lable at ScienceDirect Renewable Energy 196 (2022) 518e525 Contents lists avai Renewable Energy journal homepage: www.elsevier .com/locate/renene Geopolitical risk and renewable energy asset prices: Implications for sustainable development Anupam Dutta *, Probal Dutta School of Accounting and Finance, University of Vaasa, Finland a r t i c l e i n f o Article history: Received 21 December 2021 Received in revised form 1 July 2022 Accepted 4 July 2022 Available online 7 July 2022 Keywords: Geopolitical risk Clean energy ETFs Regime switching probabilities Volatility * Corresponding author. E-mail addresses: adutta@uwasa.fi (A. Dutta), pdu 1 This information is sourced from https://assets.b Energy-Transition-Investment-Trends-Exec-Summary https://doi.org/10.1016/j.renene.2022.07.029 0960-1481/© 2022 The Authors. Published by Elsevie a b s t r a c t This study intends to investigate the impact of geopolitical uncertainty, proxied by the geopolitical risk (GPR) index, on the volatility of renewable energy exchange traded funds (ETFs). Employing a two-state Markov regime switching model reveals that an upturn in the GPR index increases (reduces) the like- lihood of being in the low (high) volatility regime. This finding could be attributed to the fact that when the geopolitical risk increases, users of crude oil, which is highly sensitive to such risk, tend to consider clean energy as a substitute for traditional energy sources. This causes a growth in the equity prices of new energy firms, further leading to a drop in the levels of volatility. Additionally, the results of generalized autoregressive conditional heteroscedasticity (GARCH) models also confirm that higher GPR implies lower risk for these green assets. The outcomes have implications to policymakers and investors participating in clean energy markets. © 2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). 1. Introduction Over the past few years, demand for renewable energies has increased substantially due to growing concerns about climate change. Accordingly, investments in international clean energy markets have also experienced an ascending trend. For example, a recent study by Bloomberg New Energy Finance documents that global investment in transitional energy has reached $755 billion in 2021. The study also indicates that in order to achieve climate neutrality, investment in this sector would reach over $2 trillion between 2022 and 2025, and about $4.1 trillion during the period 2026e2030.1 Hence, investments in renewable energies will need to triple in the coming years with a view to reducing net carbon emissions to zero. Given that clean energy markets appear to be a relatively new class of assets to invest in, they could be highly volatile in nature [1]. Therefore, precise estimates of time-varying volatility are of paramount importance to market participants for understanding the risk of investor portfolio consisting of clean energy assets. Hence, it is crucial to identify the factors that drive the volatility of this emerging asset class. tta@uwasa.fi (P. Dutta). bhub.io/professional/sites/24/ -2022.pdf. r Ltd. This is an open access article To this end, the present study aims to investigate whether geopolitical risk (henceforth, GPR) can predict the uncertainty linked to clean energy asset returns. In doing so, we contribute to the literature in several aspects. First, based on our knowledge, it is among the preliminary studies to examine whether clean energy assets react to geopolitical uncertainty. Given that oil price varia- tions are affected by geopolitical risk through the supply and de- mand channels [2,3], GPR may exert a significant impact on clean energy assets from the intuition that oil and renewable energy prices are highly covariant [4]. In particular, geopolitical risk may affect clean energy assets via at least four channels. First, when increasing geopolitical risk has a positive effect on crude oil prices, traditional energy sources could be replaced by renewable energies [5]. This in turn improves the operating situations of new energy firms and their stock performance [6]. Second, investors’ expecta- tions for future oil supply and demand changes could influence their perspective about alternative energy industry, which is further reflected in their investment decisions [7]. Thus, variations in geopolitical risk may impact the clean energy asset returns through investor sentiment. Third, the transition to renewable energy leads to greater energy self-sufficiency and reduces geopolitical conflicts. Thus, increasing geopolitical risk will