The Energy Poverty Status of Off-Grid Rural Households: A Case of the Upper Blinkwater Community in the Eastern Cape Province, South Africa

Abstract

This paper analyses the energy poverty status in off-grid rural households and its underlying socioeconomic factors. Employing the Foster–Greer–Thorbecke Technique and Probit regression on data from 53 households, the study uncovers a diverse array of energy sources in use, including firewood, paraffin, LPG, candles, and generators. Despite this energy source diversity, the poverty line threshold, as measured by the per capita energy expenditure line (92.40 ZAR) (1 US Dollar = ZAR 18.20), reveals the prevalence of energy poverty. Approximately 15% of respondents are experiencing severe energy poverty and 22% are facing moderate vulnerability to energy poverty, while over 50% are not energy poor. This indicates that, although they may lack access to electricity, their energy usage and expenditure in other forms might still be sufficient to meet their basic energy needs. This distinction highlights the importance of assessing energy poverty, extending beyond a simplistic assessment of absolute poverty but taking into account the dynamic nature of income levels. Gender, household size, formal education, and social grants emerge as key indicators shaping the energy landscape in the area. The results clearly indicate that male-headed households and larger households are less susceptible to energy poverty, while increasing formal education and social grants increases the risk of households being exposed to energy poverty. These findings suggest that the problem of energy poverty in the area is uniquely linked to social, economic, and cultural issues. Therefore, interventions targeted at addressing energy poverty problems must address the underlying social, economic, and cultural factors.

1. Introduction

Access to affordable, reliable, and sustainable energy services is a fundamental component of modern life, with significant implications for economic development, social well-being, and environmental sustainability. However, many communities worldwide continue to grapple with the harsh realities of energy poverty, a condition that goes beyond mere aggravation and has far-reaching consequences for the livelihoods and welfare of households. South Africa, despite its reputation for an advanced economy and substantial energy infrastructure, has not been immune to these challenges. The question of energy access remains a critical concern for policy and social development as disparities in access to reliable and affordable energy persist, giving rise to the emerging concern of household energy poverty [1].

South Africa has an electricity rate of about 86.15%, with rural and urban areas having rates of 85% and 87%, respectively, making it around 3.5 times better than the rest of sub-Saharan countries [2]. Limpopo province leads in electricity access with 98.55%, whereas the Eastern Cape has the lowest at 82.00% [1]. Despite the success of the electrification program in the country, there is a noticeable number of households experiencing energy poverty. Assessing energy-related expenditure as 10% of monthly income, the Multi-dimensional Energy Poverty Index (MEPI) reveals that about 43% of South Africa’s population falls into this category, underscoring the extensive prevalence of energy poverty [3]. There is also an estimated 3.5 million households that are not connected to the main grid and lack access to electricity [4]. Approximately 60 percent of these households are situated in rural and remote areas where expanding the main grid infrastructure poses substantial logistical and financial challenges [5]. According to the main electrical energy supplier “Electricity Supply Commission (Eskom)”, expanding the grid to reach these rural communities to is costly, as their remoteness, low population density, and low-income levels make it very difficult for Eskom to recover its capital and operating costs from the tariffs alone [6].

The Upper Blinkwater community, located in the Raymond Mhlaba municipality of the Eastern Cape Province, is emblematic of such off-grid rural communities, providing a unique perspective on energy poverty [7]. Alongside other off-grid communities in the country, it faces the prospect of grid access remaining distant, projected to be unavailable for the next 8 to 15 years. Recognizing this difficulty and attempting to close the energy gap, the indigent energy policy was introduced in 2003. The Free Basic Alternative Energy (FBAE) programme. This strategy intended to help communities like Upper Blinkwater, where there is no power infrastructure and grid expansion is unlikely due to ESKOM constraints. The FBAE provides a monthly subsidy of R56.29 for alternative energy sources such as paraffin, coal, liquefied petroleum gas (LPG), and ethanol gel to indigent households [8]. However, despite the sound intentions of these policies, they have not produced significant changes in addressing energy poverty, leading to substantial discontent among researchers and communities. One important concern is the lack of a clear and standardized procedure for municipalities to implement these policies, as mandated, alongside irregularities such as inconsistent updates of indigent registers and bottlenecks in the registration process. As a result, not all low-income households benefit from the policy [1], and only about 60% of eligible households are receiving the energy subsidy [9].

Furthermore, Solar Home Systems (SHS) were implemented as a suitable temporary alternative to grid electricity, with qualified families receiving an 80% capital subsidy. However, the practical implementation of such systems in rural communities encounters significant challenges due to their financial and socioeconomic conditions. Substantial initial investment costs for SHS, ongoing operational expenses, a lack of necessary infrastructure, limited education and awareness, limited economic opportunities, and the influence of social and cultural factors all contribute to the difficulties faced in bringing these systems to fruition. Moreover, the classification of households based on socioeconomic status exacerbates the issue, impeding many from accessing the support they need and resulting in the slow adoption of SHS [5].

