Near-term increased exposure to political risk for African power trades

Abstract

Conflicts and socio-political instability plague electricity security in many African countries, hindering a transition to a net-zero emissions energy system. Established large-scale energy system planning approaches, relying on least-cost optimization models, regularly neglect the impact of political turmoil on energy security. This study assesses how six cost-optimal energy transition strategies for Africa might be affected by future socio-political instability and power trade volumes through a measure of power trade-related political risk. Our findings indicate that the most vulnerable period for Africa is the first half of the 2020-2030 decade, with socio-political risk hotspots in Western, Southern, and Central-Eastern Africa. Different socio-demographic, climatic and hydrological regime assumptions influence the distribution of socio-political risk across different regions, as countries uniquely adapt to future scenarios. These findings underscore the crucial need to integrate socio-political factors into energy systems planning, ensuring that transition strategies are flexible and cognizant of the involved risks.

This analysis of the future African power sector finds political risk to electricity trade will peak in the near-term, hitting Western, Southern, and Central-Eastern Africa hardest, calling for including political stability in energy planning.

Introduction

In the effort to close living standards gaps around the world¹, increasing electricity supply and security in African countries is a key priority. Despite its traditional reliance on fossil fuels for electricity generation, the large availability of renewable energy sources on the continent makes it a fertile ground for leapfrogging carbon-intensive technologies² and achieving a net-zero emissions power sector while ensuring socio-economic development. While hosting 18% of the world’s population, the entire African continent accounts for less than 6% of global energy use³ and only 3.8% of global greenhouse gas (GHG) emissions⁴. Nevertheless, its population and income growth will expand the demand for modern energy by a third between 2020 and 2030⁵, positioning the region as one of the major future energy consumers globally. Bearing these socio-economic trends in mind, achieving the United Nations Sustainable Development Goal (SDG) of universal access to modern energy by 2030 (SDG7) remains challenging. In 2022, in sub-Saharan Africa, 600 million people still lack access to modern, reliable, and clean energy supply, and without intensified efforts, this situation is expected to remain largely unchanged by 2030⁶.

Despite the opportunities for renewable energy and socio-economic development, the African energy transition presents some complex context-related challenges⁷. The critical levels of corruption, political instability, and the presence of ineffective governmental institutions⁸ often discourage foreign investors from tapping into the renewable energy potential of the continent⁹. Moreover, frequent events of political violence and armed conflicts¹⁰ threaten the national energy and electricity security of African countries, as power outages tend to increase during times of conflict¹¹. Power system assets, especially transmission lines, prove to be particularly vulnerable to destructive attacks^12,13. For instance, during the Mozambican civil war in 1992, the city of Johannesburg suffered severe power outages due to damage to the transmission lines that linked the Cahora Bassa hydropower plant to the South African city¹⁴. To this day, vandals’ attacks on power infrastructure continue to threaten access to electricity in many countries in sub-Saharan Africa¹⁵. In addition to the electricity supply, fuel supply is also threatened by uncertain political environments.

With few exceptions¹¹, traditional large-scale energy system planning overlooks energy security and governance considerations, which can instead be crucial in unstable and conflict-ridden political contexts¹⁶. Indeed, energy security ensures reliable and stable energy access, which is vital for economic stability¹⁷ and crisis resilience in conflict-prone regions¹⁸, while strong governance helps prevent the mismanagement of resources¹⁹, which can lead to inequalities and political turmoil. Nevertheless, policymakers frequently evaluate energy transition strategies through least-cost energy system models that primarily focus on techno-economic factors, leaving aside the environmental and socio-political implications of the energy transition^20–22. New tools have been developed and are being applied to include objectives beyond cost in energy transition planning^23–30. Recent modelling efforts centred on hydropower expansion, have integrated the planning process with an assessment of the negative impacts of dam construction on riverine ecosystems due to habitat fragmentation^31–33. However, very few studies attempt to evaluate the socio-political implications of the energy transition explicitly ^11,34–36, the disservices caused by political disruptions on energy systems^37,38 and the effects of different societal factors on the energy transition³⁹.

Large-scale energy systems planning is further complicated by uncertain future socio-economic and climatic conditions affecting future energy demand and renewable energy sources availability. Hydropower production, for one, is strictly dependent upon climatic conditions⁴⁰. This issue can be partially addressed using a scenario-based approach, usually referred to as scenario analysis^41–44. For this purpose, Shared socio-economic pathways (SSPs), framing future societal development trajectories into five quantifiable narratives, are commonly employed in energy system modelling scenario analysis^45–50. When coherently coupled with representative concentration pathways (RCP)⁵¹, they allow planning for energy systems capacity expansion under different future socio-economic and climatic trajectories.

The objective of this study is to address the limitations identified in the traditional large-scale energy systems planning approach by evaluating how least-cost energy transition strategies are impacted by future socio-political instability. More specifically, the analysis is conducted on continental Africa and takes into consideration cost-optimal power trade strategies obtained under different future socio-economic and hydro-climatic scenarios. The aim is to identify which time frames, geographic areas, and scenario assumptions are associated with substantial positive or negative risk variations, to infer relevant conclusions for policymakers.

