Examining the Impact of Fertilizer Use, Economic Expansion, Methane Emissions, and Population Growth on Food Security in Nigeria

Abstract

Food security remains a critical challenge in Nigeria. As a result, this research examines the impact of fertilizer use, economic expansion, population growth, and methane emissions on food security in Nigeria from 1970 to 2022. The methodologies used include the Autoregressive Distributed Lag (ARDL) model, Wald Test, and the Spectral Granger Causality test. The ARDL results demonstrate that in the long run, fertilizer use spurs food security, although not significantly, while population growth reduces food security insignificantly. On the other hand, economic expansion and agricultural methane emissions are positively associated with food security, likely reflecting scale effects of agricultural production rather than a direct beneficial role of emissions. In the short run, fertilizer use and methane emissions drive food security. The Wald Test also confirms the short-run findings. Furthermore, the Spectral Granger Causality test showed that fertilizer use and economic expansion Granger-cause food security in the long, medium, and short term. Population growth, however, Granger-causes food security only in the long term, while methane emissions Granger-cause food security in the medium and long term. Based on these results, policies are recommended, and their further implications are discussed.

1. Introduction

Food is vital for survival, and so any factor impeding its availability and access has to be properly examined. In addition, one of the greatest issues confronting humanity is making sure that everyone always has physical, economic, and social access to enough safe and nourishing food to satisfy their nutritional requirements and preferences for a life of activity and wellness, while also respecting ecological boundaries [1,2,3,4]. This is also the definition of food security (FOS) adopted in the World Food Summit of 1996.

The concept of FOS is multifaceted. Reference [5] identified that food security has four dimensions: availability, access, utilization, and stability. This study focuses primarily on food availability and access, which are most directly affected by agricultural production and income in Nigeria. The “supply side” of FOS is addressed by food availability, which is based on net trade, the level of inventories, and food production levels. FOS at the household level is not always ensured by a sufficient national or worldwide food supply. Due to worries about inadequate access to food, policies are now more focused on incomes, spending, markets, and prices in order to meet food security goals. Generally, utilization refers to how the body maximizes the different nutrients included in the diet. Food preparation, dietary diversity, intra-household food distribution, and proper care and feeding habits all contribute to an individual’s adequate energy and nutritional intake. This determines nutritional condition when combined with optimal biological metabolism of the food consumed. Lastly, having insufficient access to food regularly puts people’s nutritional status in danger. Thus, even if they eat enough food today, they are still deemed to be food-insecure. People’s level of FOS may be impacted by unfavorable weather, unstable political environments, or economic situations (such as joblessness or growing food costs). In addition, FOS is further subdivided into people who are food-insecure chronically (long term/persistent) or transitionally (short term/temporary).

According to ref. [6], a persistent dearth of progress in achieving the objective of Zero Hunger is revealed by the evaluation of global hunger in 2023 as determined by the rate of prevalence of undernourishment (PoU) (SDG Indicator 2.1.1). Besides hunger, there has been no improvement in four years since the high spike from 2019 to 2020 during the COVID-19 pandemic, and the worldwide prevalence of moderate to severe food insecurity (SDG Indicator 2.1.2) still much exceeds levels before the pandemic. In addition, 2.33 billion people, or 28.9% of the world’s population, were predicted to be moderately or severely food-insecure in 2023, meaning they did not regularly have access to enough food. More than 864 million people, or 10.7% of the population, are estimated to be extremely food-insecure, which means they occasionally ran out of food throughout the year and, in a worst-case scenario, have gone for a full day or more with no food. Africa has a roughly twofold higher rate of moderate or severe food insecurity (58%), while Latin America and the Caribbean (LAC), Asia, and Oceania have rates that are closer to the world average (28.2%, 24.8%, and 26.8%, respectively) [6].

In Nigeria, ref. [7] stated that 30.6 million individuals (out of a total population is 227 million) are expected to experience severe food and nutrition insecurity throughout 26 states and the Federal Capital Territory (FCT). Although this number is marginally better than the 33.1 million from the previous year, stakeholders point out that the analysis has been hampered by restricted data access. With extremely high rates of acute malnutrition and the largest number of food-insecure individuals in the world, there is a need for consistent and coordinated intervention. For a country like Nigeria, some notable factors can drive food insecurity, including climate change, a rise in the prices of goods and services, otherwise known as inflation, economic difficulties, and political instabilities mixed with violence, especially in the Northern part of Nigeria [8]. Due to these challenges, it is important to investigate the drivers of food insecurity in Nigeria so that the problems will not deteriorate, and so as not to exacerbate the poverty level.

Other factors that can drive FOS to which other studies have not paid attention are fertilizer use (FER) [9], population growth (POP) [10], economic expansion (GDP) [11], and methane (CH₄) emissions [12], which is one of the GHGs. Enhancing the yield of crops and reducing the gap between actual and attainable harvests can be accomplished through introducing and adopting various methods and technologies, e.g., proper application of fertilizers and efficient management of nutrients can play essential roles for global FOS [13]. GDP can spur FOS because the more a country produces value-added goods and services, the greater the capability of earning more revenue, while at the same time positively affecting the income of the population. According to a report by ref. [14], by 2050, the world’s population is expected to have grown to around 10 billion people. The population of Sub-Saharan Africa (SSA) is expected to double by 2050, from 227 million to 2.2 billion, an increase of over ten times compared to 1960, whereas growth will be moderate in other regions. It is essential to state that the population in Africa will still grow significantly even if women only have an average of two children, especially when life expectancy is rising and health care is improving, which lowers death rates [15]. CH₄ emissions from agriculture have a significant impact on both climate change and food security [16]. Ref. [17] opined that one of the biggest challenges facing humanity in the twenty-first century is producing the food required for a planet with a large population while also reducing CH₄ emissions. The interplay between CH₄ emissions and FOS is critical, as climate change exacerbates food production challenges, leading to increased hunger rates globally [18]. The major sources of CH₄ emissions are agriculture and fossil fuels.

Therefore, the objective of this research is to examine the impact of FER, GDP, POP, and CH₄ emissions on FOS in Nigeria from 1970 to 2022, using methods such as the Autoregressive Distributed Lag (ARDL), Wald Test, and Spectral Granger Causality.

