Analysis of the gap in enterprise access to renewable energy between rural and urban areas in Cameroon

Permanent access to energy is an essential pillar of economic development. However, there is a growing evidence that contemporary energy systems are not able to provide energy to the entire population on a sustainable basis and at affordable prices. In the face of these challenges, renewable energy can play an important role, especially in rural areas where access to centralized electricity grids is difficult. This paper aims to examine the access gaps of enterprises to renewable energy between rural and urban areas in Cameroon. The analysis is based on a sample of 209,482 enterprises, taken from the Second General Census of Enterprises in Cameroon (RGE-2) carried out by the National Institute of Statistics (NIS). The econometric estimations, obtained using the Blinder-Oaxaca decomposition, reveal that access rate to renewable energy for firms in rural areas is lower than that of firms located in urban areas. An increase in the level of education of the promoter of an enterprise, obtaining credit from banks, microfinance and savings, and the formalization of enterprises in rural areas can also contribute to reducing the gap in rural areas in terms of accessing to renewable energy. The discrimination suffered by rural enterprises related to the gender of entrepreneurs, the sector of activity, the business environment and professional experience tend to increase this gap. To reduce this gap, there is a need to promote access to finance for rural enterprises and their migration from the informal to the formal sector.


INTRODUCTION
Cameroon is a low emitter of greenhouse gases, but nevertheless aims to contribute to their reduction. That is why it signed the Paris Agreement on Climate Change. Many countries have opted for cleaner and environmentally friendly energy production technologies to cope with the horrors of climate change. Although they constitute only a small part of total global primary energy, renewable energy sources generate less CO2 emissions than non-renewable sources, which constitute a significant share of total global primary energy (Zeng et al., 2019). In Cameroon, the consumption of renewable energies during the period 2000-2017 represents between 78% and 84.5% of global energy consumption (IEA et al., 2020).
One of the overall objectives pursued by Cameroon's National Development Strategy for the period 2021-2030 (NDS30) is to put in place conditions conducive to economic growth, the accumulation of national wealth and the structural changes indispensable for the industrialization of the country. To achieve this, Cameroon plans, in particular, to increase the installed capacity of electricity

LITERATURE REVIEW
The seminal work of Sadorsky (2009) and Chang et al. (2009) laid the groundwork for the analysis of the determinants of renewable energy consumption. Over the last decade, the majority of studies considered income and prices as fundamental variables of renewable energy consumption. Li et al. (2020) show that rising GDP and energy prices are driving the growth of renewable energy consumption in OECD countries. On the other hand, Omri anad Nguyen (2014) find that over the period 1990-2011, renewable energy consumption did not increase significantly with GDP growth in 64 countries. Nyiwul (2017) obtains a similar result in a panel of 27 Sub-Saharan African countries over the period 1980-2011.
By using the GMM 1 type model, Omri and Nguyen (2014) show that higher oil prices on the world market do not stimulate demand for renewable energy. Increasing consumption of non-renewable energy reduces the consumption of renewable energy (Salim et al., 2014). Apergis and Payne (2011) show that the increase in demand for non-renewable energy does not significantly in-fluence the consumption of renewable energy in emerging countries.
Moreover, some authors have introduced environmental variables in the analysis of the determinants of renewable energy consumption. Apergis and Payne (2014) show, by using the VECM 2 model, that an increase in CO2 emissions leads to an increase in renewable energy consumption in a sample of 7 Central American countries. On the other hand, Lucas  come countries. This is not the case in high-income countries. With regard to population, Nyiwul (2017) argues that an increase in population favors the consumption of renewable energy.
In the economic literature, forecasting models are also often used to predict nuclear and hydroelectric energy consumption. Applying the CCRGM (1,1) model, Luo et al. (2020) predict that nuclear energy consumption will continue to increase over the decade 2019-2028 in North America, while hydroelectric energy consumption will peak and decline thereafter. Similarly, GRBM (1.

