A
How does the Belt and Road Initiative Influence Solar Photovoltaic Exports from China to Sub-Saharan Africa?ABSTRACT
The Belt and Road Initiative (BRI), through its “five-pronged” approach, offers opportunities for opening up trade among the member countries. Despite the potential benefits of the BRI in promoting renewable energy trade, Sub-Saharan Africa’s (SSA) share of China’s solar photovoltaic (PV) exports remains small, conflicting with the growing need for sustainable energy solutions. This study examines the impact of the BRI on the exportation of solar PV products from China to SSA. Using a dynamic DiD approach with solar PV export data from 2012 to 2022 for 37 SSA countries, we find that joining the BRI significantly increases solar PV exports, particularly for early joiners. Causal mediation analysis of both hard and soft BRI connectivity mechanisms using nonparametric bootstrapping reveals that financial coordination and trade connectivity are significant mediators of this increase; while policy coordination, people-to-people bonds and infrastructure connectivity indicate direct effects. The findings highlight the BRI’s immediate success in leveraging trade and financial tools to boost solar PV exports and suggest the need to further utilize opportunities from soft connectivity mechanisms and physical infrastructure for long-term gains.
Key words:
BRI
Solar PV exports
SSA
Hard connectivity
Soft connectivity
A
A
1. Introduction
The Belt and Road Initiative (BRI), launched in 2013, aims to enhance global connectivity through policy coordination, people-to-people ties, financial coordination, trade connectivity and infrastructure connectivity. One of the key areas of focus for the BRI is the promotion of green energy technologies, including solar photovoltaic (PV) products. As the world’s largest producer and exporter of solar PV products, China supplies over 80% of global PV components, yet faces trade barriers in traditional markets like the European Union and the United States (Chadly et al., 2024 & Zhu et al., 2023). Meanwhile, the International Energy Association (IAE, 2022 & IEA et al., 2024) highlighted that Sub-Saharan Africa (SSA) has severe energy deficits, with 43% of its population lacking electricity access. Particularly, the SSA region has a growing demand for affordable and sustainable energy sources due to both the increasing energy needs and the negative impacts of climate change on its primary energy source: hydroelectricity (Hydropower, 2024; Matiashe, 2023; & UNECA, 2024). The BRI, therefore, presents a strategic opportunity to align China’s supply of affordable solar technology with SSA’s urgent demand for clean energy.
Despite the potential benefits of the BRI in promoting renewable energy trade, SSA’s share of China’s PV exports remains small (Chiyemura, Shen & Chen, 2022; He & Gui, 2025). Although there is evidence that the BRI has led to an increase in bilateral trade from China to SSA, the specific impact on PV exports requires a systematic analysis of both hard and soft connection mechanisms. Existing studies have not systematically examined how the BRI’s five priority areas mediate PV trade flows as illustrated in the theory of change in Fig. 1.
Fig. 1
Theory of Change for Exportation of Solar PVs from China to other BRI members.
Previous research has highlighted the role of the BRI in facilitating solar PV trade from different perspectives and approaches. For example, studies by Zhu et al. (2023) and Hu et al. (2023) have shown that the BRI has led to increased exports of solar PV products from China to other BRI member countries, apart from SSA. These studies also pay limited attention to the specific mechanisms through which the BRI’s effects are mediated. In terms of policy, the BRI’s Green Silk Road and other related initiatives have notable efforts towards enhancing regional cooperation and economic development, aiming to support sustainable development goals. Among the initiatives, China’s 2018 Green Investment Principles and 2019 Green Development Coalition promote renewable energy but lack targeted PV export strategies for SSA (Green BRI Center, 2025; Secretariat of BRI International Green Center). The African Renewable Energy Initiative (AREI), which started in 2015 prioritizing solar adoption but does not effectively utilize the elements of BRI connectivity elements (Africa Renewable Energy Commission, 2025). To date, the overall energy issues faced the SSA region do not meet the requirements of the United Nations Sustainable Development Goal 7, which is to “ensure access to affordable, reliable, sustainable and modern energy for all” (UN, 2024).
From the perspectives of research, therefore, this study specifically examines the role of the BRI in China’s solar PV exports to SSA. The study contributes to the existing literature in the following aspects: (1) Focus on the underlying BRI mechanisms: this study focuses on the underlying BRI mechanisms in order to tap the needs of SSA countries and make practical recommendations for both researchers and policy makers towards promoting the BRI’s mutual benefits between China and SSA through trade in solar PV products; and (2) Innovative methodological approach: different from the previous studies that have used either static models or aggregated trade data, this research uses a dynamic Difference-in-Difference (DiD) approach combined with a causal mediation analysis. The new approach allows for an in-depth understanding of the temporal dynamics and underlying mechanisms influencing solar PV exports.
The rest of the paper is organized as follows: Section 2 presents the BRI policy and theoretical foundations, including hypotheses. Section 3 outlines methodology containing empirical models for dynamic DiD and causal mediation analyses, variables and data sources, and identification strategy in the form of a flow chart. Section 4 has empirical results, including baseline, dynamic effects, and robustness checks. Section 5 is mediation effects of hard and soft connection mechanisms and section 6 provides the conclusion, policy recommendations for enhancing BRI’s influence on PV trade between China and SSA, and limitations of the study and research directions.
2. BRI policy and theoretical analysis
2.1. Policy background and elements
Connectivity is the core objective of the BRI initiated by the People’s Republic of China in 2013. Labelled as “A Key Pillar of the Global Community Future” by the China State Council Information Office (SCIO, 2023), the BRI aims to facilitate policy coordination, infrastructure connectivity, unimpeded trade, financial integration, and closer people-to-people bonds. The initiative is designed as a non-imposed tool which allows for voluntary cooperation of members based on shared interests and values, as well as a flexible tool which can include new mutual interests over time.
Although the initial conception of BRI was to enhance connectivity between East Asia and Europe, by 2023 it had expanded to include Africa and Latin America with overall inclusion of over 150 countries globally and 30 international organizations participating through joining or signing cooperation agreements with the BRI (China SCIO, 2023). By then, the African continent had the highest proportion about 35% of BRI membership including 44 countries from Sub-Saharan Africa, and the African Union (AU) had entered into related cooperative agreements with China (Africa-China Center for Policy and Advisory, 2023). To date, almost all African countries are members of the BRI (China Green Finance and Development Centre, 2025). Further, the China SCIO (2023) noted that the BRI participating countries have also expanded practical cooperation through major platforms such as the Forum on China-Africa Cooperation (FOCAC).
One of the areas of growing focus for the BRI is green energy technology enhancement. The promotion of the Green Silk Road by China led to the release of Green Investment Principles (GIP) for the BRI in 2018 and consequently the formation of the BRI International Green Development Coalition in 2019 aimed at engaging in dialogue, exchanges, joint research, capacity building and other activities (Xiao & Yifei, 2023). For instance, Fig. 2 illustrates the consistent rise in China’s PV exports from 2007 to 2016 whereby in 2015 the exports to BRI countries started to exceed the values to the EU and US combined.