certainly encourage policy makers and investors to shift towards clean energies, which would have a positive effect on the renew- able energy asset class. Fourth, rising geopolitical risk often causes higher uncertainty leading to an immediate disruption of economic under the CC BY license (http://creativecommons.org/licenses/by/4.0/). http://creativecommons.org/licenses/by/4.0/ mailto:adutta@uwasa.fi mailto:pdutta@uwasa.fi https://assets.bbhub.io/professional/sites/24/Energy-Transition-Investment-Trends-Exec-Summary-2022.pdf https://assets.bbhub.io/professional/sites/24/Energy-Transition-Investment-Trends-Exec-Summary-2022.pdf http://crossmark.crossref.org/dialog/?doi=10.1016/j.renene.2022.07.029&domain=pdf www.sciencedirect.com/science/journal/09601481 http://www.elsevier.com/locate/renene https://doi.org/10.1016/j.renene.2022.07.029 http://creativecommons.org/licenses/by/4.0/ https://doi.org/10.1016/j.renene.2022.07.029 https://doi.org/10.1016/j.renene.2022.07.029 A. Dutta and P. Dutta Renewable Energy 196 (2022) 518e525 activities, which is further reflected in financial markets including stocks, exchange rates and commodities [8e10]. Due to a close connection with these financial markets, clean energy assets also become susceptible to geopolitical risk. Second, we employ the Markov regime switching (MRS) approach to study the impact of GPR on the probability of being in the low and high volatility regimes for different clean energy assets. Note that a number of studies find that GPR has a positive effect on global crude oil price, leading to a drop in its volatility levels [11,12]. Given that oil prices and clean energy asset returns are positively correlated [13e16], it could be postulated that GPRwould also exert a negative impact on the volatility of clean energy asset class. Assuming this indicates that when geopolitical uncertainty in- creases, the likelihood of remaining in the low (high) volatility state tends to increase (decrease). Hence, by focusing on the low (high) volatility state, we are able to examine how changes in the geopolitical risk levels impact the volatility of the clean energy asset class when stock markets are relatively calm (volatile). Doing so thus allows us to examine whether GPR drives the regime transition for clean energy asset returns. Third, we extend the discussion on the role of geopolitical risk in financial markets. More specifically, our study adds to the limited literature on the GPR-stock association at sector levels. Investi- gating such linkage is important given that there may be industry specific responses to shocks emanating from geopolitical conflicts. Since the magnitude of these responses may vary among the sec- tors, proper knowledge on how geopolitical risk impacts the vola- tility of clean energy asset class could be useful for developing appropriate hedging strategies to mitigate such risk. Hence, the information contained in our empirical results could provide insight into means of building accurate stock-valuation models and accurate forecasts of the volatility of these assets. It is worth noting that contrasting to the existing literature [6,17e19] focusing on clean energy asset class, we utilize the in- formation on renewable energy exchange traded funds (ETFs) instead of using renewable energy stock indexes. Considering these ETFs is beneficial given that unlike equities, they are not sensitive to non-synchronous trading issues. As documented by Lo and MacK- inlay [20], such issues may lead to spurious estimates when con- ducting market efficiency tests. Besides, these assets are particularly liquid and behave like a stock [21]. The empirical results might be useful for current and future investors in new energy companies. As growing concerns about environmental sustainability shift investors towards eco-friendly businesses, this research could attract those shareholders willing to decarbonize their portfolios by holding renewable energy assets. Moreover, socially responsible investments have ecological in- fluences that assure a certain degree of sustainability. Therefore, our investigation is particularly important for eco-friendly in- vestors who require sound knowledge on how to diversify the risk associated with their portfolios [22]. The findings of this study thus help such investors in detecting potential risk linked to green portfolios and gaining superior risk-adjusted returns. In sum, this strand of research offers stylized facts about socially responsible investments which market participants could consider when including clean energy assets in their