Contribution and Novelty

The contributions of this study are as follows: First, the literature on the impact of FER on FOS is very scanty, especially for a country like Nigeria. In addition, the model of this study is designed to capture some of the important factors (GDP, POP, and CH₄ emissions) that can drive FOS in Nigeria. In addition, other investigations, such as [19,20], adopted CO₂ as a proxy for GHGs, while this study used CH₄ emissions because they are related to agriculture. Likewise, the study period employed is longer (53 observations), thus providing better inference and estimates. Second, this research used the conventional ARDL method by ref. [21] in addition to the Wald Test causality method to confirm the short-run ARDL findings. To ascertain a robust long-run association, the ARDL-EC analysis proposed by ref. [22] is used. Third, to determine the direction of causality, the Spectral Granger Causality approach proposed by ref. [23] is adopted. According to refs. [24,25], this test breaks down the Granger causality between two-time processes throughout the spectrum and has good size qualities [26]. The outcome of this research will be very useful for all stakeholders, including policymakers, farmers, researchers, investors, and the public, by showing the factors that can drive food insecurity and the likely solutions to this challenge.

2. Literature Review

This section investigates the link between FER, GDP, POP, CH₄ emissions, and FOS.

2.1. FER and FOS Nexus

In 1976, ref. [27] stated that Nigeria had a subsidy scheme, which had a great impact on crop production. More specifically, farmer’s access to fertilizers improved. Ref. [13] stated that inorganic fertilizer plays an important part in global FOS; however, it should be noted that in some systems, the maximum yields are obtained by combining organic and inorganic nutrient sources. Ref. [28] emphasized that healthy soils are foundational for achieving FOS and sustainable development. To address these challenges, soil resources must receive global attention to mitigate pollution and ensure the sustainability of food systems. Ref. [29] opined that nano-fertilizers greatly enhance crop yield with high-quality fruits and grains while enhancing the quality of the soil and plant development performance. According to ref. [30], crop productivity, plant nutrient availability, and nutrient encapsulation are all enhanced with nano-fertilizers. Ref. [9] argued that poor global fertilization management results in regions with severe nutrient deficiencies in croplands associated with limited access to fertilizers that obviously restrict the production of food, as well as regions which are overfertilized, with the ensuing issues of ecological contamination that have an impact on human health.

Ref. [31] also expressed concerns about the adverse impact of rising fertilizer prices on FOS. Ref. [32] stated that Nigeria faces an FOS dilemma due to its high dependence on food imports. Projections show that by 2030, the population of Nigeria will double its 2006 estimates, which necessitates expanded production of food to meet the needs of the population and the exploration of possibilities of exports. Furthermore, Nigeria is dealing with soil deterioration because of unsuitable farming methods: erosion, deforestation, and climate change. It threatens the formerly dominating subsistent agriculture economy. The issues raised above demonstrate how urgently it is for fertilizers in food production to improve FOS. For African economies, ref. [33] confirms the role of fertilizer use in achieving FOS. Ref. [34] for India and ref. [35] for China confirmed that fertilizers, when used appropriately, contribute to FOS. Ref. [36] affirmed the importance of fertilizers in achieving FOS; the study mentioned that the market for fertilizers is intricately linked to shifting geopolitical conditions and distortions. It is important to state that excessive use of fertilizers can pollute the soil and hinder agricultural output.

2.2. GDP and FOS Nexus

Ref. [37] found that higher GDP growth influences FOS by increasing demand for agricultural commodities, leading to higher world market prices. However, developing economies may become more import-dependent, making them vulnerable to market disruptions despite improved trade balances. Ref. [38] indicated that in Ethiopia, chronic food insecurity is not primarily due to economic growth or income distribution, but rather inflationary pressures, population growth, and inadequate food storage, showing a complex relationship between GDP and food security. In selected regions, higher levels of GDP are associated with improved FOS. Countries experiencing rapid GDP per capita growth also see significant enhancements in FOS, highlighting economic growth as a crucial factor for addressing food security issues [11].

Ref. [39] found that GDP is a crucial indicator for FOS, influencing food affordability and availability. Higher GDP correlates with improved agricultural productivity and population dynamics, essential for enhancing FOS in Ukraine and other countries. Ref. [40] found that COVID-19 caused a 7.2% decrease in global GDP, leading to a 27.8% increase in food insecurity, adding 211 million food-insecure individuals in 2020. The pandemic disrupted economies, affecting food availability and accessibility, particularly in vulnerable regions. Ref. [41] establishes a negative relationship between GDP growth rate and undernutrition prevalence in developing countries, indicating that GDP is crucial for improving FOS and reducing poverty among the undernourished population across Latin America, SSA, and Asia.

Ref. [42] discovered a significant relationship between GDP and FOS, indicating that increased GDP positively impacts FOS. Enhanced scientific output and innovation also contribute to FOS in developing countries, underscoring the interconnectedness of these factors. Ref. [43] discovered that the share of food in Kazakhstan’s GDP is a critical factor in FOS, emphasizing the need for effective state regulation and support for agricultural producers to enhance economic stability and ensure domestic food market saturation. Ref. [44] found a positive and statistically significant effect of GDP on FOS in MIST countries, indicating that increases in GDP contribute to FOS. Recent studies, such as refs. [45,46], argue that GDP negatively affects FOS in WAEMU and G-7 economies, respectively.

2.3. POP and FOS Nexus

Ref. [47] opined that factors like poverty, lack of agricultural investment, and climate change exacerbate food instability, highlighting the need for biotechnological advancements to reduce undernourishment and conserve resources. Rising population, coupled with climate change effects on crop productivity and arable land, poses serious challenges to global FOS in the future [48]. Ref.’s [39] research highlights a dual connection between POP dynamics and FOS, where rapid POP increases food demand, while inadequate food affordability and availability negatively impact population size. Ref. [49] found a significant relationship between POP and FOS among refugees in Nakivale, Uganda. High fertility and influx lead to food insufficiency, while an increased working-age population enhances FOS through greater household productivity and consumption management.