AIMS
This study aims to enrich the economic literature on the determinants of access to renewable energies. Specifically, it aims to (i) analyze the access gap of enterprises to renewable energy between rural and urban areas in Cameroon and (ii) propose instruments to reduce this gap.

METHOD
The above-mentioned studies have used several macroeconometric techniques to analyze the determinants of renewable energy consumption and energy consumption forecasting models 9 . In this study, a microeconomic approach was chosen. When attempting to examine the gaps of a dichotomous dependent variable between two groups, it is often recommended to use Fairlie's (2005) decomposition method or the Blinder-Oaxaca method for non-linear models developed by Bauer and Sinning (2008). The latter is based on the decomposition of the differences between two parts: an explained part (which presents the differences between the observed characteristics) and an unexplained part (which highlights the differences in the treatment received by individuals in the field). 9 Such as GM, NGM, DGM, ARGM, ENGM, CCRGM, GRBM and Verhulst's grey clean model.

Data source
The data used comes from the Second General Census of Enterprises in Cameroon (RGE-2) conducted by the NIS (2018). The geographical scope of this operation is the entire national territory. The sampling design is stratified at two levels. At the end of this RGE-2, the balance sheet shows 209,482 listed companies and establishments, geographically locatable and operating in a fixed business premises or a developed site. This value also corresponds to the size of the sample of this study, since all companies provided information on their location and the use of clean energy.

The econometric model specification and variables description
Considering the equation for the clean energy consumption of firms estimated by the logit method independently for urban and rural areas based on the literature presented in the previous section, let's consider the following equation: • for enterprises in urban areas • for enterprises in rural areas ' .

Ri SIi SI SIi
Based on the Oaxaca and Ramsom (1999) methodology, the access gap to renewable energy can be broken down as follows: where U ERN is the rate of access of enterprises to renewable energy in urban areas, R ERN is the rate of access of enterprises to renewable energy in rural areas, and β is the estimated values of the coefficient.
represents the gap in enterprises' access to renew-able energy due to differences in average characteristics between urban and rural areas, derived using coefficients estimated for all enterprises in the study sample. For example, if the promoters of rural enterprises have a higher level of education than urban enterprises, this contributes positively to their increased access to renewable energy compared to urban enterprises. This first term is the "explained" part of the differences in access to renewable energy.
The second term in this equation, is the difference in valuation between the characteristics of firms in urban and rural areas. This second term is the "unexplained" part of the gap in access to renewable energy.
Specifically, equation (3) of the model can be rewritten as follows: and ORD is the dummy variable that provides information on computer use within the enterprise and takes the value 1 if the enterprise uses at least one computer, and 0 otherwise.
INT is the dichotomous variable that informs about the connection of the company to the internet network and takes the value 1 if the company is connected to the internet network, and 0 otherwise.
MOM is the dummy variable used to assess the adoption of mobile money by the enterprise and which takes the value of 1 if the enterprise carries out financial transactions via mobile telephony, and 0 otherwise.
AEA is the qualitative variable used to assess the business environment in Cameroon. It takes the value 1 if the enterprise does not recriminate against the business environment, and 0 otherwise.
LOC is the qualitative variable that provides information on the occupancy status of a professional premises to carry out activities. It admits two modalities: 1 if the enterprise is a tenant, and 0 if not.
EXP is the dichotomous variable that allows us to appreciate the IPU experience. It takes the value 1 if the IUP has more than 4 years of activity, and 0 otherwise.
SA is the qualitative variable that captures the IPU's sector of activity. This variable admits two modali-ties: 1 if the IPU carries out activities in the "trade" sector, and 0 if it carries out activities in the production sectors, in this case "industry" and "agriculture".
ZON is the qualitative variable that provides information on the location of the IPU. It admits two modalities: 0 if the IPU is located in an urban area, and 1 if it is located in a rural area. Table 1 shows that the observation rate of the variables varies between 65% and 100%. The proportion of enterprises using renewable energy is slightly above average (52.9%). Furthermore, it can be observed that 18% of the enterprises in the sample are located in rural areas.

