Food Security and Biofuels in Latin America and the Caribbean Region: A Data Panel Analysis on Eight Countries

Abstract

In the short, medium, and long term, a sustainable bioeconomy can help address one of the main concerns of most countries concerning the food crisis, particularly in the Latin American and Caribbean contexts, where food security and the bioeconomy are crucial for the region’s development and sustainability. However, to avoid negative impacts on the environment and food production, all sectors of the bioeconomy, especially those related to biofuel production, must be sustainable and environmentally conscious. This study analyses historical correlations between three dependent variables related to basic concepts of food security and independent variables framed by biofuel production through a panel data study in eight Latin American and Caribbean countries between 2007 and 2021. Of the three econometric models analyzed, two are statistically significant. The first shows a positive correlation between biofuel production and the food production index. The third shows a negative correlation between the percentage of undernourished people and biofuel production. In conclusion, according to the historical data evaluated in the countries studied, the impact of biofuel production on the food security variables analyzed is positive. This positive relationship can be interpreted as an opportunity for the region’s countries to generate additional bio-economic income, taking advantage of the region’s potential and providing new opportunities for producers, especially in rural areas.

1. Introduction

Since the 1990s, several international organizations have established a specific explanation of food security, where the central axes of the concept are access, availability, utilization, and stability of food resources. The World Food Summit 1996 indicated that “Food security exists when all people, at all times, have physical and economic access to sufficient, safe and nutritious food that meets their dietary needs and food preferences for an active and healthy life [1]”.

However, food security indices are different in all the world’s regions. According to the FAO’s 2022 report, the most significant increase in moderate or severe food insecurity was recorded in Africa. In regions such as Latin America and the Caribbean, food insecurity has worsened in recent years. Added to this are external factors, such as the COVID-2019 pandemic and the current Ukrainian war, which seriously affect global food security and nutrition [2].

According to the United Nations Report on the Panorama of Food and Nutritional Security in Latin America and the Caribbean, 22.5% of people in Latin America and the Caribbean need more means to access a healthy diet. In the Caribbean, 52% of the population has been affected by this situation; in Mesoamerica, this number reaches 27.8%; and in South America, this number is 18.4%. Among the variables associated with this phenomenon, the report highlights that the lack of affordability of a healthy diet has a clear relationship between a country’s income level, the incidence of poverty, and the level of inequality observed throughout the region [3].

A sustainable bioeconomy can help address one of the main concerns of most countries regarding the short-, medium-, and long-term food crisis. It can also support the transition to clean energy using bio-based materials [4]. In the case of Latin America and the Caribbean, the relationship is particularly close, as food security and the bioeconomy play a crucial role in the region’s development, sustainability, and the well-being of its people. These essential links include agricultural productivity, the diversification of food resources, sustainable resource management, rural development, and more resilient food systems in the region [5]. There are significant challenges, however, as all bioeconomy practices need to be sustainable and environmentally adequate to avoid adverse environmental and food production effects [6,7].

One of the most controversial sectors of the bioeconomy that is directly or indirectly linked to food security is biofuels. One of the most critical arguments of this link is that the increase in land under cultivation and the use of biomass for energy purposes should not be at the expense of food production [8,9].

This article analyses the relationship between two food security variables and one of the most important bioeconomy sectors (biofuels) in eight Latin American countries. The correlation is based on the three primary food security axes described above and the sustainability triple axis, which links environmental, economic, and social variables.

The paper is structured as follows. Section 2 describes the methodology applied in the empirical analysis; Section 3 explains the data used in this study and their sources; Section 4 shows the different tables and graphs and elaborates on the final results; and Section 5 illustrates the main conclusions and future research directions.

2. Materials and Methods

The objective of this paper is to analyze the relationship between food security variables and biofuels by considering indirect factors related to biofuel production that take account of economic, environmental, and social aspects in eight countries in Latin America and the Caribbean, in the light of previous research [10,11,12].

The central hypothesis to be tested in this paper is whether there is a correlation between biofuel production and the dependent variables selected for the food security assessment in the eight LAC countries under consideration.

