Rural household fuel energy transition: Evidence from

Energy for Sustainable Development 20 (2014) 30–35

Contents lists available at ScienceDirect

Energy for Sustainable Development

Rural household fuel energy transition: Evidence from Giwa LGA Kaduna State, Nigeria L.J.S. Baiyegunhi ⁎, M.B. Hassan Discipline of Agricultural Economics, SAEES — University of KwaZulu-Natal, P. Bag X01, Scottsville 3209, Pietermaritzburg, South Africa

a r t i c l e

i n f o

Article history: Received 2 July 2013 Revised 20 December 2013 Accepted 21 February 2014 Available online xxxx Keywords: Rural households Fuel energy Fuel transition Fuel stacking Multinomial logit analysis Nigeria

a b s t r a c t Rural household access to clean and affordable modern energy is critical to improving living standards in developing countries. Rural households in northern Nigeria in particular, are almost entirely dependent on fuelwood for their basic cooking needs. This has adverse effects on households' health, their productivity and environmental degradation. This paper analyzes the effect of households' socio-economic characteristics on choice of cooking fuel. A multinomial logit (MNL) model was used to estimate the determinants of fuel choice in Giwa Local Government Area of Kaduna State, Nigeria. Data analysis shows that the patterns of fuel usage are consistent with the ‘energy stacking’ theory as fuelwood are often used alongside modern fuels, and majority of the households depend largely on fuelwood as its principal cooking fuel. Modern fuels thus have failed to displace traditional fuelwood. Empirical results of MNL model shows that household head's age, educational attainment, household size, income, type of dwelling unit, the duration of food cooked and price of fuelwood are statistically significant factors influencing households' choice of cooking fuel. Implications for regional and national fuel policies are discussed. © 2014 International Energy Initiative. Published by Elsevier Inc. All rights reserved.

Introduction Access to clean, sustainable, modern, affordable and reliable energy services is an enormous challenge facing the African continent, particularly Nigeria. Energy plays a vital role in a nation's economic growth, progress and development. It is vital for socio-economic and human development as well as for poverty eradication (Eleri et al., 2012; Oyedepo, 2012). In spite of Nigeria's position as the sixth largest petroleum oil exporting nation and a leading gas exporter, the nation suffers enormous energy crisis with only about 40% of the nations' population of about 160 million people connected to the national electricity grid (Omokaro, 2008). The grid is plagued by frequent outages that last for as long as 20 h daily in places that are connected to the grid. Nigeria's current available electricity generating capacity is about 3920 MW with per capita power capacity of 28.57 W which is grossly inadequate for domestic household consumption (Ibidapo-Obe and Ajibola, 2011). Currently, 15.3 million households lack access to electricity, per capita electricity consumption has been less than 150 kWh per annum (World Bank, 2011). The North and South energy access divide is widely acknowledged. For instance, in Lagos (Southern Nigeria), almost all 1.7 million households are connected to the national grid, while only over a million households are connected in Kano (Northern Nigeria) (Eleri et al., 2012). ⁎ Corresponding author. Tel.: +27 33 260 5437. E-mail address: [email protected] (L.J.S. Baiyegunhi).

Energy poverty in Nigeria goes beyond lack of access to electricity. In 2006, fossil fuels made up of 94% of export from Nigeria with only a tiny fraction available for domestic use (Vincent-Akpu, 2012). An estimated 72% of Nigerians depend solely on fuelwood as cooking energy source (NBS/CBN/NCC, 2011). The major energy consuming activities in Nigeria's households are cooking, lighting and use of electrical appliances (Energy Commission of Nigeria, 2005). The energy consumption mix is dominated by fuelwood (50.45%) while the share of petroleum products and hydroelectricity is 41.28% and 8% respectively (Omokaro, 2008). However, over 90% of households in northern Nigeria where deforestation and desertification are most prevalent and threatening the livelihood of inhabitants still depend on fuelwood for cooking, using the traditional “three stone fires”. While most households in Nigeria still collect wood for their cooking, some especially in the wood deficient northern Nigeria tend to buy wood (NBS/CBN/NCC, 2011). Fuelwood is increasingly commercialized as an energy source, with about 38% of households in Nigeria buying fuelwood from the market (Eleri et al., 2012). More households in the South use kerosene for cooking than in the North. Per capita LPG use in Nigeria is one of the lowest in Africa despite being one of the World's leading exporters of natural gas (NBS/CBN/NCC, 2011). Today, more households are climbing down the energy ladder — moving from electricity, gas and kerosene to fuelwood and other traditional biomass (Eleri et al., 2012), and with the astronomical rise in the prices of modern fuel and the increasing level of poverty in Nigeria, the fuel choices of many of the nation's agricultural population

http://dx.doi.org/10.1016/j.esd.2014.02.003 0973-0826/© 2014 International Energy Initiative. Published by Elsevier Inc. All rights reserved.

