Determinants of Cleaner Cooking Fuel Choice: A Study of Two Riparian Communities in Ghana

1. Introduction

Globally, households decide on energy consumption by choosing a mix of diverse energy sources simultaneously due to changes in socioeconomic conditions, settings, and varying end uses [1].

In Ghana, due to a lack of access to clean and reliable sources of energy, the cost of clean fuel or the quantity of inexpensive wood fuel in the surrounding area, many households rely extensively on energy sources derived from solid fuels [2]. Solid fuels are a type of fuel that can be comprised of coal, biomass, charcoal, wood, straw, animal dung, and/or crop waste [3]. They are utilized for domestic activities such as cooking, heating, lighting, boiling water, and generating income. Three billion of the world’s poorest people live in developing countries, primarily in rural regions, and rely on solid fuels as a primary source of energy [4,5]. Wood fuel is the primary source of energy for the household (mostly for cooking), despite the possible health concerns that this form of energy can potentially bring, especially when it is used inside with microscopic soot particles that travel deep into the lungs [6]. More than 60% of households in Africa and Southeast Asia rely on solid fuels for home cooking. In Sub-Saharan Africa, about 80% of the population uses biomass for energy [7].

There are several potential benefits to households transitioning to modern clean fuels, including reduced time spent on cooking and cleaning [8]. This can be explained by the higher opportunity cost of gathering and producing fuelwood. Studies have shown that households use more wood fuel when they don’t have difficulty fetching it [9]. The energy ladder hypothesis postulates that when a family’s income rises, it will switch from using traditional biomass fuels to more contemporary alternatives. The term “energy transition” describes this change [10], which forms the basis for household fuel choice.

1.1. Energy Ladder & Fuel Stacking Concepts

The idea of the energy ladder started to emerge in parallel to the fuel-wood crisis in the 1970s and 1980s [11]. Scholars and energy experts all over the world have employed it to explain the transition in household fuel choices. The energy ladder was designed to depict a hierarchical relationship between household economic status and fuel choice for cooking and heating [12]. As a household’s income and socio-economic situation improve, it switches from dirty fuels (biomass) to intermediate fuels (charcoal, kerosene), and finally adopts cleaner and modern fuels [13]. However, the situation is quite different from the ideal in the developing world. Households do not linearly transition from biomass fuels to modern fuels, as accentuated by the energy ladder model [14,15]. Energy (fuel) stacking is a common practice in the world [16,51]. The findings from these contemporary studies are an indication that household fuel choices are influenced by a myriad of economic and non-economic factors [15].

Masera and Navia originated the concept of energy stacking as a critique of the energy ladder theory in rural Mexico [16]. The consumer economic theory recognizes that higher income necessitates higher energy consumption and vice versa. The advantages and disadvantages of fuel stacking and energy laddering are, therefore relative rather than absolute. Biomass occupies the lowest part of the energy ladder. Biomass refers to any naturally occurring combustible material. Woodfuel is a major component within this classification, but grass, crop residues, or dung also fall under biomass and are easily accessible [17,18].Charcoal is a solid fuel derived from wood through a process called pyrolysis. It involves heating the wood to burn off most of the material, leaving pure carbon [19]. Consequently, charcoal shares several advantages and disadvantages with biomass. As a historical and cultural fuel, charcoal preserves the taste of food in contrast to modern energy [20]. Like biomass, charcoal emits organic and nonorganic compounds that are harmful both to human health and the environment [19].

Kerosene is ranged above charcoal. Kerosene is derived from petroleum and is produced during the distillation of crude oil [21]. Compared to its immediate predecessor and biomass, kerosene burns more cleanly and more efficiently and is of an advantage in terms of both user health and the environment. In urban and other resource scarce settings, obtaining kerosene may require less time than gathering biomass [22]. Liquefied petroleum gas (LPG) comes next after kerosene. It is derived from petroleum. It contains propane, butane, or a combination of the two. In particular, LPG releases much less sulphur dioxide than kerosene or biomass [22]. LPG stove presents an “investment barrier” due to the cost of the equipment [16]. Households must additionally pay a deposit for the cylinders that contain the gas [23]. To combat the risks associated with distribution, a household may purchase multiple canisters, increasing the uptake costs of LPG further (Heltberg 2005). After these initial fixed costs, however, the cost of LPG every month may be less than the cost of purchasing biomass [23,24]. In this study, LPG is used as a clean fuel.