Fig. 2
Value of China’s PV Exports to Selected Groups of Countries (2007–2016).
Note
From “The Impact of the Belt and Road Initiative on Chinese PV firms’ Export Expansion” by Zhu et al., 2023.
Primarily, the BRI is based on five priorities of cooperation also known as the “five-pronged” or “wǔ tōng” approach composed of policy coordination, infrastructure connectivity, unimpeded trade, financial integration, and closer people-to-people ties (Chin, 2023). Whilst adopting the China SCIO (2023) definitions of the five priority areas, it should also be noted that the BRI operates through ‘soft’ approaches involving intangible rules, relationships and standards harmonization; as well as ‘hard’ or tangible approaches such as physical infrastructure and systems.
a.
Policy coordination: Alignment of development strategies, technological and economic policies and administration rules and standards.
b.
People-to-people ties: Promotion of friendly cooperation, cultural exchanges, tourism, education, think tank and the media, mutual learning among civilization and cultural integration and innovation.
c.
Financial integration or coordination: Facilitation of multiple forms of financial cooperation, models, channels, diversification and mechanisms for investment and financing.
d.
Unimpeded trade or trade connectivity: Promotion of trade and investment liberalization and facilitation, resolution of investment and trade barriers and improvement of business environment within the region and in all related countries.
e.
Infrastructure (facilities) connectivity: Establishment of an infrastructure network over land, marine, air and cyberspace along the BRI corridors, routes, multiple countries and ports.
Apart from BRI’s five priorities, the SCIO (2023) records new areas within the BRI’s “All Round” Connectivity. The new fields target the achievement of a healthy, green, innovative, digital Silk Road. Based on the available data, however, this research utilizes indicators representing the original five priority areas only.
2.2. Theoretical analysis and hypotheses
As noted by Feng (2019) as cited in Hu et al., 2023) there is a synergy of sustainable development goals between China and countries along the BRI. China’s search for new PV products markets emanated from production overcapacity in the PV industry and consequent trade barriers initiated in 2011 by the EU and USA against China’s rising global market competition (Zhu et al., 2023). The rest of the BRI countries, especially those in SSA, attempt to access affordable, reliable, sustainable and modern energy.
Zhu et al. (2023) employed a spatial analysis of 2009 to 2016 trade statistics using linear probability models and logit model to show that BRI markets provide an alternative market for Chinese PV enterprises. Specifically, the results indicated that BRI facilitated 4751 million more PV products conveyed from China to BRI markets from 2013 to 2016 as illustrated in Fig. 2. The research by Zhu et al. (2023), therefore, validated BRI market’s potential to cushion area Chinese PV enterprises against the EU and USA trade sanctions and that the shift to BRI market would “advance laggard countries to realize green growth” p.25779. This is consistent with the core-periphery theory which argues that movement within and between the advanced (core) and developing (periphery) levels of economy is regulated by market forces (Hryniewicz, 2014).
However, Zhu et al.’s (2023) research was more interested in China moving away its focus from the EU and USA solar PV market than in the mechanisms of the BRI to secure alternative markets, like Africa. The perspective of the research in mainly on the supply-side and is restricted to Chinese PV firms as units of analysis. This research not only uses the countries as units of analysis, but also attempts to look at the BRI connectivity elements, an approach similar to Hu et. al.’s (2023) study which focused on non-African BRI members instead. Hu et. al. (2023) highlighted the important role of the BRI “five-pronged” approach in promoting PV trade, and recommended further research in these mechanisms. Regarding the BRI connectivity index, a common challenge faced by researchers is the unavailability or inaccessibility of data which covers the sampled BRI members with respect to the indicators included in the “BRI Five Connectivity” index developed by the Peking University (Chen, J., Liu & Liu, 2020; Hu et al., 2023; Sabola, 2024). This research adopts and improves the measurement methodology used by previous researchers by introducing relevant indicators for which data is easily accessible.
In general, the BRI has led to an increase in bilateral trade between China and SSA supported by the findings of Wang et al. (2024) and Negash et al. (2024), but there is need to find out about the BRI’s specific influence on PV exports with respect to SSA countries. Kluiver (2024) pointed out that Africa has potential gain from China’s shifts of BRI towards smaller, greener and less risky projects. Similarly, Geng’s (2021) review of the BRI and its global energy implications highlighted BRI’s contribution to renewable energy development in host countries, especially in terms of solar and wind energy.
Hypothesis 1
BRI has a positive influence on solar PV exports from China to SSA.
Hypothesis 2
The influence of BRI on solar PV exports from China to SSA is mediated by the soft connectivity mechanisms.
Hypothesis 3
The influence of BRI on solar PV exports from China to SSA is mediated by the hard connectivity mechanisms.
To achieve the second specific research objective, Hypotheses 2 and 3 can be broken down into five hypotheses representing the BRI’s “five-pronged” approach.
2.2.1. Policy coordination
Using social network analysis, Hu et al. (2023) found out that policy coordination or communication promotes PV trade within the BRI framework by facilitating political communication channels among member countries. The soft approach exhibited a positive influence on PV trade from 2016 to 2020; thereafter declining in influence in 2021 and 2022 due to pandemic-induced disruptions to trade.
Hypothesis
a: Policy coordination mediates the influence of BRI on PV exports from China to SSA.
2.2.2. People-to-people ties
Unlike the case with financial connectivity, Hu et al’s (2023) study showed that people-to-people communication has a significant effect on BRI’s influence on China-SSA PV exports. Bridging the language and cultural gaps helps to reduce transaction costs and facilitate smooth progress in PV trade. Moreover, technological advancements contribute to reduction in language barriers.
Nevertheless, Hu et al (2023) found out that the influence of people-to-people connection exhibited declining significance of PV trade from 2001 to 2022; becoming non-significant in 2021 and 2022.
Hypothesis
b: People-to-people communication mediates the influence of BRI on PV exports from China to SSA.
2.2.3. Financial coordination
Wang et al. (2024) used trade gravity model to analyze the BRI and China’s Outward Foreign Direct Investment (OFDI) in SSA. The results showed that the BRI has positive effect on both FDI and trade. Sabola (2024) consolidated BRI’s positive effect on FDI for southern African countries using a Propensity Score Matching-Difference in Difference (PSM-DiD) approach. However, based on the ‘Financial Market Theory’, Hu et al. (2023) revealed that the effect of financial integration on BRI PV trade is inconsequential- likely lacking significant impact on PV trade in the presence of other strong factors.
Whilst Yasmeen et al. (2022) argued that FDI inflows can be a source of solar technology promotion, Voica et al. (2021) clarified that the impact of FDI on trade depends on the type of investment, absorptive capacity of the FDI recipient and economic development of the partner countries. Given the limited data on categorized FDI from China to SSA countries, this research uses the lump sum FDI stock recorded by the China Africa Research Initiative (CARI)- an approach utilized by other researchers like Hu et al. (2023) and Sabola (2024).