The energy systems model used for this work is the Open Source energy MOdelling SYStem (OSeMOSYS)⁵², in its implementation for the African continent, The Electricity Model Base for Africa (TEMBA)^53,54, modified to represent hydropower at the power plant level⁵⁵ using the African Hydropower Atlas (AHA) database⁵⁶. Accurate representation of hydropower is crucial for planning energy systems in sub-Saharan Africa, as several countries in the region (such as the Democratic Republic of the Congo, Zambia or Namibia) rely heavily or entirely on this resource for electricity generation⁵⁷. By integrating the capacity factors of hydropower projects reported in the AHA within the model, we capture the variability of hydropower generation in response to changing climate conditions, which can significantly impact the reliability and stability of this energy source. We solved the model considering three socio-economic and climatic scenarios, obtained by coupling SSPs⁴⁶ and RCPs⁵¹ in alignment with the most used narratives to examine global change scenarios⁵⁸. The following socio-economic and climatic scenarios were considered: SSP1-2.6 (or 'Environmental sustainability' scenario), SSP4-6.0 (or 'Inequality' scenario), SSP5-8.5 (or 'Fossil-fueled development' scenario). The model is solved by coupling each socio-economic and climatic scenario with two hydrological regime scenarios, represented by different capacity factors for hydropower plants: a median hydrological regime and a very dry hydrological regime, accounting for decision-makers’ risk aversion to potential droughts. Therefore, this approach produces six cost-optimal strategies to expand the African energy sector from 2015 to 2050, including economically viable hydropower development portfolios.

We then evaluate the vulnerability of cost-optimal energy transition strategies through a measure of power trade-related political risk (hereafter referred to as 'political risk')³⁴, quantifying the susceptibility of electricity exchanges to disruptions due to the socio-political instability of the countries involved. The risk estimation is performed at both the continental and national scales by weighing the amount of power traded across Africa based on the governance performance of the countries involved. This is expressed through a combination of three factors: the ineffectiveness of political institutions, the degree of political instability, and the level of interference of private political interests in policy decisions. Governance indicators for each country are coherently projected to 2050 according to assumptions that are consistent with the narratives of each SSP⁵⁹. From the findings, we track the temporal evolution of continental political risk over the upcoming decades across different scenarios. We pinpoint countries that could be potential hotspots of political risk and assess the differentiated impacts of socio-economic and climatic scenarios, as well as the occurrence of droughts. Finally, using an exploratory modelling approach, we quantify potential power deficits arising from deviations in cost-optimal international power trades caused by socio-political instability. The analysis assumes that countries with unstable governments, corruption, or frequent armed conflicts are more likely to deviate from the optimal level of power traded at the continental power system level. This approach allows us to identify temporal and spatial criticalities and isolate scenario assumptions that could substantially impact electricity security.

By integrating socio-political instability into energy planning, our research contributes to enhancing the robustness and adaptability of future energy systems in continental Africa. Our results show that countries exhibit diverse responses to identical scenario configurations, which can increase risk in some regions while simultaneously decreasing it in others, as adaptation strategies are determined at the national level. Distinct regional risk hotspots emerge under different scenario configurations. The underlying assumptions of these scenarios, whether related to socio-economic factors or climate, can therefore be viewed as significant contributors to the heightened risk in these areas. Our results underline the need for an integrated approach for the evaluation of energy transition strategies, encouraging further research to refine and adapt energy policies in response to evolving socio-political landscapes.

Results

Rising energy demand drives near-term political risk increase

Continental political risk is defined as the risk associated with an electricity network substantially reliant on electricity exports from countries that are politically unstable. This metric reaches its maximum value when all the network’s electricity demand is met by power exports from countries with the lowest political stability, while it approaches zero when either all exported electricity comes from countries with the highest stability, or if no electricity is exported (see Methods section and Supplementary Method 1, supplementary equations (1–11), for the complete formulation).

Results reported in Fig. 1 indicate that political risk at the continental level peaks in the first half of the 2020–2030 decade for all scenarios, subsequently decreasing steadily until 2050 (refer to Supplementary Table 1 for numerical results). Supplementary Fig. 1 displays relative changes in continental political risk across scenarios over the analysed time frame. In general, scenario SSP1-2.6 displays the highest level of risk, followed by SSP4-6.0 and SSP5-8.5, although, around 2045, the risk in scenario SSP4-6.0 surpasses that of scenario SSP1-2.6. While SSP1-2.6 is characterised by an overall improved governance performance in African countries with respect to SSP4-6.0 (see Supplementary Fig. 2), it also entails a higher overall ratio of power exports to demand at the continental level (see Supplementary Fig. 3). This greater reliance on power trades to meet energy demand offsets the benefits of improved governance, ultimately resulting in higher political risk compared to SSP4-6.0.

Fig. 1. Increased continental political risks in the decade 2020–2030.

Yearly values of political risk for the entire African network under the six scenarios considered. Political risk peaks in the first half of the 2020–2030 decade in every scenario and then declines steadily, falling below its 2020 level by 2050. SSP shared socio-economic pathway, dry capacity expansion planned against a very dry hydrological regime.

Near the peak, scenarios of energy expansion planned against very dry hydrology tend to show lower political risk compared to their analogous median hydrology scenarios. Moving towards 2050, this difference becomes less pronounced, and very dry hydrology scenarios often outperform median hydrology scenarios. This can generally be attributed to the model’s preference for energy sources such as solar and wind over hydropower when planning for dry conditions⁵⁵. Their more spatially homogeneous availability compared to hydropower results in reduced cross-border power trading compared to scenarios based on median hydrological conditions, ultimately contributing to lower overall continental political risk levels.

The surge in continental political risk observed from 2020 to 2030 stems from the increasing ratio of power exports to demand within the continent. Although both exports and demand steadily increase in absolute terms during the analyzed decades (see Supplementary Figs. 4 and 5), it is their ratio (see Supplementary Fig. 3) that closely mirrors the trajectory of continental political risk. This ratio peaks between 2020 and 2030 before declining toward 2050. During the 2020–2030 period, importing electricity from neighbours with surplus capacity offers the least-cost solution for countries with rapidly growing demand. Later, as existing power plants reach the end of their operational life, new plants tend to be built closer to demand centres⁵⁵, reducing the need for transmission and, consequently, power trade, thereby mitigating political risk.

