Understanding Success and Failure in Distributed Electrification

Abstract

The previous three chapters provided detailed information on the performance of the business models within each of the three countries included in the study. Each country had unique institutional arrangements and different types of business models, leading to quite different outcomes. However, this variation can also be used to determine whether there are overall trends in the performance of these different business models. This chapter uses the results from case studies in all three countries to draw conclusions regarding success and failure in distributed rural electrification. In the next three sections, the initial hypotheses developed in Chap. 2 are examined to determine whether they are supported by the evidence. In order to do this, we examine the impact of the business models on changes in electricity service, sustainability and replicability.

Appendix: Detailed Summary of Results

One key question has been whether distributed electrification should be accomplished through centralized operations or through decentralized actors. The results, as shown in the following tables, indicate that successes and failures can be found in both centralized and decentralized models. Highly centralized models are more likely to result in large increases in access, while more decentralized models are sometimes more limited in their effects as shown in Table 6.4. This is, in part, due to the fact that subsidies, particularly for large capital expenditures, are preferentially given to centralized models, as can clearly be seen in Table 6.5. The centralized models are those have tended to focus on broad electrification programs, geared primarily to households (e.g. the Bahia PV distribution program or the Chinese Township Electrification Program). They also tend to be geared towards more universal service at the lower end of the income scale (including providing electricity for communal structures in low income areas, as in Brazil’s PRODEEM). Less centralized options include those targeting richer households that can afford more expensive systems (e.g. those in the Inner Mongolia Autonomous Region), can provide differentiated service for different users (e.g. the Cambodian rural electricity entrepreneurs) or target productive activities (e.g. BRASUS in Brazil).

Table 6.4 Impact of centralization on access

Table 6.5 Relationship between centralization and subsidies

The relationships between sustainability and centralization and between replicability and centralization show a slightly different picture (Table 6.6 and Table 6.7). Decentralized models are more sustainable and replicable. No decentralized models examined in this study were given a low rating on either count. By contrast, the highly centralized models that are sustainable and replicable are those that are highly dependent upon subsidies.

Table 6.6 Impact of centralization on sustainability

Table 6.7 Impact of centralization on replicability

While some centralized efforts have been successful as a result of this strong support and high subsidies, that is not universally the case. Projects such as the Japanese PV/Hydro system in Cambodia only benefit initially from the high subsidies. It would also be misleading to think that centralized models with high subsidies are the best solution. As will be discussed further in the following chapter, decentralized operations can often meet niche needs and segment the market in ways that improves service over the long term and meets more expansive development needs rather than just rural household needs.

The relationship between sustainability and operating subsidies is shown explicitly in Table 6.8. Similarly, Table 6.9 shows how replicability is related to capital cost subsidies. What these results show is that models designed to recover a significant portion of capital costs and cover operations and maintenance are more sustainable and replicable. Modest subsidies that encourage the formation of markets and technology development have enhanced replicability and sustainability (e.g. China’s renewables programs). Models that subsidize all, or nearly all, capital and/or O&M costs without also creating a strong institutional support structure have not been sustainable and replicable. Examples in which this institutional support mechanism was not provided include technology donations (e.g. the Japanese PV/Hydro project in Cambodia) and the township program in China. Examples in which a significant portion of costs are not recovered but in which strong institutional support exists for continued operation include the centralized utility DG projects in Brazil and many SHP facilities in China.

Table 6.8 Relationship between operating subsidies and sustainability

Table 6.9 Relationship between capital subsidies and replicability

Distributed models with low subsidies but high sustainability are those that are able to cover their ongoing costs through tariffs or other means. This includes private solutions such as the Cambodian electricity entrepreneurs and the PV cash market in China. Some models have low operating subsidies and low sustainability, indicating that they are not able to cover their costs and do not have the subsidies to maintain their operations. These are primarily failed international donors project such as the Cambodian PV/Hydro Japanese project and the initial attempt to install PV systems in Bahia. On the other end are those that rely almost entirely on subsidies to be sustainable, primarily the centralized utility model of Brazil and centra­lized efforts in China. These will continue to operate as long as the subsidies are assured. Of course, this does mean that their sustainability can be considered to be more conditional than the models that do not rely on subsidies but are sustainable due to operating cost recovery mechanisms internal to the model. It also raises the question of whether overall results would be better if successful subsidy models could be found that included decentralized actors. This is discussed further in the following two chapters.