We have used a panel data analysis to evaluate the production of biofuels at the level of food security. According to [13], the term panel data is usually used to refer to data containing time-series observations on a number of individuals. Thus, observations in panel data have two dimensions: a cross-section, denoted by the index (i), and a time series, represented by the index (t).

The assumption of a linear econometric model that involves these two dimensions takes the following formula:

$${\gamma }_{i,t }=\beta x_{i,t}+u_{i,t} i=1,\dots \dots \dots M, t=1, \dots \dots \dots T,$$

(1)

In the formula, $\beta$ is the parameter to estimate, $u_{i,t}$ represents the error, $M$ the statistical units, and $t$ the time. Here, the error is a function of both characteristics of the individual statistical unit and the random element [14]. In panel data analysis, different techniques exist to estimate the parameter $\beta$ as the fixed effects and panel data with random effects [15].

As mentioned above, this paper aims to analyze the relationship between biofuel production and food security using a linear econometric model. The choice of the model was based on the need to analyze the evolution over time of the selected dependent variables and their relationship with biofuels in the eight selected LAC countries. To know the temporal order of the variables and to be able to follow the different trajectories over time, it is necessary to collect panel data [16]. This methodology can control for the impact of omitted variables, reveal dynamic relationships, simplify the calculation and statistical inference, and generate more accurate predictions [13,17,18].

For our study, the analysis of the econometric model is represented in the following Equation (1):

$${\gamma }_{i,t }={\alpha }_{i,t}+\beta x_{i,t}+e_{i,t}$$

where $x_{i,t}\beta$ represent the matrix of independent variables used for this study: GDP per capita, arable land, biofuel production, food prices, CO₂ emissions, net capital stocks, and growth population in a country $i$ for a period $t$. These variables have been chosen after an analysis of various research studies [19,20,21,22,23] that identify the main factors and variables that drive the production of biofuels and after taking into account the threshold axis proposed in this study. The triple-axis concept of sustainability considers the social, economic, and environmental aspects of biofuel production (see Figure 1).

The ${\alpha }_{i,t}$ indicates the specific effects unobserved in a country, and $e_{i,t}$ is the error term. The specific effects represent the historical and institutional factors of the countries studied, which, for this research, are the climate vulnerability conditions of the region, the economic classification of the countries according to international organizations (between upper, low, and middle income), and the availability of arable land in the region for biofuel production.

For our dependent variables ${\gamma }_{i,t}$, we consider three different ones related to the three axes in the food security concept: food production index, prevalence of undernourishment, and per capita food supply variability [24], using the same independent variables in the same countries. The statistical analysis was conducted in three scenarios with these three dependent variables.

Figure 1. The threshold axis is taken into account in the study for the independent variables to analyze the relationship between biofuels and food security. Own construction based on [25,26].

Figure 1. The threshold axis is taken into account in the study for the independent variables to analyze the relationship between biofuels and food security. Own construction based on [25,26].

2.1. Data Source and Data Description

This study used composite panel data for the Latin American and Caribbean region for eight countries (Argentina, Brazil, Chile, Colombia, Mexico, Paraguay, Peru, and Uruguay) for 2007–2021. Due to data availability, especially on biofuel production, which not all countries in the region have, we use this time frame and these countries. Biofuel is one of the most essential bioeconomy sectors in the region, and it is the one for which it is possible to find historical records for OECD member and non-member countries [27].

The different variables were selected by analyzing the previous empirical literature based on the authors’ criteria and were extracted from different datasets, including Food and Agriculture Organization of the United Nations (FAO Stat), World Development Indicators Database, Organization for Economic Co-operation and Development (OECD), United Nations Economic Commission for Latin America and the Caribbean, and International Energy Agency (IEA).

Table 1. Variables and data sources.

