The Effects of Education and Financial Development on Energy Poverty Reduction in Latin America

Abstract

Using data from 14 Latin American countries over the period 2000–2019, this study examines the influence of education and financial development on energy poverty reduction. Evidence from different estimation approaches, such as dynamic ordinary least squares (DOLS), fully modified ordinary least squares (FMOLS), and canonical correlation regression (CCR), showed that education substantially decreases energy poverty reduction in Latin America. The results also indicated that the nexus between financial development and energy poverty reduction is positively and significantly associated. Conversely, the linkage between education, financial development, and energy poverty reduction considerably varies among Latin American countries. We suggest that policies stimulating financial development and education will be pivotal in addressing energy poverty reduction in Latin America.

Introduction

Energy poverty and deprivation continue to dominate policy discussions and scholarly research (Apergis et al., 2022). The socioeconomic consequences of energy poverty have led recent studies to examine the drivers of energy poverty in both developed and developing economies to inform relevant national and international policies aimed at addressing the root cause of energy poverty (Khan et al., 2023; Mohsin et al., 2022; Said & Acheampong, 2023). According to the World Energy Outlook 2020 report, as of 2019, approximately 770 million people (10% of the global population) lacked access to electricity, and about 2.6 billion (34% of the global population) lacked access to clean fuel, most of whom are in developing countries. ¹ While Latin America has a higher level of electricity and clean fuels and technologies for cooking access than other developing regions such as South Asia and Africa, there is still a significant gap in access to power, with 22 million people lacking access to electricity and 80 million cooking with firewood and charcoal, using fuel-inefficient out-dated stoves (Mohsin et al., 2022). The use of contaminated cooking fuels indoors has serious health consequences, especially for women and children. In this regard, one of the Sustainable Development Goals (SDG) of the United Nations is to ensure energy access to all people, including the poor, vulnerable, and those with low incomes, so that they can meet their basic needs such as health, nutrition, and education (Mohsin et al., 2022).

Recent theoretical literature emphasizes the importance of educational attainment in reducing energy poverty. An educated population will be more efficient in managing energy use, partly as a result of the likelihood of a future fossil fuel shortage. Consequently, better education accelerates the switch from conventional fuels such as biomass and kerosene to more efficient energy sources such as renewables and natural gas, thereby alleviating energy poverty (Apergis et al., 2022). Furthermore, Khan et al. (2023) point out that energy poverty is associated with higher investments in education. A higher level of investment in education due to a higher income level will result in higher socioeconomic status and living standards for the household. As a result, an increase in living standards can aid families in improving their energy sources to better options, gradually lowering energy poverty (Figures 1 and 2).

Figure 1.

Trends of access to electricity in different world regions. Source: World Bank (2023).

Figure 2.

Trends of access to clean fuels and technologies for cooking in different world regions. Source: World Bank (2023).

On a theoretical basis, finance also contributes to the reduction of energy poverty in two main aspects (Koomson & Danquah, 2021). Firstly, energy poverty is influenced by household income. The development of financial institutions tends to improve the incomes of low-income households since they provide them with additional funds. With additional funds available, further goods and services could be created, thereby increasing incomes and profits. As a result, increasing household income can minimize energy poverty, mainly in low-income families (Khan et al., 2023). Secondly, the development of the financial sector also contributes to energy poverty reduction by promoting economic growth and energy efficacy. In Koomson and Danquah’s (2021) view, financial development leads to economic growth by accumulating savings and allocating investments to productive industries. A rise in income levels increases the accessibility of clean energy. Consistent with theory, numerous empirical studies indicate that financial development exerts a significant positive influence on energy poverty reduction (see, e.g., Fang et al., 2022; Mohsin et al., 2022; Said & Acheampong, 2023).

Despite several studies testing the nexus between financial development and energy poverty reduction, it is rare to find empirical studies that have examined the relationship between education, financial development, and energy poverty reduction, measured using access to electricity and access to clean fuels and technologies for cooking in Latin America. Developing economies such as Latin American economies are regarded as more likely than industrialized nations to fulfil the SDGs. Our main focus is on Latin American countries because these countries continue to strive to expand economic development, eradicate poverty, and improve quality of life. Therefore, this study is motivated to examine the impact of education and financial development on energy poverty reduction using data from 14 Latin American countries from 2000 to 2019. Thus, several questions arouse our interest: (1) Does education enhance access to electricity and clean fuels and technologies for cooking in Latin America? (2) Does the linkage between education and access to electricity and clean fuels and technologies for cooking differ across Latin American countries? (3) Does financial development enhance access to electricity and clean fuels, and technologies for cooking in Latin America as a panel? (4) Does the linkage between financial development and access to electricity and clean fuels and technologies for cooking differ across Latin American countries?

The novelty and the contribution of this paper to literature are discussed as follows. Firstly, to the best of our knowledge, this is the first empirical study to examine the effect of financial development and education on energy poverty within the context of Latin America. In this direction, our study contributes to the literature by complementing the existing knowledge on energy transition by unravelling the contribution of financial access and human capital investment that could facilitate access to electricity and clean cooking technologies in Latin America. In addition, this study contributes to the understanding of how investment in education and, for that matter, human capital development plays a role in enhancing access to clean and modern energy, which is at the heart of energy transition policies.

Second, this study contributes to the literature by highlighting how the impact of financial development and education on energy poverty reduction differ among countries. This study adopts panel data techniques to estimate the effect of financial development and education on energy poverty reduction in Latin American countries. However, structural and socioeconomic differences exist among Latin American countries, therefore making financial development and education have an expected heterogeneous effect on energy poverty reduction across Latin American countries. Therefore, in addition to the panel data approach, we adopted country-specific analysis to document the effect of financial development and education on energy poverty reduction, thus making this study novel.

Methodologically, this study deploys different econometric techniques such as dynamic ordinary least squares (DOLS), fully modified ordinary least squares (FMOLS) canonical correlation regression (CCR), fixed-effect, and the Driscol–Kraay estimators to empirically estimate the impact of financial development and education on energy poverty reduction. The application of these different estimators ensures the robustness of the findings.