In electrified places, because of poverty, people are not able to pay, which shows that energy poverty makes access to electricity unreliable. Meanwhile, households continue to grapple with the financial burden of securing alternative energy sources, and traditional fuels such as wood continue to remain a common practice, gathered and burned for cooking and space heating [7,10]. Mukumba and Chivanga [1] found that the Eastern Cape was among the top three provinces with a huge number of households using firewood for cooking. Relying on firewood not only poses significant health hazards due to indoor air pollution, but also has far-reaching consequences, particularly affecting the lives of women and children, who typically bear the responsibilities of cooking [11]. Similarly, the seemingly easy availability of firewood does not equate to a straightforward solution to their energy needs. For instance, women spend from approximately 2 to 9 h of their time gathering wood and managing cooking duties. This exhaustive routine limits their time for other productive activities, negatively impacting their overall quality of life [12,13]. This situation perpetuates gender inequalities and restricts opportunities for personal and economic development. Hence, women often comprise the majority of low-income people in developing countries [14,15,16]. Furthermore, the reliance on wood as a primary energy source is not only a matter of personal inconvenience, but also contributes to broader environmental challenges. The continued burning of biomass, such as wood, leads to indoor air pollution and associated health issues, causing more fatalities than tuberculosis, malaria, or other infectious illnesses [17]. For example, worldwide, approximately 4 million deaths each year are attributed to indoor household pollution, and burn-related injuries contribute to over 265,000 fatalities. Among those, 100,000 deaths involve children [18]. Moreover, the random felling of trees for fuel further exacerbates environmental degradation, contributing to broader ecological challenges [19].

Energy poverty is increasingly discussed as a subject of important research topics among scholars and an area of policymaking focusing on various dimensions, encompassing factors such as access, reliability, affordability, and socio-economic implications, particularly in terms of access to modern energy services. A study by [15] investigated energy poverty in South Africa, revealing that over 50% of the population was susceptible to energy poverty across all poverty indexes. The research identified that disparities in income distribution played a significant role in elevating energy poverty rates, with lower-income groups contributing more to overall poverty than their higher-income counterparts. In [20], it was reiterated that the prevalence of energy poverty is not just about access to electricity, but also about affordability. Their study showed that the majority of households using traditional fuels are not satisfied with their current energy situation, primarily due to issues of energy affordability. However, much of the existing research, particularly in South Africa and the Eastern Cape Province, including [2,11,15], has predominantly centred on electricity-related issues, leaving a significant gap in the literature when it comes to understanding the prevalence and state of energy poverty among marginalized communities residing off the grid, without access to electricity.

Access to and use of electricity is considered a global approach to measuring energy poverty, but this approach is subject to regional differences, especially in places not connected to electricity grids, notably in Africa. Owing to this, this research aims to close the gap in the literature by measuring the prevalence of energy poverty among off-grid households. This approach goes beyond the traditional focus on electricity availability, affordability, and reliability. It explores the broader spectrum of challenges faced by off-grid communities to understand energy poverty and measure progress towards its elimination. One key aspect that sets this research apart is that we use the energy consumption expenditure and employ the Foster–Greer–Thorbecke (FGT) approach to assess energy poverty, instead of using the fixed expenditure-based approach threshold of 10%, as [21] argues that this approach does not appear to be objective and comparable, especially where income levels differ greatly and change over time. Rather, FGT uses the energy poverty line, derived from the mean per capita consumption expenditure on energy use, representing the best indicator for measuring poverty which goes beyond the simple evaluation of absolute poverty [15]. Instead, it investigates the extent of energy poverty, allowing the identification of those vulnerable to energy poverty in the absence of electricity grid connections. This approach precisely determines how far below the energy poverty line they fall [22,23]. Drawing from [15], the FGT approach employs household energy poverty lines, defining a household as energy poor if it fails to meet the required household energy needs. Recognizing the direct link between energy poverty and socioeconomic advancement, the analysis not only categorizes households into different levels of energy poverty, but also investigates the underlying factors contributing to their energy poverty status. While these socioeconomic indicators do not directly measure energy poverty, they offer a more comprehensive understanding of correlated factors.

Understanding the state of energy poverty and the factors causing, worsening, or improving it may provide important insights for policy that aim to successfully address energy poverty and bring about transformative change in the lives of those living off the grid. For South Africa, this can serve as a gauge against which the government can assess its efforts in providing basic services and creating a harmonious environment for all citizens. This is particularly relevant given ongoing challenges and progress in service delivery in South Africa, where increasing local discontent and signs of impatience among the population regarding energy provision and access to electricity have been observed. In the body of literature, our research can serve as a foundation for future studies in several ways. Firstly, our findings provide a detailed understanding of the current energy poverty status for off-grid communities, offering a baseline for comparative analyses in other regions. Additionally, the identified socioeconomic factors contributing to energy poverty can guide researchers in designing targeted interventions and policies.

This paper is structured as follows: in Section 2, the brief background of energy access and challenges in South Africa is given. Following that, the methodology employed for data collection and analysis is described. Section 3 will present the findings of our research, and in Section 4 we will discuss the implications of these findings in the context of energy policy and community development. Finally, in Section 5 the paper concludes by summarizing the key findings and offering recommendations for future research and action.

2. Materials and Methods

2.1. Study Area

Upper Blinkwater is a small community situated near Fort Beaufort town in the Municipality of Raymond Mhlaba found in the Province of Eastern Cape, South Africa. The community is one of the less privileged rural communities in the province of the Eastern Cape, with sparsely scattered settlements, poor infrastructural development, and difficult accessibility. The community has approximately 67 households, among which the inhabitants are Xhosas and Coloureds [7].