At the top of the energy ladder is electricity. In Ghana, about 85% of the population has access to electricity, but very few use it as cooking fuel despite it being clean and the most efficient source of energy for cooking, with minimal health implications with all safety precautions in place. Electricity can be generated from large-scale sources such as nuclear, water, gas, or coal (grid-based) or non-grid sources such as small-scale oil generators or renewable energy [25]. The tasks of collecting, purchasing, or replenishing supplies largely do not apply, except for prepaid electricity, for which users must visit a utility to make payments or purchase online [23,26].

Many people continue to use cooking fuel concurrently even those who have access to clean cooking continue in a practice known as “fuel stacking”. To understand why people, use multiple fuels concurrently, a study reviewed 100 articles and concluded that fuel stacking is largely driven by the Physical Opportunity domain of the COM-B model [51]. This suggests that the persistence of biomass cooking is less culturally anchored than has previously been assumed, and highlights the importance of providing reliable, affordable access to clean fuels.

1.2. Health, Economic and Environmental Risks Associated with Cooking Fuels

The dependence on less clean fuels for cooking and heating is a major source of House Air Pollution (HAP) globally. This phenomenon is significantly associated with respiratory and cardiovascular morbidities, widening the global mortality rate [27]. Major health issues mediated by pollutants from biomass consumption include upper and lower respiratory tract infections, lung cancer, stroke, obstructive pulmonary diseases, burn injuries and premature deaths.

According to the World Bank’s 2020 report, 4 billion people worldwide lack access to clean, efficient, convenient, safe, reliable, and affordable cooking energy [28]. In low- and middle-income countries, poorly constructed cookstoves and open fires give rise to about 95% of the 300,000 global cooking-related mortalities annually [29]. Women and children bear a disproportionate share of the burden of collecting and processing firewood in developing countries [2,30]. This cultural practice takes away from productive time and reduces opportunities for women’s engagement in beneficial economic endeavours [31].

According to the 2022 world outlook energy report, the world produces 37 billion tonnes carbon dioxide (CO₂) emission yearly of which household cooking fuel combustion is thought to produce 25% of global emissions – black carbon, the second largest contributor to climate change after CO₂ [27,32]. Black carbon absorbs a million times more energy per unit mass than CO₂ but has a shorter half-life, and its effects are amplified on a regional scale in areas close to the source of emissions [27], causing global warming and disrupting regional environmental systems. According to the International Energy Agency (IEA), more than half of all wood harvested globally is used for fuel [33] and is harvested from nearby forests, depleting our forest area of some species and causing deforestation.

Governments in developing countries have enacted policy interventions focused on promoting the transition from traditional fuels by households towards the adoption of cleaner and modern fuels for HAP reduction. The promotion of LPG access, increased electricity consumption, and the adoption of improved cookstoves with chimneys are some of the interventions aimed at reducing HAP exposure [34]. Nonetheless, most people in sub-Saharan Africa are caught in a dilemma in terms of fuel choice due to socio-economic reasons [4,35,36,37,38]. The cost of using a cleaner and more sustainable source of energy plus maintaining a particular cooking culture culminates in the fuel choices made for cooking. With limited information on what may influence these choices, this study focuses on examining the determinants of cleaner fuel choice for cooking in Ghana.

2. Materials and Methods

2.1. Study Setting, Sampling, and Sample Size Calculation

The data presented here is a component of the baseline study of the Briquette for Cooking Fuel Project (B4CcokingF), a cluster randomized controlled trial in two riparian towns in the Eastern (Kpong) and Greater Accra (Asutuare) regions (Figure 1). Kpong is one of the main towns in the Lower-Manya Krobo Districts with 16 communities and mostly dominated by the Krobos, but there are also Akans, Ewes, Hausas, and Gas. Asutuare is in the Greater Accra region's Shai-Osudoku District. Six communities divide the town; the Ga-Adangme and Ga tribes make up most of the population with other tribes rediding there. Farming, fishing, or fish mongering are the main sources of income for the communities. Most households (HHs) use open fires, or charcoal stoves, or a combination of open fires and charcoal stoves, or LPG cookstoves for cooking and warming food.