The empirical findings based on the battery of the econometric techniques indicate that education and financial development increase access to clean fuels and cooking technologies and access to electricity, indicating that improving educational attainment and financial development assist in lessening energy poverty in Latin America. The results also suggest that education and financial development have a heterogeneous effect on reducing energy poverty across Latin American countries. For instance, the findings show that education reduces energy poverty significantly in Argentina, Bolivia, Costa Rica, Ecuador, El Salvador, Guatemala, Honduras, Mexico, Nicaragua, Panama, and Paraguay, whereas in Brazil, Colombia, and Peru, education does not contribute to the reduction of energy poverty. In addition, financial development significantly increases energy poverty (reduces energy access) in Brazil, Colombia, Costa Rica, Ecuador, Guatemala, Honduras, Nicaragua, and Peru, while financial development is not related to a reduction in energy poverty in Argentina, Bolivia, El Salvador, Mexico, Panama, and Paraguay. In terms of policy, these outcomes indicate that Latin America should focus on improving and maintaining educational attainment and financial development in order to achieve SDG 7.

Following is the structure of the rest of this study. In Section 2, we review the existing literature. Data descriptions, model specifications, and summary statistics are described in Section 3. Section 4 summarizes and discusses the main findings. In Section 5, we conclude our discussion and discuss policy implications.

Related Literature

Education and Energy Poverty

The relationship between education and energy poverty reduction has received significant scholarly attention in recent years. It has been argued a country’s level of education is important in reducing energy poverty. According to Edziah et al. (2021), countries’ levels of education play a crucial role in converting from inefficient to efficient technologies. He and Huang (2020) argue that a well-trained workforce and educated executives will use energy more efficiently. Oum (2019) contends that there is a connection between education and energy poverty reduction. He reveals that a substantial reduction in energy deprivation can be achieved by improving educational levels. Sharma (2016) points out that the increased expenditures on education have led to a reduction in energy poverty in India, mostly among low-income households. In addition, Crentsil et al. (2019) suggest that as the household’s attainment of education rises, energy poverty decreases in Ghana. Moreover, they argue that households with heads with no formal education are more likely to live in energy poverty than those with postsecondary education. Acharya and Sadath (2019) investigated the influence of education and economic development on energy poverty between 2004 and 2012 using a sample of 25 Indian States. The empirical evidence indicates that education reduces energy poverty more than income does. Koomson and Danquah (2021) indicate a considerable decline in Ghana’s energy poverty, which has been found to be associated with improvements in education levels. More specifically, they indicate that a 10% rise in household education attainment lessens energy poverty by around 1.85%. In addition, Rafi et al. (2021) reveal that an increase in energy poverty is associated with a decline in their enrolment in schools and their performance in examinations. Apergis et al. (2022) examine the impact of education on energy poverty reduction in the case of developing countries from 2001 to 2016. The empirical results reveal that education exerts a significant positive influence on energy poverty reduction. In particular, a 1% rise in educational level leads to a rise in the percentage of electricity access by around 0.429%. By applying the 2SLS method, Oktaviani and Hartono (2022) investigate the association between education and energy poverty reduction in the case of Indonesia. The results show that education and energy poverty reduction are positively and significantly related. Qamruzzaman et al. (2023) used an autoregressive distributed lagged (ARDL) to inspect the impact of education on energy poverty reduction in Morocco and Tunisia from 1991 to 2019. The empirical findings indicate that education has a beneficial impact on energy poverty reduction in these two countries. In particular, they found that a 10% increase in government expenditure on education led to a reduction in the level of energy poverty in Morocco by 1.261% and Tunisia by 0.889%. Khan and Ghardallou (2023) assess the impact of education on energy poverty reduction in a sample of 108 developing economies over the period 2000–2019. Based on the Pedroni co-integration technique, the empirical findings confirm the existence of a positive relationship between education and access to electricity. Shafiullah et al. (2023), using an Engel curve approach, indicate that the level of education is the crucial determinant of various energy poverty measures and energy expenditure shares across Chinese households.

The Effect of Financial Development on Energy Poverty

There has been a growing body of evidence demonstrating a connection between financial development and energy poverty reduction. It is largely acknowledged that the enhancement of financial services has drastically reduced energy poverty. For instance, Dogan et al. (2021) present evidence financial inclusion drastically lessens energy poverty among Turkish households. Koomson and Danquah (2021) investigate the relationship between financial development and energy poverty in Ghana. The results indicate that financial inclusion has a negative impact on energy poverty. Mohsin et al. (2022) look into the link between financial development and energy poverty in Latin America. The empirical findings imply that financial development is important in easing energy poverty. Further, Wang et al. (2023) examine the influence of financial development on energy poverty in the case of China. Results show that financial development alleviates ‘China’s energy poverty. Khan and Majeed (2023) stress that access to financial services plays a basic role in reducing energy poverty in a large sample of developing nations. Zhao et al. (2023) indicate that financial development positively lowers energy poverty in Korea over the period 1990–2021. Based on two-stage least squares regression (2SLS), Cheng et al. (2023) show that financial market participation significantly reduced household energy poverty in China. Said and Acheampong (2023) examined the impact of financial inclusion on energy poverty reduction in sub-Saharan African countries from 2004 through 2019, and the results revealed that financial inclusion contributes to energy poverty reduction in sub-Saharan Africa. Khan and Ghardallou (2023) assess the impact of financial development on energy poverty reduction. Using a large sample of developing economies from 2000 to 2019, the empirical evidence indicates that financial development exerts a significant positive impact on electricity access in these economies.

Although there has been a growth in the number of studies on the association between financial development and energy poverty reduction, an empirical investigation between education, financial development, and energy poverty reduction in Latin American countries is very scarce. In addition, few prior studies have assessed the effect of education on energy poverty reduction in a single country or across a panel of countries. Nevertheless, no study has examined Latin American countries individually. This study, therefore, fills the gap and contributes to the literature by exploring the nexus between education, financial development, and energy poverty reduction using panel data and country-specific analysis across 14 Latin American countries.

Figures 3 below present plots of the four indicators across Latin America: financial development, education, and energy poverty proxies by country and as a panel over time. As can be observed, all indicators exhibit increasing trends as a panel and by individual country in most nations during the period 2000–2019.

Figure 3.

Trends of the natural logarithm of education, financial development, access to clean fuels and technologies for cooking, and access to electricity across Latin America by country and as a panel over the period 2000–2019. Source: World Bank (2023).