These findings indicate that political risk in Africa is more dependent on the volume of power traded than on governance indicator values. In fact, the political risk trajectory, rising to a peak before subsequently declining, closely mirrors that of the power export to demand ratio (see Supplementary Fig. 3). In contrast, average governance performance demonstrates a consistent upward trend across all scenarios (see Supplementary Fig. 2). To further validate this observation, we conduct an ex-post analysis by weighting cost-optimal power trades in different scenarios against measured governance performance indicators for African countries from 2015 to 2024⁸. Although recorded governance conditions are more positive than projections, yielding a slightly lower overall risk across scenarios, the overall political risk trajectory remains substantially similar (see Supplementary Figs. 6–9), indicating that power exports and demand primarily drive political risk. Additionally, we carry out an uncertainty analysis by randomly varying governance parameters within a ±50% range for each African country. Specifically, we generate 100 samples of perturbed governance indicators across all scenarios and recalculate the continental political risk accordingly. The results (see Supplementary Figs. 10–15) confirm a largely unchanged trajectory for political risk, reinforcing the conclusion that power exports and demand are the driving risk factors for African countries.

These findings indicate that policymakers should take an active role in supporting African nations during their energy transition, particularly during this critical initial phase marked by heightened risks. While increasing international transmission is widely recognised as an effective approach for reducing the cost of energy systems in the short term⁶⁰, we identify that relying on it without addressing underlying governance issues presents significant security challenges. By promoting the diversification of energy sources, investing in renewable energy infrastructure, and fostering local energy production capabilities, policymakers can help mitigate the vulnerabilities associated with heavy dependence on energy and electricity exchanges and enhance the continent’s resilience to political risks. Additionally, initiatives aimed at improving governance and regional stability in the most affected areas will play a crucial role in ensuring the long-term success of the African energy transition⁶¹. Nevertheless, political risk toward the end of our analyzed time horizon is expected to decrease, under all scenarios, to even lower levels compared to 2020 ones. This suggests that a cost-optimal approach to energy systems planning can be ultimately beneficial, in the long term, also from an energy security perspective.

Regional hotspots of political risk

A country’s political risk is assessed based on both its ratio of power imports to total demand and its ratio of power exports to total generation, with both ratios weighted by the governance indicators of the trading countries involved. A country’s political risk is highest when all of its power generation is exported to nations with the lowest possible political stability, or when its energy demand is fully met by imports from highly unstable countries. Conversely, political risk approaches zero if all imports and exports are carried out with highly stable countries, or if no electricity is imported or exported (see Methods section for details and Supplementary Method 2, supplementary equations (12–31), for the complete formulation). The quantification of country-level political risk is conducted for the entire 2020–2050 period, assessing both average and maximum risk levels for each country within each decade (see Supplementary Table 2 and Supplementary Figs. 16–21 for numerical results across all scenarios). However, since the 2020–2030 decade is identified as the most critical one, our subsequent analysis of the distribution and key drivers of country-level risk is specifically focused on this period.

Figure 2 displays the maximum levels of political risk obtained for each country during the three analysed decades under the median hydrology scenarios. Results indicate that critical countries are primarily located in Western, Southern and Central-Eastern Africa. Notably, three countries exhibit very critical levels of risk (>100) across all scenarios: Liberia, Rwanda and Togo. We also perform an ex-post analysis on country-level political risk, comparing the cost-optimal power trades across countries with actual governance performance data for African nations from 2015 to 2024⁸ rather than projected ones. Despite the more positive governance indicators, resulting in a slightly lower overall continental political risk, the distribution of risk hotspots remains largely unchanged (see Supplementary Figs. 22–25), further demonstrating that power trade patterns play a dominant role in shaping political risk. Additionally, incorporating country- and technology-specific costs of capital did not substantially alter the political risk patterns in our study compared to an initial analysis using a uniform cost of capital (see Supplementary Method 4). This is primarily because investment risks across neighbouring, grid-connected countries remain relatively similar. If these trading partners had starkly different costs of capital, the effects on political risk would be greater, as the model would shift generation and exports to lower-risk (and lower-cost) nations.

Fig. 2. Western, Southern and Central-Eastern Africa are hotspots of political risk (PR).

We present the maximum levels of risk obtained in each of the three analysed decades for the three median hydrology scenarios: a SSP1-2.6, b SSP4-6.0, and c SSP5-8.5. Country-level risk follows the continental trajectory, concentrating in the 2020–2030 decade and fading markedly in the two later decades. SSP shared socio-economic pathway.

These findings underscore the potential risks associated with already precarious energy and electricity security conditions, particularly in Western Africa, which hosts two of the three countries that display critically high levels of political risk (>100): Liberia and Togo. Liberia faces challenges related to the supply of unsafe petroleum products caused by political unrest, piracy, maritime boundary disputes, oil trafficking and bunkering activities⁶². Togo, whose power supply is largely dependent on imports from neighbouring countries⁶³, is also affected by substantial power losses due to power theft and ageing infrastructure. In both countries, the root cause of critical political risk during that period can be traced back to either a very high fraction of exported power over total generation or a very high fraction of imported electricity over demand (see Supplementary Tables 2 and 3 for details). In such fragile energy landscapes, high political risk, caused by a heavy reliance on power trades, heightens the countries’ vulnerability to external shocks, increasing the risk of supply disruptions. Enhancing the reliability of electricity imports through strengthened regional governance frameworks, while simultaneously diversifying domestic generation, can help mitigate the political risk associated with the energy transition in contexts with pre-existing vulnerabilities.