2.2. Research Design and Sampling

The survey was conducted in November 2019. The quantitative methods were adopted from a cross-sectional approach. The cross-sectional research design is a useful tool in studying the complex dynamics of energy poverty, as it allows for capturing a snapshot of energy consumption patterns, expenditures, and the socioeconomic situations associated with energy poverty at a single point in time. This methodology is particularly advantageous for understanding how residents utilize energy resources and allocate their resources to meet energy needs. The design facilitates an in-depth examination of the financial aspects of energy consumption, shedding light on the economic burden imposed by energy costs. Furthermore, it enables comparisons across households, contributing to a nuanced understanding of the variations in energy poverty within the community. The practicality and efficiency of the cross-sectional approach make it a suitable choice for timely data collection.

Due to its small population size, drawing a sample may result in an even smaller sample size, making it difficult to draw any meaningful conclusions about the population. According to [24], as the population size becomes smaller and smaller, it is feasible to examine the entire population to avoid misrepresenting the data. It was therefore technically feasible for the study to examine all households in UB. However, depending on the willingness and availability of the respondents, only a total of 53 household heads were interviewed.

Data were collected using a questionnaire consisting of questions based on demographics, the type of energy used by the household for cooking, lighting and space-heating, and factors that influence energy use. The analysis of the collected data was carried out using STATA software version 15, employing descriptive statistics such as frequency distribution, percentages, and mean values.

2.3. Energy Poverty Status

In the analysis of energy poverty, two primary techniques are commonly employed. The first determines households’ energy poverty status based on energy consumption expenditure. The primary metrics and criteria used for measuring energy poverty in the study employ a unidimensional poverty measure of the FGT approach. This involves assessing households’ monthly consumption expenditure on different energy sources. However, this approach has faced criticism for potentially miscalculating energy poverty, especially among low-income households that may limit their energy consumption due to other financial priorities [15]. These households often rely on inexpensive biomass for their energy needs. As a result, estimating energy poverty in this way may lead to an underestimation of its prevalence in households. To address this concern, the FGT approach utilizes an absolute energy poverty line. To operationalize the FGT technique for energy poverty assessment, the study establishes the energy poverty line based on the mean per capita consumption expenditure on energy use within the community. By using this approach, the FGT technique considers the economic reality of the community, ensuring a locally relevant benchmark for energy poverty.

In alignment with [22,23], the energy poverty line is set at two thirds of the mean per capita consumption expenditure on energy use. This approach recognizes varying economic conditions within the community and avoids applying a one-size-fits-all threshold. That is to say, (i) any household whose monthly expenditure on energy is greater or equal to the poverty line is regarded as a non-energy-poor household; (ii) any households whose monthly expenditure on energy is less than the poverty line but greater than one third of the mean per capita expenditure on energy are regarded as moderately energy-poor households; and (iii) while those whose monthly expenditure on energy is less than one-third are regarded as extremely energy-poor households

2.4. Econometric Estimation of the Socioeconomic Factors Affecting Energy Poverty

The second technique involves determining the probability of households that are energy poor or not along with its determinants. Probit or Logit regressions are utilized for this analysis. In this context, the dependent variable is binary, taking the value of 1 if the household or individual is considered as having a low economic income and zero otherwise. In light of the above, suffice it to say that both extremely low-income and moderately low-income households can be generally classified as energy-poor households, while the counterparts are the non-energy-poor households. This conceptualization permits the use of a binary regression model to examine the factors influencing energy poverty status in the second stage. The Probit model is chosen for its ability to constrain estimated probabilities between 0 and 1, relaxing the constraints present in the Linear Probability Model (LPM) [15].

The Probit model assumes the existence of a latent, unobserved continuous variable, denoted as Y*, that determines the observed binary variable Y. Considering the per capita consumption expenditure, determining whether a household is above or below the energy poverty line. The Probit regression model used is expressed following [23] as follows:

$$\mathrm{P}\left(\mathrm{Y}=1\right)=\mathsf{\Phi }\left(\mathsf{\beta }\mathrm{X}+\mathsf{\epsilon }\right)$$

$$\mathrm{P}\left(\mathrm{Y}=0\right)=1-\mathsf{\Phi }\left(\mathsf{\beta }\mathrm{X}+\mathsf{ϵ}\right)$$

P(Y = 1) represents the probability that a household is below the energy poverty line (indicating poverty incidence) and Y* signifies the latent variable indicating whether a household’s expenditure is below the consumption poverty line, as indicated by Equation (1).

$${\mathrm{Y}}^{*}={{\displaystyle \sum }}_{\mathrm{i}=1}^{\mathrm{i}=8}{\mathsf{\beta }}_{\mathrm{i}}{\mathrm{X}}_{\mathrm{i}}+\mathsf{\epsilon }={\mathsf{\beta }}_1{\mathrm{X}}_1+{\mathsf{\beta }}_2{\mathrm{X}}_2+{\mathsf{\beta }}_3{\mathrm{X}}_3+\dots +{\mathsf{\beta }}_8{\mathrm{X}}_8+\mathsf{\epsilon }$$

(1)

The vector ${\mathsf{\beta }}_{\mathrm{i}}$ encompasses parameters to be estimated, ${\mathrm{X}}_{\mathrm{i}}$ denotes socioeconomic and institutional factors, Φ represents the cumulative distribution function of the standard normal distribution, and ε represents the stochastic error term (where ε ~ N(0, 1)). Additionally, Y can be viewed as an indicator for whether the latent variable is positive:

Y = 1 (Y* > 0),  1 if Y* > 0, i.e., (ε < β X₁), 0 otherwise.

In simpler terms, the observed variable Y takes a value of 1 if the latent variable Y* is greater than 0, and the specific condition involving ε, β, and X₁ is satisfied. If Y* is not greater than 0 or the condition is not met, Y is set to 0.