Hypothesis
a: Financial integration mediates the influence of BRI on PV exports from China to SSA.
2.2.4. Trade connectivity
Using spatial analysis to find out the impact of the BRI on Chinese PV firms’ export expansion, Zhu et al. (2023) concluded that the firms prefer to explore BRI markets due to reduced trade restrictions. The expectation of trade connectivity having a positive mediating influence on PV exports is consistent with the findings of Hu et. al (2023).
Hypothesis
b. Trade connectivity mediates the influence of BRI on PV exports from China to SSA.
2.2.5. Infrastructure connectivity
Hu et al. (2023) also explained that infrastructure expands PV trade among BRI countries by reducing distance-related challenges. Although Huang (2016) noted that energy infrastructure investments became pivotal in the early stages of BRI investments, Hu et al. (2023) argued that attention should be paid to aviation and shipping infrastructure in order to facilitate PV trade. Examples of facilities would, therefore, include modernized and expanded sea ports, air ports and roads aimed at trade and transit facilitation. This type of connectivity is important for both landlocked and coastal countries in the SSA region.
Hypothesis
c: Facilities connectivity mediates the influence of BRI on PV exports from China to SSA.
Having reached 10 years of implementation in 2023, the BRI has received a number of assessments as elaborated by the policy background and theoretical analysis. Over time, PV trade between China and other BRI members has evidently increased with a collective mediating influence from “five-pronged” approach, except for financial coordination whose effect may be inconsequential. The missing specific study on BRI’s influence on the exportation of PV products from China to SSA is an essential contribution to this research area.
3. Methodology and data
3.1. Empirical models
3.1.1. Baseline model
The effect of BRI policy implementation on PV exports from China to SSA is baselined on the following OLS regression model, pending construction of staggered DiD models.
1
where,
is PV exports from China to country
in year
is a dummy variable indicating whether a country
is treated (value 1) or not (value 0) in year
;
is a set of control variables.
The research employs a quantitative approach to conduct a dynamic Difference in Difference (DiD) analysis from 2012 to 2022 using Stata statistical package. The paper combines classical staggered DiD with doubly-robust inverse probability weighting estimators (DR-IPW) and propensity score matching (PSM) in order to ensure both causal identification and robustness to model choices. By extending the staggered DiD techniques to causal mediation analysis in R software, the research seeks to further the understand mechanisms through which the BRI influences PV exports from China to sample SSA countries. A summary identification strategy for the empirical process is illustrated in Fig. 3.
3.1.2. Dynamic effects model
The DiD method is typically used for assessing the influence of policies, like the BRI. Particularly, the dynamic DiD model is suitable for event studies where the staggered treatment implementation or the treatment effect is expected to vary across various units (Callaway & Sant’Anna, 2021; Freedman et al., 2023). The dynamic DiD model addresses shortcomings of the Two Way Fixed Effects (TWFE) model by, among other ways, handling heterogenous treatment effects by estimating the treatment effect for each group examinable at each point in time. Moreover, using techniques like inverse probability weights, Jansen (2025) noted that new DiD methods automatically ‘correct’ for the TWFE biases.
However, DiD estimation relies on a complex set of key assumptions and conditions related to the staggered treatment adoption, no anticipation effects and parallel trends assumption with ‘never-treated’ or ‘not-yet-treated’ units (Callaway & Tsyawo, 2023). Nevertheless, dynamic DiD results can reliably be used to analyze the BRI influence over time and gain deeper insights into the mechanisms and impact on PV exports.
Based on validated parallel-trends assumption, the conventional TWFE event-study model using OLS is improved by the following dynamic DiD version;
2
where,
is year in which country
signed BRI,
is event time
and
is BRI effect in
years after adoption.
Another tool emphasized by Callaway and Sant’Anna (2021) is the use of not-yet-treated control groups which change over time as more units become treated. The dynamic approach is more representative and more suitable than using static never-treated control groups for the research because almost all African countries eventually became BRI members by 2022 (China Green Finance & Development Center, 2024). The model setup eventually, allows for a robust estimation of the treatment effect by accounting for both time-invariant and time-varying confounders.
Fig. 3
Identification strategy for the empirical process.
3.1.3. Causal mediation models
A
The two-step Baron and Kenny (1986) style is used to conduct mediation analysis of both hard and soft mechanisms. Qin (2024) pointed out that causal mediation analysis is important for achieving an in-depth understanding of how a treatment generates an impact on an outcome by uncovering the fundamental pathways. The causal assumption is that there is that there is no unmeasured confounding for both the treatment-mediator and mediator-outcome relationships. Therefore, the following models were used to estimate mediation effects of the mechanisms in the analysis of the influence of BRI on PV exports, where
is mechanism
,
is BRI dummy,
is a matrix of control variables
and
is PV exports.
Step 1; Mediator model:
(3a)
Step 2; Outcome model
(3b)
In equations (3a) and (3b), β is the direct effect called Average Direct Effect (ADE) of BRI on PV exports after controlling for the mediator,
gives the indirect or mediated effect called Average Causal Mediation Effect (ACME), and
gives the total effect (ADE + ACME) transmitted via the mediator.
The research uses parametric bootstrapping to produce confidence intervals with the percentile method. Unlike traditional Sobel Tests, Rauchfleisch (2024) recommends bootstrapping techniques because the latter is more flexible for handling multiple mediators and does not assume normality of the sampling. Therefore, parametric bootstrapping suits the small sample of this research where normality may not hold and provides empirical confidence intervals with higher statistical power. The intensive computations and visuals of the bootstrapping are done using mediation package in R whereby Imai, Keele and Tingley (2010) extends the Baron and Kenny (1986) framework to allow causal interpretation and sensitivity analysis which tests the model’s sequential ignorability assumption.
3.2. Sampling
Primarily, DiD models require definition of treatment, treated group and control group. BRI membership is the treatment done through cooperation agreements between China and other countries in the same or different years. For the staggered DiD identification strategy, the control groups are made by the same countries as in the treated group but in their periods before their respective BRI signing with China (Not-yet-treated option for staggered DiD analysis).
The China Green Finance and Development Centre (2025) recorded that 48 SSA countries had signed BRI MoUs with China. Among the countries, the earliest BRI agreement with China was signed in 2015 and the latest agreement was signed in 2022. However, 2022 was also the most recent year with available data for both the control and the mediating variables which were included for analysis.
Further, according to Baker, Larcker and Wang (2021), a staggered DiD design which requires the treatment to be dynamic would involve limiting the number of years such as 3 to 5 years before and after the treatment year for each unit. Using the 3-year allowance maximizes the utilization of available data from 2012 to 2022, but reduces the number of countries for analysis to include only the 37 countries which signed BRI cooperation agreements with China from 2015 to 2019.
3.3. Variables and data sources
3.3.1. Variables
The dependent, explanatory, mediating and control variables, including their definitions or measurements are listed in Table 1.