Variables	Unit	Data Source	Time Period
Food Production Index	Index	FAO Stat	2007–2021
Per capita food supply variability	kcal/capita/day	FAO Stat	2007–2021
Prevalence of undernourishment	Percentage	CEPAL Stat	2007–2021
GDP/capita	USD per capita	World Bank	2007–2021
Arable land	percentage	World Bank	2007–2021
Biofuel production (sum of biodiesel and bioethanol production in each period per country)	million liters	OECD	2007–2021
CO2 emissions	T CO2/capita	IEA
Food prices	Index	FAO Stat	2007–2021
Net Capital Stocks (Agriculture, Forestry, and Fishing)	USD	FAO Stat	2007–2021
Population growth	percentage	World Bank	2007–2021

describes the independent variables used in our model to analyze the relationship between biofuels and food security as a threshold axis of sustainability: GDP per capita, arable land, biofuel production, food prices, CO₂ emissions, net capital stock, and population growth. For biofuel production, the value is the sum of bioethanol and biodiesel in each country per period.

As mentioned above, we decided to run three scenarios with different dependent variables in the analysis to give robustness to our results. In the first model, the dependent variable is the food production index (FPI); this variable quantifies edible and nutritious food crops (excluding coffee and tea, as they have no nutritional value) by calculating the changes in food production in a given year relative to the base year [28]. For the second model, the dependent variable was food supply, which estimates the variability in kcal/caput/day [29]. For the third model, the dependent variable was the prevalence of undernourishment, which considers the consumption of the number of calories that is insufficient to cover a person’s energy requirement for an active and healthy life [30].

Table 2. Descriptive statistics of the analyzed variables.

Variables	Mean	Std. Deviation	Min	Max
Food Production Index	96.11	11.74	57.9	122.5
Per capita food supply variability	30.83	23.68	9	176
Prevalence of undernourishment	5.3	2.82	2.5	13.6
GDP/capita	9543.55	3908.21	1206	18,825
Arable land	40.92	18.96	19	84.6
Biofuel production	4660.13	10,827.02	7.01	42,408.11
CO2 emissions	2.44	1.26	0.6	4.8
Food prices	103.42	62.99	46.82	616.24
Net Capital Stocks (Agriculture, Forestry, and Fishing)	26,344.52	27,610.02	1785	115,169
Population Growth	1	0.41	−0.08	1.91

provides a descriptive analysis that quantitatively describes the variables in our data set. Descriptive statistics aim to generalize from a sample to a larger population by describing the measure of central tendency and the dispersion or variance of the data [31,32].

To interpret Table 2, we take the mean statistics of the dependent and independent variables, which allow us to compare them with the optimum of each value and to make certain points between the countries analyzed. In the first case of the food production index, bearing in mind that it is calculated as an indicator based on 100, the region is generally below the basic average; the country with the highest index within the average is Chile, and the lowest is Paraguay. According to [33,34], Latin America is highly developed in terms of infrastructure and accessibility, allowing it to have a good average food production capacity compared to low-income countries in Africa; however, LAC suffers production impacts from greater vulnerability to climate change. Consistent with the study of [35], the region’s per capita food supply is generally below the world average, but as LAC is mostly classified as lower–middle and upper–middle income, it is still higher than low-income regions. In the case of the prevalence of undernourishment, the statistical mean of the analysis is lower than the FAO’s 2022 calculation. In contrast to our analysis, which takes only eight countries into account, the FAO report [36] looks at all the countries in the region and finds that undernourishment in Latin America and the Caribbean ranges from 7.5 percent to 9.7 percent. In terms of GDP per capita, the historical report (from 2007 to 2021) of the International Monetary Fund [37] indicates that the world average is 10.58 thousand USD. If we compare it with the statistical average of our database, it is 10% below the world average. The countries with the highest per capita income are Uruguay and Chile, and the lowest income is Peru. In the case of the percentage of arable land, the LAC countries in the study exceed the world average with 24%; in general, the region has the largest reserves of arable land on the planet, which contribute to the world’s food supply and other ecosystem services [38]. In the case of biofuels, which is the main dependent variable of the study, the historical data of the International Energy Agency [39] between 2016 and 2021 indicate an average annual production of 153 billion liters, which represents about 3% compared to the average of our database. (It is important to highlight that there is a difference of 8 years in this historical comparison). However, since we have countries such as Brazil and Argentina, which are generally important contributors to global biofuel production, both ethanol and biodiesel, the averages must be compared with the maximum and minimum values as well. Regarding the sixth independent variable, CO₂ per capita, the global average (according to statistical data [40]) over the last 15 years is about 6.35 tones, which is 39% of the average of our database. A number of scientific studies have shown that high-income regions tend to have higher carbon dioxide emissions than countries in the upper–middle and lower–middle income brackets [41,42,43]. In terms of the food prices index, historical data from the FAO [44] shows that the world average is 106.5 points over the period 2007–2021, while our database shows that the average for Latin America and the Caribbean is 3 points below the average. Finally, the average for the LAC countries in our database is 9% of the world average in terms of net capital stock in agriculture, forestry, and fisheries [45].