In Nigeria, ref. [50] concluded that POP is not the primary cause of food insecurity. Instead, ineffective agricultural policies, conflicts, displacement, climatic factors, and underutilized resources are more significant contributors to the FOS crisis than the population explosion. Ref. [51] emphasized the need for sustainable agricultural practices to address food insecurity, particularly in developing economies. Also, ref. [52] opined that the global population is expected to reach 10 billion, necessitating a 50% increase in food production. Sustainable food security requires innovative approaches to meet this demand while minimizing environmental impacts and combating hunger. Ref. [53] explained that in Nigeria, the POP significantly impacts FOS, with a growth rate of over 2.6% outpacing agricultural output growth of 3.5% from 2011 to 2020, leading to increased food insecurity, hunger, and malnutrition, necessitating government intervention in agriculture. Ref. [54] discovered that population aging in rural China has not negatively impacted FOS because artificial intelligence (AI) plays a favorable moderating role, particularly in central and western regions, enhancing FOS despite demographic changes. Targeted policies are recommended to further protect FOS.

Ref. [55] established that POP negatively impacts FOS, particularly in developing countries. The interaction between population growth and biofuel production exacerbated FOS issues across its four dimensions. Ref. [56] discovered that the rising population in Malaysia significantly impacts FOS, increasing demand for food while highlighting vulnerabilities such as dependency on imports and aging farmers. The study suggested that effective government policies, private investment, and technological advancements are essential to address these challenges and ensure sustainable food systems. Ref. [57] established that the world population is growing rapidly, intensifying pressure on food production systems, and that this growth contributes to food insecurity, with over 345 million people affected in 2023, highlighting the urgent need for sustainable agriculture to meet increasing food demands.

2.4. CH₄ Emissions and FOS Nexus

Ref. [58] highlighted that greenhouse gas emissions, particularly from red meat production, significantly contribute to methane emissions. This highlights a conflict between reducing these emissions and maintaining FOS through self-sufficiency in domestic agricultural production in Norway. Ref. [59] discovered that GHGs from agriculture can significantly impact FOS, potentially increasing undernourishment by 80–300 million people by 2050 if not effectively managed. Ref. [60] established that climate-smart agriculture practices aim to enhance food security while reducing CH₄ emissions, thereby promoting sustainable agricultural production and easing global warming impacts. Ref. [16] found that CH₄ emissions impacted FOS through methane-intensive food wastage, and can be mitigated through dietary shifts and reduced supply chain loss strategies, which are essential to balance emissions and ensure resilient food systems. The impact of CH₄ emissions on FOS contributes to climate change, which directly affects crop production and indirectly influences food availability, cost, and supply chains [18]. Ref. [61] argued that agriculture significantly contributes to CH₄ emissions, primarily from livestock and rice cultivation. Addressing these emissions is crucial for enhancing food security while transitioning to sustainable practices that mitigate climate change impacts and promote climate resilience in agricultural systems. Ref. [62] discovered a statistically negligible correlation between long-term FOS and the positive CH₄ emissions shock. Nonetheless, this variable’s negative shock has a 7.5% beneficial impact on FOS.

2.5. Gap in the Literature

First, the connection between FER, GDP, POP, CH₄ emissions, and FOS presents mixed results. Some scholars argue that these variables contribute to FOS, while others assert that the impact on FOS is negative. The inconsistencies in these findings necessitate this research to be undertaken. Second, the research on the FER and FOS nexus is quite scanty, especially for a country like Nigeria, thus necessitating the use of Nigeria as a case study. Third, none of the research employed the Spectral Granger Causality test to investigate the relationship among the variables. This is one of the unique contributions of this research.

3. Data and Methodology

3.1. Data

The data used in this research is annual data from 1970 to 2022. Food security (FOS) is proxied by the Food Production Index, which is the dependent variable. The proxy for economic growth is Gross Domestic Product (annual %) (GDP), and population is proxied by urban population growth (annual %) (POP). Urban population growth is used instead of total population or rural population because it captures the dynamics of the market and production capacity. This study also used methane (CH₄) emissions from agriculture (Mt CO₂e), which is one of the GHGs. Fertilizers used per area of cropland (FER) was also employed, which is sourced from [63]. FOS, GDP, POP, and CH₄ emissions are sourced from [64].

Table 1. Variables summary.

Variables	Description of Variables	Source
FOS	Food Production Index (2014–2016 = 100)	[64]
FER	All fertilizers used per area of cropland	[63]
GDP	GDP growth (annual %)	[64]
POP	Urban population growth (annual %)	[64]
CH4	Methane (CH4) emissions from agriculture (Mt CO2e)	[64]

presents a summary of the variables and their sources.

It is important to state that the log of the variables was taken into consideration to handle any form of outliers. However, GDP was excluded because of the negative values in the data. Furthermore, the modified model of this research can be seen in the studies of refs. [18,19]. The model of this research is presented as follows:

$${\mathrm{F}\mathrm{O}\mathrm{S}}_{\mathrm{t}}={\mathrm{f}(\mathrm{F}\mathrm{E}\mathrm{R}}_{\mathrm{t}},{\mathrm{G}\mathrm{D}\mathrm{P}}_{\mathrm{t}},{\mathrm{P}\mathrm{O}\mathrm{P}}_{\mathrm{t}},{{\mathrm{C}\mathrm{H}}_4}_{\mathrm{t}})$$

(1)

Equation (1) can be rewritten in its log form as

$${\mathrm{L}\mathrm{F}\mathrm{O}\mathrm{S}}_{\mathrm{t}}={\mathsf{\delta }}_0+{{\mathsf{\delta }}_1\mathrm{L}\mathrm{F}\mathrm{E}\mathrm{R}}_{\mathrm{t}}+{\mathsf{\delta }}_2{\mathrm{G}\mathrm{D}\mathrm{P}}_{\mathrm{t}}+{\mathsf{\delta }}_3\mathrm{L}{\mathrm{P}\mathrm{O}\mathrm{P}}_{\mathrm{t}}+{\mathsf{\delta }}_4{{\mathrm{L}\mathrm{C}\mathrm{H}}_4}_{\mathrm{t}}+{\eta }_{\mathrm{t}}$$

(2)

Based on the literature that has been reviewed, it is expected that the association between FER and FOS will be positive (${\mathsf{\delta }}_1>0$); GDP and FOS will be positive (${\mathsf{\delta }}_2>0$), POP and FOS will be negative (${\mathsf{\delta }}_3<0$), taking into consideration the population structure of Nigeria, and CH₄ emissions and FOS could be positive (${\mathsf{\delta }}_4>0$) or negative (${\mathsf{\delta }}_4<0$). The impact of these variables on FOS in Nigeria is highly dependent on the structure of the Nigerian economy.