L.J.S. Baiyegunhi, M.B. Hassan / Energy for Sustainable Development 20 (2014) 30–35

and some of rural poor households remain largely the utilization of unprocessed fuelwood (Fawehinmi and Oyerinde, 2002). The high dependence and utilization of wood for energy generation, with an estimated 27.5 million kg per day consumption have contributed to deforestation of hundreds of hectares of woodland, loss of biodiversity, soil degradation and greenhouse gas emission, resulting in desertification throughout Sub-Saharan Africa and especially in the rural areas of Nigeria (Amaewhule, undated). The growing energy poverty in Nigeria can also be strongly linked to a lack of energy law. Although a number of policy initiatives do exist (such as National Energy Policy 2003; National Policy Guidelines on Renewable Electricity 2006; Renewable Energy Master Plan 2005; National Energy Master Plan 2006), government commitment to effective implementation is lacking. The lack of energy law has reduced investors' confidence on these policies. The lack of institutions with clear vision and resources to champion universal access to both power and cooking energy for the poor and inadequate access to finance are major constraints; households and SMEs lack financial products to enable them to acquire pro-poor energy services such as clean biomass cook stoves, LPG and solar lanterns. There are also no clear service delivery models for public support for expanding access to energy services (Eleri et al., 2012). For a developing nation like Nigeria, owing to its energy crisis, issues relating to energy choices and household energy transitions are important from a policy standpoint. The choice of domestic cooking fuel in rural household in Nigeria is an issue for addressing deforestation and health hazards resulting from indoor pollution. Efforts at encouraging households to make substitutions or transitions that will result in more efficient energy use and less adverse environmental, social and health impacts are encouraged. Therefore analysis of the factors determining energy choices and use pattern in rural households in Nigeria are necessary as a first step. In the light of these facts, this study seeks to investigate the different cooking fuel choices available to rural households in Giwa Local Government Area of Kaduna State, Nigeria, and analyze the determinants of the behavior of rural household fuel choice. The rest of the paper is organized as follows: The next section discusses the theoretical and conceptual framework, followed by the research materials and methods (the description of the study area and empirical model for the study). The empirical results and discussion are presented thereafter followed by conclusions and implications for energy policy. Theoretical and conceptual framework The theory of household fuel energy choice is often based on the on the “energy ladder” model (Heltberg, 2003) and the associated fuel switching. This model placed more emphasis on household income and relative fuel prices as the basis for fuel choice (Barnes and Floor, 1999; Barnes et al., 2005). Based on household income, the energy ladder model depicts a linear three-stage switching process. The first stage involves a heavy reliance on centuries old biomass fuels, while in the second stage household moves to “transition” fuels involving the use of kerosene, coal and charcoal, and in the third stage, they switch to the use of LPG, natural gas or electricity which is a function of increased household income, and other factors such as deforestation and urbanization (Inayatullah et al., 2011). However, the simple nature of the energy ladder model placing emphasis on income wealth and substitution as a determinant of household fuel choice has been criticized by many studies (Heltberg, 2003; Masera et al., 2000; UNDP/ESMAP, 2003) for its assumption that as household income increases, the household discards the consumption of traditional fuels for the use of modern clean fuels which they can afford. These studies have shown that households often do not fully ascend the “energy ladder” but rather ‘fuel stack’, which means that with an increase in income, traditional fuels are not completely discarded, but are rather used in conjunction with modern clean fuels.