The objective of this study was to assess biofuel production’s impact on food security in eight countries in Latin America and the Caribbean through a panel data analysis, as described in the previous section.

According to different studies related to this methodology, the econometric model, as in Equation (1), provides a better robustness of the empirical results [46,47]. Equations (2)–(4) explain the three different econometric models that we have used for the fixed effect (FE) and the random effect (RE), with the variables described in Table 1. In the equations, FPI stands for food production index, PFSV is defined as per capita food supply variability, and PU indicates the prevalence of undernourishment.

$$\begin{array}{cc}\hfill FP{I}_{i,t }={\alpha }_{i,t}+& \beta GDP/percapit{a}_{i,t}+\beta Arable Lan{d}_{i,t}\hfill \\ & +\beta biofuel productio{n}_{i,t}+\beta {CO}_2 emission{s}_{i,t}\hfill \\ & +\beta food price{s}_{i,t}+\beta Net capital stock{s}_{i,t}\hfill \\ & +\beta Popullation growt{h}_{i,t}+e_{i,t}\hfill \end{array}$$

(2)

$$\begin{array}{cc}\hfill PFS{V}_{i,t }={\alpha }_{i,t}+& \beta GDP/percapit{a}_{i,t}+\beta Arable Lan{d}_{i,t}\hfill \\ & +\beta biofuel productio{n}_{i,t}+\beta {CO}_2 emission{s}_{i,t}\hfill \\ & +\beta food price{s}_{i,t}+\beta Net capital stock{s}_{i,t}\hfill \\ & +\beta Popullation growt{h}_{i,t}+e_{i,t}\hfill \end{array}$$

(3)

$$\begin{array}{cc}\hfill P{U}_{i,t }={\alpha }_{i,t}& +\beta GDP/percapit{a}_{i,t}+\beta Arable Lan{d}_{i,t}\hfill \\ & +\beta biofuel productio{n}_{i,t}+\beta {CO}_2 emission{s}_{i,t}\hfill \\ & +\beta food price{s}_{i,t}+\beta Net capital stock{s}_{i,t}\hfill \\ & +\beta Popullation growt{h}_{i,t}+e_{i,t}\hfill \end{array}$$

(4)

2.2. Data Analysis

We used two software packages that allowed us to obtain the desired results for the statistical analysis of the dataset.

The first package was SPSS version (29.0.1.0), with which we could analyze the statistical description of the variables, the plotting of the dependent variables across countries, the correlation of the independent variables, and the fixed panel data model. SPSS is a versatile package that allows many types of analysis, data transformation, and graphical and tabular output, providing researchers with academic-grade software [48].

The second statistical package used was R with the panel econometric package “plm”. This package provides functions for estimating different models and (robust) inference, aiming to simplify the estimation of linear panel models [49]. For our analysis, we test two methods for studying panel data with fixed coefficients and random effects. In fixed-coefficient models, the coefficients are allowed to vary along one dimension, and in random-coefficient models, the coefficients are assumed to vary randomly around a common mean [49,50].

3. Results

The results of the analysis can be divided into three sections. The first one is related to the correlation between the used independent variables. The second part is a visualization of the historical analysis of the dependent variables examined in the eight countries under study, including annexes with comparisons between the dependent variables and the production of biofuels in each country. Finally, the last table describes the statistical analysis of the panel data used in the three models. It compares the dependent variables related to food security and the independent variables.