Variable type | Name | Symbol | Definition |
|---|---|---|---|
Dependent | Solar PV Exports from China to Sub Saharan African countries (SSA) | PV | Annual exports of solar PV products from China to a BRI partner country: 17 HS Codes. |
Explanatory | BRI treatment | briD | 1 for signed agreement and 0 for no signed BRI agreement in a given year |
Mediating | Diplomatic Relations | briP | Number of years before or after establishing bilateral diplomatic ties (most recent year for any renewed diplomatic relations) |
Unimpeded Trade | briT | Total commodity trade between China and BRI partner country | |
Infrastructure Connectivity (briI) | briI | Gross Annual Revenues of Chinese Companies' Construction Projects | |
People-to-People Bonds (briB) | briB | Number of Chinese workers on contracted and labor services in a BRI partner country by the end of each year | |
Financial Coordination (briF) | briF | Amount FDI stock from China in a BRI country | |
Control | GDP per capita | gdp | GDP per capita (current U.S. dollars) |
Renewable energy consumption | cons | Percentage renewable energy consumption of total final energy consumption | |
Rural population | ruralpop | Percentage of total population living in rural areas | |
Tariff rate | tariff | Percentage applied simple mean of tariff on all products | |
Access to electricity | accelec | Percentage of population with access to electricity | |
Merchandise trade | merchatrade | Merchandise trade percentage of GDP |
The dependent, explanatory and mediating variables in Table 2 are mainly based on literature from Zhu et al (2023), Hu et al (2023) and Chen Y. (2019). Moreover, according to Elish and AboElsoud (2024), Hu et al. (2023) and Zhu et al. (2023), the chosen control variables are GDP per capita (lnpgdp), renewable energy consumption (lncons), rural population (ruralpop), tariff rate (tariff), access to electricity (accelec) and merchandise trade (merchatrade).
3.3.2 Data sources
In line with the sampling methodology in section 3.2, the research utilizes data from 2012 to 2022. Trade data, including PV exports, is obtained from the United Nations COMTRADE database; data for the BRI members and their respective dates of joining the BRI was obtained from the China Green Finance and Development Centre website; information on diplomatic ties between China and SSA countries was obtained from the China Ministry of Foreign Affairs website; and bilateral data for the mechanisms was obtained from the China Africa Research Initiative (CARI). Data for all the control variables was sourced from the World Bank’s World Development Indicators (WDI) database.
Several improvements been made in the choice of indicators for the variables in Table 2. Firstly, this research has refined and increased the number of HS codes for PV products (dependent variable). Zhu et. al (2023) used only one HS code, 854140, for PV products. Hu et al. (2024) extended the scope to 16 HS codes. In this research, two of 17 Hu et al.’s (2023) selected codes, 841950 and 841989 for heat exchange and temperature change materials, have been replaced by more appropriate metallic products, 730890 and 761090 used for making structures of PV products. Moreover, HS code 901380 for optical devices, appliances and instruments has been included in the PV product list. Thus, the total number of HS Codes for PV products used in this research is 17.
The second improvement in Table 2 is about the indicators for infrastructure connectivity (I) and people-to-people bonds (B) which have been selected based on relevance to this research and availability of data. For example, Hu et al. (2023) used weighted distances between countries’ capitals to measure facilities connectivity. Such an indicator is suitable for countries which are near China, unlike for this research whereby all the SSA countries have capitals which are located over 7,500 kilometers from Beijing. For people-to-people bonds and cultural integration, Hu et al. (2023) shared language as an indicator for is more applicable to countries around China than it is for countries in Africa. Thus in this research, the indicators for infrastructure connectivity and people-to-people bonds are respectively ‘Gross annual revenues of Chinese companies' construction projects’ and ‘Number of Chinese workers on contracted and labor services in a BRI partner country by the end of each year’ using data from the CARI.
3.4. Descriptive statistics for variables
Table 2 shows the descriptive statistics for the overall sample of 37 SSA countries, including the number of observations, minimum, maximum, mean and standard deviation for each variable. The logarithmic values of were used for solar PV exports, GDP per capita, energy consumption, people-to-people bonds, FDI, merchandise trade and infrastructure revenues. Missing values were recorded for GDP per capita, tariff and merchandise trade due to unavailable data. The highest solar PV exports were recorded for South Africa (US$706,101,538) in 2013; and the lowest for Mauritania (0) in the same year. Although all the standard deviations were below the mean values for all the variables, the highest variations were observed for merchandise trade, access to electricity, rural population and duration of diplomatic relations. The trends reveal relatively high variations of international trade dependence, energy access, urbanization and diplomatic relations among SSA countries.
Variable | Obs | Mean | Std. Dev. | Min | Max |
|---|---|---|---|---|---|
lnpv | 407 | 16.381 | 1.841 | 0 | 20.375 |
lnpgdp | 400 | 7.276 | .910 | 5.347 | 9.86 |
lncons | 388 | 3.921 | .817 | .588 | 4.57 |
ruralpop | 407 | 56.038 | 17.562 | 9.265 | 88.806 |
tariff | 274 | 11.532 | 3.656 | 0 | 19.97 |
accelec | 407 | 48.486 | 23.806 | 3.3 | 100 |
merchatrade | 398 | 55.834 | 34.118 | 15.208 | 263.334 |
briP | 407 | 40.189 | 14.88 | -4 | 63 |
lnbriB | 407 | 7.053 | 1.721 | 0 | 10.83 |
lnbriF | 407 | 5.362 | 2.024 | 0 | 8.919 |
lnbriT | 407 | 20.983 | 1.647 | 16.52 | 24.901 |
lnbriI | 407 | 5.777 | 1.723 | 0 | 8.93 |
Source: results by the authors. | |||||
4. Empirical results
4.1. Baseline results
Table 3 displays baseline estimates (model 1). Across all variable sets (columns 1 to 7), the coefficient (0.5 to 0.6) for BRI treatment stays robustly positive (about 0.6) and statistically significant at 1% level. This suggests that the BRI treatment has a positive and statistically significant effect on solar PV exports, even after controlling for other factors. The R-squared increases from 0.024 in the first model to 0.331 in the seventh model, indicating that adding more variables improves the model's explanatory power. The number of observations decreases as more variables are added, likely, due to missing data in some of the control variables.