31

There is also growing evidence in literature that other than household income, the distance of the household from biomass sources (Hyde and Kohlin, 2000; Pachauri, 2004) increased fuelwood availability (Narain et al., 2008), and fuelwood shortages as a result of deforestation (Veld et al., 2006) may also be an important factor influencing household fuel choices. Nonetheless, there are exceptions that need to be considered from the ‘energy ladder’ model. In rural areas of many developing countries, a large proportion of middle-income households who could in principle afford modern and convenient form of fuels continue to rely fully or partly on traditional biomass fuels (Heltberg, 2003). A number of factors such as age, family size, level of education of household head, type of food cooked and taste of food cooked with fuelwood, whether or not the household owns the dwelling units are important factors that determine household cooking fuel choice (Osiolo, 2009; Pundo and Fraser, 2006). It therefore suggests that income, although very important, is not the only determinant of household cooking energy source. Many other factors both on the demand and supply sides are now considered (Inayatullah et al., 2011). Household energy source is now explained as a portfolio choice rather than as a ladder process (Osiolo, 2009). Therefore, modeling households' fuel energy choice is considered under the general framework of consumer theory (Lancaster, 1966; Rosen, 1974), which suggests that consumers derive utility not from a commodity but from the attributes embedded in a commodity. Information at households' disposal about the various fuels influences their decisions which are driven by households' economic and non-economic constraints. The economic factors may include availability and market price of fuel, household income and expenditure, while the non-economic factors may include socio-economic characteristics such as household size, age, gender, house ownership, type of dwelling, location of residence, distance to fuel source, and access to electricity (Osiolo, 2009). This study follows evidence from literature that households choose fuel based on bundles of household socio-economic, income and agro-ecological characteristics (Heltberg, 2004; Jumbe and Angelsen, 2011; Masera et al., 2000; Osiolo, 2009; Pundo and Fraser, 2006). In this study, it is assumed that a household faces a choice among alternative fuel types, the individual household is assumed to consider the full set of offered alternative fuel types in a choice situation and has to choose the alternative that maximizes utility (Hensher et al., 2005). Consider a households' choice of a fuel type and assume that utility depends on the choice made from a set (C) i.e. the choice set that includes all the possible fuel alternatives. Thus, a household is assumed to have a utility: U ij ¼ Q Z j ; Si þ ε Z j Si

ð1Þ

where for any household i, a given level of utility will be associated with any alternative fuel choice j. The utility derived from any alternative fuel type depends on the attributes (Z) of the fuel type and other socioeconomic and agro-ecological factors affecting households' decision. Choice made among the alternative fuel types will be a function of the probability that the utility associated with a particular option (j) is higher than that associated with another alternative fuel types. The statistical model of probability Pij that alternative j is chosen by household i is given by P ij ¼ prob U ij NU ia ; a ¼ 1; 2; 3…… :: j; a≠ j:

ð2Þ

Thus if the ith household selects fuel type j, then Uij is the highest utility obtainable from among the j possible choices.

32

L.J.S. Baiyegunhi, M.B. Hassan / Energy for Sustainable Development 20 (2014) 30–35 Table 1 Explanatory variables used in the multinomial logit model. Source: Based on a priori expectations. Variables

Description

Expected sign

Age Education Household size Household income Household head occupation Dwelling unit status Type of dwelling Duration of food cooked Fuelwood price Distance to fuel source

Age of household head (in years) Years of school attendance Number of people in the household Monthly household income (in Naira) D = 1 white collar job; 0 if otherwise D = 1 if own dwelling; 0 if otherwise D = 1 if traditional; 0 if otherwise D = 1 if takes long to cook; 0 if otherwise The cost of fuelwood (in Naira/kg) Distance to fuel source (in kilometers)

− − +/− − + + + + − −

Research materials and methods

electricity. Hence, this study specifies an MNL model (discrete choice method) as follows (Greene, 2003):

