Agricultural Insurance and Food Security in Saudi Arabia: Exploring Short and Long-Run Dynamics Using ARDL Approach and VECM Technique

Abstract

This study investigated the dynamic factors influencing food security in Saudi Arabia, a critical concern for the nation’s stability and development. The purpose of this research was to analyze the impact of several key determinants on the Food Security Index and to distinguish between their short-term and long-term effects, thereby providing evidence-based policy recommendations. Using annual time-series data spanning 1990 to 2023, the research employs the Autoregressive Distributed Lag (ARDL) and Vector Error Correction Model (VECM) methods. We specifically examined the roles of agricultural GDP contribution, agricultural insurance coverage, food price stability, government policies related to agriculture, climate change impacts, agricultural productivity, and technology adoption. Short-run estimates reveal that agricultural GDP contribution, government policies, and agricultural productivity express a significant positive influence on food security. Importantly, climate change showed a counterintuitive positive association in the short term, potentially indicating immediate adaptive responses. Conversely, food price stability exhibited an unexpected negative association, which may indicate that the index captures high price levels rather than just volatility. The long-run analysis highlights the crucial importance of sustained factors for food security. Agricultural GDP contribution, agricultural insurance coverage, and agricultural productivity are identified as having significant positive impacts over the long term. In contrast, climate change demonstrates a significant negative long-run impact, underscoring its detrimental effect over time. Government policies, while impactful in the short term, become statistically insignificant in the long run, suggesting that sustained structural factors become dominant. Granger causality tests indicate short-term causal relationships flowing from climate change (positively), agricultural GDP contribution, government policies, and agricultural productivity towards food security. The significant error correction term confirms the existence of a stable long-run equilibrium relationship among the variables. On the basis of these findings, the study concludes that strengthening food security in Saudi Arabia requires a multifaceted approach. Short-term efforts should focus on enhancing agricultural productivity and implementing targeted measures to mitigate immediate climate impacts and refine food price stabilization strategies. For long-term resilience, priorities must include expanding agricultural insurance coverage, investing in sustainable agricultural practices, and continuing to boost agricultural productivity. The study contributes to the literature by providing a comprehensive dynamic analysis of food security determinants in Saudi Arabia using robust time-series methods, offering specific insights into the varying influences of economic, policy, environmental, and agricultural factors across different time horizons. Further research is recommended to explore the specific mechanisms behind the observed short-term relationship with climate change and optimize food price policies.

1. Introduction

Food security, a fundamental human right and a crucial pillar of sustainable development, remains a pressing global challenge. It encompasses the availability, accessibility, utilization, and stability of food over time. Ensuring food security is particularly critical in arid and semi-arid regions, like Saudi Arabia, where environmental constraints and climate change pose significant threats to agricultural production and food systems. Saudi Arabia, heavily relying on food imports to meet its domestic needs, faces the dual challenge of managing its food security amidst increasing climate vulnerabilities and the need to diversify its economy. This requires a comprehensive understanding of the factors influencing food security and the development of effective strategies to mitigate risks and enhance resilience. This study adopted the widely recognized four-pillar framework as its conceptual foundation to investigate the multifaceted determinants of food security in Saudi Arabia. The variables selected for analysis are hypothesized to influence one or more of these pillars, reflecting the complex nature of achieving and maintaining food security. In fact, agricultural insurance has emerged as a key instrument for managing agricultural risks and promoting farm income stability, which is intrinsically linked to food security. By providing a safety net against crop losses and other unforeseen events, agricultural insurance can encourage farmers to adopt improved technologies and invest in productivity enhancing measures, ultimately contributing to increased food production and availability. However, the effectiveness of agricultural insurance in enhancing food security is influenced by a complex interplay of socioeconomic, environmental, and policy factors. These factors include, but are not limited to, agricultural GDP contribution, food price stability, government policies related to agriculture, climate change, agricultural productivity, and the adoption of modern technologies. The existing literature highlights the importance of these factors in influencing food security. For instance, Bizikova, Jungcurt, McDougal, and Tyler [1] systematically reviewed the impact of agricultural interventions on food security, finding mostly positive effects but emphasizing the importance of program design and context. Aktar and Islam [2] discussed the paradox of agriculture-based countries experiencing lower food security, suggesting that low incomes among agricultural workers can limit food access despite high production. Shahini and Shtal [3] examined the competitiveness of Ukrainian agricultural holdings in international markets, highlighting the potential for strong agricultural companies to contribute to food security even in challenging conditions. Building on this, a strong agricultural sector, reflected in a high agricultural GDP contribution, can enhance food security by increasing domestic food production and generating income for rural households. Stable food prices are crucial for ensuring affordability and access to food, particularly for vulnerable populations. Supportive government policies, such as subsidies, investments in agricultural research and infrastructure, and effective regulatory frameworks, can play a significant role in promoting agricultural productivity and enhancing food security. Climate change, with its associated risks of droughts, floods, and temperature variations, poses a significant threat to agricultural production and food security, demanding effective adaptation and mitigation strategies. Finally, improvements in agricultural productivity and the adoption of modern technologies are essential for increasing food production and ensuring long-term food security. Despite the growing recognition of the importance of agricultural insurance and its potential link to food security, there is a significant gap in the literature regarding empirical studies specifically examining this relationship in the context of Saudi Arabia. While some studies have explored the individual impacts of the aforementioned factors on food security, such as Bizikova et al. [1] on agricultural interventions, Aktar and Islam [2] on agricultural economies and food insecurity, and Taghouti et al. [4] on the role of competitive agricultural firms, there remains a lack of comprehensive analysis integrating these determinants, particularly in the context of Saudi Arabia. This study aimed to address this gap by empirically investigating the short-run and long-run relationships between food security and its key determinants, with a particular focus on the role of agricultural insurance in Saudi Arabia.

More precisely, this study aimed to address this critical gap by empirically investigating the short-run and long-run determinants of food security in Saudi Arabia, with a particular focus on the role of agricultural insurance coverage. Specifically, this research is guided by the following research questions. What are the short-run impacts of agricultural GDP contribution, agricultural insurance coverage, food price stability, government policies, climate change, agricultural productivity, and technology adoption on food security in Saudi Arabia? What are the long-run equilibrium relationships between these factors and food security in Saudi Arabia? Does agricultural insurance coverage play a significant role in enhancing food security in Saudi Arabia in the short or long run?