3.2. Methodology

Autoregressive Distributed Lag (ARDL)

The ARDL model, presented by [21], has several advantages, including the following: (i) it can be applied when variables show different levels of integration, with the exception of I(2); (ii) it applies to data with small sample sizes; (iii) it permits both short- and long-term analysis; (iv) it is statistically efficient and permits model flexibility; and (v) it can produce an unbiased long-term model [65]. Consequently, the ARDL model is shown as follows:

$$\begin{array}{ll}{∆\mathrm{L}\mathrm{F}\mathrm{O}\mathrm{S}}_{\mathrm{t}}={\varphi }_0+& {\sum }_{\mathrm{i}=1}^{\mathrm{p}}{\varphi }_1{∆\mathrm{L}\mathrm{F}\mathrm{O}\mathrm{S}}_{\mathrm{t}-1}+{\sum }_{\mathrm{i}=1}^{\mathrm{q}}{\varphi }_2{∆\mathrm{L}\mathrm{F}\mathrm{E}\mathrm{R}}_{\mathrm{t}-1}+{\sum }_{\mathrm{i}=1}^{\mathrm{q}}{\varphi }_3{∆\mathrm{G}\mathrm{D}\mathrm{P}}_{\mathrm{t}-1}\\ & +{\sum }_{\mathrm{i}=1}^{\mathrm{q}}{\varphi }_4{∆\mathrm{L}\mathrm{P}\mathrm{O}\mathrm{P}}_{\mathrm{t}-1}+{\sum }_{\mathrm{i}=1}^{\mathrm{q}}{\varphi }_5{{∆\mathrm{L}\mathrm{C}\mathrm{H}}_4}_{\mathrm{t}-1}+{\mathsf{\gamma }}_1{\mathrm{L}\mathrm{F}\mathrm{O}\mathrm{S}}_{\mathrm{t}-1}\\ & +{\mathsf{\gamma }}_2{\mathrm{L}\mathrm{F}\mathrm{E}\mathrm{R}}_{\mathrm{t}-1}+{\mathsf{\gamma }}_3{\mathrm{G}\mathrm{D}\mathrm{P}}_{\mathrm{t}-1}+{\mathsf{\gamma }}_4{\mathrm{L}\mathrm{P}\mathrm{O}\mathrm{P}}_{\mathrm{t}-1}+{\mathsf{\gamma }}_5{{\mathrm{L}\mathrm{C}\mathrm{H}}_4}_{\mathrm{t}-1}+{\mathsf{\mu }}_{\mathrm{t}}\end{array}$$

(3)

The difference operator is represented by $∆$; the lags of the regressand and regressors are represented by p and q, respectively; the intercept is represented by ${\varphi }_0$; the short-run coefficients vary from ${\varphi }_1$ to ${\varphi }_5$, the long-run coefficients range from ${\mathsf{\gamma }}_1$ to ${\mathsf{\gamma }}_5$, and the error term is represented by ${\mathsf{\mu }}_{\mathrm{t}}$. The model appears as follows if cointegration takes place:

$$\begin{array}{ll}{∆\mathrm{L}\mathrm{F}\mathrm{O}\mathrm{S}}_{\mathrm{t}}={\varphi }_0+& {\sum }_{\mathrm{i}=1}^{\mathrm{p}}{\varphi }_1{∆\mathrm{L}\mathrm{F}\mathrm{O}\mathrm{S}}_{\mathrm{t}-1}+{\sum }_{\mathrm{i}=1}^{\mathrm{q}}{\varphi }_2{∆\mathrm{L}\mathrm{F}\mathrm{E}\mathrm{R}}_{\mathrm{t}-1}+{\sum }_{\mathrm{i}=1}^{\mathrm{q}}{\varphi }_3{∆\mathrm{G}\mathrm{D}\mathrm{P}}_{\mathrm{t}-1}\\ & +{\sum }_{\mathrm{i}=1}^{\mathrm{q}}{\varphi }_4{∆\mathrm{L}\mathrm{P}\mathrm{O}\mathrm{P}}_{\mathrm{t}-1}+{\sum }_{\mathrm{i}=1}^{\mathrm{q}}{\varphi }_5{{∆\mathrm{L}\mathrm{C}\mathrm{H}}_4}_{\mathrm{t}-1}+{\mathsf{\gamma }}_1{\mathrm{L}\mathrm{F}\mathrm{O}\mathrm{S}}_{\mathrm{t}-1}\\ & +{\mathsf{\gamma }}_2{\mathrm{L}\mathrm{F}\mathrm{E}\mathrm{R}}_{\mathrm{t}-1}+{\mathsf{\gamma }}_3{\mathrm{G}\mathrm{D}\mathrm{P}}_{\mathrm{t}-1}+{\mathsf{\gamma }}_4{\mathrm{L}\mathrm{P}\mathrm{O}\mathrm{P}}_{\mathrm{t}-1}+{\mathsf{\varpi }\mathrm{E}\mathrm{C}\mathrm{T}}_{\mathrm{t}-1}\\ & +{\mathsf{\gamma }}_5{{\mathrm{L}\mathrm{C}\mathrm{H}}_4}_{\mathrm{t}-1}+{\mathsf{\mu }}_{\mathrm{t}}\end{array}$$

(4)

$\mathsf{\mu }$ is the Error Correction Term (ECT) coefficient.

4. Analysis and Discussions

4.1. Summary Statistics

The data of this research is summarized in

Table 2. Descriptive statistics.

	LFOS	LFER	GDP	LPOP	LCH4
Mean	3.906086	1.636722	3.814211	1.551582	3.426426
Median	4.046379	1.816452	4.204831	1.572778	3.414183
Maximum	4.786241	2.942859	25.00724	1.827098	4.076396
Minimum	3.052113	−1.386294	−13.12788	1.267351	2.727735
Std. Dev.	0.595964	0.935246	6.223279	0.137961	0.378480
Skewness	−0.138144	−1.492625	0.148951	−0.104345	−0.000154
Kurtosis	1.518698	5.276859	5.277634	2.386623	2.080505
Jarque–Bera (JB)	5.014220	31.12825	11.65197	0.927021	1.867083
Probability	0.081503	0.000000	0.002950	0.629071	0.393159
Observations	53	53	53	53	53

Note: This table reports descriptive statistics for the variables over the period 1970–2022, including the mean, standard deviation, minimum, and maximum values. The statistics provide an overview of the scale, dispersion, and distributional properties of the variables prior to econometric analysis.