Study area and data

0 exp β j X i with j ¼ 1; 2; 3…… ProbðY i ¼ jÞ ¼ X4 ″ exp βk X i k¼1

The study was conducted in Giwa Local Government Area of Kaduna State, Nigeria. Giwa Local Government Area is located in the northern guinea savannah and on the plains of Northern Kaduna. It lies between the latitude 11° 20′–11°52′ N and longitude 7° 0′–7° 5′ E. The area is characterized by alternating dry and wet seasons with a mean annual rainfall that varies from 1000 to 1300 mm and temperature of between 76 °F and 78 ° F. The major source of livelihood in this area is farming, engaging about 70% of the rural population. Farming is traditional in nature with emphasis on the cultivation of crops which include maize, millet, groundnut, cowpea etc. and keeping of livestock. The study utilizes a cross sectional data collected in 2010/2011 through personal interview of 120 households who are randomly selected from the 11 political wards namely, Galadima, Gangara, Danmahawayi, Giwa, Idasu, Kadage, Kakangi, Kidandan, Panhauya, Shika and Yakawada using a structured questionnaire. Only married women were sampled and interviewed, because of the belief of Nigerian rural population in this region that cooking is the sole responsibility of women. In these rural communities fuel procurement for cooking is largely the responsibility of women rather than men. To a large extent, from empirical and field observations, only women and girls collect fuelwood and prepare food. For this reason, this study targeted women rather than men and to be interviewed, the woman has to be a housewife (i.e. married). The main question of the survey required the respondents to indicate the primary fuel choice mostly used for cooking. Data were also collected on household socio-economic characteristics. The dependent variable in the empirical estimation was fuel choice from the set of fuel types. In this study ‘fuel choice’ was used to mean preferences among the fuel types and fuelwood. The choice of explanatory variables used was based on theory, previous studies and data availability. Since the explanatory variables included in the multinomial logit model are the outcome of ex-ante expectations, no unambiguous predictions on the signs of these variable effects on fuel choices can be made. The explanatory variables and the hypotheses of how each influences fuel choice are presented in Table 1. Empirical model for the study In this study, households' choice from among the four fuel types was estimated within the multinomial logit (MNL) framework (Gujarati and Porter, 2009). The MNL model was used to analyze the factors affecting households' choice of cooking fuel (fuelwood, kerosene, natural gas and electricity). The MNL model has been commonly applied to analyze discrete choice data (Farsi et al., 2007; Osiolo, 2009; Rao and Reddy, 2007). It is suitable because it allows the analysis of decisions across more than two types of fuels. The response variable includes four distinct unordered alternatives: fuelwood, kerosene, natural gas and

ð3Þ

where Yi is the dependent variable representing the fuel type chosen by a household and takes the values of 1, 2, or 3 if the household chooses kerosene, natural gas or electricity. Fuelwood is used as the reference category. Xi represents a vector of explanatory variables that includes socio-economic characteristics, fuel attributes, market and agroecological factors affecting a household fuel choice. βi represents vectors of estimated coefficients. The results of MNL model are interpreted in terms of the odds ratios, i.e. the ratios of the probability of choosing one outcome category over the reference category. These ratios are defined as: ln

P ij ¼ X i β j −βk ¼ X i β j P ik

if k ¼ 1:

ð4Þ

A positive parameter indicates that the relative probability of choosing other fuel type over fuelwood increases relative to the probability of choosing fuelwood over the other fuel types (kerosene, natural gas and electricity). Empirical results and discussion Household socio-economic characteristics The socio-economic characteristics of the respondents are presented in Table 2. All the sample household heads are female and are married. Table 2 shows that the average age of household head is 46 years, while the average number of year of school attendance is 5.2 years. The average household size is 5.9 with the majority of household heads are engaging in ‘blue collar’ jobs such as farming, trading, and arts, or are full-time housewives. Only 15% of the respondents are civil servants. The average monthly income is N9543 (about $60), indicating that a household subsists on an average of $2 per day. Majority (68%) of the respondents live in traditional houses built with mud bricks, although some of these are cemented. Only 24% of the households own their dwelling units. Household portfolio of cooking energy The distribution of household by their main fuel choice is presented in Table 3. The domestic fuels used by households are fuelwood, kerosene, gas and electricity. In the study area, all sampled households use fuelwood. Table 3 shows that fuelwood and kerosene are the main energy sources, with 63.3% of the households having their cooking energy needs met from fuelwood, and 23% from kerosene, while 5% and 8.7% of

L.J.S. Baiyegunhi, M.B. Hassan / Energy for Sustainable Development 20 (2014) 30–35 Table 2 Socio-economic characteristics of households in the survey. Source: Field survey data.

Table 3 Average share of household's useful cooking energy by principal cooking fuel. Source: Field survey data.

Household socio-economic characteristics No. of respondents Percentage Mean Age (years) 25–39 40–59 50–69 Educational status No education (0 years) Primary & adult education (1–6 years) Secondary (7–12 years) Tertiary (N12 years) Household size 1–5 6–10 10–15 Occupation Farming Traders Artisans Civil servant Unemployed/full-time house wife Average income per month (N) b/ = 5000 5001–10,000 10,001–20,000 N20,000 Dwelling status Own dwelling place Do not own dwelling place Type of main dwelling unit Traditional Modern