Explicitly, the model can be expressed as:

Y = Energy poverty status of households (1 for energy-poor households; 0 otherwise)

X₁ = Gender of household head

X₂ = Age of household head

X₃ = Household size

X₄ = Formal education status

X₅ = Employment status

X₆ = Dwelling type

X₇ = Log of social grants amount

X₈ = Index of non-productive assets

The selection of socioeconomic indicators is guided by the previous literature on energy poverty and socioeconomic development [25]. These indicators contribute to a more complete understanding of the complex socio-economic landscape surrounding energy poverty. Furthermore, a post-estimation analysis (marginal effects) was carried out for ease of interpretation, as expected.

3. Results

3.1. Descriptive Results of Energy Sources

Table 1. Descriptive results of the existing energy sources among respondents.

Energy Source	Energy Source				Energy Usage
	Freq.	Percent	Mean	Std. Dev.	Cooking	Heating	Lighting
Wood	51	96.2	0.96	0.192	73.6	79.2	-
Paraffin	49	92.5	0.92	0.267	41.5	18.9	98.1
LPG	45	84.9	0.85	0.478	79.2	-	-
Candles	50	94.3	0.94	0.233	-	-	94.3
Diesel (Generator)	7	13.2	0.13	0.361	-	-	7.5
Solar	2	3.8	0.04	0.342	-	-	3.8

presents the descriptive findings detailing the current energy sources utilized for domestic purposes in the Upper Blinkwater community. That is, firewood, paraffin, liquefied petroleum gas (LPG), candles, and generators. Predominantly, firewood emerged as the most widely utilized energy source (n = 15 households, 96.2%, SD = 0.192), serving the dual purposes of space heating for 79.2% of households and cooking for 73.6% of respondents. Paraffin ranked second in prevalence, employed by n = 49 households (92.5%, SD = 0.267). Its primary applications included lighting (98.1% of households), cooking (41.5% of households), and space heating (18.9% of households). Candles were predominantly used for lighting in approximately n = 50 households (94.3%, SD = 0.233). Despite some adoption of cleaner energy sources such as diesel for generators and solar energy, the community’s reliance on these remained negligible. Only n = 7 households (13.2%, SD = 0.361) utilized a diesel generator, and a mere n = 2 households (3.8%, SD = 0.342) employed solar energy for lighting and charging cell phones.

The descriptive findings highlight a prevalent reliance on traditional energy sources, particularly firewood and paraffin, for essential domestic functions in the Upper Blinkwater community. The limited adoption of cleaner energy sources suggests potential barriers or preferences for traditional methods despite the availability of alternative technologies.

3.2. Descriptive Results of the Socioeconomic Characteristics

The descriptive results of the key variables related to the energy poverty of the respondents are presented in

Table 2. Descriptive results of key demographic and socioeconomic characteristics.

Variable	Mean	Std. Dev.	Min.	Max.
Age	52.4	17.57348	24	101
Household Size	3.4	2.061641	1	8
Wages	552.83	967.6704	0	3500
Grant	1093.6	1310.334	0	5500
Remittance	454.72	1073.02	0	4300
Own Business	220.75	1127.516	0	8000
Total Monthly Income	2321.9	1990.098	0	8890
Wood Expenditure	48.5	21.69688	0	100
Paraffin Expenditure	50.4717	49.86274	0	200
Liquefied Gas Expenditure	104.15	107.3892	0	375
Candles Expenditure	34.415	21.56692	0	125
Generator Expenditure	10.23	33.83821	0	150
Total Energy Expenditure	247.75	160.485	25	814
Per Capita Energy Expenditure	92.40	78.16768	10	316

. The data reveal that, on average, the respondents were approximately 54.4 years old, with a standard deviation of 17.5, suggesting a diverse age range but a prevalent middle-aged demographic. The household sizes among the respondents were relatively small, averaging 3.4 members, with a standard deviation of 2.061641, indicating some variability in household sizes.

Income sources among the respondents varied, with those who worked earning an average monthly salary of about 552.83 ZAR and various contributions, such as 454.72 ZAR from remittances and at least 220.75 ZAR a month for those who were self-employed. The data further show that their total monthly income amounted to at least 2321.9 ZAR, with the highest income reaching 8890 ZAR, while some had no income at all. This indicates disparities in the financial resources among the respondents.

The detailed analysis of energy expenditures illustrates the respondents’ reliance on a mix of traditional and conventional energy sources. The respondents spent approximately 48.49 ZAR of their income per month on acquiring wood, 50.47 ZAR per month on paraffin, and an average of 104.15 ZAR per month on LPG. This indicates that some households in the community had access to cleaner energy sources, which can be more efficient and environmentally friendly. Additionally, candles were used for lighting and incurred an average expenditure of approximately 34.42 ZAR per month, while generator expenditure was relatively low, with an average of 10.23 ZAR, suggesting that only a few households owned and operated generators.

The overall energy expenditure averaged 247.75 ZAR per month, indicating a significant a substantial financial burden in meeting their energy needs. On a per capita basis, the average energy expenditure was 92.40 ZAR, which gives an indication of the economic challenges within households. Thus, the per capita energy expenditure line was (92.40 ZAR) (1 US Dollar = ZAR 18.20).

3.3. Energy Poverty Status of the Respondents

Table 3. Energy poverty status of the respondents.