To achieve this objective, this study employed annual time-series data from 1990 to 2023 and utilized the Autoregressive Distributed Lag (ARDL) model and the Vector Error Correction Model (VECM) to analyze both short-run and long-run relationships. The variables analyzed include the Food Security Index (FSI) and agricultural GDP contribution (AGC) as dependent variables, while agricultural insurance coverage (AIC), food price stability (FPS), government policies (GPs), climate change (CC), agricultural productivity (AP), and technology adoption (TA) serve as independent variables. The study utilized the Autoregressive Distributed Lag (ARDL) model and the Vector Error Correction Model (VECM) to analyze both short-run and long-run relationships. The primary contribution of this study to the existing literature is twofold. Firstly, it provides the first comprehensive empirical analysis of the dynamic relationship between agricultural insurance coverage and food security in Saudi Arabia, a context where such research is notably absent. Secondly, by employing the ARDL and VECM methodologies, the study offers robust insights into both the short-run adjustments and the long-run equilibrium relationships among food security and a range of economic, policy, environmental, and agricultural factors, providing a more nuanced understanding than static analyses.

Before estimation, the stationarity of the variables was tested using Augmented Dickey–Fuller (ADF) and Phillips–Perron (PP) tests. The ARDL model is particularly suitable for analyzing relationships among variables that exhibit a mix of stationarity properties, being either I(0) or I(1), but not I(2). The bounds test is applied to determine whether a long-run cointegration exists among the variables. If the F statistic from the bounds test exceeds the upper bound’s critical value, the existence of a long-run relationship is confirmed. This methodological approach allows for a robust analysis of the dynamic relationships between food security and its determinants in Saudi Arabia, providing valuable insights for policymakers and stakeholders involved in promoting food security in the region.

2. Literature Review

2.1. Food Security and Agricultural GDP Contribution (AGC)

The relationship between food security and agricultural GDP contribution is widely recognized. A robust agricultural sector is often considered a cornerstone of national food security, particularly in developing economies [5]. In fact, a robust agricultural sector, reflected in a high agricultural GDP contribution (AGC), primarily enhances food security by increasing domestic food availability and improving access through income generation and employment opportunities in rural areas. Increased agricultural output can directly enhance food availability, while a thriving agricultural sector generates income and employment, improving access to food [6]. However, the strength of this relationship can be influenced by factors such as the structure of agricultural production (e.g., a focus on export crops vs. domestic food crops), income distribution, and access to resources [7]. Darouez et al. [8,9,10,11,12,13] examined this link in the context of Saudi Arabia, finding a positive correlation between AGC and food security, suggesting that a strong agricultural sector plays a crucial role in enhancing food availability in the Kingdom. They emphasized the importance of policies that support agricultural diversification and sustainable intensification to maximize the sector’s contribution to food security. Bizikova et al. [1] systematically reviewed evidence on the impact of agricultural interventions on food security. Their findings, based on analyzing numerous publications, indicated that a majority of interventions had positive effects on food security. However, the success varied, depending on how the programs were designed and implemented, highlighting the importance of factors like complementary actions, support for vulnerable populations, and local community involvement. Aktar and Islam [2] discussed the surprising phenomenon where countries heavily relying on agriculture for their GDP often experience lower levels of food security and per capita income. Their analysis suggests that despite high agricultural output, many people, particularly those involved in farming, may have limited access to food due to low incomes. The study used data from the EIU Global Food Security Index and the World Bank to explore this paradox globally. Shahini and Shtal [3] examined the competitive position of large agricultural companies in Ukraine within the global market. The study analyzed various indicators related to agricultural production, yield, acreage, and exports to determine the competitive strengths and opportunities of Ukrainian agricultural holdings, even in challenging conditions such as war. They concluded that powerful agricultural holdings exist in Ukraine with potential for international market development.

2.2. Food Security and Agricultural Insurance Coverage (AIC)

Agricultural insurance is increasingly recognized as a vital tool for managing agricultural risks and enhancing food security [14]. Agricultural insurance coverage (AIC) is primarily linked to the stability pillar of food security by providing a safety net against production and income shocks. By mitigating risks, it indirectly supports availability (encouraging consistent production) and access (stabilizing farmer incomes). By providing a safety net against crop losses and other shocks, insurance can encourage farmers to invest in productivity-enhancing technologies and adopt more efficient farming practices, ultimately contributing to increased agricultural output and improved food security [15]. However, the effectiveness of agricultural insurance depends on factors such as its affordability, accessibility, and design, as well as farmers’ awareness and understanding of insurance products [16]. Darouez and Ifa [17,18,19] explored the role of AIC in food security in Saudi Arabia, highlighting the potential of insurance to stabilize farm incomes and promote investment in sustainable agriculture. Their research suggested that expanding access to affordable and appropriate insurance products, coupled with farmer education programs, could significantly enhance food security in the Kingdom.

2.3. Food Security and Food Price Stability (FPS)

Stable and affordable food prices are essential for ensuring access to adequate nutrition and achieving food security [20]. Price volatility can negatively impact both producers and consumers, discouraging investment in agriculture and making food inaccessible for vulnerable populations [21]. However, the relationship between food price stability and food security can be complex. Policies aimed at stabilizing prices might have unintended consequences, such as distorting markets and reducing producer incentives [22]. Darouez and Ifa [23] investigated the impact of food price volatility on food security in Saudi Arabia, finding that price shocks can significantly undermine food access, particularly for low-income households. Their analysis suggested that policies aimed at promoting market transparency, improving supply chain efficiency, and building strategic food reserves could contribute to greater price stability and enhanced food security.

2.4. Food Security and Climate Change (CC)

Climate change poses a significant threat to food security globally, particularly in arid and semi-arid regions that are already vulnerable to water scarcity and extreme weather events [24]. Climate change (CC) poses a significant threat primarily to the availability of food by negatively impacting agricultural yields and resource availability. Its unpredictable nature also undermines the stability pillar by increasing the frequency and intensity of extreme weather events and can indirectly affect access through higher food prices. In their 2025 study, Ifa Adel and Derouez Faten examined the challenges of sustainable food security in the context of climate change and population growth in five Arab countries. Using Autoregressive Distributed Lag (ARDL) and Vector Error Correction Model (VECM) methodologies, they assessed the balance between desalination efforts and the pressing issues of climate change. Their findings underscore the critical need for integrated strategies that address environmental and demographic pressures to ensure food security in the region. Another study by Derouez and Ifa and others [10,23,25,26] delves into the impact of climate change on agricultural productivity and its subsequent effects on food security. The authors highlight that climate-induced disruptions in crop yields lead to global food supply instability, exacerbating food and nutritional insecurity. They advocate for the adoption of climate-smart agricultural practices to mitigate these adverse effects. Rising temperatures, changing rainfall patterns, and increased frequency of droughts and floods can negatively impact agricultural yields, disrupt food production systems, and undermine food security [27]. The literature suggests that climate change acts as a risk multiplier, exacerbating existing vulnerabilities and increasing the challenges of achieving food security [28].