, where the variable with the largest mean is LFOS (3.906086), followed by GDP (3.814211), LCH₄ (3.426426), LFER (1.636722), and LPOP (1.551582). For skewness, LFOS (−0.138144), LFER (−1.492625), LPOP (−0.104345), and LCH₄ (−0.000154) are negatively skewed, while GDP (0.148951) is positively skewed. In addition, all variables except LFER are normally skewed because they have the value of zero. Relating to kurtosis, LFOS (1.518698), LPOP (2.386623), and LCH₄ (2.080505) are platykurtic because the values are less than 3, while LFER (5.276859) and GDP (5.277634) are leptokurtic because the values are more than 3. Furthermore, as reported by the probability values of the JB test, LFOS (0.081503), LPOP (0.629071), and LCH₄ (0.393159) show evidence of normal distribution because the probability values are greater than 5%, while LFER (0.000000) and GDP (0.002950) are not normally distributed. The summary of statistics is further presented in a graphical form in Figure 1. The graphs show that in recent times, LFOS and LCH₄ are upward sloping, while LFER, GDP, and LPOP are sloping downwards. It illustrates a steady upward trend in food production (LFOS) alongside methane emissions, while fertilizer use intensity (LFER) shows a relative decline in recent years, suggesting a widening gap between input availability and production needs.

4.2. Stationarity and Chow Breakpoint Tests

When determining the kinds of methods to use, it is crucial to investigate the stationarity of the data. This study employed the Augmented Dickey–Fuller (ADF) [66] and Phillips–Perron (PP) [67] tests, presented in

Table 3. Stationarity tests.

	ADF		PP
Variables	I(0)	I(1)	I(0)	I(1)	Outcome
LFOS	1.539976	−4.731075 *	−2.918587	−7.734490 *	I(1)
LFER	−3.612101 *	−8.284668 *	−3.829104 *	−8.288239 *	I(0)
GDP	−5.780762 *	−10.97036 *	−5.782298 *	−12.33414 *	I(0)
LPOP	−2.983223	−1.870828 ***	−3.012536	−7.409805 *	I(1)
LCH4	−2.934795	−8.632501 *	−2.934795	−8.926273 *	I(1)

This table reports Augmented Dickey–Fuller (ADF) and Phillips–Perron (PP) unit root test results. The null hypothesis is the presence of a unit root. * and *** denote rejection of the null at the 1% and 10% significance levels, respectively. I(0) and I(1) denote stationarity at levels and first differences.

. The outcome showed that LFOS, LPOP, and LCH₄ are stationary at I(1), while LFER and GDP are stationary at I(0). This result demonstrates a mixed order of integration. It is, however, crucial to note that these tests do not account for structural breaks, thus necessitating the use of the Chow breakpoint test, presented in

Table 4. Chow breakpoint test.

F-statistic	16.06073	Prob. F(5,43)	0.0000
Log likelihood ratio	55.83284	Prob. Chi-Square(5)	0.0000
Wald Statistic	80.30365	Prob. Chi-Square(5)	0.0000

This table reports the Chow breakpoint test for structural stability. The null hypothesis is that there is no structural break at the specified breakpoint. Rejection of the null indicates a significant structural change in the relationship among the variables.

. The test shows that there is a presence of structural break because all the p-values are significant, requiring the use of a dummy variable. This research chose the year 2008 as the structural break year because it was ascribed to the global financial crisis, which led to total economic turmoil, affecting various sectors of the economy, including the food sector.

4.3. ARDL Results

4.3.1. ARDL Bounds Test and ARDL-EC Test

In

Table 5. ARDL bounds test and ARDL-EC test.

F-Statistic	5.915445
Value	1%	5%	2.5%	10%	p-Value
I(0)	3.06	2.39	2.7	2.08	0.009
I(1)	4.15	3.38	3.73	3	0.051

This table reports the ARDL bounds test for cointegration. The null hypothesis is no long-run relationship among the variables. The reported F-statistic is compared to the critical bounds of Pesaran et al. (2001) [21]. Rejection of the null indicates the existence of a long-run equilibrium relationship.

, the ARDL bounds test results indicate that the computed F-statistic is higher than the critical values for 1%, 2.5%, 5%, and 10%, respectively. This indicates that the alternative hypothesis of a long-term connection will be accepted, while the null hypothesis of a long-term connection will be rejected. Given the significance of the results, the ARDL-EC p-values also demonstrate a long-term association. The ARDL bounds test uses EViews’ lag selection (EViews 12) (1, 2, 0, 0, 1, 0), but the ARDL-EC uses Stata’s lag selection (State 17) (1, 2, 0, 0, 1) (see Table 5). The long- and short-run results are presented in

Table 6. Long- and short-run ARDL results.

Long-Run Results
Variables	Coefficient	Std. Error	t-Statistic	Prob.
LFER	0.239882	0.189097	1.268567	0.2118
GDP	0.034180 **	0.015721	2.174149	0.0355
LPOP	−0.538668	0.463913	−1.161138	0.2523
LCH4	1.612437 *	0.341473	4.722012	0.0000
DUM	−0.411404 ***	0.209009	−1.968358	0.0558
C	−1.050955	1.463124	−0.718295	0.4766
Short-Run Results
Variables	Coefficient	Std. Error	t-Statistic	Prob.
D(LFER)	0.051952 *	0.014343	3.622175	0.0008
D(LFER(−1))	−0.057788 *	0.014731	−3.922925	0.0003
D(LCH4)	0.712413 *	0.131469	5.418861	0.0000
CointEq(−1) *	−0.117457	0.017048	−6.889691	0.0000
Wald Test
Variables	F-statistic	Prob.	Chi-square	Prob.
LFER	147.3704	0.0000	294.7409	0.0000
LCH4	16.07653	0.0003	16.07653	0.0001

Long-run ARDL: This table presents long-run coefficient estimates from the ARDL model. *, **, and *** denote statistical significance at the 1%, 5%, and 10% levels, respectively. Standard errors are reported in parentheses. Coefficients can be interpreted as long-run elasticities. Short run: This table reports short-run dynamic coefficients from the ARDL Error Correction Model (ECM). The Error Correction Term (ECT) captures the speed of adjustment toward long-run equilibrium following a shock. A negative and statistically significant ECT confirms model stability and convergence. Wald Test: This table reports Wald Test statistics for short-run and long-run causality. The null hypothesis is no causality. *, **, and *** denote rejection of the null at the 1%, 5%, and 10% levels, respectively.