portfolios to moderate climate related risk. 2. Literature review We now provide a brief review of prior studies that deal with new energy equity markets. This literature can be divided into two segments. The first strand of literature investigates the price spill- over linkage between dirty and clean energy assets. Important contributions in this segment include Henriques and Sadorsky [13] 519 Broadstock et al. [14], Reboredo et al. [15], Dawar et al. [16], Managi and Okimoto [23], Bondia et al. [24], among others. Henriques and Sadorsky [13], for instance, adopt a vector autoregression (VAR) process to explore whether clean energy stock prices react to oil price and technology stock price shocks. The authors show a sig- nificant connection amongst these markets. Using an asset pricing model, Broadstock et al. [14] find that the linkage between oil and Chinese clean energy stocks appears to be strong during the turmoil periods. The study also demonstrates a positive association between these variables implying that an upturn in oil prices would encourage investments in alternative energy firms. Moreover, Bondia et al. [24] also provide evidence of price spillover effects between clean and dirty energy assets. Employing the Granger causality method, the study shows that oil prices seem to rule the stock prices of renewable energy firms. Applying a continuous wavelet approach, Reboredo et al. [15] document a robust long- term relationship between oil and clean energy stocks, albeit such linkage is found to be fragile in the short run. A recent study by Saeed et al. [25] provides empirical evidence on the return spill- overs between clean and dirty energy assets in lower and upper quantiles. More recently, Dawar et al. [16] reveal that clean energy asset prices respond differently to new information on West Texas Intermediate (WTI) oil prices under diverse market conditions. The second line of literature examines the volatility linkage between commodity prices and new energy asset returns. Sadorsky [26], for example, finds volatility cross effects between fossil fuel and green stocks. Methodologically, the study considers a number of multivariate GARCH models and concludes that clean energy assets are a good hedge for portfolios comprising oil or dirty assets. Using the time-varying copula approach, Reboredo [27] shows that oil price volatility significantly contributes around 30% to downside and upside risk of new energy firms. Ahmad [28] uses the direc- tional spillover method, proposed by Diebold and Yilmaz [29], to document that technology and renewable energy stocks are the dominant emitters of volatility spillovers to the US crude oil mar- ket. Besides, Dutta [17] shows that crude oil implied volatility (hereafter, OVX) exerts a significant impact on the realized volatility of alternative energy stocks. A similar study by Ahmad et al. [1] finds gold and OVX as a good hedge for clean energy equities. Dutta et al. [30] document similar findings as well. Furthermore, Bouri et al. [31] show that crude oil along with gold appear to be safe- haven assets for new energy firms amid the turmoil periods. In addition, Xia et al. [32] demonstrate that fossil energyerelated products such as oil, gas, coal, electricity and carbon emit volatility to clean energy assets. More recently, Yahya et al. [33] combine a two-regime threshold vector error correction model with the DCC- GARCH process to document a long-term volatility linkage between oil and clean energy assets. 3. Materials and methods 3.1. Data The daily GPR data, recently created by Caldara and Iacoviello [34], are retrieved from http://www.policyuncertainty.com. Next, we collect the information on clean energy ETFs from Thomson Reuters DataStream database. Three different ETFs are studied in our research: Invesco WilderHill Clean Energy ETF (henceforth, PBW), Invesco Global Clean Energy ETF (henceforth, PBD) and Invesco Solar Energy ETF (henceforth, TAN). All these indexes allow investors to have an exposure to clean energy investments. Our sample covers the period from April 15, 2008 to March 10, 2020, yielding 2997 daily data points. The starting point of our sample is dictated by the availability of the clean energy data. It is also noteworthy that the daily GPR data are not available afterMarch 10, http://www.policyuncertainty.com Table 1 Descriptive statistics of clean energy ETFs and GPR. Index PBW PBD TAN GPR Mean �0.0372 �0.0212 �0.0687 109.56 Standard deviation 2.095 1.828 2.8546 85.51 Skewness �0.4747 �0.6818 �0.3886 2.6107 Kurtosis 9.356 13.52 10.07 18.05 Jarque-Bera