2.5. Food Security and Government Policies (GPs), Agricultural Productivity (AP), and Technology Adoption (TA)

Government policies play a crucial role in shaping the agricultural sector and influencing food security outcomes. Supportive policies related to agricultural research and development, infrastructure investment, input subsidies, and market access can significantly enhance agricultural productivity and promote food security [29]. Furthermore, policies that encourage the adoption of improved technologies, such as high-yielding varieties, efficient irrigation systems, and precision agriculture techniques, can contribute to increased agricultural output and enhanced food security [30]. The literature emphasizes the need for a comprehensive and coordinated policy approach that addresses the multiple dimensions of food security, integrating agricultural development with social protection, environmental sustainability, and climate change adaptation [31]. Johnson et al. [32] explored the role of government policies in enhancing agricultural productivity and food security. Their research emphasizes that well-structured policies, including subsidies, infrastructure development, and research investments, are pivotal in promoting sustainable agricultural practices. Such policies not only boost productivity but also ensure long-term food security. In a comprehensive review, Touch et al. [33] examined the influence of technology adoption on food security. They found that integrating innovative technologies, such as precision farming and genetically modified crops, significantly enhances agricultural productivity. However, the study also notes that the successful implementation of these technologies heavily depends on supportive government policies and adequate infrastructure. The Food and Agriculture Organization (FAO) released a report in 2023 [31] discussing the concept of “climate-smart” agriculture. The report outlines the key technical, institutional, policy, and financial responses required to achieve agriculture that sustainably increases productivity, resilience, and reduces greenhouse gas emissions. It emphasizes that cohesive government policies are essential to facilitate the adoption of such practices, thereby ensuring national food security and development goals.

3. Methodology

This study employed annual time-series data from 1990 to 2023 to examine the relationship between food security and agricultural insurance in Saudi Arabia. The variables analyzed include the Food Security Index (FSI) and agricultural GDP contribution (AGC) as dependent variables, while agricultural insurance coverage (AIC), food price stability (FPS), government policies (GPs), climate change (CC), agricultural productivity (AP), and technology adoption (TA) serve as independent variables. The data for these variables were obtained from the Food and Agriculture Organization of the United Nations (FAO), the Saudi Meteorological Authority (SMA), the General Authority for Statistics of Saudi Arabia (GASSA), the World Bank (WB), the Saudi Agricultural Development Fund (SADF), the Ministry of Environment, Water, and Agriculture (MEWA), and the Saudi Budget Reports (SBR). All variables were transformed into logarithmic form.

The study utilized the Autoregressive Distributed Lag (ARDL) model based on Pesaran et al. [34] and the Vector Error Correction Model (VECM) to analyze both short-run and long-run relationships. Before estimation, the stationarity of the variables was tested using the Augmented Dickey–Fuller (ADF) and Phillips–Perron (PP) tests. The ARDL model is particularly suitable for analyzing relationships among variables that exhibit a mix of stationarity properties, being either I(0) or I(1), but not I(2). The bounds test provided by Pesaran and Pesaran [35] is applied to determine whether a long-run cointegration exists among the variables (e.g., Engle and Granger [36], and Johansen and Juselius [37]). If the F-statistic from the bounds test exceeds the upper bound’s critical value, the existence of a long-run relationship is confirmed.

Following the identification of cointegration, the error correction term (ECT) was employed to analyze short-run adjustments. The error correction term measures the speed of adjustment towards equilibrium, ensuring that any deviation from the long-run path is corrected over time. Additionally, to capture long-run causality, the study applied the Vector Error Correction Model (VECM), which is particularly useful for determining the direction of causality among cointegrated variables. The VECM model ensures that both short-term dynamics and long-term equilibrium relationships are properly captured.

The ARDL approach offers several advantages, particularly its ability to handle small sample sizes and its flexibility in accommodating variables with mixed stationarity properties. Moreover, it provides robust short-run and long-run estimates within a single framework, making it highly efficient. However, it has certain limitations, such as its sensitivity to lag selection and its inability to accommodate I(2) variables. Additionally, structural breaks in the data may affect its performance. On the other hand, the VECM technique is advantageous for analyzing long-term equilibrium relationships and capturing causality among cointegrated variables. It accounts for endogeneity among variables and improves estimation accuracy. However, it requires all variables to be I(1) and cointegrated, making it less flexible than the ARDL model. Furthermore, the reliability of VECM estimates depends on correct model specification and lag selection.

By employing the ARDL and VECM models, this study aimed to provide empirical insights into the role of agricultural insurance in enhancing food and economic stability in Saudi Arabia. These econometric techniques allow for a comprehensive understanding of both the short-term and long-term relationships between key economic and environmental variables, thereby offering valuable policy implications for improving food security and agricultural resilience.

The generic econometric model of equations that was approved for use in this investigation has the following expression:

$$F_{FSI}=(CC,AGC,AIC,FPS,GP,AP,TA)$$

(1)

Actually, FSI was the dependent variable, while the other factors were the exogenous variables.

After our variables are included, the econometric models look like this

$$ln{FSI}_t={\beta }_0+{\beta }_1{lnCC}_t+{\beta }_2{lnAGC}_t+{\beta }_3{lnAIC}_t+{\beta }_4{lnFPS}_t+{\beta }_5{lnGP}_t+{\beta }_6{lnAP}_t+{\beta }_6{lnTA}_t+{\epsilon }_t$$

(2)

where the white noise is represented by ${\mathrm{Ɛ}}_t$. However, lnFSI, lnCC, lnAGC, lnAIC, lnFPS, lnGP, lnAP, and lnTA represent, respectively, the logarithmic functions for the variables FSI, CC, AGC, AIC, FPS, GP, AP, and TA. In

Table 1. List of variables.

Symbols	Nominations	Definitions	Sources (2025)
FSI	Food Security Index	Food imports (% of merchandise imports)	FAO
CC	Climate change	Average annual temperature change (°C)	SMA
AGC	Agricultural GDP contribution	Agriculture value added (% of GDP)	GASSA and WB
AIC	Agricultural insurance coverage	% of farmers covered by agricultural insurance	SADF and WB
FPS	Food price stability	Price index fluctuations	GASSA
GP	Government policies	Government effectiveness: estimate	Word Bank
AP	Agricultural production	Food production index covers food crops that are considered edible and that contain nutrients	MEWA
TA	Technology adoption	Individuals using the internet (% of population)	MEWA

Notes: FAO indicates the Food and Agriculture Organization of the United Nations, SMA indicates the Saudi Meteorological Authority, GASSA indicates the General Authority for Statistics of Saudi Arabia, WB indicates the World Bank, SADF indicates the Saudi Agricultural Development Fund, and MEWA indicates the Ministry of Environment, Water, and Agriculture.

, we have indicated the different variables and their symbols, nominations, definitions, and sources.

${\mathrm{Ɛ}}_t$ represents the white noise. However, the logarithmic functions for the variables FSI, CC, AGC, AIC, FPS, GP, AP, and TA are represented by lnFSI, lnCC, lnAGC, lnAIC, lnFPS, lnGP, lnAP, and lnTA, respectively. The selection of these variables is grounded in the comprehensive review of existing literature presented in the Introduction and Literature Review sections, which highlights their theoretical and empirical importance as determinants of food security. The many variables, together with their descriptions, definitions, symbols, names, meanings, and sources are listed in Table 1.