Empirical Methodology

Data Descriptions

The key aim of this paper is to study the linkage between education, financial development, and energy poverty reduction by taking into account economic growth, industrialization, and trade openness. The list of Latin American countries includes Argentina, Bolivia, Brazil, Colombia, Costa Rica, Ecuador, El Salvador, Guatemala, Honduras, Mexico, Nicaragua, Panama, Paraguay, and Peru. Because of missing data for Venezuela, Cuba, and Haiti, we excluded these countries from the panel. We use the data from the web page of World Development Indicators (World Bank, 2023) with respect to energy poverty reduction variables, financial development, economic growth, industry, and trade openness, while the data for education is taken from Penn World Table 10.1. Table 1 describes the variables used in this study.

Table 1.

Detail of Variables.

Variable name	Unit of Measurement	Definition
Access to electricity (ATE)	(% of population)	Access to electricity is the percentage of population with access to electricity.
Access to clean fuels and technologies for cooking (CFTC)	(% of population	Access to clean fuels and technologies for cooking is the proportion of total population primarily using clean cooking fuels and technologies for cooking.
Education (EDU)	Index	Human capital index based on years of schooling and returns to education
Financial development (FD)	(% of GDP)	Measured through the proxy ‘deposit money banks as a share of GDP’.
GDP per capita growth (GDPPCG)	(Annual %)	Annual percentage growth rate of GDP per capita based on constant local currency.
Industrialization (INDUS)	(% of GDP)	Industry (including construction), value-added.
Trade openness (TO)	(% of GDP)	Trade is the sum of exports and imports of goods and services measured as a share of the gross domestic product.

Model Specification

This study follows the empirical work of Mohsin et al. (2022) and Said and Acheampong (2023) to investigate the effects of education and financial development on energy poverty reduction. Therefore, the empirical models to examine the effects of education and financial development on energy poverty reduction are given by

$\mathit{A}\mathit{T}\mathit{E}=\mathit{f}\left(\hspace{0.17em}\mathit{E}\mathit{D}\mathit{U},\mathit{F}\mathit{D},\mathit{G}\mathit{D}\mathit{P}\mathit{P}\mathit{C}\mathit{G},\mathit{I}\mathit{N}\mathit{D}\mathit{U}\mathit{S},\mathit{T}\mathit{O}\right)$

(1)

$\mathit{C}\mathit{F}\mathit{T}\mathit{C}=\mathit{f}\left(\mathit{E}\mathit{D}\mathit{U},\mathit{F}\mathit{D},\mathit{G}\mathit{D}\mathit{P}\mathit{P}\mathit{C}\mathit{G},\mathit{I}\mathit{N}\mathit{D}\mathit{U}\mathit{S},\mathit{T}\mathit{O}\right)$

(2)

where ATE is access to electricity, CFTC is access to clean fuels and technologies for cooking, EDU denotes education, FD represents financial development, GDPPCG is real GDP per capita growth, INDUS is industry (including construction), value-added, and TO is trade openness. To decrease the sharpness of the data, we take the natural log of all the variables. In contrast with a simple linear transformation, the log-linear model offers empirically consistent and efficient results. ² We present the augmented multivariate production function for our linear-log model as

${\hspace{0.17em}\mathit{l}\mathit{n}\mathit{A}\mathit{T}\mathit{E}}_{\mathit{i}\mathit{t}}=\mathit{\beta }+{\mathit{\beta }}_{\mathbf{1}}{\mathit{l}\mathit{n}\mathit{E}\mathit{D}\mathit{U}}_{\mathit{i}\mathit{t}}+{{\mathit{\beta }}_{\mathbf{2}}{\mathit{l}\mathit{n}\mathit{F}\mathit{D}}_{\mathit{i}\mathit{t}}+\mathit{\beta }}_{\mathbf{3}}{\mathit{G}\mathit{D}\mathit{P}\mathit{P}\mathit{C}\mathit{G}}_{\mathit{i}\mathit{t}}+{{\mathit{\beta }}_{\mathbf{4}}{\mathit{l}\mathit{n}\mathit{I}\mathit{N}\mathit{D}\mathit{U}\mathit{S}}_{\mathit{i}\mathit{t}}+\mathit{\beta }}_{\mathbf{5}}\hspace{0.17em}\mathit{ln}\hspace{0.17em}{\mathit{T}\mathit{O}}_{\mathit{i}\mathit{t}}{+\mathbf{\varepsilon }}_{\mathit{i}\mathit{t}}$

(3)

${\mathit{l}\mathit{n}\mathit{C}\mathit{F}\mathit{T}\mathit{C}}_{\mathit{i}\mathit{t}}=\mathit{\alpha }+{\mathit{\alpha }}_{\mathbf{1}}{\mathit{l}\mathit{n}\mathit{E}\mathit{D}\mathit{U}}_{\mathit{i}\mathit{t}}+{{\mathit{\alpha }}_{\mathbf{2}}\hspace{0.17em}\mathit{ln}\hspace{0.17em}{\mathit{F}\mathit{D}}_{\mathit{i}\mathit{t}}+\mathit{\alpha }}_{\mathbf{3}}{\mathit{G}\mathit{D}\mathit{P}\mathit{P}\mathit{C}\mathit{G}}_{\mathit{i}\mathit{t}}+{{\mathit{\alpha }}_{\mathbf{4}}{\mathit{l}\mathit{n}\mathit{I}\mathit{N}\mathit{D}\mathit{U}\mathit{S}}_{\mathit{i}\mathit{t}}+\mathit{\alpha }}_{\mathbf{5}}\hspace{0.17em}\mathit{ln}\hspace{0.17em}{\mathit{T}\mathit{O}}_{\mathit{i}\mathit{t}}{+\mathbf{\varepsilon }}_{\mathit{i}\mathit{t}}$

(4)

Where β and α represent coefficients in equations (3) and (4), respectively. i denotes the countries involved in estimation (1, 2, …, 14), t shows the period of analysis (2000–2019), and ε is the residual. β1, β2, β3, β4, and β5 are the coefficients of EDU, FD, GDPPCG, INDUS, and TO in equation (3), while α1, α2, α3, α4, and α5 are the coefficients of EDU, FD, GDPPCG, INDUS, and TO in equation (4).