33

18 62 40

15 51.6 33.4

35 46 58

40 58

33.3 48.3

– 3.6

14 8

11.7 6.7

8.2 9

42 66 12

35 55 10

3.7 6 8

51 30 18 15 6

42.5 25 15 12.5 5

– – – – –

48 26 18 26

40 21.7 15 23.3

1550 5075 10,500 21,050

44 76

36.7 63.3

– –

82 38

68.3 31.7

– –

the households depend on natural gas and electricity respectively. The share of fuelwood, kerosene, natural gas and electricity in the cooking energy mix of the households that have them as their principal fuel source respectively is 72%, 86%, 69.9% and 35% respectively. The above descriptive analysis shows that the observed patterns in the data are consistent with in part with the ‘energy stacking’ theory, which states that households continue to use more than one fuel type and not simply switch to a new fuel as their income increases. Fuel staking is seen to dominate in rural areas, as modern fuel is used alongside traditional fuelwood. Modern fuel has therefore failed to displace fuelwood. Evidence from other studies in different countries suggests that adoption of modern fuel often brings multiple fuel use, resulting in household consuming a portfolio of energy sources (Heltberg, 2005; Masera et al., 2000; Osiolo, 2009). Determinants of household cooking fuel choice: multinomial logit regression result The MNL model was estimated with robust errors using STATA 11 software program (Long and Freeze, 2006) and the results are presented in Table 4. The result of chi-square test shows that the likelihood ratio statistics are highly statistically significant (p b 0.000) suggesting that the MNL model has a strong explanatory power. The estimated MNL correctly predicted about 73.4%, 18.8%, 2.6% and 5.2% of households' preference for using fuelwood, kerosene, natural gas and electricity respectively. The variance inflation factors (VIF) was to test for multicolinearity, the VIFs for all the variables included in the model were less than 10, indicating the absence of multicolinearity. The parameter estimates of the MNL model only provide the direction of the effect of the independent variables on the dependent (response) variable (Table 4). Also, these estimates do not represent the actual magnitude of change or probabilities. Therefore, the marginal effects from the MNL model, which measures the expected change in probability of a

Principal fuel used

Average percentage (%) share of cooking energy Fuelwood

Kerosene

LNG

Electricity

Fuelwood Kerosene Natural gas Electricity

72 8.2 1.2 29

27 86 23 28

– 1.6 69.9 8

3 4.2 6.2 35

Household (%)

63.3 23 5 8.7

particular choice being made with respect to a unit change in the independent variable, are reported in Table 5 for the significant variables. Estimated coefficients for all the fuel choice alternatives are compared with fuelwood, a reference choice as the base choice. The estimated coefficient for household head's age is negative and statistically significant for the probability of household choice of natural gas (Table 4), implying that an increase in the age of household head is less likely to influence the choice of natural gas relative to fuelwood. The marginal effects suggest that a year's increase in the age of the household head is likely to decrease the choice of natural gas by 32% relative to fuelwood (Table 5). The reason could be that the older the household head is the more likely she has built her taste around fuelwood that is available and affordable rather than fuels that are necessarily efficient. This is consistent with the findings of Heltberg (2005) that households will continue to use fuelwood through developed loyalty, traditional cooking methods or taste preference. The estimated coefficient for household head's education is positive and statistically significant for the probability of a household choice of kerosene and natural gas as their main cooking fuels relative to fuelwood (Table 4). This implies that an increase in household head's education is likely to influence the choice of kerosene and natural gas over fuelwood. The marginal effect suggests that a unit increase in the years of household head's education increase the likelihood of the transition from fuelwood to kerosene and natural gas by about 18% and 17% respectively (Table 5). A possible explanation is that, increased level of education improves households' income, taste, knowledge of fuel attributes and preference modern clean fuels. A highly educated woman is likely to lack time to collect fuelwood because of the opportunity cost of her time and would rather purchase fuelwood alternatives which are cleaner but expensive ceteris paribus. This is consistent with the findings of Peng et al. (2010); Farsi et al. (2007); Heltberg (2005), that when women education level is higher, they use less of fuelwood and more of commercial fuels possibly because their opportunity cost of fuelwood collection is increasing. The estimated coefficient for household size is negative and statistically significant for the probability of household transition from fuelwood to natural gas (Table 4). This implies that an increase in household size reduces the probability of the transition from fuelwood to natural gas. The marginal effect suggests that a unit increase in household size decreases the likelihood of the choice of natural gas over fuelwood by about 28% (Table 5). Theoretically, household size is expected to have a negative influence on the use of fuelwood alternatives (kerosene, natural gas, and electricity), because households with many members may have larger labor input for fuelwood collection. They are more likely to have extra labor in the form of child labor that can be used to collect fuelwood freely from open fields; free collection of fuelwood lowers its price and makes it cheaper to cook for many people using fuelwood relative to its alternatives. The estimated coefficient for household income is positive and statistically significant for the transition from fuelwood to kerosene, natural gas and electricity (Table 4). This implies that an increase in household income increases the probability of the choice of kerosene, natural gas and electricity. The marginal effect suggests that an increase