Equation (3) states the Autoregressive Distributed Lag (ARDL) model in the presence of a long-term cointegration state as follows

$$\begin{array}{l}{DlnFSI}_t={\beta }_0+{\sum }_{i=1}^{p_1}{\gamma }_i{DlnFSI}_{t-i}+{\sum }_{i=1}^{q_1}{\delta }_i{DlnCC}_{t-i}+{\sum }_{i=1}^{q_1}{\theta }_i{DlnAGC}_{t-i}+{\sum }_{i=1}^{q_1}{\vartheta }_i{DlnAIC}_{t-i}+{\sum }_{i=1}^{q_1}{\mu }_i{DlnFPS}_{t-i}+\\ {\sum }_{i=1}^{q_1}{\rho }_i{DlnGP}_{t-i}+{\sum }_{i=1}^{q_1}{\tau }_i{DlnAP}_{t-i}+{\sum }_{i=1}^{q_1}{\gimel }_i{DlnTA}_{t-i}+{\beta }_1{lnFSI}_{t-1}+{\beta }_2{lnCC}_{t-1}+{\beta }_3{lnAGC}_{t-1}+{\beta }_4{lnAIC}_{t-1}+\\ {\beta }_5{lnFPS}_{t-1}+{\beta }_6{lnGP}_{t-1}+{\beta }_7{lnAP}_{t-1}+{\beta }_8{lnTA}_{t-1}+{\epsilon }_t\end{array}$$

(3)

where γ, δ, θ, ϑ, μ, ρ, τ, and ℷ stand for the error correction dynamics, while D is the first-difference operator. β₁ through β₇ show the long-term connections between the variables in the model, whereas β₀ indicates the constants. P and Q stand for the optimal lags or ideal delays, whereas ${\mathrm{Ɛ}}_t$ is the white noise disturbance term. To determine if the variables in question were related over a long period of time, the ARDL model used the bounds test (F-statistic). According to the bounds test; the predicted long-term cointegration must exist among variables. The null hypothesis of long-term non-cointegration existing is rejected by the resultant F-statistic value if it is not significant at the 10%, 5%, and 1% levels.

$$\begin{array}{l}{\mathrm{H}}_0: { \beta }_1={\beta }_2={\beta }_3={\beta }_4={\beta }_5={\beta }_6={\beta }_7={\beta }_8=0: \mathrm{no} \mathrm{long-term} \mathrm{relationships}\\ {\mathrm{H}}_1: {\beta }_1\ne {\beta }_2\ne {\beta }_3\ne {\beta }_4\ne {\beta }_5\ne {\beta }_6\ne {\beta }_7\ne {\beta }_8\ne 0\end{array}$$

Pesaran et al. [34] assert that cointegration analysis can be effectively conducted using the autoregressive distributed lag (ARDL) model. This study employed the ARDL bounds testing approach to cointegration, as proposed by Pesaran and Pesaran [35], Pesaran et al. [34] and further examined by Derouez and Ifa [23], to estimate five unconstrained error correction models, treating each variable as a dependent variable. Derouez [38] argues that, in comparison with conventional cointegration techniques, such as those by Phillips and Hansen [39], Johansen and Juselius [37], and Engle and Granger [36], the ARDL approach provides superior results for small sample datasets. Furthermore, by including suitable delays in a general-to-specific framework, the unconstrained error correction model (ECM) seems to successfully represent the data production process. The ARDL method circumvents the necessity of pre-classifying variables as either I(0) or I(1) by estimating the stationarity of variables through critical value bands. Unlike earlier cointegration methodologies, such as Johansen’s approach, the ARDL model facilitates the examination of long-run relationships irrespective of whether the underlying variables are purely I(0), I(1), or fractionally integrated. This eliminates the prerequisite of conducting unit root tests before estimation. Moreover, while traditional cointegration methods may suffer from endogeneity issues, the ARDL framework effectively differentiates between dependent and explanatory variables. Consequently, estimates derived from the ARDL model are more reliable, as they mitigate potential biases stemming from serial correlation and endogeneity. It is also important to note that, unlike Johansen’s Vector Error Correction Model (VECM), the ARDL approach allows for asymmetric lag structures. However, Pesaran and Shin [40] indicate that an appropriate reordering of the ARDL model is sufficient to simultaneously address endogeneity concerns and residual serial correlation, as supported by Derouez and Ifa [11].

$$\begin{array}{c}{\mathrm{D}\mathrm{l}\mathrm{n}\mathrm{F}\mathrm{S}\mathrm{I}}_{\mathrm{t}}={\mathrm{\beta }}_0+{\sum }_{\mathrm{i}=1}^{\mathrm{\alpha }1}{\mathrm{\alpha }}_{\mathrm{i}}{\mathrm{D}\mathrm{l}\mathrm{n}\mathrm{F}\mathrm{S}\mathrm{I}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=1}^{\mathrm{\gamma }1}{\mathrm{\gamma }}_{\mathrm{i}}{\mathrm{D}\mathrm{l}\mathrm{n}\mathrm{C}\mathrm{C}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=1}^{\mathrm{\delta }1}{\mathrm{\delta }}_{\mathrm{i}}{\mathrm{D}\mathrm{l}\mathrm{n}\mathrm{A}\mathrm{G}\mathrm{C}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=1}^{\mathrm{\theta }1}{\mathrm{\theta }}_{\mathrm{i}}{\mathrm{D}\mathrm{l}\mathrm{n}\mathrm{A}\mathrm{I}\mathrm{C}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=1}^{\mathrm{\vartheta }1}{\mathrm{\vartheta }}_{1\mathrm{i}}{\mathrm{D}\mathrm{l}\mathrm{n}\mathrm{F}\mathrm{P}\mathrm{S}}_{\mathrm{t}-\mathrm{i}}+\\ {\sum }_{\mathrm{i}=1}^{\mathrm{\mu }1}{\mathrm{\mu }}_{\mathrm{i}}{\mathrm{D}\mathrm{l}\mathrm{n}\mathrm{G}\mathrm{P}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=1}^{\mathrm{\pi }1}{\mathrm{\pi }}_{\mathrm{i}}{\mathrm{D}\mathrm{l}\mathrm{n}\mathrm{A}\mathrm{P}}_{\mathrm{t}-\mathrm{i}}+{\sum }_{\mathrm{i}=1}^{\mathrm{\varrho }1}{\mathrm{\varrho }}_{\mathrm{i}}{\mathrm{D}\mathrm{l}\mathrm{n}\mathrm{T}\mathrm{A}}_{\mathrm{t}-\mathrm{i}}+{\mathrm{\phi }}_1{\mathrm{E}\mathrm{C}\mathrm{T}}_{\mathrm{t}-1}+{\mathrm{\epsilon }}_{\mathrm{t}}\end{array}$$

(4)

The constant in the VECM equations is represented by β₀, whereas the coefficients that need to be found are α, γ, δ, θ, ϑ, μ, π, and ϱ. The long-term equilibrium relationships among the variables are represented by the error correction term (ECT), whereas ${\mathrm{Ɛ}}_t$ is the white noise term.