test 5155.12*** 14056.45*** 6312.43*** 31700.67 ADF test �51.93*** �55.43*** �50.97*** �11.33*** PP test �51.88*** �55.42*** �50.87*** �42.40*** Notes: This table presents the descriptive statistics for different ETFs and GPR index. We consider the log-returns for these ETFS, while the GPR data are at levels. ***p < 0.01, **p < 0.05, *p < 0.10. A. Dutta and P. Dutta Renewable Energy 196 (2022) 518e525 2020. Fig. 1 depicts the time-series plots of different indexes consid- ered in this research. As exhibited in this graph, we identify a substantial fall in the clean energy asset prices during the current COVID-19 pandemic period. At the same time, the geopolitical risk has also increased significantly. We report the summary statistics in Table 1. Negative mean returns are observed for all the ETFs with TAN being more volatile than the rest. Besides, none of the time series including GPR sat- isfies the normality assumption. Finally, the augmented Dickey- Fuller (ADF) and Phillips-Perron unit root tests confirm that the return indexes along with GPR are stationary. 3.2. Empirical method This paper employs a two-step methodology. Step I involves estimating a two-state Markov regime switching regression approach to generate the state probabilities. While doing so, we consider the information on crude oil volatility index (OVX) as a covariate. Many recent papers conclude that renewable energy equities respond significantly to the changes in the level of oil market volatility [17e19]. Uddin et al. [19], for instance, show that when analyzing the dependence structure between renewable energy and other asset classes, it is crucial to control for the effect of OVX. Next, step II consists of regressing the probabilities of remaining in the low (high) volatility regime on the GPR data. Our purpose is to observe how changes in GPR index affect the risk linked to new energy assets when the volatility levels are low. We frame the MRS regression model as follows: Ri;t ¼ai;rt þ bi;rt Ri;t�1 þ gi;rtDOVXt�1 þ ui;t (1) where, Ri;t denotes the logarithmic difference for the i-th ETF index at time t, rt refers to a discrete regime variable, ai;rt is the regime- dependent intercept and bi;rt and gi;rt are regime-dependent slope coefficients. At time period t, the transmission probability from regime 1 to regimem at time period t þ 1 is dependent on the regime at time period t entirely. In addition, the transition proba- bilities are given as: Fig. 1. Time series of clean 520 pjk ¼Prðrtþ1 ¼ kjrtþ1 ¼ jÞ; pjk � 0; XM k¼1 pjk ¼1 (2) In our empirical analysis, we consider two regimes in order to obtain the estimates for low and high volatility states. Following Uddin et al. [19], we also use the regime classification measure (RCM) for evaluating the accuracy of our regime switching process: RCMðrÞ¼100r2ð1 = TÞ XT t¼1 Yr i¼1 bpi;t (3) The above statistic lies between 0 and 100. Note that our MRS process appears to be a good-fitting model if the RCM statistic is close to 0. Next, we propose the following regression model to investigate the impact of GPR on the volatility of clean energy ETFs: asin ffiffiffiffiffi pt p ¼ l0 þ Xn i¼0 li logðGPRÞt�i þ εt (4) where, pt indicates the filtered probability of lying in the low (high) volatility regime at time t. The asin (arcsine function) is used to transform the probabilities so that they can be used in a linear regression model [35]. energy ETFs and GPR. A. Dutta and P. Dutta Renewable Energy 196 (2022) 518e525 4. Empirical findings The results of our MRSmodel, presented in Table 2, demonstrate that all the new energy ETFs react negatively to OVX shocks. Hence, when oil market is highly volatile, a fall is witnessed in clean energy asset returns. This result is not startling, since an upward swing in energy market uncertainty tends to induce a drop in the level of energy prices, which further causes the prices of clean energy as- sets, close substitutes of oil assets, to decline [19]. We also report statistically significant results for all the sigma coefficients, which would suggest a swapping between the low and high volatility regimes. In addition, for all the ETF indexes, the RCM statistic indicates that the regime switching approach can be considered to be a good- fitting model. Figs. 2 and 3 demonstrating the filtered probabilities of remaining in low and high volatility regimes also reveal a strong switching pattern for each of the clean energy assets. Note that for PBW index, regime 2 remains the high volatility state, while regime 1 being the high volatility state for other 2 indexes as evidenced