The Lower Volta Lake (LVL) extends from the Akosombo Dam in southeastern Ghana to the town of Yapei in the Central Gonja District, Northern Region of Ghana, some 400 kilometres (250 miles) to the north. The southern end of the LVL is home to Asutuare and Kpong. Aquatic weeds cover about 70% of the entire LV basin. Kpong and Asutuare areas, thus disrupting the water ecosystem, social, cultural, and economic activities, as well as the education of the girl child.

In 2018, the Volta River Authority and the School of Public Health at the University of Ghana launched the B4CcokingF with the goals of (i) managing invasive aquatic weeds through the production of biomass briquettes for cooking/heating and (ii) offering the communities an alternative fuel source and livelihood (i.e., biomass briquettes from aquatic weeds) to minimize the reliance on wood fuel. In 2019, we conducted a baseline census of all 22 communities in Kpong and Asuture towns. We identified 8,825 HHs, totaling 1907 compound houses. A " household" is a collective of individuals who share the same space, eat from the same pot, adhere to the same housekeeping practices, and receive care as a single unit. We used the term "compound house" to refer to HHs that share a common space. We restricted the B4CcokingF study to compound houses with a maximum of four households, to prevent air pollution from nearby HHs from contaminating the study houses. This restriction will enable us to supply briquettes to each HH in a compound house during the project's intervention period. Therefore, we considered 1104 of the 1907 compound houses that met the above requirement.

We estimated the sample size by considering the following factors: a 30% decrease in PM_2.5 levels from 850 g/m3 [39]; a 0.1 cluster effect; an intra-cluster correlation of 0.25; a 95% confidence interval; 5% type I error rate; 80% power; 359 variability; and a 15% rate of respondents who did not respond or dropped out. We determined the sample size for the RCT to be 210 for each group (420 overall) using a method described in [40]. In the current analysis, all 420 baseline participants were considered. Ethical approval for this study was obtained from the Ethics Review Committee, Ghana Health Service through School of Public Health, University of Ghana and community leaders of Asutuare and Kpong township (GHS-ERC:015/03/21).

2.2. Data Collection

Trained field workers, led by a community representative, visited eligible HHs in each community and conducted survey interviews with primary cooks with a REDCap (Research Electronic Data Capture) device to gather data from participants using Android phones. Participants were interviewed with a slightly modified WHO questionnaire [41]. The online questionnaire is divided into six sections: HH characteristics, HH socioeconomic characteristics, water and sanitation, fuel storage for cooking, fuel devices for heating and lighting, and health outcomes. 2

2.2.1. Measures

Outcome Variables

The outcome variable was primary household fuel type / choice- the use of charcoal (yes, no), or wood fuel (yes, no), or LPG (yes, no) as a primary fuel. This was transformed into a nominal variable, where 0 was coded as wood fuel, 1 as charcoal and 2 as LPG. The use of wood fuel served as the base or reference category for the model.

Explanatory Variables

The explanatory variables are age (coded as <40, 41-64, >65), sex (coded as 0=male, 1=female), education (coded as 0=primary, 1=secondary, 2=tertiary), civil status (coded as 0=single, 1=married, 2=ever married), religion (coded as 0=Christian, 1=Muslim, 2=Agonist), ethnicity (coded as 1=Ga-Adangbe, 2=Ewe, 3=Akan, 4=others), age of household head (coded as 0=<40, 1=41-64, 2=>65), household head education (coded as 0=primary, 1=secondary, 2=tertiary), family size (coded as 0= <3, 1=3-5,2=>5), dwelling type (coded as 0=Urban, 1= Peri-urban), toilet facility type (coded as 0=improved, 1=unimproved), water source (coded as 0=improved, 1=unimproved), waste disposal method (coded as 0=burning, 1=indiscriminate disposal, 2=collection), wealth index (coded as 0=Lower class, 1= middle class, 2=upper class), and household average monthly income.