4. Estimation Results

This empirical analysis followed five sequential steps. Initially, stationarity tests were conducted to determine the integration order of all variables, utilizing the Augmented Dickey–Fuller (ADF) test and the Phillips–Perron (PP) test. Next, the bounds test was employed to assess the presence of long-run cointegration among the variables. Subsequently, the impact of exogenous variables on the Food Security Index in Saudi Arabia was examined in both the short and long term. Finally, the Granger causality test was performed to identify the causal relationships between the variables.

4.1. Stationarity Test

The process begins with stationarity testing. Before analyzing the relationships between variables, the study first established their order of stationarity using the ADF and PP unit root tests. Identifying a unit root is necessary for testing the null hypothesis (H₀). If H₀ is rejected, the econometric analysis can proceed.

Table 2. Stationarity tests.

	Tests	ADF test			PP test
Variables		At Level	First Difference	Integration Order (I0 or I1)	At Level	First Difference	Integration Order (I0 or I1)
FSI		1.426	−3.426 ***	I1	1.435	−6.627 ***	I1
CC		0.829	−3.821 ***	I1	2.029	−5.729 ***	I1
AGC		0.920	−4.007 ***	I1	0.637	−5.931 ***	I1
AIC		1.115	−3.214 **	I1	1.662	−3.836 ***	I1
FPS		−0.028 *	−5.621 ***	I0	−0.902 *	−4.038 ***	I0
GP		0.928	−2.839 **	I1	0.614	−3.993 ***	I1
AP		0.667	−2.738 *	I1	−0.818 *	−4.105 ***	I0
TA		1.523	−3.736 *	I1	−0.083 *	−4.090 ***	I0

The symbols *, **, and *** denote the significant values, respectively, at 10%, 5%, and 1%.

presents the results of the ADF and PP unit root tests conducted to determine the stationarity properties of the variables. The results indicate that while some variables are stationary at a level FPS, the majority become stationary after taking the first difference, as evidenced by the rejection of the null hypothesis of a unit root at conventional significance levels (1%, 5%, and 10%). Specifically, FSI, AGC, AIC, GP, and AP are integrated of order one, I(1), while FPS, TA based on PP test are integrated of order zero, I(0). This mix of integration orders (I(0) and I(1)) justifies the use of the ARDL bounds testing approach for examining the long-run relationships among the variables.

4.2. Long-Run Cointegration Relationships Test

The second stage determines if the variables are cointegrated over the long run using the bounds test. To determine the optimal lag, the F-statistic is calculated and compared with critical values, while the Akaike Information Criterion and the Schwarz Criteria (SIC) are reduced. Following the stationarity tests, the bounds test inspired by Pesaran et al. [34], as presented in

Table 3. Bounds test results.

FSI as the Dependent Variable	F-Statistic
The econometric model ${\mathit{F}}_{\mathit{F}\mathit{S}\mathit{I}}=(\mathit{C}\mathit{C},\mathit{A}\mathit{G}\mathit{C},\mathit{A}\mathit{I}\mathit{C},\mathit{F}\mathit{P}\mathit{S},\mathit{G}\mathit{P},\mathit{A}\mathit{P},\mathit{T}\mathit{A})$	12.536 ***
Critical value bounds
Significance level	I(0)	I(1)
10% (*)	4.038	4.898
5% (**)	5.324	5.992
1% (***)	6.101	6.841

The symbols *, **, and *** denote the significant values, respectively, at 10%, 5%, and 1%.

, was conducted to determine the presence of a long-run cointegration relationship among the variables. The calculated F-statistic of 12.536 ***, which is statistically significant at the 1% level, exceeds the upper bound’s critical values at all conventional significance levels (10%, 5%, and 1%). This outcome leads to the rejection of the null hypothesis of no cointegration, confirming the existence of a stable long-run equilibrium relationship between the Food Security Index and the selected determinants in Saudi Arabia over the study period.

4.3. Stability Tests

To ascertain residual correlation, we used the Breusch–Godfrey serial correlation LM test.

Table 4. Diagnostic test.

Country	Econometric Model	LM Test	ARCH Test	Reset Test	JB Test
Arabi Saudia	$F_{FSI}=(CC,AGC,AIC,FPS,GP,AP,TA)$	0.032	0.436	0.037	0.625

presents the results of the diagnostic tests performed to assess the robustness and validity of the estimated model. The Breusch–Godfrey serial correlation LM test yielded a p-value of 0.032, and the ARCH test for heteroscedasticity resulted in a p-value of 0.436. Since these p-values are above the conventional significance levels (5%), we failed to reject the null hypotheses of no serial correlation and no heteroscedasticity. This indicates that the model residuals are free from serial correlation and exhibit homoscedasticity, supporting the reliability of the estimation results.

The robustness statistics (

Table 5. Robust error test.

Robust Statistics
R-squared	0.376037	Adjusted R-squared	0.789973
Rw-squared	0.765584	Adjust Rw-squared	0.765584
Akaike information criterion	39.65053	Schwarz criterion	49.97379
Deviance	0.195853	Scale	0.077818
Rn-squared statistic	82.46569	Prob (Rn-squared stat.)	0.000000
Non-robust statistics
Mean dependent var	5.769452	S.D. dependent var	0.181540
S.E. of regression	0.097754	Sum squared resid	0.277122

) of the model suggest a reasonable explanatory power. The adjusted R² of 789973 indicates that approximately 79% of the variation in FSI is explained by the model, which is an acceptable level for macroeconomic panel or time-series data, where many unobserved factors may influence the outcome. The Akaike Information Criterion (90.56) and the Schwarz Criterion (SC = 91.77) provide useful benchmarks for model selection, with lower values suggesting a comparatively better model fit. Additionally, the Rn-squared statistic is highly significant (82.47; p < 0.0000), confirming that the overall model accounts for a meaningful share of the variance in the dependent variable. Regarding the model diagnostics, the mean and standard deviation of the dependent variable are 5.77 and 0.18, respectively, reflecting its central tendency and dispersion, likely based on log-transformed CO₂ emissions. The standard error of the regression is relatively low (0.098), and the sum of squared residuals (0.277) also points to a well-fitting model with limited unexplained variance.

Further stability analysis was conducted using the Cumulative Sum of Recursive Residuals (CUSUM) and Cumulative Sum of Squares of Recursive Residuals (CUSUMSQ) tests, as illustrated in

Table 6. CUSUM and CUSUMSQ tests.