Table 3. Correlation analysis.

Control Variables	GDP/Capita	Arable Land	Biofuel Production	CO2 Emission	Food Prices	Net Capital Stocks	Population Growth
GDP/capita	1.000
Arable land	0.386	1.000
Biofuel production	−0.121	−0.255	1.000
CO2 emissions	0.470	−0.074	−0.233	1.000
Food prices	0.045	−0.004	−0.022	0.034	1.000
Net Capital Stocks	−0.002	−0.323	0.850	0.030	0.058	1.000
Population Growth	−0.438	−0.567	−0.156	0.092	0.004	−0.032	1.000

shows the correlation analysis of the independent variables. With the results below, we can confirm that there is no collinearity in the econometric model. According to [51], collinearity can exacerbate any model misspecification bias. However, it does not induce any measurable bias when the model is correctly specified and the correlations are less than 0.99. It is consistent with the present analysis, as the parameters are below the lower values.

It should be noted that, among the independent variables, one considered in this study as a relevant input for producing biofuels is the percentage increase in arable land in the different countries studied. According to [52], the region is rich in natural resources, including more than 5 million km² of arable land and 28% of potential new arable land. Despite droughts and scarcity in some sub-regions, it also has the highest proportion of renewable water resources [53]. Given the availability of arable land and favorable natural conditions in many LAC countries [54], several research studies point to the possibility of producing sustainable biofuels without compromising the food supply [55]. These include technological optimization of biomass production and consideration of second-generation biofuel production from cellulosic sources, which are already abundant in the region. Indeed, sustainable biofuels are crucial for transitioning to low-carbon energy systems that can help countries secure energy while ensuring sustainable land use [56].

Comparing the food production indices of the countries studied, a positive historical evolution can be observed in most cases. Countries such as Chile and Paraguay have shown more stable figures over the years (Figure 2). However, it must be considered that the region has suffered a series of major climatic events, such as the El Niño phenomenon (high rainfall) or droughts, which affect food production indices. The impact of these phenomena is felt in the northern Pacific, the Andes, and the Caribbean departments. According to the [57] report, “El Niño exacerbates conditions of vulnerability in the region by causing a poor spatial and temporal distribution of rainfall, which affects the food and nutrition of the population.” This phenomenon significantly reduces the area used for agricultural production, causes losses in growing crops, reduces crop yields, and represents millions of dollars in losses for the affected countries [58].

We presented a visualization of the comparison between the bioeconomy sector under study, such as biofuels, and one of the dependent variables, such as food production indices. Supplementary Figure S1 shows the historical evolution of these variables in the countries studied. In the two countries with the highest biofuel production (Argentina and Brazil), biofuel production does not exceed food production rates, while in the other countries, the relationship is the opposite in specific years.

Figure 3 shows the historical comparison between countries in the region regarding the prevalence of undernourishment over the 15 years of the study. Uruguay, Chile, and Argentina have the lowest historical percentages, while countries such as Peru, Colombia, and Paraguay have the highest recorded rates. The average prevalence of destruction is 5.3 percent for all the historical data reported. This assessment is consistent with the report [59], which shows that most moderate or severe food insecurity in LAC increased by 20.5 percentage points between 2014 and 2020, while Mesoamerica increased by 7.3 percentage points over the same period.

The historical relationship between the second dependent variable and biofuel production is shown in Supplementary Figure S2. The intersection of the biofuel production curve is higher than the prevalence of undernourishment in countries such as Brazil, Paraguay, Colombia, and Peru from 2010.

Finally, the historical evolution of the variability of the per capita food supply in the LAC countries is shown in Figure 4. In general, all countries have an average of 30 points. Only Paraguay shows a significant high peak between 2011 and 2015 based on the data set analyzed. According to the report of [60], a series of public policies and programs were implemented in 2009–2012 as part of the national food safety plan.