2.3. Data Analysis

Descriptive, bivariate and multivariate analyses were used to identify the predictors of cleaner cooking fuel choice. Descriptive statistics were used to report the frequency of the socio-demographic characteristics of participants (e.g., primary cook, household head and household characteristics) and cooking systems (e.g., stove and fuel usage, fuel usage by primary, secondary and tertiary fuel choice). This was followed by a bivariate analysis that assessed the relationship between household demographics by fuel type. Due to the nominal nature of the outcome variable, a multinomial logistic regression model was used to compute a multivariable model for the determinants of fuel choices. Multicollinearity was not an issue as all the model covariates had a variance inflation factor of less than 10 (The range of VIF from the model was 1.5 - 4.7) [42].

3. Results

3.1. Characteristics of Study Participants

Table 1 is a summary table of the characteristics of study participants. A descriptive analysis of the data shows that most of the participants were below 40 yrs (59.2%), females (93.5%), had primary education (79.0%), married (50.6%), belonged to the Christian religion (92.1%), and had a normal BMI (38.1%). Also, most of the household heads were above 65 yrs (94%) and were male (67.1%). The majority of the households (78.8%) were situated in peri-urban settings, with family sizes exceeding six (39.6%). Additionally, 96.1% had improved water sources, 80.2% had improved toilet facilities, and 53.7% disposed of their waste indiscriminately. The households belonged to the lower-class wealth index, accounting for 67.8%.

3.2. Bivariate Association Between Participant Characteristics and Fuel Type

Table 2 is a description of the bivariate association between participant characteristics and use of fuel type. The results from Table 2 shows that the sex of the participants was significantly associated with household fuel types (p=0.019). That is, the highest (64.6%) and lowest (15.0%) fuel types used by female participants were wood fuel and LPG respectively. The marital status of the participants was significantly associated to household fuel types (p=0.001). That is, the highest (61.5%) and lowest (10.3%) fuel types used by married participants were wood fuel and LPG respectively. The educational level of the participants was significantly associated to household fuel types (p<0.001). That is, the highest (69.8%) and lowest (11.2%) fuel types used by participants who had primary education were wood fuel and LPG respectively. The ethnicity of the participants was significantly associated to household fuel types (p=0.024). That is, the highest (58.7%) and lowest (17.5%) fuel types used by Ewe participants were wood fuel and LPG respectively.

The sex of the household head was significantly associated to household fuel types (p=0.016). That is, the highest (61.0%) and lowest (13.1%) fuel types used in houses with male household heads were wood fuel and LPG respectively. The educational level of the household head was significantly associated to household fuel types (p<0.001). That is, the highest (74.5%) and lowest (9.8%) fuel types used in houses with household heads having primary education were wood fuel and LPG respectively.

The income level of households was significantly associated to household fuel types (p=0.022). That is, the highest (61.4%) and lowest (15%) fuel types used in households with a monthly income of GHC 501-1000 were wood fuel and LPG respectively. The household ownership status was significantly associated to household fuel types (p=0.004). That is, the highest (64.0%) and lowest (8.6%) fuel types used by participants who live in rented houses were wood fuel and LPG respectively. Household dwelling type was significantly related to household fuel types (p=0.001). That is, the highest (75.0%) and lowest (8.3%) fuel types used by participants who live in houses occupied by multiple households were wood fuel and LPG respectively. Family size was significantly associated to household fuel types (p<0.001). That is, the highest (69.7%) and lowest (10.1%) fuel types used by participants whose family size was 4-6 were firewood and LPG respectively.

3.3. Predicting the Association Between Study Characteristics and Fuel Choice

Table 3 presents the results of the multinomial logistic regression estimate of household characteristics by fuel type. In this model, being ever married decreased the likelihood of using LPG (adjusted odds ratio (AOR) = 0.12, confidence interval (CI), 0.02 – 0.71) and charcoal (AOR: 0.18, 0.04-0.83) compared to wood fuel. Household head having a tertiary education increases the likelihood of utilizing LPG (AOR: 11.66, 95%CI: 1.57,12.08). Having waste collected by a waste management company increases the likelihood of using LPG (AOR: 8.95 ,95%CI: 1.20- 66.48). Having 6 or more family members in a household decreases the likelihood of using LPG (AOR: 0.18, 95%CI: 0.06-0.51). Relatively, rich households are more likely to use LPG (AOR: 3.65, 95%CI: 1.489.03) and charcoal (AOR: 11, 95%CI: 2.62-46.26).