Summary Statistics

Table 2 illustrates the descriptive statistics of all the variables. There are wide variations in the percentage of residents who have access to electricity, ranging from 63.14% to 100%; the mean is 91.01%. Moreover, the mean value is lower than the median value of 94.95%, which shows that the data is skewed to the left. Furthermore, access to clean fuels and technologies for cooking ranges from 30.7% to 99.9% of the population. As the mean value of CFTC is 73.57%, the median is 82.85%, showing that the data is also skewed. Consequently, a significant percentage of Latin American households lack access to electricity, clean fuels, and cooking technologies. In terms of the education index, the series ranges from 1.62 to 3.1, with a mean value of 2.44 and a median value of 2.53 (just close to the 50th percentile), indicating a uniform distribution. As with financial development, the sample also shows considerable variation, with a minimum of 12.59 and a maximum of 114.78. The mean value (44.81) is higher than the median value (39.99), indicating that the data are skewed to the right. There is a considerable reliance on finances for those who benefit from financial development. As a result of this alteration, it is fascinating to explore the impact of financial development and education on energy poverty reduction in Latin America as an overall panel as well as for individual countries.

Table 2.

Summary Statistics.

Variable	Obs	Mean	Median	Std. dev.	Min	Max
ATE	280	91.01	94.95	9.14	63.14	100
CFTC	280	73.57	82.85	20.7	30.7	99.9
EDU	280	2.44	2.53	0.95	1.62	3.1
FD	278	44.81	39.99	20.76	12.59	114.78
GDPPCG	280	2.05	2.09	2.87	−11.85	9.93
INDUS	280	27.46	26.83	5.1	17.74	42.17
TO	280	67.09	64.99	30.42	21.85	166.7

Table 3 presents the associations among the study indicators, specifically education (EDU), financial development (FD), and energy poverty reduction (ATE and CFTC). A high correlation between ATE and CFTC has been shown (0.76). Additionally, there is a strong correlation between EDU and CFTC (0.72) and a moderate correlation between EDU and ATE (0.47). Finally, the correlation between trade openness (TO) and (ATE) is moderate (−0.42), while the correlation between financial development (FD) and CFTC is weak (0.28).

Table 3.

Correlations.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)
ATE (1)	1.00
CFTC (2)	0.76	1.00
EDU (3)	0.47	0.72	1.00
FD (4)	0.12	0.28	0.19	1.00
GDPPCG (5)	−0.05	0.02	0.08	−0.01	1.00
INDUS (6)	0.18	−0.05	0.12	−0.58	0.05	1.00
TO (7)	−0.42	−0.38	−0.19	0.16	0.21	−0.17	1.00

Then, we computed variance inflation factors to quantify the effect of multicollinearity among regressors in Table 4. The results show that each variable’s VIF is below 10; therefore, the data series does not pose a risk of multicollinearity.

Table 4.

VIF Analysis Variable.

	VIF	1/VIF
lnEDU	1.16	0.86
lnFD	1.58	0.63
GDPPCG	1.06	0.94
lnINDUS	1.56	0.64
lnTO	1.13	0.89
Mean VIF	1.3

Results and Discussion

Cross-Sectional Dependence

After conducting this preliminary analysis, we performed tests to check for cross-sectional dependence (CD). An inability to consider CD may cause size distortions and efficiency losses. This study, therefore, uses the Friedman (1937) method to examine whether the selected panel members are interdependent. The outcomes tabulated in Table 5 indicate that the null hypothesis is not rejected, which confirms the residuals are cross-sectionally independent under the fixed-effect specification.

Table 5.

Results of the Cross-Sectional Dependence Tests.

Test	Statistics	Prob.
Dependent variable: LnATE Friedman’s test	8.807	0.787
Dependent variable: LnCFTC Friedman’s test	12.847	0.460

Panel Unit Root Test

The confirmation of cross-sectional independence encouraged us to employ the Im et al. (2003) unit root test. Table 6 shows that all the variables have a unit root. After taking the first difference of all the variables, the stationarity is implemented, indicating that all the variables are integrated in order 1. It is necessary to establish the stationary of variables before estimating them empirically.

Table 6.

Results of Panel Unit Root Test With Trend.

	Level		First Difference
Variable	Z[t bar]	p-Value	Z[t bar]	p-Value
lnATE	−0.491	0.312	−8.014***	0.000
lnCFTC	0.668	0.748	−3.436***	0.000
lnHC	3.687	0.999	−1.354*	0.088
lnFD	−0.824	0.205	−2.056**	0.020
GDPPCG	−1.153	0.124	−2.077**	0.019
lnINDUS	−0.411	0.341	−5.673***	0.000
lnTO	−0.841	0.200	−5.399***	0.000

*, **, *** demonstrate significance level at 10%, 5%, and 1%, respectively.

Panel Co-integration Test

Afterwards, it is crucial to investigate whether the variables have a co-integrated link. In this study, Pedroni (2004) is used, and the testing results are shown in Table 7. Three separate statistics tests are included in the Pedroni test: Modified Phillips–Perron, Phillips–Perron, and Augmented Dickey–Fuller (ADF). In all panels, the null hypothesis is that there is no co-integration, and the alternative hypothesis is that there is co-integration. With the three test statistics, the null hypothesis is rejected in favour of the alternative hypothesis. The results signify that all the variables have a long-run relationship because of the statistical significance of all the statistics.

Table 7.

Results of Panel Co-integration Tests.

No	Co-integration Test	Statistic	p-Value	Statistic	p-Value
Pedroni test		Equation 1		Equation 2
Ho: No co-integration – Ha: All panels are co-integrated
1	Modified Phillips–Perron t	2.652***	0.004	5.014***	0.000
2	Phillips–Perron t	−4.970***	0.000	2.352***	0.009
3	Augmented Dickey–Fuller t	−4.278***	0.000	3.342***	0.000

***Significant at 1% level.

Estimation of Long-Run Coefficients

As estimation techniques, we use fully modified ordinary least squares (FMOLS) and dynamic ordinary least squares (DOLS). Fully modified ordinary least square regression technique was developed by Pedroni (2001), which is a residual-based test that provides efficient results when co-integrated variables are involved. In addition, FMOLS is considered reliable when the sample size is small and eliminates endogeneity and serial correlation issues. Additionally, Hamit-Haggar (2012) developed a DOLS approach which provides better results than FMOLS and eliminates correlation among regressors. Furthermore, we use Canonical Co-integrating Regression (CCR) estimation as a robust estimation to validate FMOLS and DOLS results. CCR procedure reveals simple mixture distributions, confirms asymptotic Chi-square validation, and resolves the problem of non-scalar disturbances (Park, 1992).