Regarding the relationship between biofuel production and food supply per capita variability, Supplementary Figure S3 shows a significant historical variance in all countries, making it difficult to distinguish between positive and negative relationships in the curves. The only exception is Brazil, where the per capita supply is higher and equal to the production of biofuels over time.

Table 3 summarizes the results of the three different panel data analyses for each econometric model. Coefficients, their p-values, standard errors (in brackets), and summary statistics are presented in the table.

Table 4. Panel data results.

Dependent Variable: Food Production Index	FE Estimation (2)	RE Estimation (3)	Dependent Variable: Per Capita Food Supply Variability	FE Estimation (3)	RE Estimation (4)	Dependent Variable: Prevalence of Undernourishment	FE Estimation (5)	RE Estimation (6)
GDP/capita (USD)	1.157 × 10−03 * (4.8469 × 10−04)	1.164 × 10−03 *** (3.447 × 10−04)		9.1002 × 10−04 (1.311 × 10−03)	3.4031 × 10−04 (1.122 × 10−03)		−6.684 × 10−05 (5.21 × 10−05)	−1.0997 × 10−04 * (5.034 × 10−05)
Arable land (percentage)	9.224 × 10e−01 (7.325 × 10−01)	1.210 × 10−03 (6.657 × 10−02)		−1.671 × 10+00 (1.982 × 10+00)	3.294 × 10−02 (3.473 × 10−01)		−4.20 × 10−01 *** (7.879 × 10−02)	−7.40 × 10−02 *** (1.530 × 10−02)
Biofuel production (million liters)	1.700 × 10−03 ** (5.131 × 10−04)	4.878 × 10−04 * (1.898 × 10−04)		5.576 × 10−04 (1.388 × 10−03)	1.8229 × 10−04 (7.569 × 10−04)		−1.231 × 10−04 * (5.519 × 10−05)	−1.142 × 10−04 *** (3.375 × 10−05)
CO2 emissions (CO2 per capita)	−2.7054 × 10+00 (4.764 × 10+00)	−1.675 × 10−01 (1.048 × 10+00)		−1.323 × 10+01 (1.289 × 10+01)	−1.16 × 10+01 * (4.951 × 10+00)		−1.004 × 10+00 (5.125 × 10−01)	1.423 × 10+00 *** (2.187 × 10−01)
Food prices (Index)	7.863 × 10−02 *** (1.833 × 10−02)	7.678 × 10−02 *** (1.455 × 10−02)		−1.207 × 10−01 * (4.961 × 10−02)	−8.479 × 10−02 * (4.073 × 10−02)		−6.217 × 10−03 ** (1.972 × 10−03)	−2.285 × 10−03 (1.831 × 10−03)
Net Capital Stocks of Agriculture, Forestry, and Fishing (UDS)	3.545 × 10−05 (1.609 × 10−04)	−1.464 × 10−04 * (7.3472 × 10−05)		−5.0876 × 10−04 (4.356 × 10−04)	−1.924 × 10−04 (2.908 × 10−04)		−2.713 × 10−05 (1.73 × 10−05)	6.338 × 10−06 (1.298 × 10−05)
Growth population (Percentage)	1.1252 × 10+01 * (4.362 × 10+00)	1.061 × 10+01 *** (3.024 × 10+00)		5.226 × 10+00 (1.1803 × 10+01)	3.91 × 10+00 (9.383 × 10+00)		−3.213 × 10−01 (4.6918 × 10−01)	−6.02 × 10−01 (4.213 × 10−01)
R-squares	0.39563	0.33629		0.062199	0.088989		0.42855	0.52143
Adj. R-squares	0.31504	0.29481		−0.062841	0.032051		0.35236	0.49152
p-value	2.2763 × 10−09	6.7089 × 10−10		0.43922	0.14124		1.4551 × 10−10	2.22 × 10−16

Note: *, **, *** indicate 10%, 5%, and 1% significant levels, respectively.

shows the results of our analysis. The mechanics of the analysis are based on determining, for each econometric model, the relationship between dependent variables related to food security and independent variables directly associated with biofuel production. Column (1) examines the relationship between the independent variables and the food production index using fixed effects panel data analysis, while column (2) uses random effects for the first econometric model. The results of the second econometric model with the dependent variable variability of food supply per capita are presented in columns (3) and (4). Finally, the last two columns show the third analysis. Here, the dependent variable related to food security is the prevalence of undernourishment.