4. Discussion

This study aimed to examine the determinants of cleaner cooking fuel choice among two riparian peri-urban communities in Ghana. Findings from this study show that most households used charcoal as their primary fuel (64.49%), followed by LPG (20.77%) and wood fuel (14.73%). A family size of 6 or more and having been previously married reduced the likelihood of using LPG as the primary fuel. Household heads with tertiary education increased the likelihood of utilizing LPG and having or being relatively rich, as indicated by the wealth tertile increased the likelihood of using LPG and charcoal compared to wood fuel as a primary fuel. Additionally, the majority of rich households with heads having tertiary education and having their waste collected by waste management companies increased the likelihood of using LPG. It is worth noting that although increasing wealth increased the likelihood of using cleaner fuel (LPG), this did not stop the wealthy from using dirty fuels such as charcoal more extensively. The findings from this study highlight the challenge of fuel stacking, where changes in behavior towards the use of only clean sources of fuel do not hold for most households.

This study has several strengths. First, data were derived from a baseline of a cluster randomized control trial, which randomly assigned households to the intervention and control arms. Thus, each household, or participant by inference, had an equal chance of recruitment into any of the arms. Second, a census was taken of all the communities to ensure representativeness, thus, reducing selection bias. Finally, this study employed a structured questionnaire on RedCapp using a slightly modified version of the WHO questionnaire as such information bias may not explain these findings. The WHO questionnaire for monitoring household energy use and SDG indicators has been used in several studies across the Africa region, an example is Parvizi et al., 2024.

The study findings showed that most households used charcoal as their primary and secondary cooking fuel and LPG as their main tertiary fuel. This result is consistent with a study among urban households in Ghana, which established that most households rely largely on wood and charcoal and occasionally on LPG or electricity [43]. Also, relying on aggregate data using the GLSS, it was reported that wood fuels continue to be the most common energy source used for cooking by households in Ghana [44,49]. In Nigeria, a population-based research report suggests that almost 70% of households rely on biomass for cooking [45].

The novelty of this study lies in the findings that demographic factors such as higher educational level, having waste collected by waste management companies, and relatively rich households are the main determinants of LPG use. This is analogous to a study among households in Malawi, which showed that families choose one type of cooking fuel over another depending on the age of the household head, income, location and household head education status [48]. Another related study in Uganda showed that demographic characteristics such as level of education, age, and income influence a family’s choice of cooking fuel [50]. Furthermore, corroborating studies conducted among households in China and Kenya found that family size, economic status, prices of fuel, gender and education are the determinants of household fuel choices in urban settings [46,47}. In sum, the extant literature has emphasized general household fuel choice and not cleaner sources of energy. Thus, this study shows that such factors may similarly influence the use of cleaner sources of energy, such as liquified petroleum gas.

5. Conclusions

The findings from this study highlight the need for important policy and programmatic changes aimed at increasing education on the use of cleaner fuels and resourcing households with the means to follow through with this education. Education should focus on enabling household heads to make fuel substitutions that will improve energy efficiency and have fewer negative effects on the environment, society, and people’s health. Effective interventions should include the advancement of the Government of Ghana’s efforts to raise the overall level of education, education on environmental risk factors, and support the most vulnerable transition out of poverty. This will contribute to a reduction in the use of charcoal from wood, which will indicate a reduction in the pressure that is placed on wood supplies and will help contribute towards mitigating deforestation. Furthermore, stakeholders and policies in the energy sector should advance and promote the use of renewable, clean technologies to reduce the impact on the ecosystem, reduce pollution, and meet household demand. This should include enhancing initiatives such as the National LPG Promotion Program and the Rural LPG Promotion Program, increasing subsidies for LPG, extending the LPG Cylinder Recirculation Model to rural regions, and addressing the issue of LPG supply bottlenecks [52].