Results and Discussions

Education, Financial Development, and Energy Poverty Reduction: Aggregated Sample Results

We use two indicators of energy poverty reduction: access to electricity (% of population) (ATE) and access to clean fuels and technologies for cooking (% of the population) (CFTC). It is interesting to mention that the results are quite robust under different specifications. Table 8 (Panel A) displays the panel estimations’ outcomes. The results indicate that education is positively and significantly associated with ATE in all specifications. Specifically, a 1 % increase in educational attainment is associated with an increase of ATE within a range of 0.231%–0.232%. This result implies that energy poverty reduction is mainly influenced by the level of education in Latin America. The result does not agree with Doğanalp et al.’s (2021) findings, which indicate that education wields no significant impact on energy poverty reduction in the case of BRICS. In contrast, our result is in line with the findings of Crentsil et al. (2019) for the case of Ghana, Acharya and Sadath (2019) for the case of India, Apergis et al. (2022) for a sample of 30 developing countries; Oktaviani and Hartono (2022) for the case of Indonesia; Khan and Ghardallou (2023) for the case of 108 developing countries; Shafiullah et al. (2023) for the case of China; and Khan et al. (2023) for a large sample of emerging economies, which contend that education plays a basic role in minimizing energy poverty. Education positively affects access to electricity since it can boost a country’s economic growth and further increase access to electricity (Zhang et al., 2019). In addition, through education, individuals can improve their skills, creativity, and productivity; they can better understand their potential as workers, and as a result, energy poverty is alleviated (Hassan et al., 2022).

Table 8.

Panel Estimation Results of the Impact of Education and Financial Development on Energy Poverty Reduction.

	Model 1	Model 2	Model 3
	DOLS	FMOLS	CCR
Panel A/Dependent variable: ATE
EDU	0.231**	0.232***	0.231***
	(0.106)	(0.066)	(0.067)
FD	0.052	0.050**	0.050*
	(0.043)	(0.025)	(0.026)
GDPPCG	−0.002	0.000	0.000
	(0.008)	(0.003)	(0.004)
INDUS	0.136	0.134**	0.135**
	(0.102)	(0.061)	(0.064)
TO	−0.074**	−0.069***	−0.069***
	(0.034)	(0.021)	(0.021)
Constant	3.971***	3.958***	3.958***
	(0.469)	(0.272)	(0.286)
Number of observations	275	277	277
Number of countries	14	14	14
Panel B/Dependent variable: CFTC
EDU	1.454***	1.354***	1.349***
	(0.339)	(0.248)	(0.251)
FD	0.159	0.117	0.113
	(0.138)	(0.096)	(0.099)
GDPPCG	0.006	0.003	0.003
	(0.027)	(0.013)	(0.016)
INDUS	−0.019	−0.063	−0.072
	(0.327)	(0.230)	(0.237)
TO	−0.202*	−0.223***	−0.224***
	(0.108)	(0.080)	(0.080)
Constant	3.264**	3.744***	3.795***
	1.507	1.023	1.069
Number of observations	275	277	277
Number of countries	14	14	14

*, **, *** demonstrate significance level at 10%, 5%, and 1%, respectively.

The findings further denote that financial development exerts a positive impact on ATE in specifications 2–3, revealing the favourable influence of finance in lowering energy poverty. A 1 % increase in financial development brings about a 0.05% increase in ATE. Our result contradicts the findings of Zhao et al. (2023), which indicates that financial development exerts an insignificant impact on energy poverty reduction in Korea but confirms the results of Mohsin et al. (2022) and Said and Acheampong (2023) in the cases of Latin America and sub-Saharan Africa, respectively. Financial development positively enhances access to electricity because investment in electricity infrastructure requires funding (Acheampong et al., 2022; Zhang et al., 2019). A developed financial system will channel savings and allocate funds for electricity infrastructure investment. In addition, the results exhibit that the coefficient of industry is positive and significant. A 1 % increase in industry contributes to a 0.134% to 0.135 % increase in ATE. Finally, the findings reveal that trade openness has a significant negative impact influence on ATE. This outcome does agree with Zhao et al. (2022), who reveal that foreign trade increases energy poverty but contradicts the results of Yan et al. (2023), which confirm that trade openness reduces energy poverty in the case of China.

Table 8 (Panel B) presents the outcomes of the influence of education on energy poverty, proxied by access to clean fuels and technologies for cooking (CFTC). The empirical findings show that education is significantly and positively related to CFTC in all specifications, suggesting that improving Latin America’s level of education will efficiently expedite the process of elimination of energy poverty. A 1 % rise in education increases CFTC within a range of 1.349 %–1.454%. This mitigating effect of education on energy poverty illustrates Latin American countries’ progress towards sustainable development; these findings are in line with the results of Apergis et al. (2022) and Sharma (2016).

Education and Financial Development and Access to Electricity: Country-specific Results

As can be seen in Table 9, the connection between education and energy poverty reduction, measured using access to electricity (ATE), varies across countries. The findings indicate that the link between education and ATE is positively and significantly associated with 7 nations (Argentina, Bolivia, Costa Rica, Ecuador, El Salvador, Guatemala, Honduras, Mexico, Nicaragua, Panama, and Paraguay). Conversely, the outcomes display that this nexus is negatively and significantly related to the case of Colombia. In the cases of Brazil and Peru, however, an insignificant relationship exists between ATE and education. Additionally, the results point out that financial development has a significant positive impact on 8 countries. Financial development in Brazil, Colombia, Costa Rica, Ecuador, Guatemala, Honduras, Nicaragua, and Peru increases ATE. The outcomes show that the connection between financial development and ATE is negatively correlated with 6 nations (Argentina, Bolivia, El Salvador, Mexico, Panama, and Paraguay. Furthermore, the findings reveal that GDP per capita growth (GDPPCG) has a favourable impact on ATE in 3 nations (Guatemala, Mexico, and Panama), negative and significant in 3 nations (Colombia, Ecuador, and El Salvador), with no significant relationship in the remaining countries. The findings further reveal that industry has a beneficial effect on ATE in 6 countries. The results, however, find that industry has a significant negative impact on ATE in 4 nations. Finally, trade openness is positively and significantly related to 2 nations (El Salvador and Peru), negatively and significantly related to 8 nations, and insignificant for the rest of nations.