RESULTS
According to Table 2, the consumption of renewable energy by enterprises located in urban areas is higher than that of enterprises operating in rural areas. There appears a difference of 0.031 points, significant at the 1% threshold.
In addition, it is seen that in urban areas about 54.3% of business promoters have at least secondary education compared to 35.3% in rural areas. This suggests that the level of school dropout or discontinuation in rural areas is very high. In the same vein, it can be observed that in both urban and rural areas, there is a predominance of men at the head of enterprises. However, it seems to be more pronounced in rural areas. Besides, the rate of access of enterprises to credit from banks and microfinance is too low in both urban and rural areas. However, the constraints on access to financing are more acute in rural areas.
It would appear that enterprises located in urban areas have a higher level of ICT connectivity than their rural counterparts. As an illustration, 6.2% of enterprises have access to the Internet in urban areas compared to 3% in rural areas. In the same vein, 14.2% of enterprises have at least one computer in urban areas against 8.7% in rural areas. However, this level of connectivity of enterprises to ICTs appears low.
The results recorded in Table 3 reveal that improving the business climate (AEA), belonging to the formal sector (SF), professional experience (EXP), professional premises occupancy status (LOC), access to credit from a bank (BAN), microfinance (EMF) or savings (SAV) have a positive and significant influence on access to renewable energy. In the same vein, increasing the education level of business promoters (EDU), their gender (SEX) and marital status (MS) also contribute to increasing access to renewable energy.
On the other hand, enterprises operating in the "commercial" sector have relatively more difficulty in gaining access to renewable energy. The same is true for enterprises connected to the Internet network, as well as those living in rural areas. Table 4 presents the results of the Oaxaca-Blinder decomposition of the renewable energy consumption differential by enterprises. The estimated average rate of renewable energy consumption by businesses in rural areas is lower than that of businesses located in urban areas. In fact, the estimated average consumption rate of renewable energy by enterprises in rural areas is approximately 69.22%, while that of enterprises located in urban areas is approximately 77.9%. This corresponds to a difference of about 8.71 points.
The gap linked to the explained share equation reflects the increase in the rate of access to renewable energy by enterprises in rural areas, if they had the same characteristics as those located in urban areas. This is the gap that would remain in the absence of the disparities between rural and urban areas, i.e. the characteristics of enterprises in the two above-mentioned areas are identical. The results show that the explained part related to the differences recorded in terms of individual characteristics represents about 1.29% of the total gap.
The explained share of the gap in enterprises' access to renewable energy between urban and rural areas is significantly and positively influenced by the variables representing membership in the formal sector (SF), occupation status of a business premises (LOC), access to financing from a bank (BAN), microfinance (EMF) and savings (SAV), computer ownership (ORD) and level of educa- In addition, the second gap related to the unexplained (see equation (3)) part equation represents 98.71% of the total gap. This result reveals that the discrimination suffered by enterprises located in rural areas in Cameroon in terms of access to renewable energy predominates when trying to explain the gaps between urban and rural areas. In principle, the delay of realization of Low Voltage connection by ENEO is 5 working days in urban areas against 15 days in rural areas 10 . According to the facts on the ground, these delays are still very long in rural areas.
This discrimination is significantly and positively influenced by the assessment of the business environment in Cameroon (AEA), sector of activity (SA), professional experience (EXP), status of occupation of a business premises (LOC) and the gender of the company promoter. On the other hand, it is significantly and negatively influenced Note: *, ** and *** represent significance at 10%, 5% and 1%, respectively.
by access to financing from banks (BAN) and savings (SAV), use of a computer (ORD) and the level of education of the company promoter (EDU).