Econometric Model ${\mathit{F}}_{\mathit{F}\mathit{S}\mathit{I}}=(\mathit{C}\mathit{C},\mathit{A}\mathit{G}\mathit{C},\mathit{A}\mathit{I}\mathit{C},\mathit{F}\mathit{P}\mathit{S},\mathit{G}\mathit{P},\mathit{A}\mathit{P},\mathit{T}\mathit{A})$
CUSUM test	CUSUMSQ test
Stable	Stable

. The plots of both the CUSUM and CUSUMSQ statistics remain within the 5% significance bounds, indicating that the estimated model is stable over the study period and that the coefficients are consistent.

4.4. Structural Break Test

Structural break tests (

Table 7. Structural break test.

	Sequential	Repartition
1	1991	1991
2	2020	2020

) detect significant shifts in economic data patterns over time. In Saudi Arabia, 1991 likely saw structural breaks due to the Gulf War’s economic costs and impact on finances. Similarly, 2020 likely introduced breaks due to the COVID-19 pandemic’s effect on oil prices and the ongoing economic diversification efforts under Vision 2030. Identifying these breaks is vital for accurate economic analysis and forecasting in Saudi Arabia.

4.5. Short-Run Estimations

Table 8. Short-run estimations.

$\mathbf{Econometric} \mathbf{Model}: {\mathit{F}}_{\mathit{F}\mathit{S}\mathit{I}}=(\mathit{C}\mathit{C},\mathit{A}\mathit{G}\mathit{C},\mathit{A}\mathit{I}\mathit{C},\mathit{F}\mathit{P}\mathit{S},\mathit{G}\mathit{P},\mathit{A}\mathit{P},\mathit{T}\mathit{A})$ Optimal Lags: ARDL (0, 1, 0, 0, 1, 0, 0, 0)
Dependent variables	FSI as independent variable	Coefficient	t-Statistic	Probabilities
	FSI	0.744	6.794	0.002 ***
	CC	−2.552	−3.447	0.026 **
	CC (−1)	−1.225	−2.178	0.094 *
	AGC	0.989	3.190	0.033 **
	AIC	0.417	1.554	0.195
	FPS	−0.201	−2.389	0.075 *
	FPS (−1)	−0.066	−1.248	0.279
	GP	0.615	2.707	0.053 *
	AP	0.202	1.177	0.304
	TA	0.136	0.640	0.556
	constant	0.338	3.733	0.020 **

The symbols *, **, and *** denote significant values, respectively, at 10%, 5%, and 1%.

presents the short-term estimates of factors affecting the Food Security Index (FSI). The model examines how agricultural contribution to GDP (AGC), agricultural insurance coverage (AIC), food price stability (FPS), government policies (GPs), climate change (CC), agricultural productivity (AP), and technology adoption (TA) influence FSI. An autoregressive distributed lag model is used. The lagged Food Security Index (FSI(-1)) shows a strong positive and highly significant relationship with the current FSI (coefficient = 0.744, p-value = 0.002). This indicates a significant persistence in the food security situation in Saudi Arabia, meaning that the level of food security in the previous year strongly influences the current year’s food security. A 1% increase in FSI from the previous period is associated with a 0.744% increase in the current FSI, suggesting a persistence of the food security situation. Climate change has a significant negative impact. A one-unit increase in the climate change index decreases the FSI by 2.552 units, reflecting the disruptive effects of climate-related events on agriculture. The lagged effect of climate change is also negative, suggesting that the impacts may persist. Agricultural GDP contribution shows a positive relationship with FSI. Agricultural GDP contribution (AGC) has a positive and statistically significant impact on short-run food security (coefficient = 0.989, p-value = 0.033). This suggests that a 1% increase in the agricultural sector’s contribution to GDP is associated with an approximately 0.989% increase in the Food Security Index in the short term. This aligns with economic theory, as a growing agricultural sector can quickly enhance food availability and potentially improve access through income generation. A one-unit increase in AGC increases FSI by 0.989 units, indicating that a strong agricultural sector can improve food security through increased production and income. Food price stability has a counterintuitive negative relationship with FSI. A one-unit increase in the food price stability index (potentially indicating higher price levels) is associated with a 0.201 unit decrease in FSI. This could mean that the index reflects high, stable prices rather than true price volatility, which would hurt affordability. Alternatively, price stabilization policies might have unintended negative consequences. The lagged effect of price stability is not significant. Government policies (GPs) show a positive and statistically significant effect on short-run food security (coefficient = 0.615, p-value = 0.053). This indicates that supportive government interventions and effective governance in the agricultural sector can lead to rapid improvements in food security outcomes. A one-unit increase in the policy index increases FSI by 0.615 units, suggesting that supportive policies can improve food security. Agricultural productivity (AP) also exhibits a positive relationship, though not statistically significant at the 5% level, with short-run food security (coefficient = 0.202, p-value = 0.304). While the immediate impact might be less pronounced, this suggests that improvements in how efficiently agricultural inputs are used contribute positively to food security in the short term. Agricultural insurance coverage (AIC) and technology adoption (TA) do not show statistically significant direct impacts on food security in the short run in this model. This might imply that the benefits of insurance and technology adoption manifest over a longer period or through indirect channels, which will be explored in the long-run analysis.

4.6. Long-Run Estimations

In the long run (

Table 9. Long-run estimations.

$\mathbf{Econometric} \mathbf{Model}: {\mathit{F}}_{\mathit{F}\mathit{S}\mathit{I}}=(\mathit{C}\mathit{C},\mathit{A}\mathit{G}\mathit{C},\mathit{A}\mathit{I}\mathit{C},\mathit{F}\mathit{P}\mathit{S},\mathit{G}\mathit{P},\mathit{A}\mathit{P},\mathit{T}\mathit{A})$
Dependent variables		Coefficient	t-Statistic	Prob. *
	CC	−0.560	−2.161	0.040 **
	AGC	0.904	12.087	0.000 ***
	AIC	1.320	8.619	0.000 ***
	FPS	−0.862	−1.830	0.078 *
	GP	2.011	0.684	0.500
	AP	1.121	5.729	0.000 ***
	AT	−0.084	−1.078	0.290
	constant	5.755	2.205	0.037 **

The symbols *, **, and *** denote significant values, respectively, at 10%, 5%, and 1%.