Energy Poverty Level	Frequency	Percentage
Extremely Energy Poor	8	15.1
Moderate Energy Poor	12	22.6
Non-Energy Poor	33	62.3
Total	53	100

illustrates the distribution of energy poverty levels among the respondents, indicating the varying class of energy poverty within the off-grid households. Table 3 distinguishes respondents into three distinct groups based on the severity of their energy poverty status: “Extremely Energy Poor”, “Moderate Energy Poor”, and “Non-Energy Poor”.

The data reveal that 15.1% of the respondents’ fell into the category of “Extreme Energy Poor”. This suggests that the respondents in these households had a heavy reliance on traditional and less reliable energy sources, indicating a high level of vulnerability due to their limited financial resources to meet their basic energy needs. The implication is that these respondents faced significant challenges in meeting their basic energy needs, potentially impacting various aspects of their daily lives and well-being.

About 22.6% of the respondents were categorised as being in the moderately energy-poor group, suggesting that they faced a less severe yet still substantial level of energy hardship. This means that, while they might have had relatively better access to energy sources, they still encountered notable energy access challenges and therefore still relied on traditional energy sources.

The majority of respondents, totalling 62.26%, were found to be non-energy poor, meaning that they experienced relatively better access to energy services and a reduced reliance on traditional or less reliable energy sources, which implies a considerable adoption of conventional energy sources. This majority reflects a positive sign, indicating that a significant portion of the community enjoyed a relatively favourable energy situation.

3.4. Energy Poverty Levels and Related Respondents’ Characteristics

Table 4. Energy poverty levels and socio-economic features of the respondents.

	Energy Poverty Level
Variable	Extremely Energy Poor		Moderate Energy Poor		Non-Energy Poor		Total
Gender
Male	2	(3.8%)	7	(13.2%)	14	(26.4%)	23	(43.4%)
Female	6	(11.3%)	5	(9.43%)	19	(35.9%)	30	(56.6%)
Age
≤30	1	(1.9%)	1	(1.9%)	4	(7.55%)	6	(11.3%)
31–40	1	(1.9%)	1	(1.9%)	5	(9.43%)	7	(13.2%)
41–50	1	(1.9%)	1	(1.9%)	8	(15.1%)	10	(18.87%)
51–60	5	(9.43%)	6	(11.3%)	6	(11.3%)	17	(32.1%)
≥61	0	(0%)	3	(5.66%)	10	(18.87%)	13	(24.53%)
Education
No Schooling	3	(5.66%)	1	(1.9%)	4	(7.55%)	8	(15.1%)
Primary	3	(5.66%)	9	(17%)	19	(35.9%)	31	(58.49%)
Secondary	2	(3.77%)	2	(3.77%)	9	(17%)	13	(24.53%)
Tertiary	0	(0%)	0	(0%)	1	(1.9%)	1	(1.9%)
Household Income Sources
No Income	2	(3.77%)	3	(5.66%)	3	(5.66%)	8	(15.1%)
Wages	1	(1.9%)	1	(1.9%)	13	(24.5%)	15	(28.3%)
Grant	5	(9.43%)	9	(17%)	22	(41.5%)	36	(67.9%)
Remittance	1	(1.9%)	1	(1.9%)	7	(13.2%)	9	(17%)
Own Business	0	(0%)	0	(0%)	4	(7.5%)	4	(7.5%)
Total Monthly Income
≤1000	6	(11.3%)	4	(7.55%)	5	(9.43%)	15	(28.3%)
1001–3000	2	(3.77%)	7	(13.2%)	12	(22.6%)	21	(39.6%)
3001–5000	0	(0%)	1	(1.9%)	11	(20.7%)	12	(22.6%)
≥5001	0	(0%)	0	(0%)	5	(9.43%)	5	(9.43%)

presents an analysis of the demographic and socioeconomic factors associated with the varying levels of energy poverty among the respondents. The table distribute the respondents into their three distinct groups, as seen above, with each representing a different level of energy poverty.

The findings revealed that a larger percentage of females (35.9%) did not experience energy poverty, compared to 24.6% of males. This suggests that, in general, a higher proportion of females had more reliable access to energy sources. However, among those experiencing extreme energy poverty, females were more represented at 11.3%, while only 3.8% of the male respondents fell into this category. On the other hand, a higher percentage of males (13.2%) were moderately energy poor compared to a smaller proportion of females (9.43%) falling into this group. This reveals that, while females were more prone to extreme energy poverty, males were more prevalent in the moderate energy poverty category.

The analysis of age reveals that respondents aged 51–60 constituted the majority (approximately 9.43%) of those experiencing extreme energy poverty. This age group also accounted for 11.3% of the moderately energy poor category. In contrast, other age groups, especially those above 60 years, had lower proportions (about 1.9% and 5.66%) in both the extreme and moderate energy poverty categories compared to energy poverty compared to those in other age brackets. This suggests that the respondents in the 51–60 age group were more susceptible to energy poverty than the respondents in other age brackets.

Education and energy poverty: Respondents with primary education or no education dominated all levels of energy poverty. Approximately 5.66% of those without education and those with primary school education were classified as extremely energy poor, in contrast to 3.77% of those with secondary education. None of those with tertiary education fell into the extremely energy poor category. Additionally, those with primary education were more likely to be moderately energy poor compared to the other groups. These findings suggest that respondents with lower levels of education were more vulnerable to severe energy poverty than those with secondary education, and individuals with tertiary education were less affected by energy poverty. These results indicate the significant relationship between educational attainment and energy poverty, with higher education levels correlating well with not experiencing energy poverty.