Highlighting countries with high levels of political risk does not inherently designate them as unsuitable for energy investments. Such an evaluation would require additional considerations beyond the scope of this analysis, including local policy frameworks, regulatory environments, resource availability and financing mechanisms for renewable energy. Instead, our aim in highlighting these countries is to pinpoint regions where targeted interventions, particularly investments in renewable energy technologies, could bolster energy security and thereby help mitigate political risk. An understanding of the root causes of political risk in critical areas is therefore essential for developing effective strategies to mitigate these risks. According to its formulation (see the Methods section and Supplementary Method 2, supplementary equations (12–31)), a country’s political risk can be mitigated by reducing either its fraction of exported power over total generation or its fraction of imported electricity over total demand. Therefore, countries that can expand their electricity systems through local power generation are better positioned to reduce the impact of rising energy demand on political risk. However, while reducing reliance on power trades mathematically lowers risk exposure in our model, a more sustainable long-term strategy involves addressing the instability of trading partners to unlock the economic benefits of regional integration. In fact, for energy transitions to be just and durable in such contexts, they must be guided by a transformative political economy approach, addressing historical inequalities and resource distribution conflicts⁶¹.

Implications of hydrological and socio-economic assumptions

To assess how scenario assumptions influence political risk, we calculate changes in risk levels across scenarios. Our analysis reveals that the same scenario assumptions produce different political risk effects among African countries. This is a result of each country’s unique energy resource availability, which leads to specific cost-optimal energy system expansion strategies for meeting the projected demands. Figure 3a shows absolute variations in the maximum country-level political risk between median and very dry hydrology scenarios during the 2020–2030 decade, characterised by the most critical levels of political risk at the continental scale. The results indicate that, even when planning energy systems against very dry hydrology, drought events result in increased political risk, particularly in Southern Africa, compared to median hydrology scenarios.

Fig. 3. Country-level political risk variations.

Obtained by subtracting: a risk of median hydrology scenarios from risk of very dry hydrology scenarios (assuming same shared socio-economic pathway (SSP) and representative concentration pathway (RCP)) b risk of more sustainable scenarios from risk of less sustainable scenarios (assuming median hydrology). Risk values refer to the maximum levels of political risk for each country in the network in 2020–2030 (displayed in Fig. 2). Identical scenario assumptions yield opposing effects across countries, reflecting their unique cost-optimal strategies.

For example, when the expansion of the energy system is planned under very dry hydrology, countries in the Zambezi area, such as Mozambique, tend to reduce their reliance on hydropower for self-consumption, due to its increased cost with respect to an expansion planned under median hydrology (see Supplementary Table 3 for details). Consequently, they are forced to increase their electricity imports, leading to heightened political risk under drought scenarios. For Namibia, the lower hydropower availability is compensated through increased domestic generation from solar. Nevertheless, despite the reduction in electricity exports, this does not align with the decrease in power generation, resulting in a higher export-to-generation ratio, which contributes to increasing political risk in accordance with its formulation (see the Methods section and Supplementary Method 2, supplementary equations (12–31)). As for Mozambique and Namibia, their high reliance on hydropower can potentially heighten political risks under drought scenarios. However, hydropower generation appears to be more reliable in some countries (for example, the Democratic Republic of the Congo), where, also in the presence of very dry hydrology, negligible reductions in hydropower generation are observed (see Supplementary Figs. 35 to 37). The higher prevalence of run-of-river hydropower schemes in these countries and the low sensitivity of precipitation distribution patterns to scenario assumptions in these regions might explain this result.

Conversely, countries that experience substantial decreases in hydropower generation during droughts may not necessarily face heightened political risks due to increased electricity imports. This dynamic is particularly observable in the Nile River basin. Egypt, for example, effectively manages this risk by increasing the contribution of solar energy to the mix of generation, thus diminishing its dependence on imports and exports. Likewise, Sudan and Ethiopia achieve comparable results by expanding the adoption of natural gas. In contrast, Rwanda can successfully mitigate risk by more than halving its electricity exports through increased generation from geothermal. Kenya, by enhancing its wind generation, can achieve a similar risk reduction (refer to Supplementary Table 3 for numeric details). Therefore, while the relationship between very dry hydrology scenarios and political risk is substantial, it is not uniform across all countries. Diversified generation mixes represent effective strategies for mitigating political risk. Grasping these dynamics is essential for policymakers dealing with the challenges of electricity security and political stability amid hydrological uncertainties.

As for the hydrological regime, socio-economic development and climate policy assumptions substantially affect the political risk. Figure 3b illustrates the absolute variations in maximum country-level political risk, during the 2020–2030 decade, between more and less sustainable scenarios. Our findings reveal that scenario SSP1-2.6 is associated with a slight increase in political risk, particularly noticeable in Southern Africa, compared to SSP4-6.0 and SSP5-8.5. Namibia, Zimbabwe and Botswana, for instance, exhibit notably lower risks in the SSP4-6.0 and SSP5-8.5 scenarios compared to SSP1-2.6. The socio-economic trajectories associated with the less sustainable scenarios are characterised by accelerated population growth, leading to heightened power demands and, subsequently, increased power generation needs. However, in scenarios with less stringent climate regulations, these nations can meet the augmented power requirements primarily through inexpensive fossil fuels (mostly coal). The additional generation also curtails their reliance on foreign electricity imports and exports, thereby mitigating their political risk (refer to Supplementary Table 3).

In this context, a notable anomaly arises with Togo, diverging from the continental trend by experiencing heightened political risk in less sustainable scenarios. The shift towards cheaper coal as a power source in these scenarios prompts Togo to transition from having no coal in its generation mix to sourcing over half of its generation mix from it. Conversely, the use of natural gas, previously relevant in Togo’s electricity portfolio, diminishes substantially or entirely in less sustainable scenarios (see Supplementary Table 3). The increased power generation demanded by the socio-economic assumptions underpinning these scenarios also leads to a surge in both electricity imports and exports for the country. Once again, Togo emerges as a regional electricity dispatch intermediary, utilising its additional coal-based generation primarily for export purposes, which increases its exposure to power trade-related political risk.