.

4.3.2. Long- and Short-Run ARDL Results

The interpretation of the long-run ARDL results is as follows: A 1% increase in LFER spurs LFOS by 0.23%, although not significantly. Secondly, as GDP rises by 1%, LFOS increases by 0.03%, significant at a p-value of 5%. Thirdly, a 1% increase in LPOP reduces LFOS by 0.53%, although insignificantly. Lastly, as LCH₄ increases by 1%, LFOS rises by 1.61%, significant at the 1% p-value. It is important to state that LCH₄ has the most significant impact among all the variables. Furthermore, the dummy variable is significant at the 10% p-value.

In the short run, as LFER increases by 1%, LFOS increases by 0.05%, significant at the 1% p-value. However, it is observed that at the first lag, the association between LFER and LFOS is negative. This showed that a 1% rise in LFER decreases LFOS by 0.05%. LCH₄ is also confirmed to have a positive link with LFOS, just as in the long run. A 1% increase in LCH₄ increases LFOS by 0.71%, significant at the 1% p-value. The result in the short run also demonstrates that LCH₄ has the most significant impact on LFOS among all the variables.

In addition, at a speed of 11.7%, the term of adjustment (−0.117457) will be reversed to reach equilibrium in the long run. The R-Squared is 99.6%, while the Adjusted R-Squared is 99.5%. Also, the Durbin–Watson value is 1.996026. The F-statistic is 1238.127 (p = 0.000000).

To ascertain if the short-run results can be verified, the Wald Test is employed. The outcome of the test shows that LFER has a short-run causal impact on LFOS, and LCH₄ has a short-run causal impact on LFOS. The p-values of LFER and LCH₄ are significant at 1%.

4.3.3. Diagnostic Tests

Table 7. Diagnostic tests.

Residual Diagnostics	F-Stat	p-Value	Outcome
Normality test	0.244155	0.885080	Normally distributed
Serial Correlation LM test	0.060431	0.9414	No serial correlation
Heteroskedasticity Test	1.973876	0.1505	Homoskedastic
Ramsey Reset Test	2.109536	0.1349	Stable

This table reports diagnostic test statistics for serial correlation, heteroskedasticity, normality, and functional form. The null hypotheses correspond to no serial correlation, homoskedasticity, normal residuals, and correct functional form. Failure to reject the null indicates that the model is well specified.

summarizes the diagnostic and stability tests performed in this investigation. The experiments indicate that the model is structurally stable. All probability values are above the 5% significance level. The model has a normal distribution of 0.88%. In addition, Figure 2 displays the CUSUM and CUSUM of Squares graphs, both of which are inside the 5% critical bound, confirming the model’s stability in this investigation.

4.4. Spectral Granger Causality Test

In

Table 8. Spectral Granger Causality test.

Variable	Long Term	Medium Term	Short Term
	$\mathit{\omega }\mathit{i}=0.01$	$\mathit{\omega }\mathit{i}=1.00$	$\mathit{\omega }\mathit{i}=2.00$
LFER → LFOS	14.2028	4.8604	4.9976
	0.0008 **	0.0880 ***	0.0822 ***
GDP → LFOS	8.8879	5.2343	8.6500
	0.0117 **	0.0730 ***	0.0132 **
LPOP → LFOS	9.2110	1.1823	1.6695
	0.0100 **	0.5537	0.4340
LCH4 → LFOS	6.2885	6.2583	0.6122
	0.0431 **	0.438 **	0.7363

This table reports the Spectral Granger Causality test results, which examine whether causality between variables varies across different frequency bands. Low frequencies correspond to long-run (structural) causality, while high frequencies capture short-run (transitory) causality. Rejection of the null hypothesis at specific frequencies indicates frequency-dependent causal effects. ** and *** denote rejection of the null at the 5% and 10% levels, respectively.

, this test proposed by [23] showed that LFER Granger-causes LFOS in the long, medium, and short term. The medium and short term are significant at the p-value of 10%, while the long term is significant at 5%. Also, GDP Granger-causes LFOS in the short, medium, and long term. The long- and short-term impacts are significant at 5%, while that of the medium-term is significant at 10%. However, LPOP Granger-causes LFOS only in the long term, significant at 5%. Lastly, LCH₄ Granger-causes LFOS in the medium and long term, both significant at the p-value of 5%. The graph is further presented in Figure 3.

4.5. Discussions

Figure 3 presents the Spectral Granger Causality results, illustrating how the causal relationship between the explanatory variables and food security varies across different frequencies. In this framework, low frequencies correspond to long-run or structural relationships, whereas high frequencies capture short-run or transitory dynamics. When the estimated causality curve exceeds the 5% critical value line, the null hypothesis of no causality is rejected at the corresponding frequency band. This approach, therefore, allows us to distinguish whether a variable influences food security primarily in the short run, the long run, or both. For instance, fertilizer use and methane emissions display stronger causal effects at higher frequencies, indicating predominantly short-term influences, while GDP exhibits more pronounced causality at lower frequencies, suggesting a dominant long-run relationship with food security in Nigeria.

For the LFER and LFOS nexus, this research ascertained that LFER drives LFOS positively in the short run. However, in the first lag, this association is negative. The immediate short-term positive impact demonstrates that fertilizers enhance the productivity of crops, leading to increased food availability. Ref. [13] opined that fertilizers, both organic and inorganic, play a crucial role in global food security. Another reason for this result is that the Nigerian government supports the subsidization of fertilizers for its users, especially the farmers. This is evident in the subsidy scheme of 1976. Ref. [27] asserted that the supply and consumption of fertilizer are positively associated with agricultural output. In addition, the subsidy level of the inputs also contributes to input utilization, which is favorably connected to agricultural output value (agricultural GDP). Furthermore, the timeliness of the rain and conducive atmospheric conditions positively contribute to food security. This means that if fertilizers are used in the right way and the atmospheric conditions are suitable, output grows tremendously. Concerning food security in Africa, ref. [33] argued that significant enhancements in soil health are required to boost agricultural output and lower crop failure rates. Ref. [68] also had a similar assertion for the Ethiopian economy. Ref. [36] stated that global food security and agricultural productivity depend on fertilizers. Furthermore, agronomic factors can also enable LFER to spur LFOS. Agronomic mechanisms refer to the biological and physical processes that occur in the field to turn fertilizer into food. Since Nigeria’s soil structures are often depleted, fertilizers have a boosting impact on LFOS through various channels. First, they replenish essential nutrients like nitrogen and phosphorus, which act as a fuel for the plant to grow larger leaves and stronger roots. Second, they close the yield gap by allowing a single hectare of land to produce significantly more grain than it would naturally. Third, the use of fertilizer improves soil structure and water retention, making crops more resilient to Nigeria’s erratic rainfall.