When considering household income sources, it was evident that respondents who received grants were predominantly affected by all levels of energy poverty. Approximately 9.43% were classified as extremely energy poor, a rate even higher than those with no income, comprising 3.77%, and those receiving a salary at 1.9%, as well as remittances at 1.9%. Furthermore, the grant recipients constituted 17% of the group moderately affected by energy poverty, in contrast to 5.66% of those without income, and 1.9% of those earning a salary or receiving remittances. On the other hand, individuals who were self-employed were not susceptible to either extreme or moderate energy poverty. This implies that, while those dependent on grants were vulnerable to severe energy poverty levels, they also made up the largest group of non-energy-poor individuals.

Furthermore, when examining the distribution of respondents by their income levels, it becomes apparent that respondents with monthly incomes below 1000 ZAR were severely affected by energy poverty. Those with incomes between 1001 and 3000 ZAR were moderately energy poor, while those with incomes above 3000 ZAR were either only moderately exposed to energy poverty or, in the case of those with incomes above 5000 ZAR, not affected by energy poverty at all. These results suggest that a higher income or greater financial resources enable individual households to access cleaner energy sources more easily than those in lower-income groups.

3.5. Estimates of Energy Poverty Situation

Table 5. Regression estimates of energy poverty situation.

Energy-Poverty	Coefficient	AME (dy/dx)	Std. Err.	z	p > |z|
Constant	−10.8036	--	--	--	--
Gender	−1.2911	−0.3030	0.1648	−1.84 ***	0.066
Age	−0.0122	−0.0028	0.0061	−0.46	0.643
Household size	−0.5417	−0.1271	0.0366	−3.47 *	0.001
Educational status
No formal education	0.4537	0.0988	0.1806	0.55	0.584
Formal education	1.9484	0.4152	0.2371	1.75 ***	0.080
Employment status	−0.0792	−0.0185	0.2241	−0.08	0.934
Dwelling type (base = mud housing)
Mud-cement housing	−0.0379	−0.0077	0.1799	−0.04	0.966
Brick housing	−1.1397	−0.2461	0.1603	−1.53	0.125
Index of non-productive assets	0.0787	0.0184	0.0486	0.38	0.704
Log of social grant amount	2.0242	0.4750	0.1153	4.12 *	0.000

Note: dy/dx for factor levels is the discrete change from the base level; AME—Average Marginal Effects; *** and *—Significance level at 10% and 1%, respectively. Source: Data analysis, 2023.

presents the estimates from a regression model used to examine the effects of some independent variables on households’ energy poverty status in the study area. However, the interpretations, as expected, rely on the average marginal effects estimates (dy/dx) obtained from the post-estimation.

From the results, the coefficient for gender is negatively significant with households’ energy poverty status (−0.303, p < 0.1), suggesting that, on average, for a change in gender, specifically being a male, there was a decrease of 0.303 in the probability of being energy poor. In practical terms, this suggests that, within the context of the study, being in a male-headed household was linked with a reduced likelihood of facing energy poverty compared to households headed by females. The negative coefficient suggests that ceteris paribus, there is a protective or beneficial effect associated with having a male household head in relation to energy poverty. This finding may reflect societal roles and expectations wherein male-headed households, compared to women who are ascribed household chores roles, men are often considered as the primary breadwinners or decision makers, and are better positioned to afford and provide for basic energy needs, thus decreasing vulnerability to energy poverty.

For the age of household head, the results indicated that, on average, as ageing set in, there was a decrease of approximately 0.0029 in the probability of being energy poor. However, this change or relationship was not statistically significant at any significant level. Therefore, changes in age are unlikely to be economically meaningful in explaining the variations in the probability of the being energy poor or non-poor. This means that the observed decrease in the probability of energy poverty associated with aging is so small that it cannot be considered as economically meaningful or significant. Therefore, the age of the household head, based on these results, does not appear to be a substantial or influential factor in explaining the variations in the likelihood of being energy poor or non-poor. In terms of household size, the result returned a negative estimate, which invariably suggests that, on average, for every one-unit change in household size, there was a decrease of approximately 0.1271 in the probability of experiencing energy poverty. This relationship was statistically significant at p < 0.01 probability level. This could be as a result of responsibility sharing among the economically active household members. As these members contributed collectively to meeting basic energy needs, it helped to lift the household away from the cycle of energy poverty. The implication is that a larger number of individuals within a household might contribute to a more efficient allocation of resources and efforts, making it less likely for the household to face challenges related to energy poverty. The result reveals the potential benefits of a larger household size in fostering a cooperative and resourceful approach to addressing energy needs.

Furthermore, educational status exerted a positive and significant relationship (−0.303, p < 0.1) with households’ energy poverty status. Specifically, having formal education had a positive and significant relationship (4152, p < 0.1) with households’ energy poverty status, suggesting that, with a unit increase in households being educated, the more likely they were to be energy poor than those with no formal education. These results completely go against expectations, and a good reason why this appears is the strong cultural attachment to the use of traditional and accustomed (coal and/or wood) energy sources in this society. It is possible that households with higher levels of formal education may still maintain a preference for these traditional energy sources due to cultural practices or preferences.