34


in household income is likely to increase the transition from fuelwood to kerosene, natural gas and electricity by 11%, 53% and 45% respectively (Table 5). A possible explanation is that as households' income increases they shift to the use of clean and modern fuel. This is consistent in part with the ‘energy ladder’ theory (Heltberg, 2005). The estimated coefficient for household type of dwelling place is negative and statistically significant for the probability of household transition from fuelwood to natural gas and electricity as its main cooking fuel (Table 4). This implies that households living in traditional houses (those without internal kitchen room) are less likely to choose natural gas and electricity over fuelwood. The marginal effect suggests that for households living in traditional houses the likelihood of transition from fuelwood to natural gas and electricity decrease by about 37% and 38% respectively (Table 5). A reason may be that most rural households use external kitchen; this does not encourage the use of modern clean fuels that require expensive equipment. The estimated coefficient for duration of food cooked is negative and statistically significant for the probability of household transition from fuelwood to natural gas and electricity as its main cooking energy (Table 4). This implies that the longer the duration of the food cooked the less likely a household will choose natural gas and electricity as its main cooking fuel. The marginal effect suggests that the likelihood for such household to choose natural gas and electricity over fuelwood decreases by 16% and 13% respectively (Table 5). A possible reason may be that fuelwood alternatives have higher relative costs per unit of time and fuelwood is cheaper and readily available. If a household cook food that takes longer time to prepare, the household is more likely to use fuelwood ceteris paribus. The estimated coefficient for the price of fuelwood is positive and statistically significant for the probability of household transition from fuelwood to kerosene as its main cooking fuel (Table 4). This implies that an increase in the price of fuelwood is likely to increase the

Table 4 Parameter estimates of the multinomial logit model explaining household fuel choice. Variables

Kerosene

Natural gas

Electricity

Age

−0.41 (−0.72) 0.22 (0.05)⁎⁎⁎ 0.02 (0.03) 0.68 (0.34)⁎⁎

−1.03 (0.24)⁎⁎⁎

0.18 (0.51) 0.02 (0.29) −0.03 (0.08) 1.22 (0.28)⁎

Education Household size Ln household income Household head occupation Ownership of dwelling Type of dwelling Duration of food cooked Ln fuelwood price Ln distance to fuel source Constant

0.03 (0.05) 0.71 (0.65) 0.02 (0.17) 0.95 (1.28) 0.033 (0.01)⁎⁎⁎ 0.72 (0.91) −3.74 (1.01)⁎⁎⁎

0.021 (0.01)⁎⁎ −0.20 (0.10)⁎⁎ 1.87 (0.35)⁎⁎⁎ 0.01 (0.06) 0.09 (0.52) −0.30 (0.12)⁎

0.07 (0.08) 0.59 (0.43) −3.02 (0.09)⁎

−0.30 (0.13)⁎⁎

−0.02 (0.01)⁎⁎

0.02 (0.01) 0.68 (1.00) −5.85 (1.55)⁎⁎⁎

0.07 (0.05) 0.21 (0.92) −4.92 (1.25)⁎⁎⁎

No of observation 120 Pseudo-likelihood −329.53 2 134.67 Wald Chi 0.0000 Prob N Chi2 0.68 Pseudo R2 Fuel choice accuracy (correctly predicted): Fuelwood = 73.4%; kerosene = 18.8%; natural gas = 2.6%; electricity = 5.2% Note: Standard error in parentheses. ⁎⁎⁎ Denotes statistical significance at the 1% probability level. ⁎⁎ Denotes statistical significance at the 5% probability level. ⁎ Denotes statistical significance at the 10% probability level.

probability of household choice of kerosene as cooking fuel. The marginal effect suggests that the likelihood for such household to choose kerosene over fuelwood increases by 28% (Table 5). A possible reason could be because kerosene appears to be a principal substitute for fuelwood in the study area. So an increase in the price of fuelwood leads to a decrease in its demand, therefore causing the demand for kerosene to increase i.e. as the market price of fuelwood increases, households shift to modern clean fuels. This is consistent with the economic theory; the demand for a commodity is a decreasing function of price of that commodity.

