressive distribution lag model estimation

4.4. Diagnostic and stability tests

4.5. Novel dynamic ARDL simulations

4.6. Kernel-based regularized least squares

4.7. Robustness and moderating effects analysis

4.8. Mechanism analysis

5. Conclusion and policy implications

Declarations

References

Footnotes

Additional Declarations

Status:

Version 1