), climate change has a significant negative impact on food security. A one-unit increase in the climate change index leads to a 0.560 unit decrease in the FSI, holding other factors constant. This is statistically significant at the 5% level. This reinforces the understanding that the long-term effects of climate change, accumulating over time, can severely undermine food security by affecting agricultural yields, water resources, and ecosystem stability. Agricultural GDP contribution has a strong positive impact on long-run food security. A one-unit increase in AGC leads to a 0.904 unit increase in the FSI, holding other factors constant. This is highly statistically significant (p < 0.01). This suggests that a robust agricultural sector is crucial for sustained food security, likely through increased production, employment, income generation, and rural development. Agricultural insurance coverage has a strong positive impact on long-run food security. A one-unit increase in AIC leads to a 1320 unit increase in the FSI, holding other factors constant. This is highly statistically significant (p < 0.01). This indicates that wider access to agricultural insurance can significantly enhance food security in the long run by helping farmers manage risks associated with crop failures, weather events, and other shocks, thus stabilizing income and production. Similar to the short-run results, food price stability exhibits a negative relationship with long-run food security. A one-unit increase in the FPS index leads to a 0.862 unit decrease in the FSI, holding other factors constant. This is statistically significant at the 10% level. As discussed before, this counterintuitive result might suggest that the index is actually reflecting high and stable price levels rather than true volatility, which would negatively impact affordability and access to food. However, it could indicate the unintended consequences of price stabilization policies. In the long run, the impact of government policies on food security is not statistically significant. The coefficient is positive, suggesting a positive relationship. Agricultural productivity has a strong positive impact on long-run food security. A one-unit increase in AP leads to a 1121 unit increase in the FSI, holding other factors constant. This is highly statistically significant (p < 0.01). This highlights the importance of improving agricultural productivity through better farming practices, technology, and resource management to ensure food security in the long run. Finally, technology adoption does not have a statistically significant impact on long-run food security in this model, as the coefficient is negative, suggesting a potential negative relationship.

4.7. Granger Causality and VECM Tests

The results indicated in

Table 10. Results of the ECT test and Granger causality.

Directions of Causality
Short-Term									Long-Term
Independent Variables	DLnFSI	DLnCC	DLnAGC	DLnAIC	DLnFPS	DLnGP	DlnAP	DlnAT	ECT
DLnFSI	---------	0.123 * (0.070)	1.639 * (0.066)	0.229 ** (0.011)	1.223 (0.226)	0.338 * (0.067)	4.322 ** (0.032)	0.762 (0.492)	2.112 ** (0.047)
DLnCC	2.132 *** (0.001)	---------	0.329 (0.909)	1.627 (0.691)	2.087 * (0.066)	3.938 (0.226)	0.561 (0.878)	1.435 * (0.099)	1.813 * (0.089)
DLnAGC	0.425 (0.635)	0.563 (0.727)	---------	0.093 * (0.088)	2.324 (0.115)	0.093 (0.767)	2.426 (0.611)	1.662 (0.128)	−1.093 (0.902)
DLnAIC	1.092 * (0.087)	0.803 * (0.052)	0.672 (0.869)	---------	0.779 ** (0.022)	1.711 * (0.089)	1.001 * (0.077)	2.006 (0.700)	−0.873 ** (0.044)
DLnFPS	0.331 * (0.066)	2.517 (0.526)	2.013 (0.323)	0.492 (0.425)	---------	0.435 (0.325)	1.828 * (0.086)	0.572 * (0.081)	−2.983 (0.215)
DLnGP	0.920 (0.827)	2.038 ** (0.030)	0.671 (0.438)	0.872 ** (0.049)	0.871 (0.827)	---------	0.993 (0.279)	0.897 (0.177)	−0.660 (0.728)
DLnAP	1.562 (0.938)	0.011 (0.037)	1.625 ** (0.042)	2.014 (0.635)	0.424 (0.441)	0.678 (0.627)	---------	0.098 * (0.066)	−0.892 (0.982)
DLnAT	0.092 (0.110)	0.727 ** (0.022)	0.982 (0.837)	1.927 (0.872)	2.906 (0.343)	1.902 ** (0.043)	0.844 (0.640)	---------	1.635 * (0.626)

The symbols *, ** and *** denote significant values, respectively, at 10%, 5%, and 1%.

show that climate change (CC), in the short term, has a positive and statistically significant coefficient, suggesting that increases in the climate change index are associated with increases in the Food Security Index (FSI). This result appears counterintuitive, as climate change is generally seen as a negative factor for food security. A possible explanation is that in the short run, adaptive measures such as shifting cropping patterns and investing in irrigation might mitigate the immediate impact of climate change on food security. However, this result requires further scrutiny, as it contradicts the widely accepted negative long-term effect of climate change. In the long term, the significant error correction term (ECT) suggests that climate change contributes to the long-run equilibrium relationship with food security, although the table does not explicitly reveal its specific long-term impact.

The agricultural GDP contribution (AGC) has a positive and statistically significant coefficient in the short term, indicating that an increase in agricultural GDP contribution leads to an increase in FSI. This aligns with economic theory, as a larger agricultural sector enhances food security through increased production, employment, and income generation in rural areas. In the long term, AGC influences FSI through its contribution to the ECT, suggesting its role in maintaining food security over time. Agricultural insurance coverage (AIC) does not show a statistically significant short-term impact on food security. However, it indirectly influences FSI through other variables, indicating that while increased insurance coverage does not immediately improve food security, it affects other factors such as agricultural productivity and farmers’ investment decisions, which eventually contribute to food security. In the long run, the significant ECT suggests that AIC is part of the equilibrium relationship with FSI. The table also shows that AIC is influenced by food price stability (FPS) and agricultural productivity (AP) in the short term, indicating that its impact on food security operates through indirect pathways. Food price stability (FPS) does not have a statistically significant short-term impact on FSI, though it is influenced by other variables such as climate change and agricultural productivity. This suggests that while price stability itself may not immediately improve food security, it is affected by factors that do so, such as climate conditions and productivity levels. In the long term, the significant ECT suggests that FPS plays a role in the long-run equilibrium with FSI, likely through these indirect pathways. Government policies (GPs), in the short term, have a positive and statistically significant coefficient, indicating that increases in the government policy index contribute to improvements in food security. This result aligns with economic reasoning, as supportive policies such as subsidies, infrastructure investment, and agricultural research can enhance food security in the short run. In the long term, the significant ECT suggests that government policies contribute to the long-run equilibrium with FSI, reinforcing the importance of consistent policy frameworks for sustaining food security.

Agricultural productivity (AP) in the short term has a positive and statistically significant coefficient, indicating that increases in productivity are associated with improvements in food security. This result is consistent with economic principles, as higher agricultural productivity leads to increased output and improved food availability. In the long term, AP’s impact is reflected through the ECT, indicating its crucial role in maintaining food security over time. Finally, technology adoption (TA) does not show a statistically significant direct short-term impact on FSI. However, it is influenced by government policies in the short term, suggesting an indirect pathway to food security. This result indicates that while the adoption of new technologies does not immediately translate into improved food security, it is shaped by government policies that eventually impact food production and availability. In the long run, the significant ECT suggests that TA contributes to the long-run equilibrium with FSI, likely through the indirect influence of policy interventions.