In the present study, the sample size (N) is 120 observations with 10 variables in each of the three analyses. The results of the Hausman’s statistical test reflect the following results:

FIRST ANALYSIS: F = 2.3646, df1 = 7, df2 = 105, p-value = 0.02771

SECOND ANALYSIS: F = 6.3131, df1 = 7, df2 = 105, p-value = 0.148

THIRD ANALYSIS: F = 1.5846, df1 = 7, df2 = 105, p-value = 3.41 × 10⁻⁶

The first analysis shows a positive relationship between the food production index and biofuel production in the countries analyzed, with an average of medium significance levels (10% in fixed effects and 1% in random effects) in the two statistical models used. Food prices are another variable with a high level of significance and a direct relationship. The first and last variables, GDP per capita and population growth, respectively, also show a positive correlation with a medium level of significance. Finally, the only variable that appears to be inversely related to the dependent variable is CO₂ emissions, but it is not statistically significant. In general, if we look at the p-values of the first model, they are statistically significant, which confirms the rejection of the null hypothesis in this study.

In the second econometric model, most variables are negatively linked with the dependent variable. Food prices have a low significance between 10% in both fixed and random effects, while CO₂ gas emissions have a low significance only in the random analysis. For the production of biofuels, the relationship is positive but not significant in the two factors analyzed. The remaining variables have no significance where no more detailed level is carried out. Overall, the statistical significance values for this model were not significant, which is reflected in the R-squared and p-value numbers.

In the third econometric analysis, where the prevalence of undernourishment is assessed as a dependent variable, it can be observed that most independent variables have a negative correlation in both statistical models. Specifically for the biofuel production variable, the significance is low by 1% in fixed effects and high by 10% in random effects, respectively. Arable land has a high significance in both factors analyzed. While food prices are only medium in the fixed factor, CO₂ emissions and GDP per capita are medium and low, respectively, in the random analysis. Generally, this econometric model is as statistically significant as the first one, which allows us to reject the null hypothesis.

4. Discussion

Our results confirm several previous studies on the relationship between biofuel production and some variables related to food security. The first analysis shows a positive relationship between the food production index and biofuel production in the countries analyzed, with an average of medium significance levels (10% in fixed effects and 1% in random effects) in the two statistical models used. This positive relationship is subject to several factors that need to be analyzed in more detail, such as the implementation of subsidy policies for the production of biofuels that encourage producers to produce more biomass, the substitution effect at the consumption level and production level and the main links that biofuels have with the food and energy markets. According to [12,61], biodiesel production positively affects food security through increased daily per capita energy consumption and the food production index. For [62], biofuel development can provide employment opportunities for smallholder farmers and workers in biofuel conversion and processing. At the same time, it can generate income and expand agricultural production technology. In the specific case of Latin America and the Caribbean [54,63], there is evidence that biofuels, especially ethanol and biodiesel, have contributed to the increase in demand for agricultural commodities over the last two decades. Higher agricultural prices in the biofuel scenarios, which provide supply-side incentives to intensify production and increase and reallocate land, also contribute to this increase in demand [64]. Regarding the negative relationship between the CO₂ emitted and food produced in the study of [65,66], the authors found that agricultural production indices directly affect carbon dioxide equivalent emissions, especially in distressed countries. Within the same perspective [67], CO₂ emissions affect food production in developing countries, concluding that supporting climate-smart agricultural practices is crucial to increase food security sustainability. In the case of food prices, several studies indicate the stretch relation between the supply and demand of food balance on agriculture prices [68,69].