(1) | (2) | (3) | (4) | (5) | (6) | (7) | |
|---|---|---|---|---|---|---|---|
Variables | lnpv | lnpv | lnpv | lnpv | lnpv | lnpv | lnpv |
1.treat | 0.570*** | 0.579*** | 0.509*** | 0.486*** | 0.521*** | 0.625*** | 0.641*** |
(3.16) | (3.19) | (2.85) | (2.70) | (2.63) | (2.92) | (3.33) | |
lnpgdp | 0.149 | 0.851*** | 0.741*** | 0.845*** | 0.997*** | 1.380*** | |
(1.50) | (5.74) | (4.14) | (4.15) | (4.22) | (6.32) | ||
lncons | 1.053*** | 1.035*** | 1.149*** | 1.107*** | 0.998*** | ||
(6.35) | (6.22) | (5.25) | (5.01) | (5.00) | |||
ruralpop | -0.008 | -0.018** | -0.019** | -0.035*** | |||
(-1.10) | (-2.05) | (-2.20) | (-4.24) | ||||
tariff | -0.021 | -0.014 | -0.109*** | ||||
(-0.66) | (-0.42) | (-3.42) | |||||
accelec | -0.009 | -0.028*** | |||||
(-1.27) | (-3.91) | ||||||
merchatrade | -0.033*** | ||||||
(-7.76) | |||||||
Constant | 16.107*** | 15.046*** | 5.799*** | 7.108*** | 6.819*** | 6.289*** | 8.616*** |
(128.50) | (20.44) | (3.55) | (3.52) | (2.90) | (2.63) | (3.97) | |
Observations | 407 | 400 | 382 | 382 | 260 | 260 | 260 |
R-squared | 0.024 | 0.031 | 0.116 | 0.118 | 0.167 | 0.172 | 0.331 |
| Note: t-statistics in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.1 | |||||||
Overall, the BRI treatment consistently shows a positive and significant effect on solar PV exports across all models. The control variables also provide insights into other factors affecting solar PV exports to be explored further in line with the policy analysis in section 2. Although the coefficients for other control variables are consistent with international trade theories, the negative coefficients for merchandise trade would be influenced by a combination of factors such as the trade composition and solar PV policy and regulatory environment that require further analysis.
4.2. Robustness checks
4.2.1. Dynamic effects test
Firstly, the parallel trends assumption is validated by making a reference to the pre-treatment coefficients which are not significantly different from zero for 2012 and 2013 coefficients as shown in Fig. 4. In the absence of the BRI, the treated and control groups would have followed similar paths. To create visual symmetry around the treatment year, pre_1 has been omitted. In the proceeding specifications, pre_1 will also not be included in order to avoid perfect collinearity of the baseline with the constant.
Fig. 4
Parallel trends test result
Note
The figure compares differences in the trend of changes of Chinese solar PV exports to SSA countries before and after joining the BRI.
In the treatment period (post_0), the effect is still around zero, suggesting that the immediate impact of the BRI on solar PV exports at the times of signing was not substantial or detectable. Post-treatment periods starting from post_1 show that the effects become positive, indicating an increase in solar PV exports. Although the individual post-treatment coefficients are not always significant as disaggregated in Table 4 (Column 1), the effects are positive and large ranging from 52% to 106%.
The results show a significant increase in solar PV exports starting from 2014, with the largest effect observed in 2016. It means that BRI impact on solar PV exports to SSA is not immediate. It builds over two to three years and persist, supporting dynamic, delayed, but durable treatment effect. Table 4 shows detailed results of dynamic effects model and other robustness tests.
Variables | (1) Dynamic effects | (2) PSM-alternative clustering | (3) PSM-Control variables | (4) Effectiveness test (Deposition- DR-IPW) |
|---|---|---|---|---|
lnpv | lnpv | lnpv | lnpv | |
1.treat | -0.033 to 0.304 | 0.108 | 0.136 | 0.114 |
(0.125 to 0.816) | (0.163) | (0.237) | (0.302) | |
lnpgdp | 1.811*** | 1.822*** | 1.838*** | |
(0.361) | (0.348) | (0.364) | ||
lncons | 0.226 | 0.232 | 0.265 | |
(1.028) | (0.982) | (1.035) | ||
ruralpop | -0.089 | -0.087 | -0.084 | |
(0.061) | (0.061) | (0.063) | ||
tariff | 0.006 | 0.005 | 0.005 | |
(0.033) | (0.034) | (0.033) | ||
accelec | -0.017 | -0.017 | -0.017 | |
(0.011) | (0.012) | (0.011) | ||
merchatrade | 0.021** | 0.021** | 0.021** | |
(0.008) | (0.008) | (0.008) | ||
_cons | 6.352 | 15.690*** | 6.241 | 5.877 |
(8.153) | (0.215) | (7.654) | (8.194) | |
R-squared | 0.330 | 0.109 | 0.324 | 0.324 |
Observations | 260 | 407 | 260 | 260 |
Note: Standard errors in parentheses, * p < 0.1, ** p < 0.05, *** p < 0.01 | ||||
4.2.2. Propensity-Score Matching
a.
Alternative clustering
There is a large, significant and durable BRI impact across the 2014–2022 window with alternative clustering. When errors are clustered differently for the propensity-score matched sample, the point estimates are slightly higher but still positive. With reference to Table 5 (Column 2), for instance in 2014, alternative clustering gives has 0.632 against 0.520 for the dynamic effect. The standard errors also improve turning more years significant at 1%. The parallel trend assumption still holds under the new clustering. The downward bias in the unmatched sample is modest and the main conclusion survives after re-clustering and re-weighting. The significant constant in the model suggests a strong baseline level of solar PV exports which is not as pronounced after the inclusion of control variables and the use of alternative methods of estimation with smaller sample sizes.
b.
Control variables
After balancing observables, the BRI still delivers a 50–90% solar PV export lift (Column 3 of Table 5). However, adding the controls lowers precision as the sample shrinks to 260 observations and the effects lose significance after 2017. The point estimates stay positive and significant at 10% for 2014–2017 period. It is important to note that R-squared jumps from 11% in Column 2 to 32% in Column 3; indicating that the controls absorb noise and improve the explanatory power of the whole model.
The control-variable pattern shows that GDP per capita and merchandise trade, now showing a positive relationship under the dynamic analysis, play significant roles in increasing solar PV exports from China to SSA. Thus, richer countries and trade-open countries import more solar PV products from China. However, the significant roles of GDP per capita and merchandise trade do not explain away the BRI effect.
4.2.3. DR-IPW
a.
Effectiveness (Deposition) test
The DR-IPW confirms that the BRI effect is neither a result of matching nor of parametric assumptions. Column 4 of Table 4 gives virtually identical estimates to those in PSM-controls (
difference) showing that results are not model driven. The standard errors are slightly larger, but 2014–2017 and 2020–2022 coefficients remain significant and without sign changes. Moreover, the pre-treatment years, 2012–2013 are near or equal to zero confirming no anticipation and no hidden trends assumptions for staggered DiD analysis.
Overall, BRI membership causes a sustained 50–100% increase in Chinese solar PV exports to SSA. This result is robust to sample construction, error structure, covariate adjustment, and double-robust estimation.
b.
CSDID
Using the DR-IPW estimator again for CSDiD analysis further points out that the effect of the timing of BRI signing on Chinese solar PV exports to SSA countries. The effect is concentrated among the first SSA countries (including Cameroon, Comoros, Somalia and South Africa in 2015) to sign. The later joiners still require enough post-period data to reveal estimable impact from the sample data. Table 5 shows a summary of CSDiD results analyzed using Treatment-on-the-Treated (ATT) by cohort, parallel –trend (placebo) checks and inverse-probability weights. The control for the CSDiD model is the ‘not-yet-treated’ years for each country in the sample.