Our results reveal that, although continental trends can be identified, it is crucial to account for unique economic attributes and available resources of individual countries when examining the effects of climate change mitigation policies. Indeed, these unique attributes may result in diverse pathways for energy system development, each associated with different socio-political risks. Our results, however, should not be interpreted as a recommendation for African countries to pursue less stringent climate change mitigation policies and unbridled economic development as less risky pathways for development. Our modelling framework and scenario assumptions, which do not fully capture the real-world implications of energy transitions, should not be considered a predictive tool. Rather, the value of this analysis lies in identifying risk trends over time and determining the factors whose mitigation can effectively support a more socially sustainable transition toward a lower-carbon and more equitable energy system. Moreover, we intend to highlight that socio-political impacts for each country largely depend on its ability to adapt to changes in its economically viable power mix. Least-cost energy system models can help identify plausible development pathways for such power mixes, but their outcomes are still characterised by high uncertainty and should be contextualised to better understand real-world implications.

Poor governance is the main driver of power deficits

Continental power deficits are evaluated with an exploratory approach, assuming that countries with unstable governments, corruption, or conflicts may deviate from optimal power trade. A deviation factor reduces electricity exchange on each transmission line based on the exporter’s instability. As the instability of the exporting country increases, the range for sampling its deviation factor widens, reflecting the heightened risk of disrupted power transmission in more conflict-prone areas. From the deviated power transmission, the resulting deficit for countries participating in trade can then be calculated (see Methods section for details and Supplementary Method 3, supplementary equations (32–37), for the formulation).

Results, summarised in Table 1, highlight that, in alignment with political risk, 2020–2030 is the most critical decade for power deficits. This holds considering both average deficits, which don’t exceed 1.5%, and maximum deficits, which don’t exceed 12%. Moreover, SSP4-6.0 scenarios display higher levels of deficit than SSP1-2.6 and SSP5-8.5 scenarios. Nevertheless, it is plausible to assert that the predominant driver of increased deficit in SSP4-6.0 scenarios is the overall poorer governance status of African countries in those scenarios (as reflected in Supplementary Figs. 38–40).

Table 1.

Average and maximum continental power deficits due to operational deviations

	SSP1-2.6				SSP4-6.0				SSP5-8.5
	Median		Very dry		Median		Very dry		Median		Very dry
Average deficit
2020–2030	1.15%		1.15%		1.45%		1.31%		1.07%		0.93%
2030–2040	0.17%		0.18%		0.60%		0.60%		0.15%		0.17%
2040–2050	0.02%		0.02%		0.26%		0.24%		0.02%		0.01%
Total	0.45%		0.45%		0.77%		0.72%		0.41%		0.37%
Maximum deficit
2020–2030	LR	9.48%	LR	11.84%	LR	10.76%	ZM	10.49%	ZM	8.18%	ZM	9.17%
2030–2040	RW	2.14%	RW	2.06%	RW	8.68%	RW	8.09%	ZM	2.41%	ZM	2.56%
2040–2050	EG	0.81%	EG	0.78%	RW	1.99%	RW	2.00%	EG	0.43%	EG	0.34%

Countries displaying maximum risk values are reported (LR = Liberia, ZM = Zambia, EG = Egypt, RW = Rwanda, CM = Cameroon). Deficits are largest during the 2020–2030 decade across all scenarios, with the highest values occurring under the SSP4-6.0 scenarios due to the poorer projected governance performance of African nations.

SSP shared socio-economic pathway.

Figure 4, on the other hand, displays results from the country-level power deficit assessment. Results in Fig. 4a indicate that hotspots of maximum deficit during the most critical decade would be mainly located in Central and Western Africa under all scenarios. Fig. 4b, instead, shows the difference between the maximum and minimum deficits registered over the entire ensemble of samplings obtained with the exploratory approach employed (see Methods Supplementary Method 3, supplementary equations (32–37), for the methodology). The regions with the greatest variation in deficit largely coincide with those experiencing the highest deficit levels, with additional areas emerging in both Western and Eastern Africa. The countries experiencing the most significant deficit fluctuations are often in regions marked by high political instability, leading to greater variability in power exchanges due to potential operational deviations. Supplementary Figs. 26–31 display average country-level power deficits under the different scenarios.

Fig. 4. Country-level power deficits assessment.

a Maximum country-level power deficits due to operational deviations in 2020–2030 in median hydrology scenarios. Results refer to the average yearly deficits calculated over n = 100,000 samplings. b Maximum deficit variations in 2020–2050 in median hydrology scenarios. The variation is calculated between the maximum and minimum deficits registered over an ensemble of n = 100,000 samplings. Deficit hotspots lie mainly in Central and Western Africa and largely coincide with the areas of greatest deficit variability. SSP shared socio-economic pathway.

Deficit hotspot locations do not show substantial regional overlap with political risk hotspots (Fig. 2), except for some of the most critical countries, such as Liberia and Rwanda. Specifically, Liberia, similarly to Togo, functions as a regional electricity trade intermediary, importing surplus power to be exported to its neighbouring countries. However, countries that trade electricity with Liberia, namely, Côte d’Ivoire, Guinea, and Sierra Leone, face greater instability than those whose grid is connected with Togo (Benin, Ghana, Niger, and Nigeria). Consequently, Liberia is more vulnerable to potential power deficits, despite sharing comparable levels of political risk with Togo. In general, countries characterised by high potential deficits are associated with either a power mix heavily dependent on electricity imports or exports or a regional framework marked by considerable and widespread socio-political instability.

Discussion

This study investigates how least-cost energy system transition strategies across continental Africa could be impacted by future socio-political instability. To this aim, we integrate projections of governance indicators with energy system modelling under six scenarios harmonising socio-economic development, climate policy, and hydrologic regime assumptions. By analysing two metrics of risk (power trade-related political risk and potential conflict-driven power deficits), we assess the effects on the electricity security of African countries. Despite wide variations in national adaptation strategies, we identify key risk hotspots across the continent and pinpoint potential risk factors within the scenario assumptions.