Conclusion and implications for energy policy Data from rural households in northern Nigeria show that the transition from traditional fuelwood to clean modern commercial fuels is still slow, given that fuelwood is the main source of cooking energy for the majority of the rural households. Household fuel consumption pattern provides a useful insight in distinguishing between different conceptual models of energy transition process. The data from the study area confirmed that fuel stacking is a more accurate description of household energy than the energy ladder model, as fuelwood is often used alongside modern fuels. The majority of the households depend largely on fuelwood as its principal cooking fuel. Modern fuels thus have failed to displace traditional fuelwood. Empirical results of the MNL model show that household head's age, educational attainment, household size, income, type of dwelling unit, the duration of food cooked and price of fuelwood are statistically significant factors influencing households' choice of cooking fuel. Education can play a role in the transition from traditional fuelwood to modern commercial fuels. Investment in the education of the rural dwellers will improve household's access to better job opportunities which will increase their productivity and income. With increasing income, rural households will be able to dwell in modern houses; also, as changes occur in the educational status of rural residents (especially young women), the opportunity cost of their time in collecting fuelwood and raising children increases. Thus, additional shifts in fuel use to clean and modern fuels could be seen. Moreover, while the shift away from fuelwood is occurring, the commercial energy source which appears to be the principal substitute

Table 5 Marginal effects from the multinomial logit model explaining household fuel choice. Variables

Kerosene

Natural gas

Electricity

Age

0.21 (0.69) −0.18 (−0.52) −0.632 (0.002)⁎⁎ −0.117 (0.003)⁎⁎

0.32 (0.47) 0.17 (0.02)⁎ −0.281 (0.004)⁎⁎ 0.534 (0.013)⁎

0.19 (0.82) 0.20 (0.29) 0.321 (0.008)⁎⁎ 0.483 (0.017)⁎⁎

0.011 (0.006) 0.346 (0.018)⁎⁎ 0.37 (0.06)⁎⁎⁎ 0.162 (0.009)⁎⁎

0.016 (0.005) −0.145 (1.15) 0.38 (0.14)⁎⁎ 0.132 (0.006)⁎⁎

0.113 (0.48) 0.048 (0.99)

0.181 (0.86) 0.007 (1.14)

Education Household size Ln household income Household head occupation Ownership of dwelling Type of dwelling Duration of food cooked Ln fuelwood price Ln distance to fuel source

0.008 (0.007) −0.271 (0.008)⁎⁎ 0.087 (0.84) 0.15 (1.02) 0.281 (0.81) 0.113 (0.05)⁎

Note: Standard error in parentheses. ⁎⁎⁎ Denotes statistical significance at the 1% probability level. ⁎⁎ Denotes statistical significance at the 5% probability level. ⁎ Denotes statistical significance at the 10% probability level.


for fuelwood is kerosene; given that kerosene is a major contributor of air pollution in Nigeria, a further switch to clean and modern fuels such as natural gas and electricity is important. Improving rural electricity grids and expanding rural access to natural gas and electricity can raise households' productivity and increase their income. Therefore, ensuring greater availability and accessibility to a variety of energy source will play a crucial role in the transition process. References Amaewhule, K. (undated). What are the constraints facing the development and implementation of energy efficiency policy in Nigeria and how can these be overcome? http:// www.dundee.ac.uk/cepmlp/car/html/car7_article7.pdf. Accessed on 20/04/2013. Barnes DF, Floor WM. Biomass energy and poor in developing countries. J Int Aff 1999; 53(1):237–59. Barnes DF, Krutilla K, Hyde W. The urban energy household transition: social and environmental impacts in the developing world. Washington, DC: Resources for the Future Press; 2005. Eleri EO, Ugwu O, Onuvae P. Expanding access to pro-poor energy services in Nigeria. Abuja, Nigeria: International Center for Energy, Environment and Development; 2012. Energy Commission of Nigeria. Renewable energy master plan. Abuja: Federal Republic of Nigeria; 2005. Farsi M, Filippini M, Pachauri S. Fuel choices in urban Indian households. Environ Dev Econ 2007;12:757–74. Fawehinmi AS, Oyerinde OU. Household energy in Nigeria: the challenge of pricing and poverty in fuel switching. J Energy Dev 2002;27(2):277–84. Greene WH. Econometric analysis. 5th ed. Upper Saddle River, New Jersey: Prentice Hall; 2003. Gujarati DN, Porter DC. Basic econometrics. 5th ed. New York: McGraw-Hill/Irwin; 2009. Heltberg R. Household energy use in developing countries: a multi-country study. Oil and gas policy division, discussion paper. Washington, DC, New York: World Bank; 2003. Heltberg R. Fuel switching: evidence from eight developing countries. Energy Econ 2004; 26:869–87. Heltberg R. Factors determining household fuel choice in Guatemala. Environ Dev Econ 2005;10:337–61. Hensher D, Rose J, Greene W. Applied choice analysis: a primer. Cambridge: Cambridge University Press; 2005. Hyde WF, Kohlin G. Social forestry reconsidered. Silva Fennica 2000;34(3):285–314. Ibidapo-Obe O, Ajibola OOE. Towards a renewable energy development for rural power sufficiency. Proceedings of International Conference on Innovations in Engineering and Technology. Akoka–Yaba, Lagos, Nigeria: University of Lagos; 2011. p. 842–50.