Table 9.

Country-specific Estimates on the Relationship Between Education, Financial Development, and Access to Electricity.

Variables	EDU	FD	GDPPCG	INDUS	TO	Constant
Argentina
FMOLS	0.367***	−0.009***	0.000	−0.007	0.005	4.235***
	(0.021)	(0.003)	0.000	(0.013)	(0.004)	(0.047)
CCR	0.357***	−0.013*	0.000	0.003	0.002	4.239***
	(0.028)	(0.007)	0.000	(0.024)	(0.008)	(0.065)
Bolivia
FMOLS	2.627***	−0.181***	−0.002	0.459***	−0.135***	1.529***
	(0.080)	(0.035)	(0.003)	(0.067)	(0.032)	(0.332)
CCR	2.628***	−0.191***	−0.001	0.450***	−0.146**	1.638***
	(0.101)	(0.054)	(0.006)	(0.089)	(0.071)	(0.564)
Brazil
FMOLS	0.037	0.05**	0.000	0.038*	−0.015	4.267***
	(0.043)	(0.021)	0.000	(0.020)	(0.010)	(0.077)
CCR	0.026	0.055**	0.000	0.033*	−0.013	4.265***
	(0.043)	(0.023)	0.000	(0.018)	(0.012)	(0.085)
Colombia
FMOLS	−0.364***	0.133***	−0.002***	−0.019***	−0.021**	4.540***
	(0.032)	(0.008)	0.000	(0.006)	(0.010)	(0.033)
CCR	−0.365***	0.133***	−0.002***	−0.018*	−0.021	4.536***
	(0.064)	(0.016)	(0.001)	(0.009)	(0.019)	(0.062)
Costa Rica
FMOLS	0.147***	0.018***	0.000	0.050***	−0.004	4.253***
	(0.031)	(0.002)	0.000	(0.009)	(0.003)	(0.045)
CCR	0.139**	0.019***	0.000	0.050***	−0.003	4.250***
	(0.056)	(0.004)	0.000	(0.012)	(0.005)	(0.056)
Ecuador
FMOLS	0.316***	0.008**	−0.001***	0.056***	−0.022**	4.130***
	(0.039)	(0.003)	0.000	(0.018)	(0.009)	(0.018)
CCR	0.317***	0.008*	−0.001***	0.058**	−0.022*	4.127***
	(0.027)	(0.005)	0.000	(0.022)	(0.012)	(0.021)
El Salvador
FMOLS	0.594***	−0.153***	−0.005***	−0.127***	0.138***	4.524***
	(0.008)	(0.013)	0.000	(0.016)	(0.007)	(0.104)
CCR	0.596***	−0.152***	−0.005***	−0.123***	0.143***	4.484***
	(0.011)	(0.016)	(0.001)	(0.022)	(0.012)	(0.167)
Guatemala
FMOLS	0.389***	0.065***	0.010***	0.228***	−0.366***	4.719***
	(0.053)	(0.013)	(0.001)	(0.034)	(0.034)	(0.122)
CCR	0.381***	0.065***	0.010***	0.233***	−0.374***	4.742***
	(0.054)	(0.017)	(0.002)	(0.060)	(0.053)	(0.087)
Honduras
FMOLS	0.742***	0.229***	0.000	−1.093***	−0.085*	6.877***
	(0.111)	(0.032)	(0.002)	(0.084)	(0.045)	(0.411)
CCR	0.705***	0.234***	0.000	−1.104***	−0.0972	6.980***
	(0.131)	(0.037)	(0.003)	(0.124)	(0.064)	(0.726)
Mexico
FMOLS	0.505***	−0.036***	0.001***	0.026***	−0.036***	4.292***
	(0.019)	(0.002)	0.000	(0.002)	(0.003)	(0.008)
CCR	0.502***	−0.036***	0.001***	0.027***	−0.035***	4.289***
	(0.017)	(0.002)	0.000	(0.004)	(0.004)	(0.016)
Nicaragua
FMOLS	1.151***	0.023***	−0.001	−0.007	−0.059***	3.731***
	(0.053)	(0.005)	0.000	(0.021)	(0.005)	(0.051)
CCR	1.182***	0.022***	0.000	−0.020	−0.061***	3.761***
	(0.094)	(0.007)	(0.001)	(0.040)	(0.006)	(0.097)
Panama
FMOLS	1.273***	−0.017*	0.001***	−0.011	−0.054***	3.557***
	(0.057)	(0.009)	0.000	(0.012)	(0.006)	(0.068)
CCR	1.273***	−0.019	0.001**	−0.010	−0.053***	3.560***
	(0.056)	(0.018)	0.000	(0.018)	(0.009)	(0.084)
Paraguay
FMOLS	0.338***	−0.020***	0.000	−0.169***	0.004	4.928***
	(0.027)	(0.005)	0.000	(0.030)	(0.022)	(0.173)
CCR	0.324***	−0.016*	0.001	−0.158***	0.013	4.851***
	(0.023)	(0.010)	(0.001)	(0.043)	(0.039)	(0.331)
Peru
FMOLS	−0.005	0.268***	−0.002	0.025	0.351***	2.107***
	(0.201)	(0.031)	(0.002)	(0.190)	(0.105)	(0.462)
CCR	−0.054	0.268***	−0.001	0.005	0.358**	2.200***
	(0.255)	(0.046)	(0.006)	(0.277)	(0.165)	(0.662)

*, **, *** demonstrate significance level at 10%; 5% and 1%, respectively.