by the sigma values. Table 3 displays the findings of Equation (4) for the low volatility state. We report that li, measuring the effect of geopolitical un- certainty, appears to be statically significant for each clean energy ETF. Hence, we empirically show that new energy ETFs are sensitive to geopolitical risk. More importantly, the impact is positive indi- cating that with an increase in the GPR index, the likelihood of remaining in the low volatility state seems to be increasing. Next, the findings of Table 4 are consistent with those shown in Table 3. We now observe that the impact of GPR is negative revealing an inverse association between the regime probabilities and geopolitical risk. Therefore, when there is an upsurge in the GPR index, the likelihood of remaining in the high volatility state decreases. This finding could be attributed to the fact that when the geopolitical risk increases, users of crude oil, which is highly sen- sitive to such risk, tend to consider clean energies as a substitute for traditional energy sources. This causes a growth in the equity prices of new energy firms, further leading to a drop in the levels of volatility. Moreover, it is evident from Tables 3 and 4 that the lagged values of GPR are statistically significant. For instance, in Table 4, we find significant values up to lag 7when looking at the results for the PBW index. A possible explanation for this finding is that investors might react to information at different points in time, or have dif- ficulty in evaluating the effect of GPR on the returns and act with a delay. Table 2 Estimates of MRS approach. Panel A: Estimated coefficients Index State Constant AR(1) PBW S1 0.0226 (0.0293) 0.0704*** ( S2 �0.3872** (0.1969) - 0.0014 (0. PBD S1 �0.3042* (0.1716) �0.0611 (0. S2 0.0282 (0.0222) 0.0242 (0.02 TAN S1 �0.1872 (0.1964) 0.0457 (0.03 S2 �0.0272 (0.0405) 0.0723*** ( Panel B: Transition probabilities and expected durations Index P11 P12 P21 P22 PBW 0.9952 0.0048 0.0294 0.9706 PBD 0.9706 0.0294 0.0056 0.9944 TAN 0.9388 0.0612 0.0171 0.9829 Notes: This table displays the estimates of our MRS model provided in equations (1)e(3 ***p < 0.01, **p < 0.05, *p < 0.10. 521 In sum, our findings confirm that green assets react to geopo- litical risk significantly and thus the GPR index has emerged as an important factor affecting the investment decisions. In other words, we document that GPR can predict the uncertainty associated with clean energy stock returns. 5. Additional tests This section studies the effect of GPR on the conditional vola- tility of clean energy assets. Such inspections will further verify our previous finding that increasing GPR has a negative effect on the volatility levels of new energy ETFs. That is, we aim to investigate if higher GPR implies lower risk for these green assets. To serve this purpose, we estimate both symmetric and asym- metric generalized autoregressive conditional heteroscedasticity (GARCH) models. We first define the mean equation as follows: Rt ¼pþ 4Rt�1 þ εt (5) where, Rt denotes the logarithmic returns for a specific ETF index at time t. The residual term εt is assumed to follow the normal or Student's t distribution. Notably, the AR(1) process has been selected based on the values of Akaike information criterion (AIC) and Bayesian information criterion (BIC) criteria. Regarding the conditional variance equation, we consider an extended GARCH (1,1) model, defined as: h2t ¼uþ aε2t�1 þ bh2t�1 þ qDGPRt�1 (6) with u; a and b being the parameters of GARCH (1,1) process. In addition, h2t refers to the conditional variance and ε 2 t�1 represents the effect of news or shocks. Note that ða þ b) captures the volatility persistence for ETF returns. Next, we consider estimating an asymmetric version of GARCH model. In particular, we employ an extended GJR-GARCH process which is given by: h2t ¼uþ aε2t�1 þ dε2t�1St�1 þ bh2t�1 þ qDGPRt�1 (7) where, St�1 denotes a dichotomous variable that equals 1when εt�1 is negative and 0 otherwise. The persistence of volatility amounts to aþ bþ 1 =2 d. The results of our GARCH analyses are reported in Tables 5 and 6. In both exhibits, we find that the impact of GPR on the conditional volatility is significant and negative as indicated by the corre- sponding parameter q. Thus, both symmetric and asymmetric OVX shocks Sigma c2 test 0.0194) �0.2710*** (0.0162) 0.9739*** (0.0214) 204.03** 0475) �0.2872*** (0.0458) 3.6933*** (0.0469) 0450) �0.2411*** (0.0332) 3.4541*** (0.0405) 315.53*** 01) �0.2464*** (0.0156) 0.2538*** (0.0196) 96) 0.3250*** (0.0591) 4.2568*** (0.0497) 195.22*** 