35

Inayatullah J, Khan H, Hayat S. Determinants of rural household energy choices: an example from Pakistan. Pol J Environ Stud 2011;21(30):635–41. Jumbe CBL, Angelsen A. Modelling choice of fuelwood source among rural households in Malawi: a multinomial probit analysis. Energy Econ 2011;33(2011): 732–8. Lancaster KJ. A new approach to consumer theory. J Polit Econ 1966;74:132–57. Long JS, Freeze J. Regression model for categorical and limited dependent variables using STATA. 2nd ed. College Station, TX: Stata Press; 2006 [224 pp.]. Masera OR, Saatkamp BD, Kammen DM. From linear fuel switching to multiple cooking strategies: a critique and alternative to the energy ladder model. World Dev 2000; 28(12):2083–103. Narain U, Gupta S, Veld K. Poverty and resource dependence in rural India. Ecol Econ 2008;66(1):161–76. NBS/CBN/NCC. 2010 Annual socio-economic survey on Nigeria. Abuja, Nigeria: Nigeria Bureau of Statistic–Central Bank of Nigeria–Nigerian Communication Commission collaborative study; 2011. Omokaro O. Energy development in a fossil fuel economy: the Nigerian experience. In: Uyigue E, Agho M, Edevbaro A, editors. Report of national dialogue to promote renewable energy in Nigeria. Benin City, Nigeria: Community Research and Development Centre; 2008. Osiolo HH. Enhancing household fuel choice and substitution in Kenya. Discussion paper no. 102. Kenya: Kenya Institute for Public Policy Research and Analysis (KIPPRA); 2009. Oyedepo SO. Energy and sustainable development in Nigeria: the way forward. Energy Sustain Soc 2012;2(15):1–17. Pachauri S. An analysis of cross-sectional variations in total household energy requirements in India using micro survey data. Energy Policy 2004;32(15):1723–35. Peng W, Hisham Z, Pan J. Household level fuel switching in rural Hubei. Energy Sustain Dev 2010;14(2010):238–44. Pundo MO, Fraser GCG. Multinomial logit analysis of household cooking fuel choice in rural Kenya: a case of Kisumu District. Agrekon 2006;45(1):24–37. Rao MN, Reddy BS. Variation in energy use by Indian households: an analysis of microlevel data. Energy 2007;32:143–53. Rosen S. Hedonic prices and implicit markets: product differentiation in pure competition. J Polit Econ 1974;82:34–55. UNDP/ESMAP. Access of the poor to clean household fuels in India. Joint United Nations Development Program (UNDP)/World Bank Energy Sector Management Assistance Program (ESMAP), South Asia Environment and Social Development Department. Washington, D.C.: World Bank; 2003 Veld VK, Narain U, Gupta S, Chopra N, Singh S. India's firewood crisis reexamined. RFF Discussion Paper 06-25. Washington D.C.: Resource for the Future (RFF); 2006 Vincent-Akpu I. Renewable energy potentials in Nigeria. International Association for Impact Assessment (IAIA). 32nd annual conference proceedings. 27 May–June 2012, Portugal; 2012. World Bank. World development indicators — 2011. Washington, D.C.: World Bank; 2011