Category (years) | Key ATT coefficients and weight values | Standard error | |
|---|---|---|---|
ATT paths by cohort | g2015 | t_2014_2015 = + 0.247 | 1.563 |
g2017 | t_2016_2018 = + 0.747 | 1.350 | |
g2018 | t_2017_2018 = + 0.496 | 1.340 | |
Parallel-trend diagnostics | t_2012_2013 | -0.031 to 0.321 | 1.045 to 1.474 |
t_2013_2014 | -0.239 to 0.839 | 1.963 to 1.309 | |
Inverse-probability weights | w2015_* | 0.020*** | 0.007 |
w2017_* | 0.015** | 0.006 | |
w2018_* | 0.131*** | 0.017 | |
Note: * p < 0.1, ** p < 0.05, *** p < 0.01 | |||
Generally, the early signers (g2015) drive all the statistically meaningful gains in solar PV exports from China to SSA with a steady build-up of solar PV exports by about 28% (e^0.0247-1). Although the aggregate ATT (0.465, SE
0.949) does not show heterogeneity consistent with the low but balanced inverse-probability weights, the disaggregated CSDID analysis shows that the early adopters of BRI have big gains in Chinese solar PV exports. The parallel-trend diagnostics further confirm no differential and systematic pre-treatment effects. However, the 2019 group was excluded from aggregate ATT due to no post-treatment years for construction.
5. Mediation analysis for Hard and Soft Connections
Each of the five BRI priority areas of cooperation has positive and equivalent total effect (0.377) on BRI’s impact on solar PV exports from China to SSA; but the hard mechanism of financial coordination and trade connectivity are the only mechanisms with significant mediating effects. Table 6 shows two-stage causal mediation results using nonparametric bootstrapping 500 simulations of 170 clean observations. The mediation (ACME), direct (ADE) and total effects of each mechanism are measured based on model 3a and model 3b whereby the control variables are gdp per capita (pgdp) and merchandise trade percentage of gdp (merchatrade) with reference to their significance in the dynamic effects analysis.
Mediator effects | Soft Mechanisms | Hard Mechanisms | |||
|---|---|---|---|---|---|
Diplomatic relations | People-to-people bonds | Financial coordination | Trade connectivity | Infrastructure connectivity | |
ACME | -0.033 | -0.099* | 0.299** | 0.212* | 0.009 |
[-0.080, 0.011] | [-0217,-0.001] | [0.102,0.495] | [0.026,0.388] | [-0.116,0.134] | |
ADE | 0.410** | 0.476*** | 0.077 | 0.165 | 0.367** |
[0.114, 0.667] | [0.163,0.731] | [-0.170,0.349] | [-0.079,0.411] | [0.074,0.629] | |
Total Effect | 0.377* | 0.377* | 0.377* | 0.377* | 0.377* |
[0.075,0.657] | [0.071,0.656] | [0.064,0.680] | [0.063,0.653] | [0.053,0.673] | |
Prop. Mediated | -0.088 | -0.263 | 0.794* | 0.562* | 0.025 |
[-0.444,0.038] | [-1.925,0.017] | [0.370,2.298] | [0.092,1.750] | [-0.851,0.393] | |
Sample size used: 170 Simulations: 500 | |||||
| Note: 95% Lower and 95% Upper Confidence Intervals in parentheses. * p < 0.1, ** p < 0.05, *** p < 0.01 | |||||
5.1. Soft Connection mechanisms
5.1.1. Diplomatic Relations
The strong direct effect (0.410) of diplomatic relations implies that BRI’s diplomatic relation efforts may boost solar PV exports such as through high-level agreements or bilateral partnerships. In line with Hu et al.’s (2023) findings, improved diplomatic relations harmonize trade objectives, facilitates technology transfer and strengthens bilateral cooperation; in the process reducing trade barriers and facilitating access for Chinese solar PV products in SSA countries.
However, the negative but insignificant mediation effect (-0.033) of diplomatic relations suggests that inconsistent policy implementation or lack of alignment between BRI policies and local renewable energy needs in SSA may not effectively mediate Chinese solar PV exports to SSA.
5.1.2. People-to-people bonds
The people-to-people bonds mechanism has the strongest significant direct effect (0.476) likely due to the BRI’s direct investments in human capital, including training and exchange programs which could enhance trade relationships, but not specifically for solar PV exports. The significant but negative mediation effect (-0.099) suggests that the cultural or exchange programs might divert attention and resources away from solar PV trade whilst SSA communities prioritize other goods over solar technology.
Thus, the insignificant proportion (-26.3%) mediated by people-to-people bonds supports that this mechanism does not positively mediate solar PV exports. While Hu et al. (2023) found significant effect of people-to-people bonds for solar PV trade between China and its neighboring countries, cultural and institutional distance affect the mechanism’s significance. Therefore, it is likely that the mediating effect of people-to-people bonds on solar PV trade is not immediate or easily quantifiable in the short to medium terms for SSA countries.
5.2. Hard Connection mechanisms
5.2.1. Financial coordination
This research shows that financial coordination is the primary driver of Chinese solar PVs to SSA. Contrary to the results of Hu et al. (2023), financial coordination has a significant and positive mediating effect (0.299) on Chinese solar PV exports to SSA. Although the direct effect (0.077) is not significant, BRI’s financial mechanisms such as FDI, loans and currency swaps likely facilitate solar PV exports by easing payment barriers and reducing transaction costs and risks associated with cross-border trade in solar PV products.
A
The critical role of financial coordination is revealed by its significant high proportion (79%) of mediated effect. Chen X., Zhang and Yang (2025) confirm BRI’s impact on exports through financial support and reduction of capital constraints. With respect to creating more opportunities for increasing trade in SSA countries, Wang et al (2024) also emphasized existing and required financial support and investment in SSA countries whose majority have low income. Financial integration could reduce transaction costs and risks associated with cross-border trade in solar PV products.5.2.2. Trade connectivity
The mediation effect (0.212) of trade connectivity is positive, significant and accounts for about 56.2% of its total effect. BRI-driven trade connectivity such as streamlined customs, reduced tariffs and logistics improvements actively facilitates Chinese solar PV exports to SSA. Although trade connectivity might face broader ‘third-party’ friction effects within the trade network inferred from the insignificant direct effect (0.165), it entirely promotes solar PV exports through harmonized trade relations and improved infrastructure (Liu et al., 2021, as cited in Zhu et al., 2023).
The trade connectivity effects also reveal interrelatedness of the mechanisms. For instance, trade barriers may be overcome through enhanced diplomatic relations and infrastructure connectivity; leading to joint long run effects from tariff reductions, market integration and enhanced supply chain efficiency for increased solar PV exports.