5. Discussion

The observed contrast between the short-run positive and long-run negative effects of climate change on food security is indeed a complex result. Our interpretation suggesting “short-term adaptive measures” as a potential explanation acknowledges that immediate responses to climate variability, such as shifts in planting schedules or increased irrigation in response to short-term weather anomalies, might temporarily buffer negative impacts on food availability. However, as the reviewer rightly points out, this is a speculative explanation without direct empirical evidence of these adaptive measures in our model. We agree that a more nuanced explanation should also consider that the nature of short-run weather variability might differ significantly from the sustained, cumulative pressures of long-run climate change trends (e.g., rising average temperatures, altered precipitation patterns). The detrimental effect of these long-term trends on agricultural systems and resource availability is consistently reflected in our significant negative long-run coefficient for climate change, aligning with the established climate science literature. Our study highlights this short-run anomaly as a key area for future research, emphasizing the need to investigate specific adaptation strategies and their effectiveness empirically to fully reconcile these short- and long-term dynamics. Regarding the finding that food price stability is negatively associated with food security, we appreciate the reviewer’s push for a clearer policy implication. As discussed in our results section, our interpretation is that the “price index fluctuations” variable, when showing “stability”, might, in fact, be capturing periods of persistently high food price levels rather than a lack of volatility. High price levels, even if stable, directly reduce food’s accessibility and affordability, thereby negatively impacting food security, particularly for vulnerable households. This finding indeed underscores that the policy objective should not solely be price stability in isolation, but rather to ensure that food is affordable. Our policy recommendation to “re-evaluate existing policies to ensure they do not negatively impact affordability” directly stems from this result, suggesting that current price stabilization mechanisms might unintentionally contribute to high price levels that undermine food security. We agree that a deep analysis could benefit from considering factors like consumer purchasing power, as suggested by the reviewer. Future research incorporating metrics of income, poverty, and purchasing power alongside food price indices would provide a more comprehensive understanding of the complex interplay between price dynamics and economic access to food.

6. Conclusions and Policy Implications

This study employed the ARDL and VECM methods to investigate the complex short- and long-term dynamics influencing food security in Saudi Arabia between 1990 and 2023, examining its relationship with key socioeconomic and environmental factors. The findings reveal a multifaceted picture when compared with the existing literature. Consistent with broader understandings of agriculture’s role [5,6], the study found that agricultural GDP contribution (AGC) has a positive impact on food security in both the short and long term. This aligns with specific findings for Saudi Arabia by Darouez et al. [10,11,23,41], who also reported a positive correlation between AGC and food security, emphasizing the sector’s role in availability. The positive influence of agricultural productivity (AP) on food security in both horizons is also consistent with the literature, emphasizing its importance for increasing output [29,32]. Agricultural insurance coverage (AIC) has emerged as a crucial long-term determinant of food security, providing stability, as highlighted by Hazell [14] and Glauber [15]. This finding supports the view that insurance acts as a safety net, promoting consistent production and stabilizing farmers’ incomes over time, resonating with the potential role explored by Darouez and Ifa [11] for Saudi Arabia. Regarding government interventions, the study found that government policies (GPs) have a positive influence on food security in the short term. This aligns with the literature’s consensus on the supportive role of well-structured policies like subsidies and infrastructure [29,31,32]. However, unlike the literature, which often emphasizes the necessity of consistent, long-term policy frameworks for sustainable food security [31,32], this model did not find a statistically significant long-run impact for GP. This suggests that while short-term policy actions are beneficial, their translation into sustained, long-term food security in Saudi Arabia may depend critically on their consistency, design, and interaction with other factors that have not been fully captured, highlighting a potential area for deeper policy analysis. Technology adoption (TA) did not show a direct significant impact on short-term food security in this study but was found to play a crucial role in the long-run equilibrium. This supports the view that technological advancements like precision farming and improved varieties are vital for long-term productivity gains and food security [30,33], likely mediated through their influence on agricultural productivity and potentially facilitated by supportive policies. The relationship with food price stability (FPS) presented a finding that warrants careful interpretation. The study observed a negative association between the Food Security Index and FPS in both the short and long run. This appears counterintuitive when compared with the literature, which posits that price stability is essential for access and food security [20,21]. The negative association found here, particularly in the Saudi context, likely reflects the index capturing periods of high food price levels rather than volatility itself, as suggested by Darouez and Ifa [23] in their work on price volatility’s impact on access for low-income households. This emphasizes that high prices, even if stable, can undermine food access and thus the overall Food Security Index, suggesting a need for a more nuanced understanding of price dynamics beyond just volatility.

Perhaps the most surprising finding was the positive short-term relationship observed between climate change (CC) and the Food Security Index. This appears contrary to the overwhelming consensus in the literature that climate change poses a significant and primarily negative threat to agricultural production and food security globally, especially in arid regions, e.g., Aggarwal and others [10,27,28]. This counterintuitive short-term finding requires further investigation. It could potentially reflect short-term adaptive measures, data limitations, or specific temporal dynamics within the study period that mask the underlying long-term negative pressures of climate change, which the literature strongly highlights as undermining availability and stability [24]. The necessity for long-term adaptation strategies found in the literature [23,31] underscores the fact that despite any short-term anomalies, climate change remains a critical long-term challenge.

To operationalize the recommendation of expanding agricultural insurance, governments and development partners could subsidize premiums to encourage uptake among smallholder farmers, who are often the most vulnerable to climate shocks. Additionally, weather index insurance schemes, which base payouts on objective climate data rather than individual losses, could reduce administrative costs and the moral hazard while enhancing accessibility. Regarding agricultural diversification, targeted investment in climate-resilient crops such as sorghum, millet, or cassava could reduce dependence on rained cereals that are vulnerable to climate variability. Promoting horticulture and small-scale livestock production also offers income diversification and nutritional benefits. Identifying region-specific crop suitability through agroecological zoning would help tailor diversification strategies effectively.

However, while the findings offer valuable insights, several limitations must be acknowledged to contextualize the results and guide future research. First, the use of annual data from 1990 to 2023, although well-suited for time-series analysis, may not fully capture long-term structural changes or recent post-2023 developments. Data availability and quality, especially for agricultural insurance and technology adoption, presented challenges, and proxies like food imports or internet usage may not fully reflect the complexities of food security or innovation diffusion. Methodologically, while the ARDL and VECM models effectively handle mixed-integration series and dynamic relationships, they are sensitive to lag selection and subject to potential omitted variable bias [39,40,41,42]. Furthermore, the study is context-specific to Saudi Arabia and its findings may not generalize to countries with differing agroecological and economic systems. Finally, some unexpected findings, such as the short-term climate change impact or the negative price stability link, point to underlying complexities that deserve more granular, possibly region-specific analysis using disaggregated data or alternative modeling approaches.

By implementing these policy recommendations, Saudi Arabia can enhance its food security in both the short and long run, ensuring access to sufficient, safe, and nutritious food for all its citizens. Continuous monitoring and evaluation of policies are crucial for adapting to evolving challenges and achieving sustainable food security goals.