The second model analyzed is not statistically significant overall. Biofuel production and population growth are the two variables that are positively correlated with the dependent variable. These results align with the study by [11,70], which points out that the growth in biofuel production can explain the increase in food supply and prices and the demand for alternative energy sources. The study of [71] predicts a rise of 12% in the cultivated area in Latin America and the Caribbean between 2005 and 2030. This would lead to an estimated 34% increase in productivity and a 23% increase in area per capita, resulting in food supply per capita increasing by 65% over the period studied.

In the last model considered, the overall statistic is significant. The four variables, the percentage of arable land, biofuel production, CO₂ emissions, and food prices, are negatively correlated with the dependent variable with a certain degree of significance. The interplay between several local and global issues, such as energy, food, land use, and development, analyzes the relationship between biofuels and food security dimensions complex [72,73]. Therefore, biofuel production and the policies that support its development can positively and negatively impact each of the four dimensions of food security—availability, access, nutrition, and stability. According to [74,75,76], biofuel production harms food security because the shift from food to biofuel production increases competition for access to production factors such as labor, water, and land. Regarding the environment, the study of [73] shows that the negative impact of CO₂ emissions has led to a decline in food security in the countries studied.

5. Conclusions

This study analyses the historical correlation between three dependent variables related to the basic concepts of food security and independent variables framed by the production of biofuels through a panel data study in eight countries of Latin America and the Caribbean between 2007 and 2021.

Of the three econometric models analyzed, two are statistically significant. The first indicates a positive correlation between biofuel production and the food production index, while the third indicates a negative correlation between the percentage of undernourished people and biofuel production. The significance of the rest of the independent variables falls within the threshold axis considered in the study to analyze the relationship between biofuels and food security.

The results obtained are directly related to the potential of the region in terms of the availability of arable land that has the potential to expand biofuel production without compromising food production rates on a larger scale, but this implies taking into account the socio-economic conditions of the country and the public policies that guarantee its sustainability in the long term.

The correlation for CO₂ emissions in the first model is negative due to increased land cover for food and biofuel production and population growth. In the second model, the correlation is also negative since more food production and supply are needed to reduce the number of undernourished people in the countries under consideration. However, it is important to stress that this comparison does not directly relate biofuel production to CO₂ emissions. Both factors are treated as independent variables in the econometric models.

In conclusion, the impact of biofuel production on the food security variables analyzed is positive, according to the historical data evaluated in the countries studied. This positive relationship can be interpreted as an opportunity for the region’s countries to generate additional bio-economic income, taking advantage of the region’s potential and providing new opportunities for producers, especially in rural areas. At the same time, the production of bio-energy can be seen as a stabilizer for the prices of food commodities, which generally show a high degree of volatility due to global economic and natural factors. Consequently, it is recommended to promote the development of sustainable biofuels (not only first generation) that contribute to the maintenance of and increase in food production while improving the quality of the environment, especially in the most socio-economically vulnerable countries. In this sense, the governments of countries in the region must ensure that biofuel strategies consistently reduce emissions and contribute to food production.

6. Limitations

The study has a significant limitation in terms of access to historical data concerning the production of biofuels for the countries in Latin America and the Caribbean. This limitation has an impact on the number of years over which the analysis is possible, as it is the most relevant independent variable in this study. Only eight countries could be the subject of analysis, about 24% of the region’s total. However, regarding the bioeconomy sector under study, such as biofuels, the research embraces the central producer countries in the region, for example, Argentina, Brazil, Uruguay, and Mexico.

Another significant limitation was the selection of food security variables for the study. While it is true that there are several indicators established by the FAO in this area, not all of them have more than 15 years of historical information, which would allow a complete panel analysis of the data. Therefore, one of the criteria established by the authors to determine the dependent variables was the historical access to information in the countries under study.

This is directly related to the source of the data (see Table 1). As explained above, the subtracted information of the different variables has a limited time span of 15 years in terms of cross-sectional comparisons that allow a more complete panel analysis. However, given that the current time series of the data is relatively short, the reliability of the panel data estimators is low in the conventional sense.

The independent variables examined in the study were limited to factors and variables related to biofuel production according to studies analyzed by the authors, but we are aware that they do not cover all factors that could be considered in future research.