Table 4. Blinder-Oaxaca decomposition results
Source: Authors. Note: *, ** and *** represent significance at 10%, 5% and 1%, respectively. Table 4 shows that increasing the number of formal sector firms in rural areas can help close the gap in renewable energy consumption between urban and rural areas. Informal sector enterprises are often at a disadvantage in terms of access to renewable energy. For a connection to the ENEO electricity grid, companies are required to be registered in the Trade and Personal Property Credit Register. This requirement is prohibitive for informal sector enterprises. Enterprises in the formal sector generally benefit from easier access to standard utility connections such as electricity, water and telephone (Ingram et al., 2007). In addition, these enterprises are more inclined to invest in renewable energy because they have easier access to bank financing and land. Based on a representative sample of informal enterprises in seven West African cities, Grimm et al. (2013) find no evidence of a systematic and significant contribution of access to electricity to business performance. However, focusing on a more homogeneous sample of tailors in the city of Ouagadougou, Burkina-Faso, these authors find that electricity increases firm performance by stimulating the adoption of modern machinery and extending working hours.

DISCUSSION
Improving rural enterprises' access to credit from banks, MFEs, and tontines can help reduce the gap in renewable energy consumption between urban and rural areas. As an emerging industry, renewable energy can be very risky, and different sources of investment are needed to finance different projects. Venture capital and equity financing may be more appropriate for small projects with new technologies, while bank or debt financing is needed for larger projects (Ji & Zhang, 2019). Al-Mulali et al. (2015) show that credits to the economy have mixed effects on renewable energy sources in Europe over the period 1990-2013. Indeed, an increase in credits stimulates the production of five renewable energy sources that are renewable energy fuels or waste production, hydroelectric production, nuclear production, solar production and wind production. But this impact is only significant for wind generation. Anton  The least experienced enterprises are also discriminated against in terms of access to renewable energy. This result suggests that in order to ensure the connection of an enterprise to its electricity grid, it would seem that ENEO takes into account the experience of the enterprise itself.
Women entrepreneurs seem to be discriminated against in terms of access to renewable energy. This result reinforces the World Bank's 2019 result, which shows that Cameroon is ranked 159th in the world in terms of gender equality in employment and entrepreneurship. Specifically, this discrimination, according to the World Bank, is perceptible at the level of: • "travel", there are still impediments to women's freedom of movement, since there are still cases where women cannot freely decide where to go, travel and live; • "access to employment", women still face barriers to getting a job or practicing a trade or profession without authorization; • "marriage", since some women are not protected by law when they are subjected to domestic violence; • "business management", since there are some women who encounter obstacles to starting and running a business, particularly when registering businesses, opening bank accounts and signing contracts; • "asset management", there are gender differences in ownership and succession.

CONCLUSION
The potential positive impact of renewable energies on sustainable development seems to be the subject of consensus in the economic literature. Reducing the gap in access to renewable energy between rural and urban enterprises is of paramount importance for rural productivity, poverty alleviation, unemployment and rural-urban migration. This study highlights the factors likely to influence the gap in enterprises' access to renewable energy between rural and urban areas in Cameroon. At the methodological level, the paper used descriptive statistical techniques and the Oaxaca-Blinder decomposition using the database resulting from the 2nd general census of enterprises.
Statistical and econometric results reveal that enterprises' access to renewable energy in rural areas is lower than in urban areas. Moreover, there is a predominance of unexplained gap, which suggests that there is discrimination against enterprises in rural areas. Also, they show that in rural areas, improving the level of education of the business promoter, access to computers, obtaining credit from banks, micro-finance and savings, and formalizing businesses can also help reduce the gap in access to renewable energy. Discrimination against rural enterprises based on the gender of the entrepreneurs, the sector of activity, the occupation status of the business premises, the business environment and professional experience tend to increase this gap.
Therefore, to reduce the access gap to renewable energy between urban and rural areas, it might be advisable in rural areas to strengthen the capacities and skills of entrepreneurs, promote access to credit from banks, micro-finance and tontines, promote the migration of enterprises from the informal to the formal sector, and encourage the use of computers. In the same vein, measures should be taken to reduce the discrimination suffered by rural enterprises in terms of access to renewable energies by focusing on the gender of entrepreneurs, the business climate, the sector of activity, professional experience and the occupation status of business premises.