Our findings highlight that political stability and governance of countries in the African continent could pose substantial risks to the reliability of their energy systems. Measures improving these factors could support a safe expansion of international power trade between these countries and mitigate political risks to energy supply, particularly in regions frequently affected by socio-political instability. These results align with the priorities outlined in the African Union Agenda 2063⁶⁴, which identifies the long-term security of energy supply as a fundamental objective⁷. However, achieving energy security necessitates overcoming substantial barriers, particularly the insufficient private and public investment in the energy sector^65,66. Addressing this financial gap remains a key issue in international climate policy discussions, including those at the 29th Conference of the Parties (COP29)⁶⁷, where mechanisms for enhancing climate financing for developing countries continue to be a focal point.

While our results are robust across the scenarios examined, this work could be improved by addressing the following limitations and exploring alternative assumptions. A first category of limitations pertains to the optimisation modelling framework used. Recent literature has highlighted that optimisation frameworks do not adequately represent real-world energy transitions^68–71. Specifically, our model relies on the common assumption that decision-makers possess perfect foresight. This is an unrealistic premise that fails to account for disruptive events that could have unforeseen consequences on energy security⁷². Nevertheless, by comparing projected power generation outputs from our modelling framework against measured data from EMBER⁷³ for the years 2015 to 2024, we show that the projected trajectories do not differ significantly in magnitude or final energy mix composition (see Supplementary Fig. 42). Small variations can be attributed to the effects of the SSP-RCP scenario assumptions, the unforeseen effects of COVID-19 and other sources of uncertainty introduced by our model, which is intended as a decision-support tool rather than a predictive one⁶⁸.

Another key limitation is the assumption of full cooperation among countries to minimise overall continental network costs, which does not necessarily reflect real-world decision-making processes. Future research could explore the effects on energy transition strategies if a non-cooperative approach is employed (which would be aligned with the assumptions of SSP4) and if the minimisation of power trade-related political risk is included as an optimisation objective. The latter would provide a valuable method for identifying cost-effective measures to mitigate political risk during the energy transition. This analysis could be conducted within a multi-objective framework or, as in this study’s single-objective framework, by constraining the model to acceptable political risk levels and solving it iteratively for different thresholds. Finally, future assessments should also consider integrating political risk factors into investment decisions for new transmission infrastructure, to capture how stability concerns might influence the physical expansion of the grid network.

Furthermore, our model does not fully capture the effects and implications of socio-political instability on Africa’s energy transition. The expansion of energy infrastructure, whether fossil-based or renewable, is often intertwined with extractive violence⁷⁴ and structural inequalities⁷⁵, which are not accounted for in our framework and could be the focus of future modelling efforts. Additionally, economic factors such as the declining return on energy investments⁷⁶ and local currency depreciation⁷⁷ could significantly influence cost-optimal planning and merit future exploration. Country-specific and time-varying capacity factors for wind power plants could also be incorporated in future work, although, since wind speed patterns are projected to change primarily on a regional scale through 2050^78,79, these effects are likely to be averaged out at the national level.

Another limitation pertains to the uncertainty introduced by the SSP-RCP scenario assumptions, which investigate only a limited range of possible futures. This also affects the SSP-driven projections of countries’ socio-political instability, which are limited to the considered SSP scenarios and are biased towards historical socio-political instability. Future studies could introduce a broader range of scenario assumptions to investigate the differentiated impacts of socio-economic and climatic factors on political risk within the network. Moreover, future research endeavours should concentrate on investigating the individual contribution of socio-economic and climatic assumptions on political risk variations. Since different socio-demographic assumptions (expressed by SSPs) are coupled with varying levels of mitigation efforts (associated with RCPs), our analysis does not allow us to definitely determine which of these factors drives the increased political risk observed in SSP1-2.6 scenarios. Socio-demographic assumptions influence energy demand across African countries, which may shape power trade patterns, while mitigation efforts and emission constraints impact energy costs, potentially encouraging greater reliance on power trade at the expense of domestic power generation. Decoupling these scenario assumptions in future analyses would help isolate their respective effects on political risk. For instance, the well-established notion that climate policies can enhance electricity security, leading to reduced dependence on imports^58,80,81, should correlate with a decrease in political risks in scenarios where stricter climate mitigation policies are implemented while socio-demographic assumptions remain unchanged.

Moreover, despite the challenges associated with predicting them, different methods for projecting socio-political instability and conflict could also be explored. Our analysis suggests that power trade volume is a more significant driver of political risk than governance-related events, but we acknowledge that real-world factors such as geopolitical conflicts, resource disputes, and civil unrest can have unforeseen impacts on energy security. These dynamics are not always fully captured by governance indicators and SSP narratives, making them highly unpredictable. However, future research should focus on developing new indicators to better account for their impact on energy systems, as they may lead to unexpected disruptions. This is especially true for scenarios associated with SSP4, which are characterised by widening disparities in economic opportunities and political power⁸². In such scenarios, increasing inequalities could have unpredictable effects on socio-political stability across the African continent, possibly triggering civil unrest and disruptions. Conversely, it is equally important to note that rapid improvements in institutional governance, unexpected technological breakthroughs, or accelerated regional cooperation could foster far more optimistic, stable, and resilient environments for the energy transition than presented in this paper.

Finally, our methodological framework for evaluating power trade-related political risks, which stems from the work by ref. ³⁴, leads us to conclude that a higher volume of power traded across unstable countries results in higher exposure to political risks. However, we acknowledge that our framework does not quantify the potential positive externalities arising from increased grid integration. International power trade can be crucial to reduce system costs, improve reliability, and spur greater economic interdependence^83,84 that may effectively reduce the potential for conflict. A pertinent historical example is the post-war economic integration of the European Union, where resource sharing contributed significantly to long-term regional stability⁸⁵. While these feedback loops are difficult to quantify and integrate into a linear optimisation framework, they represent an important variable for future research. Although this evidence suggests that energy isolationism is not necessarily the only way to reduce political instability, our conclusion stands: increasing endogenous generation within African countries is crucial to ensure energy demand growth is reliably satisfied.