0.0215) �0.2937*** (0.0234) 1.5893*** (0.0263) DU1 DU2 RCM 210.07 34.06 17.52 33.96 178.86 19.01 16.32 58.63 18.41 ). Values in parentheses indicate standard errors. Fig. 2. Filtered probabilities for low volatility regime. Note: The filtered probabilities are derived from the Markov regime switching regression. The probabilities refer to the likelihoods of remaining in the low volatility states for the PBW, PBD and TAN indexes. The X-axis indicates the timeline, while the Y-axis shows the filtered probabilities. Fig. 3. Filtered probabilities for high volatility regime. Note: The filtered probabilities are derived from the Markov regime switching regression. The probabilities refer to the likelihoods of remaining in the high volatility states for the PBW, PBD and TAN indexes. The X-axis indicates the timeline, while the Y-axis shows the filtered probabilities. A. Dutta and P. Dutta Renewable Energy 196 (2022) 518e525 GARCH specifications conclude the same. Such inverse relations demonstrate that when an increment is observed in geopolitical risk index, the volatility of these ETFs seems declining. 6. Discussion A number of studies [28,30,36] have examined the effects of crude oil price uncertainty, metal price uncertainty and climate policy uncertainty on the volatility levels of clean energy assets. However, whether these assets react to geopolitical uncertainty remains under-studied. This is surprising given that geopolitical 522 conflicts trigger environmental pollution, which could be reduced with the transition to alternative energies [6]. Thus, with the in- crease in geopolitical uncertainty in recent years, many countries have invested more in renewables. Such initiatives are also re- flected in the asset prices of clean energy companies. It is also worth noting that oil and clean energy markets are close substitutes. As oil price variations are highly sensitive to geopolitical turmoil [11,12], clean energy prices, which exhibit similar dynamics like traditional energy prices, might react signif- icantly to the rise and fall in geopolitical risk. This would cause the asset prices of alternative energy firms to be varying in response to Table 3 Impact of geopolitical risk (Low volatility state). Estimate Standard error Decision Panel A: PBW Constant 0.3465 0.0547 *** logðGPRÞt�1 0.5257 0.0119 *** logðGPRÞt�2 0.4635 0.0123 *** logðGPRÞt�3 0.0367 0.0124 *** logðGPRÞt�4 0.0385 0.0122 *** logðGPRÞt�5 0.0410 0.0118 *** F-statistic 62.16 *** Adj. R2 (%) 9.62 Panel B: PBD Constant 0.1818 0.0609 *** logðGPRÞt�1 0.0488 0.0130 *** logðGPRÞt�2 0.0398 0.0134 *** logðGPRÞt�3 0.0327 0.0136 ** logðGPRÞt�4 0.0388 0.0135 *** logðGPRÞt�5 0.0423 0.0133 *** logðGPRÞt�6 0.0452 0.0129 *** F-statistic 55.43 *** Adj. R2 (%) 10.27 Panel C: TAN Constant 0.0642 0.0593 Insignificant logðGPRÞt�1 0.0363 0.0123 *** logðGPRÞt�2 0.0259 0.0127 ** logðGPRÞt�3 0.0255 0.0129 ** logðGPRÞt�4 0.0363 0.0129 ** logðGPRÞt�5 0.0402 0.0129 *** logðGPRÞt�6 0.0303 0.0128 ** logðGPRÞt�7 0.0444 0.0127 *** logðGPRÞt�8 0.0508 0.0124 *** F-statistic 54.45 *** Adj. R2 (%) 13.15 Notes: In this Table, estimates of Equation (4) are presented. We select the number of lags based on the values of Akaike information criterion (AIC) and Bayesian in- formation criterion (BIC). ***p < 0.01, **p < 0.05, *p < 0.10. Table 4 Impact of geopolitical risk (High volatility state). Estimate Standard error Decision Panel A: PBW Constant 1.2774 0.0570 *** logðGPRÞt�1 �0.0459 0.0120 *** logðGPRÞt�2 �0.0388 0.0124 *** logðGPRÞt�3 �0.0314 0.0125 ** logðGPRÞt�4 �0.0304 0.0126 ** logðGPRÞt�5 �0.0271 0.0125 ** logðGPRÞt�6 �0.0272 0.0123 ** logðGPRÞt�7 �0.0261 0.0119 ** F-statistic 45.93 *** Adj. R2 (%) 10.01 Panel B: PBD Constant 1.2890 0.0582 *** logðGPRÞt�1 �0.0606 0.0129 *** logðGPRÞt�2 �0.0519 0.0133 *** logðGPRÞt�3 �0.0509 0.0133 *** logðGPRÞt�4 �0.0617 0.0138 *** F-statistic 75.09 *** Adj. R2 (%) 9.28 Panel C: TAN Constant 1.5878 0.0586 *** logðGPRÞt�1 �0.0397 0.0123 *** logðGPRÞt�2 �0.0260 0.0128 ** logðGPRÞt�3 �0.0311 0.0129 ** logðGPRÞt�4 �0.0359 0.0129 *** logðGPRÞt�5 �0.0477 0.0128 *** logðGPRÞt�6 �0.04382 0.0127 *** logðGPRÞt�7 �0.0473 0.0122 *** F-statistic 59.41 *** Adj. R2 (%) 12.58 Notes: In this Table, estimates of Equation (4) are presented. We select the number of lags based on the values of Akaike information criterion (AIC) and Bayesian in- formation criterion (BIC). ***p < 0.01, **p < 0.05, *p < 0.10. Table 5 GARCH estimates. Parameters Y Error distribution is normal Error distribution is Student's t Estimates Standard error Estimates Standard error u 0.049*** 0.010 0.049*** 