5.2.3. Infrastructure connectivity
The direct effect (0.367) of infrastructure connectivity is positive and significant. BRI’s general infrastructure investments may benefit exports by improving overall business environments in SSA. However, the indirect effect (0.009) of this mechanism is not significant, possibly because physical infrastructure projects such as roads and power grids are not yet operational to meaningfully facilitate solar PV trade. Moreover, IEA (2023) noted that BRI’s infrastructure faces misaligned priorities. Despite the difference between the direct and indirect effects, improved facilities make transportation of solar PV products easier and cheaper to increase exports in the long-term. Chen, X., et al. (2025) and Zhu et al. (2023) support that infrastructure is a ‘precursor’ for trade.
5.3. Sensitivity analysis
Mediation sensitivity analysis is used for evaluating the robustness of mediation effects in the presence of unmeasured confounding between the treatment and the mediator. The ACME as a function of sensitivity parameter
), is used to assess how sensitive the estimated mediation effect is to potential violations of the assumption that the mediator and outcome are conditionally independent given the treatment and covariates (Imai et al., 2010). The sensitivity graphs for each mechanism are presented in Fig. 5 including a summary table of critical values (
at which ACME = 0 or where the solid line intersects the dotted line). If the ACME confidence bands do not overlap the dotted line across the range of
, then mediation effect is highly robust. Moreover, if the bands cross the dotted line at low
values, then the result is sensitive to minor confounding. Since the analysis assumes no unmeasured confounding at
, robustness of a given mechanism increases as its absolute critical value increases.
Figure 5 confirms that the causal interpretation is robust at different levels for the five mechanisms. Specifically, the negative critical value (
) for diplomatic relations suggests that the diplomatic relations mechanism is relatively sensitive because hidden factors like geopolitical tensions which may affect both BRI diplomatic relations and solar PV exports would need weak correlation in order to reduce the mediation effect. People-to-people bonds and infrastructure connectivity have moderate robustness, each having
. Financial coordination and trade connectivity have high robustness each with
; and it is not surprising that these two mechanisms show significant mediating effects.
In summary, the causal mediation analysis shows that the BRI’s influence on solar PV exports is primarily mediated by the hard mechanisms of financial coordination and trade connectivity, respectively contributing 79.4% and 56% to their own total effects. All the mechanisms indicate positive total effects averaging 0.377. However, diplomatic relations, people to people bonds and infrastructure connectivity play significant roles in the direct effects of BRI impact on solar PV exports. These findings suggest that while some mechanisms like financial coordination and trade connectivity enhance the BRI’s impact, others might not be as effective or could lessen the overall positive effects. The mediation sensitivity analysis confirms that financial and trade linkages under the BRI are the most reliable mediators of Chinese solar PV exports to SSA, while diplomatic relations is the least robust due to its sensitivity to minor confounding. Nevertheless, both hard and soft mechanisms are important for facilitating Chinese solar PV exports to SSA.
Fig. 5
Mediation sensitivity analysis for Average Causal Mediation Effect
Note
Results obtained after conducting nonparametric bootstrapping using R software.
6. Conclusions, Recommendations and Limitations
6.1. Conclusions
The paper employs complementary dynamic DiD frameworks followed by causal mediation analysis of BRI mechanisms in order to analyze 2012 to 2022 data on solar PV exports from China to SSA. The results show that:
6.1.1. The overall impact of BRI membership is positive.
BRI membership significantly increases Chinese solar PV exports to SSA countries. The magnitude of this impact varies depending on the country’s time of joining the BRI framework. Early members experience substantial and sustained gains while the impact on later members requires more recent data to confirm significance. These findings are economically meaningful and robust across alternative estimators, including DR-IPW and PSM.
6.1.2. The BRI influence is transmitted mainly through hard mechanisms.
A causal mediation analysis of both hard and soft mechanisms shows that financial coordination and trade connectivity have significant influence on solar PV exports, demonstrating high robustness to unmeasured confounding. These hard mechanisms are the primary channels through which the BRI membership impacts Chinese solar PV exports to SSA.
6.1.3.
Infrastructure connectivity and soft mechanisms do not significantly mediate the BRI impact in the short term.
While infrastructure connectivity, diplomatic relations and people to people bonds contribute to the direct effect of BRI membership on solar PV exports, their respective mediation effects are not significant within the period analyzed. In addition to infrastructure connectivity, the soft mechanisms are important but they do not appear to mediate the BRI impact to a substantial degree in the short term.
6.2. Recommendations
The following recommendations are made based on the above conclusions:
6.2.1. Prioritizing and monitoring energy-specific infrastructure.
Renewable energy mini-grids, off-grid solar systems, and other energy-specific infrastructure are important in SSA (World Bank, 2022). Investment in such infrastructure is key to realizing BRI’s influence on complementary Chinese solar PV exports to SSA. Apart from generic infrastructure projects like roads, railways and ports, it is also important to monitor and ensure that logistical and transportation benefits from facilities connectivity promote solar PV export growth. This research supports Hu et al.’s (2023) recommendation to establish a dedicated BRI Solar Trade Center in China due to the increasing global attention towards solar energy. An equivalent institution should be established in Africa with respect to its growing demand for affordable and sustainable energy.
6.2.2. Reassessing and targeting soft mechanisms towards promoting trade in solar PV products.
BRI policy alignment and cultural exchange interventions should focus on strengthening financial mechanisms and improving trade relationships. For instance, the BRI could negotiate BRI trade facilitation agreements with clear provisions for renewable energy and expand green financing such as provision of concessional loans for African solar projects tied to Chinese PV exports. Moreover, it is important to align local solar energy adoption policies with cultural exchanges that include renewable energy training programs.
6.2.3. Enhancing the availability or accessibility of up-to-date data on BRI indicators.
To achieve both policy and research recommendations, however, it is important to make available or accessible up-to-date data on BRI indicators, including disaggregated FDI for renewable energy sectors. Availability of information on a majority of indicators put forward by the Peking (2018) Index would enhance the monitoring and evaluation of the influence of the BRI on solar PV trade and other areas of interest. Specifically, a web-based indicator platform for tracking the progress of the BRI five priorities would be useful for both researchers and policy experts.
6.3. Limitations and Future Research Directions
Although the results survive relatively strict doubly-robust checks, the aggregate precision is limited by short follow-up ending in 2022 and small matched sample for propensity scores. Therefore, it is important that further research should include more recent years for later BRI joiners and other developing countries beyond the SSA region and Africa. Secondly, this research (section 4.2) has only provided insights on temporal heterogeneity for early versus late adopters of BRI in terms of the effect on solar PV exports from China to SSA countries. Further research can explore how the impact may vary across other subgroupings within the SSA region. In addition, the causal mediation analysis assumes linearity and no mediator-outcome interactions. Babin, Cano-Sancho, Vigneau, and Antignac (2023) argued that considering all mediators separately may fail when the mediators interact with each other. Further researcher should, therefore, consider using adjusted approaches and non-linear models such as Structural Equation Modelling (SEM) in order to further study the BRI mechanisms and their possible interactions.