However, it is essential to state that the excessive use of fertilizers can pollute the soil or lead to soil degradation. This impact is observed in the association between LFER and LFOS in the first lag, in the short run. Other factors responsible for this outcome include excessive dependency on fertilizers, poor quality of fertilizers, and delayed fertilizer distribution, which could sometimes be driven by fertilizer supply shocks. Ref. [69] claimed that enterprises pass input costs on to consumers when supply shocks occur, leading to an instantaneous increase in the prices of outputs in both the local and international markets. Although the fertilizer subsidy scheme in Nigeria has been impactful, ref. [27] stated that there are still challenges that need to be solved, such as rising import prices and the seeming incapacity of the federal and state budgets to cover these expenses, and the total absence of the initiative of the private sector in the acquisition and distribution of the input. Ref. [70] claimed that support programs and subsidies helped businesses expand and ensure that farmers could afford a sufficient supply of fertilizers. However, the artificially cheap fertilizer prices also contributed to a national trend of fertilizer abuse, which resulted in nutrient contamination. Further studies that support this assertion include [71,72,73].

In the long run, the impact is favorable, but insignificant. This shows that the impact of the policies implemented by the Nigerian government has not been fully effective in the long run. Some of the factors responsible for this include structural constraints, inequality in fertilizer access, inconsistencies in policies, and environmental impact. Structural challenges entail inadequate/poor rural infrastructure and inadequate access to markets. Regarding inequality in fertilizer access, the more connected or wealthier the farmers are, the easier it will be for them to have access to fertilizers. However, those who lack the connections will also lack access to the fertilizers they need, thus decreasing long-term aggregate gains. Inconsistencies in policies are sometimes driven by tribal or religious clashes and an unstable political climate or political instabilities, which are very prevalent within the Nigerian economy, especially in areas related to agriculture. These challenges lead to the suspension of policies in those farming areas, thus limiting the positive impact on food security. It is important to state that some of these policy failures can be attributed to corruption, and inadequate training on how the fertilizer will be used effectively. Furthermore, although this research used fertilizers used per area of cropland as a proxy for fertilizer use, some fertilizers are more dominant than others in the Nigerian context, such as nitrogen and phosphorus. This could also be one of the main causes of the insignificant relationship. For example, if a farmer keeps using the same type of fertilizer every year without variation, the soil eventually breaks down through several negative mechanisms. Also, fertilizer alone cannot boost yields without improved seeds, irrigation, and extension services, which are often missing in smallholder systems from developing countries; additionally, continuous synthetic fertilizer use without organic matter addition is known to degrade soil fertility, which could lead to an insignificant effect.

In the short and long run, the impact of LCH₄ on LFOS is positive, and has the highest impact on LFOS in this study. It is important to emphasize that methane emissions are not a productive input into agriculture, but rather a by-product of agricultural activities, particularly livestock production and rice cultivation. Methane (CH₄) emissions are used as a proxy for agricultural intensity due to agriculture being the dominant CH₄ source in Nigeria. Therefore, the positive long-run association observed between CH₄ emissions and food security should not be interpreted as methane improving food security per se. Instead, this relationship reflects scale effects in agricultural production, whereby increases in food output are accompanied by higher methane emissions. This interpretation aligns with the broader climate-food literature and avoids conceptual confusion regarding the role of greenhouse gases in food systems.

According to ref. [74], agricultural methane’s shorter atmospheric duration makes it less likely to damage crops over time. Ref. [12] found that LCH₄ improves LFOS, unlike other types of GHGs that negatively impact food security. In Nigeria, Ref. [75] further established that LCH₄ improves the productivity of agriculture. This assertion is contrary to existing studies, which found that LCH₄ harms food security. Ref. [76] stated that LCH₄ leads to agricultural productivity challenges. Ref. [62], however, found a positive but insignificant link between LCH₄ and LFOS.

In the long run, it shows that GDP increases LFOS. Firstly, an economy that grows sustainably assists the Nigerian government in generating more revenue, while at the same time attracting Foreign Direct Investment (FDI). This equips the government to invest robustly in the agricultural sector, in the areas of improved inputs (such as fertilizers and access to high-quality seeds), mechanization, irrigation infrastructure, research and development (R&D), and improved market and food access. Aside from the government having to invest in the agricultural sector, economic growth also has an impact on the purchasing power of individuals. If the purchasing power is high, then the demand for food will increase, leading to food availability in Nigeria, thus contributing to guaranteed food security. These studies affirmed that economic growth can be seen as a crucial factor in addressing food security issues [10,11,43]. On the other hand, some scholars disagree with this argument. Ref. [45] stated that economic expansion may make food insecurity worse, suggesting that if economic disparities continue, growth by itself cannot end food insecurity. Ref. [38] also argued that chronic food insecurity is not primarily due to economic growth or income distribution, but rather inflationary pressures, population growth, and inadequate food storage.