Despite these limitations, our findings highlight the critical importance of incorporating socio-political considerations into energy systems planning, to define energy transition strategies that are robust and adaptable to political risk. In particular, rather than discouraging investment in these regions, our identification of high-risk countries is intended to highlight areas where targeted interventions, especially investments in renewable energy technologies, can strengthen energy security and, in turn, mitigate political risk. Since many African countries are operating within legacy structural inequalities imposed during the 20th Century, helping them overcome these deep-rooted historical burdens will be a fundamental priority in any transition scenario, as sustainable energy development cannot be decoupled from equitable socio-economic progress.

Methods

OSeMOSYS-TEMBA AHA

The modelling framework employed for this work is OSeMOSYS-TEMBA^53,54,86, modified to represent hydropower at the power plant level⁵⁵ based on the African Hydropower Atlas (AHA)⁵⁶. OSeMOSYS -TEMBA is a regional version of the global model, OSeMOSYS Global^87,88, representing the electricity generation and supply system for 48 African countries. It is an open-source optimisation model that finds the least-cost energy generation, capacity expansion and transmission strategies, based on energy sources availability, to satisfy given trajectories of energy demands over time. The model solves a linear programming problem assuming cooperative centralised decision-making (minimising costs at the continental level) and with perfect foresight over the entire modelling period.

To enhance the level of detail in its representation of hydropower plants, the model is integrated with the AHA⁵⁶, a database that compiles information on the capacity factors of 633 existing and planned hydropower projects in Africa. Capacity factors are incorporated along with their historical trends and projected values up to 2050, considering three distinct socio-climatic and hydrological regime scenarios. Constraints are introduced to ensure that the construction of each new hydropower plant is governed by a binary decision variable, representing whether a plant is built or not⁵⁵. The model is solved with a time horizon starting in 2015 to generate projections for the period from 2020 to 2050.

To ensure a coherent representation of the financing costs associated with capital investments across countries and technologies, we incorporated country- and technology-specific yearly cost of capital data from ref. ⁸⁹, following the methodology employed in the GNU MathProg implementation of OSeMOSYS⁹⁰. For details regarding the cost of capital implementation, please refer to Supplementary Method 4, supplementary equations (38–40).

Scenario definition

Energy demands definition in the energy system model is based on scenarios from the Inter-Sectoral Impact Model Intercomparison Project (ISIMIP2b)⁹¹. These scenarios include datasets on future final energy demands, land-use changes, and climate impacts on the hydrological cycle that are consistent with shared narratives. In fact, within each ISIMIP2b scenario, the definition of the energy demand is based on a single shared socio-economic pathway (SSP)⁴⁶, describing the future socio-economic and political environment. Within ISIMIP2b, each SSP also is associated with a representative concentration pathway (RCP)⁵¹, describing a radiative forcing trajectory that corresponds to a specific mitigation effort.

All SSP-RCP combinations, therefore, reflect specific levels of challenges to adaptation, challenges to mitigation, and actual efforts for climate change mitigation at the global scale. The three scenarios considered for this study, in alignment with the most used narratives to examine global change scenarios⁵⁸, are SSP1-2.6 ('Environmental sustainability' scenario), which accounts for sustainable development and constrained carbon emissions in a global effort to limit long-term global warming to 2 °C, SSP4-6.0 ('Inequality' scenario), which focuses on heterogeneous economic development among regions and is not associated with specific climate mitigation efforts, SSP5-8.5 ('Fossil-fueled development' scenario), which accounts for high fossil-fuelled economic growth and is, therefore, associated with high greenhouse gas emissions.

The final energy demands in this study are obtained by combining projections available in OSeMOSYS-TEMBA with SSP-informed projections, considering that the first is more reliable in the short term while the latter is more relevant in the long term (see Supplementary Information in ref. ⁵⁵ for additional details).

African hydropower atlas

In the AHA⁵⁶, historical capacity factors for existing hydropower projects are estimated monthly via the soil and water assessment tool (SWAT+)⁹² with meteorological data covering the 1980–2016 time period. Capacity factor projections for the 2020–2050 period are also available for the three ISIMIP2b scenarios, using as input bias-adjusted projected variables obtained from four global climate models: the Geophysical Fluid Dynamics Laboratory Earth System Model 2M (GFDL-ESM2M), the Hadley Centre Global Environment Model version 2 Earth System (HadGEM2-ES), the Institut Pierre-Simon Laplace Climate Model 5A Low Resolution (IPSL-CM5A-LR), and the Model for Interdisciplinary Research on Climate version 5 (MIROC5), forced with the concentrations of the RCPs associated with each SSP.

Capacity factors in AHA also reflect different hydrological regimes, as they are based on inflow profiles for each month under normal, very wet, and very dry conditions. Normal capacity factors are derived as monthly median, while very wet and very dry are obtained by multiplying the monthly median profiles by the ratio of the 5th and 95th percentiles of annual generation to multi-annual average generation. This study considers normal and very dry scenarios. Specifically, the latter serve to account for the risk aversion of policymakers to drought events. Additional details on the integration of the AHA in OSeMOSYS-TEMBA are reported in ref. ⁵⁵.

Power trade-related political risk

The evaluation of the political risk associated with the cost-optimal energy strategies is based on the work by ref. ³⁴, which provides a formulation of the preventable political risk associated to the African power network. In Trotter’s work, political risk, defined as the political risk that can be avoided by designing the network differently, is obtained as a linear combination of six factors, each describing a different risk driver: external political pressure/commitments, national energy security/sovereignty, efficacy of political institutions, political instability, private political interests and politicised land access. We conducted a preliminary calculation of risk, following the methodology in Trotter’s work, using as inputs the cost-optimal power strategies obtained from our model optimisation and the socio-political data from the author’s original work. The analysis showed that only three factors have a substantial impact on the overall political risk, contributing to more than 90% of the total risk on average: efficacy of political institutions, political instability and private political interests. Thus, for the calculation of the network-level political risk, we opted for focusing our attention only on the three said factors, disregarding the non-relevant ones.