A
Author Contribution
Wenjia WANG: Conceptualization, Methodology, Resources, Writing - review & editing. Blessings ZITTA: Data Curation, Writing - original draft. Xiaoxia SHI: Mechanism analysis. Jing SHUAI (Corresponding author): Writing - review & editing, Funding acquisition. Chuanmin SHUAI (Corresponding author): Supervision.
References
Africa-China Centre for Policy & Advisory (2023), November A decade of China’s BRI: assessing its impact & implications for Africa’s Agenda 2063
A
AREI (2025) Africa Renewable Energy Initiative. AU-AFRECA
Babin É, Cano-Sancho G, Vigneau E et al (2023) A review of statistical strategies to integrate biomarkers of chemical exposure with biomarkers of effect applied in omic-scale environmental epidemiology. Environ Pollut330 (2023):121741Baker AC, Larcker DF, Wang CCY (2021) How much should we trust staggered difference-in-differences estimates? Harvard Business School Working Paper: 21–112
Baron RM, Kenny DA (1986) The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. J Pers Soc Psychol 51:1173–1182
Callaway B, Sant’Anna PHC (2021) Difference-in-differences with multiple time periods. J Econom 225(2):200–230
Callaway B, Tsyawo ES (2023) Treatment effects in staggered adoption designs with non-parallel trends. J Econom. Advance online publication
Chadly A, Moawad K, Salah K et al (2024) State of global solar energy market: overview, China's role, challenges, and opportunities. Sustain Horiz 11:1–30
Chen J, Liu Y, Liu W (2020) Corrigendum to Investment facilitation and China's outward foreign direct investment along the belt and road. China Econ Rev 62:101507
Chen X, Zhang M, Yang G (2025) The Belt and Road Initiative and risk transmission mechanisms in China’s manufacturing global industry chain: a structural vulnerability analysis. J Glob Dev 10(2):123–145
Chen Y, Wu Y (2019), April 18 Report on policy coordination index-five connectivity index report (2018). Taihe Institute
Chin I (2023), May 31 What are the 5 priorities of the BRI cooperation? Our China Story
China Green Finance and Development Centre (2025) Countries of the Belt and Road Initiative (BRI)
China State Council Information Office (2023), October Full text: the Belt and Road Initiative: a key pillar of the global community of shared future
Chiyemura F, Shen W, Chen Y (2021) Scaling China’s green energy investment in Sub-Saharan Africa: challenges and prospects. The African Climate Foundation
Elish E, AboElsoud ME (2024) Assessing the Belt and Road Initiative’s environmental footprint: an impact evaluation analysis of African member countries. Humanit Soc Sci Commun 11(1):989
Freedman SM, Hollingsworth A, Simon KI et al (2023) Designing difference in difference studies with staggered treatment adoption: Key concepts and practical guidelines (Working Paper 31842). National Bureau of Economic Research
Geng Q (2021) The Belt and Road Initiative and its implications for global renewable energy development. Curr Sustain/Renew Energy Rep 8:40–49
Green BRIC (2025) The Green Investment Principle (GIP) for the Belt and Road Initiative. Green Belt and Road Initiative Center
He HY, Gui CX (2025), April 30 China PV Industry Review and Outlook
Hryniewicz J (2014) Core-periphery: An Old Theory in New Times. Eur Polit Sci 13
Hu L, Hu J, Huang W (2023) Evolutionary analysis of the solar photovoltaic products trade network in Belt and Road Initiative Countries from an economic perspective. Energies 16(16):6371
Huang Y (2016) Understanding China's Belt & Road initiative: motivation, framework and assessment. China Econ Rev 40:314–321
Hydropower (2024) December). Africa hydropower regional profile
IEA (2022) Special report on solar pv global supply chains
IEA (2023), June 9 Key findings of IEA's renewable energy market update - outlook for 2023 and 2024
A
IEA IRENA, World Bank., WHO (2024) Tracking SDG 7: The energy progress report. World Bank, Washington DCImai K, Keele L, Tingley D (2010) A general approach to causal mediation analysis. Psychol Methods 15(4):309–334
Jansen A, November (2025) Difference-in-difference (DiD). Asjad Naqvi
Kluiver J (2024) July 24). Africa has much to gain from a more contained BRI. ISS Africa
Matiashe FS (2023) November 22). Malawi, other countries hard-hit by climate change count on new loss and damage fund. The Africa Report
Negash ES, Wang ZK, Knerly Y et al (2024) The impact of the Belt and Road Initiative and Chinese outward foreign direct investment on the bilateral trade between China and Sub-Saharan Africa countries. Singap Econ Rev 69(1):35–59
A
Peking University (2018) World premier of the Belt and Road Initiative: 2018 report on five connectivity indexes at Taihe Civilizations Forum. Peking University Institute of Ocean ResearchSabola GA (2024) The link between Chinese belt and road initiative and foreign direct investment inflows in Southern Africa. Borsa Istanb Rev 24(2):256–262
Secretariat of BRI International Green Center (2019) Green BRI Green Review. Secretariat of BRI International Green Development Coalition, Beijing, China
UNECA (2024), November 17 out of the 20 countries most threatened by climate change are in Africa, but there are still solutions to this crisis
United Nations (2024), December The 17 Goals
Rauchfleisch A (2024) Multivariate statistics [Lecture Notes]. Department of Political Science, University of Zurich
Wang Z, Lu Y, Zhang S et al (2024) Analysis of the BRI and China’s OFDI in Sub-Saharan Africa. Singap Econ Rev 69(1):35–59
World Bank (2022) World Bank annual report 2022: World Bank lending (Fiscal 2022)
Xiao A, Yifei D (2023), May Green Silk Road drives sustainable development in BRI landscape. China Daily
Zhu X, Gu Z, He C, Chen W (2023) The impact of the Belt and Road Initiative on Chinese PV firms’ export expansion. Environ Dev Sustain 26:25763–25783
Acknowledgements
This research is supported by the Key Project (No. 25AGL032) and General Project (No. 23BGL215) of the National Social Science Fund of China.
Competing interests
The authors declare no competing interests.
Ethical approval
This study does not involve human participants or their data so no informed consent and other forms of ethical approve is needed.
Informed consent
No informed consent was needed for this paper. This article does not contain any studies with human participants performed by any of the authors.
1
Angola, Benin, Burundi, Cabo Verde, Cameroon, Chad, Comoros, Côte d'Ivoire, Djibouti, Equatorial Guinea, Ethiopia, Gabon, The Gambia, Ghana, Guinea, Kenya, Lesotho, Liberia, Madagascar, Mali, Mauritania, Mozambique, Namibia, Nigeria, Rwanda, Senegal, Seychelles, Sierra Leone, Somalia, South Africa, South Sudan, Sudan, Togo, Uganda, Tanzania (URT), Zambia and Zimbabwe.
2
A total of 17 HS codes were selected for the solar PV product list: 700991, 700992, 711590, 730890, 732290, 761090, 830630, 841280, 841919, 841990, 850239, 850440, 854140, 900190, 900290, 900580 and 901380.