Education and Financial Development and Access to Clean Fuels and Technologies for Cooking: Country-specific Results

Table 10 exhibits the findings of the effects of education and financial development on energy poverty, measured using access to clean fuels and technologies for cooking (CFTC). As shown in Table 10, education has a significant positive effect on CFTC in all countries in the sample. However, the link between financial development and CFTC varies throughout countries. For instance, the results show that financial development has a significant positive impact on CFTC in 6 countries (Brazil, Costa Rica, Honduras, Nicaragua, Paraguay, and Peru). In contrast, the results point out that financial development has a significant negative impact on the CFTC in 7 nations (Argentina, Colombia, Ecuador, El Salvador, Guatemala, Mexico, and Panama). In the case of Bolivia, no significant relationship is found. Additionally, the results show that GDPPCG exerts a positive influence on CFTC in 5 nations (Colombia, Guatemala, Mexico, Panama, and Paraguay), negative and significant in 2 nations (Argentina and Ecuador), and insignificant for the remaining countries. Furthermore, the findings reveal that the nexus between industry and CFTC is positively and significantly related in 5 nations (Bolivia, Brazil, Colombia, Ecuador, and Mexico), negative and significant for 4 countries (El Salvador, Honduras, Panama, and Paraguay); insignificant for the remaining countries. Finally, the findings suggest that open to foreign trade improves CFTC in 6 nations (Argentina, Bolivia, El Salvador, Nicaragua, Panama, and Peru). However, the linkage between trade openness and CFTC is negatively and significantly related to 4 nations (Brazil, Costa Rica, Guatemala, and Mexico).

Table 10.

Country-specific Estimates on the Relationship Between Education, Financial Development, and Access to Clean Fuels and Technologies.

Variables	EDU	FD	GDPPCG	INDUS	TO	Constant
Argentina
FMOLS	0.289***	−0.028***	−0.0002***	0.001	0.012***	4.333***
	(0.012)	(0.002)	0.000	(0.008)	(0.002)	(0.027)
CCR	0.283***	−0.031***	0.000	0.009	0.009**	4.331***
	(0.015)	(0.004)	0.000	(0.015)	(0.005)	(0.036)
Bolivia
FMOLS	1.916***	0.003	0.001	0.118***	0.028**	1.874***
	(0.034)	(0.015)	(0.001)	(0.028)	(0.014)	(0.142)
CCR	1.901***	−0.006	0.002	0.105***	0.020	1.994***
	(0.044)	(0.021)	(0.002)	(0.037)	(0.027)	(0.216)
Brazil
FMOLS	0.136***	0.048**	0.000	0.068***	−0.025**	4.071***
	(0.042)	(0.021)	0.000	(0.020)	(0.010)	(0.076)
CCR	0.126***	0.051***	0.000	0.063***	−0.025**	4.079***
	(0.036)	(0.019)	0.000	(0.016)	(0.010)	(0.074)
Colombia
FMOLS	0.936***	−0.008***	0.0003***	0.025***	−0.001	3.585***
	(0.011)	(0.003)	0.000	(0.002)	(0.004)	(0.012)
CCR	0.935***	−0.007	0.000	0.025***	0.000	3.583***
	(0.022)	(0.005)	0.000	(0.003)	(0.007)	(0.023)
Costa Rica
FMOLS	0.279***	0.039***	0.000	−0.011	−0.038***	4.309***
	(0.090)	(0.007)	0.000	(0.027)	(0.009)	(0.129)
CCR	0.301**	0.041***	0.000	−0.002	−0.039***	4.257***
	(0.136)	(0.010)	(0.001)	(0.032)	(0.013)	(0.147)
Ecuador
FMOLS	0.598***	−0.012***	−0.0002**	0.016***	−0.003	3.943***
	(0.011)	(0.001)	0.000	(0.005)	(0.003)	(0.006)
CCR	0.602***	−0.012***	−0.0002*	0.015**	−0.003	3.941***
	(0.008)	(0.001)	0.000	(0.007)	(0.004)	(0.007)
El Salvador
FMOLS	1.929***	−0.776***	−0.014***	−1.055***	0.392***	7.759***
	(0.041)	(0.068)	(0.002)	(0.083)	(0.039)	(0.546)
CCR	1.913***	−0.780***	−0.014***	−1.062***	0.371***	7.895***
	(0.057)	(0.095)	(0.003)	(0.128)	(0.060)	(0.940)
Guatemala
FMOLS	0.354*	−0.216***	0.013***	−0.154	−0.457***	6.624***
	(0.195)	(0.048)	(0.005)	(0.125)	(0.125)	(0.446)
CCR	0.280	−0.233***	0.021***	0.019	−0.637***	6.889***
	(0.204)	(0.056)	(0.008)	(0.210)	(0.200)	(0.341)
Honduras
FMOLS	0.603***	0.277***	0.003	−0.815***	−0.030	4.974***
	(0.170)	(0.049)	(0.003)	(0.130)	(0.070)	(0.632)
CCR	0.546***	0.285***	0.004	−0.802***	−0.0594	5.083***
	(0.195)	(0.056)	(0.005)	(0.187)	(0.098)	(1.098)
Mexico
FMOLS	0.554***	−0.048***	0.0003**	0.055***	−0.056***	4.121***
	(0.061)	(0.006)	0.000	(0.006)	(0.010)	(0.027)
CCR	0.535***	−0.047***	0.000	0.058***	−0.052***	4.105***
	(0.056)	(0.006)	0.000	(0.010)	(0.012)	(0.047)
Nicaragua
FMOLS	2.314***	0.032***	0.000	0.004	0.039***	1.791***
	(0.043)	(0.004)	0.000	(0.017)	(0.004)	(0.041)
CCR	2.307***	0.033***	0.000	0.008	0.039***	1.783***
	(0.076)	(0.006)	0.000	(0.032)	(0.005)	(0.078)
Panama
FMOLS	0.961***	−0.021***	0.0003**	−0.016**	0.024***	3.500***
	(0.036)	(0.006)	0.000	(0.007)	(0.004)	(0.043)
CCR	0.960***	−0.018*	0.000	−0.017*	0.023***	3.497***
	(0.033)	(0.010)	0.000	(0.010)	(0.005)	(0.054)
Paraguay
FMOLS	0.910***	0.114***	0.001**	−0.219***	−0.036	3.819***
	(0.039)	(0.007)	0.000	(0.043)	(0.032)	(0.248)
CCR	0.884***	0.114***	0.002**	−0.238***	−0.043	3.940***
	(0.033)	(0.013)	(0.001)	(0.057)	(0.054)	(0.439)
Peru
FMOLS	2.040***	0.412***	−0.003	−0.029	0.590***	−1.452**
	(0.312)	(0.049)	(0.003)	(0.295)	(0.163)	(0.716)
CCR	1.964***	0.413***	−0.002	−0.053	0.596**	−1.322
	(0.400)	(0.071)	(0.009)	(0.433)	(0.258)	(1.034)

*, **, *** demonstrate significance level at 10%, 5%, and 1%, respectively.