The relevant factors in question characterise different dimensions of risk that arise when electricity exchanges occur between unstable countries. The network’s total amount of exported power is weighted on the governance performance of countries that carry out such trades. Each of the three risk factors refers to a different dimension of countries’ governance: the ineffectiveness of political institutions, the degree of political instability, and the level of interference of private political interests in policy decisions. For their complete formulation, refer to Supplementary Method 1, supplementary equations (3–11).

The characterisation of countries’ governance performance is based on the World Bank’s Worldwide Governance Indicators (WGI)⁸ and on the degree of autocracy obtained from the Polity IV scale⁹³. Governance indicators are projected to 2050, according to the SSP assumptions behind every scenario, following the methodology discussed in ref. ⁵⁹ and described in the following section.

We also define a country-level formulation of political risk to identify regional hotspots of political risk in Africa. Each of the relevant continental risk factors described above is split into two country-level factors: one considering power imports and one considering power exports. Export-related factors account for the ratio of electricity exports to total generation, while import-related factors consider the ratio of electricity imports to total demand, with both ratios weighted by the specified governance indicators. The six defined factors are combined linearly, with equal weights, to obtain the total country-level political risk. For the complete formulation, refer to Supplementary Method 2, supplementary equations (12–31).

Governance indicators projections

The projection of governance indicators to 2050, in alignment with the SSP assumptions, is performed according to the methodology described by ref. ⁵⁹. In that work, the aggregate governance indicator from the WGI database is modelled as a function of key socio-economic factors, including gross domestic product (GDP) per capita, the share of the population with post-secondary education, and the gender gap in mean years of schooling. The model is calibrated using a panel dataset of country-level governance observations from 1995 to 2015, applying a multiple regression framework that accounts for both cross-sectional and temporal variations. To generate future governance projections, the estimated regression coefficients from historical data are applied to SSP-consistent projections of GDP, education, and the gender gap in education, ensuring alignment with plausible socio-economic development scenarios.

The aggregate WGI indicator incorporates six dimensions of governance. For our political risk formulation, only three of them are relevant: 'government effectiveness', 'political stability and absence of violence' and 'control of corruption'. Therefore, the approach used by Andrijevic et al. on the aggregate governance indicator is replicated only on such relevant factors. Historical time series of each indicator are used to calibrate the model, and the estimated coefficients are imposed on the above-mentioned projected socio-economic variables. Supplementary Table 6 shows variations in yearly values of continental political risk obtained by including SSP-projected governance indicators compared to historical ones.

Power deficit calculation

The methodology employed for this analysis, summarised in Supplementary Fig. 41, stems from the work by ref. ⁹⁴, analysing the impact of operational deviations on the robustness of regional water supply portfolios. Adapting such methodology to this case study, the procedure employed consists of the implementation of a tolerance analysis. This includes the sampling of plausible operational deviations from the cost-optimal power trade strategies and the evaluation of the vulnerability of cost-optimal solutions via the calculation of yearly power demand deficits.

We assume that future national decision-makers will determine the amount of electricity to be exchanged with neighbouring countries, following a cost-minimisation perspective. In that case, the amount of electricity to be exchanged between two countries in a given year and on a given international transmission line is assumed to equal the optimal amount obtained from our least-cost energy systems model. In this framework, we suppose that countries characterised by unstable governmental institutions, corruption, or frequent armed conflicts may deviate from the level of power trade exchange, which is optimal at the level of the continental power system.

Therefore, we reduce the amount of electricity exchanged on each transmission line by a deviation factor that is randomly sampled within a range that depends on the degree of instability of the electricity exporter (see supplementary equation (32) in Supplementary Method 3). Higher levels of instability in the exporting country correspond to wider sampling ranges for its deviation factor, reflecting the idea that greater instability increases the likelihood of conflict events that may disrupt power transmission. Consequently, for each transmission line, we determine the reduced amount of power traded, which allows us to calculate the resulting energy deficit for each country in the network (see supplementary equations (33–37) in Supplementary Method 3).

Author contributions

T.B., A.Car., M.G. and A.Cas. conceptualised the study and developed the methodology. T.B. and A.Car. curated the data and conducted the formal analyses. T.B. created the visualisations and wrote the original draft. T.B., A.Car., M.G. and A.Cas. reviewed and edited the manuscript. A.Cas. acquired the funding.

Peer review

Peer review information

Nature Communications thanks Florian Egli, Mashael Yazdanie and the other anonymous, reviewers for their contribution to the peer review of this work. A peer review file is available.

Funding

T.B., M.G. and A.Cas. were supported by the SOS-WATER project funded by the European Union’s Horizon EUROPE programme (Grant Agreement Number 101059264).

Data availability

Cost-optimal power trades for the different SSP-RCP scenarios, including data and scripts to reproduce them, are available in the OSeMOSYS-TEMBA AHA Zenodo repository from ref. ⁵⁵ (https://zenodo.org/record/7931050). Worldwide Governance Indicators (WGI)⁸ are publicly available on the World Bank’s DataBank (https://databank.worldbank.org/source/world-development-indicators). Data on the degree of autocracy, obtained from the Polity IV scale⁹³, are available at the Integrated Network for Societal Conflict Research (INSCR) Data Page (http://www.systemicpeace.org/inscrdata.html). Data and scripts used for the Governance Indicators projections are available on the GitHub repository from ref. ⁵⁹ (https://github.com/marina-andrijevic/governance2019). Source data are provided with this paper.

Code availability

All the data and processing scripts to reproduce the results in this study are openly available on Code Ocean (https://doi.org/10.24433/CO.9402060.v2).

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-77362-x.