The coefficient for employment status indicates that, on average, a one-unit change in employment status (being employed) was associated with a decrease of approximately 0.0186 in the probability of being energy poor. Moreover, this change status did not appear to have a statistically significant impact on the probability of being energy poor or non-poor. The observed decrease in the probability of energy poverty with employment status change suggests a potential positive correlation between employment and energy well-being. That is, individuals with employment may have a more stable income, making it easier for them to meet their basic energy needs, and thus decreasing their likelihood of falling into energy poverty. The lack of statistical significance, however, indicates that this relationship may not be strong enough to draw definitive conclusions about its impact on energy poverty status, implying that other factors, not captured in the analysis, may contribute to the overall dynamics of energy poverty.

In the same manner, these coefficients attached to dwelling unit types were not statistically significant at any probability level. This simply suggests that the dwelling type (be it mud housing, mud-cement, and brick housing type) households live in did not have any significant influence on households’ energy poverty status. The same can be said of households’ assets, because the findings revealed that asset index did not have any statistically significant influence on the probability that households would be energy poor or not. The lack of statistical significance in these coefficients suggests that, within the studied context, the physical characteristics of the dwelling units and the overall asset wealth of households did not play a substantial role in determining whether a household was categorized as energy poor. This implies that, despite variations in dwelling types and asset levels, these factors may not be strong indicators of households’ vulnerability to energy poverty.

On the other hand, the coefficient for the natural log of social grants amount had a positive effect on households’ energy poverty status, and was significant statistically (0.4751, p < 0.01). This suggests that, on average, a one-unit change in the amount of grants accessed by households was likely to induce an increase of approximately 0.4751 in the probability of being energy poor. The unexpected association between larger grant amounts and a higher probability of experiencing energy poverty, contrary to a priori expectations, raises significant concerns about the effective utilization of social grants by beneficiary households. This finding suggests that, despite receiving more substantial financial assistance through grants, households were not experiencing the anticipated improvement in their energy well-being. Instead, the larger grant amounts appeared to be linked to an increased likelihood of energy poverty.

4. Discussion

Based on the study’s findings, the per capita energy expenditure reveals the prevalence of energy poverty among households in the Upper Blinkwater. The results indicate that, on average, these households spend a total of 247.75 ZAR per month on energy. When compared to other low-income households and even some middle-income households, such as those in the Delft and Joe Slovo areas, who spend between ZAR 158 and 178 for gas and 187 for paraffin monthly [8], it is evident that Upper Blinkwater households face a significant financial burden in meeting their energy requirements. Notably, this expenditure exceeds even the electricity spending of low-income households in the Delft and Joe Slovo areas, reported to be slightly above R100 a month [8]. However, on overall, majority of these households are not classified as energy poor. Therefore, this energy expenditure may reflect a favourable energy situation with reduced reliance on traditional or less reliable sources.

The distribution of socioeconomic indicators revealed some interesting results. Specifically, the gender dimension of energy poverty in this study reveals that male-headed households are less likely to experience energy poverty, reflecting distinct gender roles. Within households, women are traditionally assigned responsibilities centred around the kitchen and resource access, while men, if economically active, are tasked with providing for the household [2,15]. This assignment of roles means male-headed households are less susceptible to energy poverty. In contrast, women often face restricted access to resources and decision-making power, which limits their influence over energy-related matters [2]. Consequently, energy poverty disproportionately affects women and girls, as they are often stereotyped as the primary users of unclean energy sources.

The findings related to the age of household heads challenge the common belief that older individuals are more susceptible to energy poverty, contrary to [26]. The results suggest that older household heads are more likely to have secure access to energy compared to their younger counterparts. However, a non-significant coefficient implies that the variable of “age” may not have had a direct impact on the energy poverty status of the respondents. Instead, other characteristics of household heads or household members may have been more influential. For instance, in this area and other rural regions of South Africa, social grants provided to older citizens may play a significant role in offering financial stability, enabling them to afford cleaner energy sources like paraffin and LPG. Thus, age may not directly affect energy poverty; the type and amount of grants they become eligible for as they age may have a more substantial role in determining a household’s energy security.

Contrary to the perspectives of [15,27], who viewed household size and energy poverty from a consumption angle, suggesting that larger households consume more resources, this study presents a different view. It reveals that larger households are less likely to experience energy poverty. As Ismail and Khembo [27] notes, larger households often benefit from the economic contributions of more members, leading to a higher income and better access to modern energy sources, which effectively meets their energy needs.

Surprisingly, the study finds that higher educational attainment increases susceptibility to energy poverty, challenging the expectation that education should reduce energy poverty. Higher education is often associated with improved job prospects and higher incomes, enabling households to afford modern energy sources and alleviate energy poverty [28]. The findings suggest that education fosters energy awareness among educated individuals, making them conscious of the inefficiencies of their current energy sources. The absence of electricity, a modern energy source, exposes them to energy poverty, supporting energy transition theories.

Conversely, Bardazzi and Pazienza [26] argue that deeply rooted traditions and cultural norms dictate the use of specific energy resources within societies, often resisting change, even when individuals are educated and economically stable. This cultural adherence may overshadow the potential for adopting more modern and efficient energy alternatives, contributing to the increased likelihood of being categorized as energy poor.

The study further reveals that employment significantly reduces the risk of energy poverty, while unemployment increases vulnerability. Stable employment provides a reliable source of income, enabling individuals and households to afford essential energy services and technologies, aligning with energy transition theories. Koomson et al. [29] also noted that employment precarity can lead to financial instability, making energy expenses a substantial burden for those in precarious employment.