Robustness Checks

A further check was carried out to examine the sensitivity of the results to alternative estimation strategies and methods. We use the robust standard errors proposed by Driscoll and Kraay (1998) for panel regressions as a robustness check. The results in Table 11 are similar to those reported in Table 8.

Table 11.

Panel Estimation Results of the Impact of Education and Financial Development on Energy Poverty Reduction.

	Model 1	Model 2
	Fixed effects (Driscoll–Kraay)	Random effects (Driscoll–Kraay)
Panel A/Dependent variable: ATE
EDU	0.669***	0.616***
	(0.044)	(0.050)
FD	0.024	0.032*
	(0.015)	(0.015)
GDPPCG	0.000	0.001
	(0.001)	(0.001)
INDUS	0.143***	0.138***
	(0.028)	(0.027)
TO	0.000	−0.005
	(0.031)	(0.021)
Constant	3.355***	3.412***
	(0.107)	(0.138)
Number of observations	278	278
Number of countries	14	14
Panel B/Dependent variable: CFTC
EDU	0.872***	0.911***
	(0.081)	(0.079)
FD	0.112***	0.107***
	(0.033)	(0.025)
GDPPCG	0.000	0.001
	(0.001)	(0.001)
INDUS	0.082*	0.092***
	(0.043)	(0.023)
TO	0.096**	0.077**
	(0.044)	(0.035)
Constant	2.402***	2.436***
	(0.127)	(0.154)
Number of observations	278	278
Number of countries	14	14

*, **, *** demonstrate significance level at 10%, 5%, and 1%, respectively.

Conclusion and Policy Implications

This study examines the impact of financial development and education on energy poverty reduction using panel data from 14 Latin American countries from 2000 to 2019. This study deploys DOLS, FMOLS, and CCR as the main estimation techniques and further uses fixed effect and Driscol–Kraay estimators to test the robustness of the results. The findings of this study can be summarized as follows.

First, our analysis confirmed the existence of a positive and statistically significant effect of financial development and education on energy poverty reduction in Latin America.

Second, the results showed that education has a heterogeneous effect on reducing energy poverty across Latin American countries. For instance, the results indicate that education significantly increases access to electricity in Argentina, Bolivia, Costa Rica, Ecuador, El Salvador, Guatemala, Honduras, Mexico, Nicaragua, Panama, and Paraguay, while it significantly reduces access to electricity in Colombia. For Brazil and Peru, education does not affect access to electricity. Also, education was found to significantly increase access to clean cooking fuels and technologies across all the countries.

Third, the results showed that financial development has a heterogeneous effect on reducing energy poverty across Latin American countries. Financial development significantly increases access to electricity in Brazil, Colombia, Costa Rica, Ecuador, Guatemala, Honduras, Nicaragua, and Peru but significantly reduces access to electricity in Argentina, Bolivia, El Salvador, Mexico, Panama, and Paraguay. Also, financial development significantly increases access to clean cooking fuels and technologies in Brazil, Costa Rica, Honduras, Nicaragua, Paraguay, and Peru. In contrast, the results point out that financial development has a significant negative impact on clean cooking fuels and technologies in Argentina, Colombia, Ecuador, El Salvador, Guatemala, Mexico, and Panama. In the case of Bolivia, no significant relationship exists between financial development and clean cooking fuels and technologies.

These findings have two main implications for policies for attaining SDG 7 in Latin America. Firstly, the study calls on policymakers in the Latin America region to invest significantly in education since it contributes significantly to access to electricity and, clean cooking fuels, and technologies. Also, for education to act as an enabler of access to clean and modern energy, governments need to implement educational programmes that aim at training households on energy-saving strategies and also raise awareness on the benefits of using clean and modern energy such as LPG, natural gas, and solar energy. Consistent with the recommendation of Acheampong et al. (2022), for education to drive SDG 7, policymakers should prioritize technical education and train individuals to acquire the technical know-how needed to operate and maintain electrification infrastructure and to produce clean cooking technologies.

Secondly, the significant positive effect of financial development on electricity access underscores the need for policymakers to implement financial inclusion strategies to ensure easy financial access in the region. Easy access to finance is key to reducing energy deprivation as it enables poor households to access electricity and clean cooking technologies and fuels (Acheampong, 2023; Koomson & Danquah, 2021; Said & Acheampong, 2023). Also, this study calls on policymakers to implement policies that ensure the stability and efficiency of the financial system. A stable and efficient financial system is important since it enables clean cooking and off-grid technologies producers, distributors, and retailers to have easy access to finance to support their investment in clean energy and off-grid technologies that are instrumental for the attainment of SGD 7. Finally, financial institutions could support efforts to enhance access to clean and modern energy in Latin American countries by providing affordable loans to clean cooking and off-grid technology manufacturers as well as poorer households.

Despite the contribution of this study, there are still some avenues for future research. This study can be extended in many ways. While this study only focused on Latin American countries, future studies could examine the link between education, financial development, and energy poverty reduction in energy-poor regions such as sub-Saharan Africa and South Asia. Also, our study did not consider the role of institutional quality in the effect of financial development and education on energy poverty reduction. Therefore, future studies can benefit and advance knowledge by examining the impact of political institutions on the relationship between education, financial development, and energy poverty. Also, this study adopted a narrow definition of energy poverty, which emphasizes the lack of access to electricity and clean cooking fuels and technologies. To advance knowledge, future studies should adopt multidimensional energy poverty indicators to re-examine this problem considered in this study. Finally, a comparative analysis focussing on the level of the country’s economic development could provide a better understanding of the role of financial access and human capital in accessing clean and modern energy.