Lastly, food security is threatened because of the rising population in Nigeria. It is crucial to state that, in this research, the impact is not significant, which demonstrates the effectiveness of Nigerian government policies. It is, however, essential to examine the damages that a rising population can have on FOS. LPOP’s negative impact on LFOS is attributed to increased consumption of food resources, land competition, inequality, and ecological damage. Firstly, the demand for food may rise because of a rising population. Food supplies and infrastructure may be under stress as a result, particularly if agricultural productivity does not keep up with a growing population. It is important to note that the population in Nigeria is more than 200 million, and it is still growing [64]. Ref. [7] stated that even though the number of people who have escaped food insecurity has improved, Nigeria still has the largest number of people without food worldwide. Secondly, the demand for land for agriculture, as well as other uses like housing and development, rises in tandem with the population. Farmers may find it more challenging to operate and produce food if land prices rise. Third, a growing population may cause ecological damage through erosion of soil, contamination of water, and deforestation. Lower crop yields and damage to agricultural land are possible outcomes of this. Lastly, because the benefits of economic expansion might not be spread equally among the population, LPOP can worsen inequality. Some people will end up with enough food, while others will lack food. Furthermore, aside from the negative impact of a growing population, urban demographic pressure in Nigeria also creates a structural crisis that weakens the country’s farming capacity. First, it triggers a labor drain, where the youngest and most productive workers move to the cities, leaving an aging rural population to manage labor-intensive farms without the help of modern machinery. Second, a dietary shift towards urban staples like imported rice and wheat discourages local farmers from growing traditional crops because the national economy begins to prioritize foreign imports over fixing local supply chains. This research outcome agrees with the studies of refs. [10,15,55,77,78], and is in contrast with the studies of refs. [10,19].

5. Conclusions and Policy Recommendation

5.1. Conclusions

Food security remains a critical challenge in Nigeria. As a result, this research examines the impact of FER, GDP, POP, and CH₄ emissions on FOS in Nigeria from 1970 to 2022, using methodologies such as ARDL, Wald Test, and Spectral Granger Causality. The ARDL results demonstrate that in the long run, FER spurs FOS, although not significantly, while POP reduces FOS insignificantly. On the other hand, GDP and CH₄ emissions positively contribute to FOS. In the short run, FER and CH₄ emissions drive FOS. The Wald Test confirms the short-run findings. Furthermore, the Spectral Granger Causality test showed that LFER and GDP Granger-cause FOS in the long, medium, and short term. POP, however, Granger-causes FOS only in the long term, while CH₄ emissions Granger-cause FOS in the medium and long term.

5.2. Policy Recommendations

Although the empirical results show a positive association between methane emissions and food security, this should not be interpreted as methane emissions having a direct beneficial effect on food security. Rather, methane emissions in Nigeria largely reflect the scale and intensity of agricultural activity, particularly livestock production and rice cultivation, which simultaneously increase food availability and emissions. The positive coefficient therefore captures a scale effect of agricultural expansion rather than a productive role of methane itself. Based on the empirical finding that fertilizer use has a positive but statistically insignificant long-run effect on food security, and only a short-run positive effect, we recommend that fertilizer policy should focus on improving efficiency rather than merely increasing quantity. This study’s recommendations are as follows: (1) Since the use of fertilizers contributes to FOS, the Nigerian government must ensure that it continues to subsidize agricultural inputs so that the farmers can afford them, which directly contributes to FOS. However, the government should not stop at subsidizing agricultural inputs like fertilizers; farmers should also be trained adequately on how to use these inputs so as not to lead to soil pollution. One other important point the policymakers should note is that their policies will be stifled if a conducive political climate is not created. Policies should prioritize nutrient use efficiency and integrated soil health programs to mitigate the lagged negative effects of synthetic fertilizer. Riots, political clashes, and tribal wars should be avoided at all costs, especially in agriculturally abundant areas. Political instabilities hinder economic progress, which conclusively affects food security. (2) Given the strong and statistically significant positive long-run effect of economic growth on food security, we recommend policies that promote inclusive and agriculture-linked economic growth. In macroeconomics, one of the major goals of any government is for its economy to continue to grow. Growth allows any nation to generate enough revenue to meet its populace’s needs, such as investments in agriculture and infrastructure, agro-processing, market access, and value chains that translate macroeconomic growth into improved household food availability and access. It also demonstrates rising individual incomes. The implication of this is that as individuals’ incomes rise, they will be able to afford more food, increasing its demand, thus necessitating additional food production and distribution. Therefore, the policymakers and all key stakeholders should ensure that they continue to create sustainable measures that will drive economic expansion, and this can be achieved by creating a suitable environment that will attract FDI, making Nigeria an agricultural hub of the world by investing heavily in agriculture, and investing in domestic industries such that they will be able to compete with multinational corporations of the world. (3) Since urban population growth shows a negative and insignificant long-run association with food security, reflecting demand pressure and rural–urban labor shifts, we recommend policies that manage demographic transitions rather than attempt population control. There is nothing wrong with a rising population. However, a rising population should be commensurate with the food supply that will be able to cater to the population. Thomas Malthus stated in one of his theories that if population rises faster than the food supply, it could lead to economic problems such as poverty. Therefore, by encouraging more women to enter the workforce through gender equality, funding education and training, assisting smallholder farmers, and funding agricultural research and development, the Nigerian government should make policies that promote education, women’s empowerment, and youth employment in agri-food systems to manage demographic pressure while supporting food production. As a result, these policies are expected to enhance nutrition, boost agricultural investments, lessen the impact on the environment, promote food access, and ease the strain on resources. (4) Lastly, although this research found that methane emissions show a positive association with food security, this reflects increased agricultural activity rather than a beneficial role of emissions themselves. More policies relating to renewable energy development and adoption should be promoted; we recommend promoting climate-smart agricultural practices that reduce methane emissions per unit of output, such as improved livestock feeding practices, better manure management, and sustainable rice cultivation, so that food production can increase without worsening environmental externalities.

Each of these recommendations is directly derived from the empirical results of this study and is intended to translate the statistical findings into practical and sustainable policy guidance.

5.3. Limitations and Future Research Suggestions

Food security is a multidimensional concept encompassing availability, access, utilization, and stability. In this study, food security is proxied by a food availability indicator due to data availability and consistency over the sample period. However, alternative indicators such as average dietary energy intake or the prevalence of undernourishment could capture different dimensions of food security. While the core relationships identified in this study are expected to remain qualitatively similar, the magnitude and statistical significance of the coefficients may vary with the choice of proxy. Future research could extend this analysis by employing alternative food security measures to test the robustness of the results.

The main focal point of this research is Nigeria, which implies that results generated from this study cannot be generalized. It can serve as a policy guide for other African/Advanced economies that have similar economic conditions like Nigeria. Thus, other studies can carry out investigations based on a panel of countries or other individual African/Advanced economies where agriculture contributes significantly to their GDP. In addition, it is observed that other variables like human resource capability, R&D, and agricultural labor force can contribute significantly to food security. Thus, these variables should be taken into consideration while combining them with advanced econometric techniques.