Regarding dwelling type, whether mud, mud-cement, or brick housing, does not significantly impact energy poverty. However, individuals living in brick or mixed houses are less vulnerable to energy poverty than those in mud houses. In agreement with [30], the type of house often reflects socio-economic status, influencing overall well-being. Modern brick houses signify higher living standards and access to essential resources, while traditional mud houses imply lower incomes and limited resources. Hence, respondents living in more modern houses are less likely to experience energy poverty compared to those in mud houses.

Similarly, household assets reflect wealth and living standards [31]. This implies that individuals with greater wealth have the freedom to select energy sources, often choosing cleaner options. However, in Upper Blinkwater, respondents with more assets are still susceptible to energy poverty, suggesting factors like tradition and cultural attachment may play a more substantial role in influencing energy poverty outcomes.

The findings indicate that households receiving larger grant amounts are more prone to energy poverty. At first glance, this may appear counterintuitive, since social grants are aimed at reducing poverty [32], and it would be expected that increased income would alleviate energy poverty, suggesting that higher grant amounts might lead to a heightened demand and affordability of energy, including cleaner sources. This is substantiated by the descriptive analysis, which indicates that the families receiving grants were largely not experiencing energy poverty. In this context, the findings align with the hierarchy of needs and the energy ladder theory. Both theories posit that individuals prioritize their needs, which evolve with their income levels [25]. The argument is that, as grant recipients’ financial situations improve with each additional grant amount, there is a shift in their energy demand from traditional sources to modern ones, ultimately transitioning towards cleaner sources like electricity. However, the findings suggest otherwise, indicating that, despite receiving larger amounts of grant money, households remain energy poor. The increasing grant amount may imply an increasing number of grant recipients, which may also imply a rise in household members, resulting in an increased demand for energy and, consequently, a higher susceptibility to energy poverty.

5. Conclusions

This paper aimed to determine the prevalence of energy poverty among households in the Upper Blinkwater community. The findings indicate that, based on the per capita energy poverty line ZAR 92.40, energy poverty was prevalent among households in the Upper Blinkwater. However, only a smaller proportion of households lived in severe energy poverty, which also indicates a positive shift toward more sustainable and reliable energy sources. Overall, the study highlighted that households in Upper Blinkwater may be electricity poor due to their off-grid status, although about 37% were energy poor, with 15% living in extreme energy poverty; the majority of households may not be considered as energy poor. That is, despite their lack of access to electricity, their energy expenditure and usage in other forms might still be sufficient to meet their basic energy needs. In light of the aforementioned, while electricity is undoubtedly an essential component in the broader context of the country’s energy poverty, the challenges faced by the Upper Blinkwater go beyond a simple lack of access to electricity. Arguably, the evidence and observations suggest that the problem of energy poverty in the area is uniquely linked to social, economic, and cultural issues. By addressing these underlying factors, interventions can be more targeted and sustainable.

Several key findings challenge conventional assumptions while emphasizing the importance of specific factors in understanding and addressing energy poverty. Notably, gender, household size, education, and social grants are variables that significantly explain the prevalence of energy poverty in the Upper Blinkwater community.

The results clearly demonstrate that the gender–energy-poverty dynamic can be attributed to gendered divisions of labour within households, where women often bear responsibility for energy-related tasks. Larger households are less likely to experience energy poverty due to the economic contributions of more members, resulting in higher incomes and improved access to energy sources. The unexpected finding that higher education levels correlate with higher energy poverty, underscoring the role of traditions and cultural norms in energy usage patterns. Cultural factors often lead individuals to use traditional energy resources, regardless of their educational and economic circumstances.

These findings underscore the importance of understanding these complex relationships in efforts aimed at addressing energy poverty phenomenon and enhancing the well-being of vulnerable households. Such measures should include gender-inclusive energy programs that empower women, support for larger households to ensure access to modern energy sources, and a focus on cultural sensitivity when implementing energy transition initiatives. Education on energy efficiency and the monitoring and evaluation of programs and community engagement in the design and implementation of such programs should be key components of the strategy. Further research will be needed in three aspects so as to completely understand energy dynamics in the Upper Blinkwater rural community:

• Analysis of energy dynamics: further research on the specific challenges and advantages associated with each energy source, considering factors such as availability, cost, environmental impact, and community preferences.
• Longitudinal assessment energy poverty dynamics, that is, a research that tracks changes in the energy poverty over time, the establishment of effectiveness of the interventions, and the impact of external factors, etc.
• Cultural and social factors in energy poverty, to explore community-specific beliefs, beliefs, practices, and social structures that shape energy access and energy patterns.

Although addressing electricity poverty should be a priority, simply providing access to electricity may not be sufficient, but understanding the interplay between social dynamics, economic conditions, and cultural aspects is key. That is, any intervention should consider and address the underlying social, economic, and cultural factors contributing to energy poverty in an area. However, addressing electricity poverty remains crucial, as it directly stimulates economic opportunities, fostering job creation and general well-being and the development of individuals and communities. Most importantly for South Africa, it promotes social inclusion and reduces disparities.

The study, however, faces data limitations due to the small size of the community. While the findings provide valuable insights, caution is necessary when generalizing the results due to the inherent data limitations. Furthermore, the restricted data may compromise the statistical power of the study, hindering the drawing of definitive conclusions regarding the state of energy poverty in the community. To address these limitations, future research should consider adopting a mixed-methods approach that integrates both qualitative and quantitative data, incorporating a more diverse range of variables that could impact energy access, which is essential. This will likely provide a richer and more dynamic understanding of the factors influencing energy poverty, even beyond the Upper Blinkwater community.