0.014 a 0.080*** 0.007 0.082*** 0.010 b 0.907*** 0.008 0.906*** 0.011 q �0.001** 0.0005 �0.001** 0.0005 Persistence 0.987 0.988 Log-likelihood �5875.63 �5850.30 AIC 3.927 3.911 BIC 3.931 3.925 Notes: q measures the effect of GPR. The persistence of volatility amounts to.aþ b: ***p < 0.01, **p < 0.05, *p < 0.10. A. Dutta and P. Dutta Renewable Energy 196 (2022) 518e525 the fluctuation in GPR index. Moreover, several recent studies [8,10] also argue that increasing geopolitical risk leads to higher uncer- tainty which could be harsh to the systematic development of na- tional economy, further being reflected in traditional financial markets such as stocks and oil. The clean energy asset market could receive similar shocks from geopolitical uncertainty as it maintains a high correlation with other asset classes including equities and commodities. Furthermore, it is also expected that both investor sentiment and trading decisions are significantly influenced by the changes in the level of geopolitical risk which might drive the volatility of clean energy asset class. Hence, geopolitical risk might play a crucial role in understanding the volatility dynamics of clean energy asset class. To this end, our study joins the literature on clean energy asset markets by documenting a negative association between geopo- litical risk and the volatility of clean energy assets. As discussed earlier, the significant linkage between these variables could be attributed to the high correlations between oil and new energy markets, rapid transition to renewable energy, investors’ enthua- sim for renewables and financial market integration. Given that geopolitical risk has emerged as a key determinant of clean energy asset returns, market participants could use its information content to predict the volatility of renewable energy firms. Hence, along with other uncertainty measures (e.g., climate policy uncertainty, 523 equity and commodity market volatility indexes), geopolitical risk also deserves more attention from investors, researchers and policymakers. In sum, our study suggests that policymakers and investors participating in alternative energymarkets should pay considerable attention to the impact of geopolitical risk on this new asset class when forecasting future volatility, assessing portfolio risk and making appropriate hedging decisions. Understanding the perfor- mance of green assets is of utmost importance to ethical investors Table 6 GJR-GARCH estimates. Parameters Y Error distribution is normal Error distribution is Student's t Estimates Standard error Estimates Standard error u 0.061*** 0.010 0.061*** 0.014 a 0.036*** 0.009 0.035*** 0.012 b 0.908*** 0.008 0.905*** 0.011 g 0.072*** 0.011 0.079*** 0.017 q �0.0009** 0.0004 �0.001** 0.0005 Persistence 0.9800 0.9795 Log-likelihood �5862.89 �5839.39 AIC 3.919 3.904 BIC 3.933 3.920 Notes: qmeasures the effect of GPR. The persistence of volatility amounts to.aþ bþ 1 =2g: ***p < 0.01, **p < 0.05, *p < 0.10. A. Dutta and P. Dutta Renewable Energy 196 (2022) 518e525 given that these companies produce ecological goods and provide climate-friendly services [37e39]. Besides, proper knowledge on the risk linked to clean energy asset class is essential for socially responsible investors as they focus not only on the environmental performance of a firm but also consider its financial performance. Therefore, our results offer key implications to both investors and policymakers. 7. Conclusions Earlier studies find that oil price volatilities are sensitive to geopolitical uncertainty. However, the effect of such risk on the volatility of renewable energy assets, which exhibit similar dy- namics like the oil assets, is yet to be explored. This paper intends to fill this gap by examining the effect of geopolitical uncertainty, measured by the geopolitical risk (GPR) index, on the volatility levels of renewable energy exchange traded funds. Methodologically, we employ the Markov regime switching process and different forms of GARCHmodels to serve our purpose. Looking at the filtered probabilities obtained from the MRS re- gressions, we find that an upturn in the GPR index increases (re- duces) the likelihood of being in the low (high) volatility regime. This result suggests that when GPR mounts, consumers of crude oil would shift towards clean energies in order to replace traditional Table A1 Acronyms Acronym Meaning AIC Akaike information criteria ADF Augmented Dickey-Fuller BIC Bayesian information criteria DCC-GARCH Dynamic conditional correla ETF Exchange traded fund GARCH Generalized autoregressive c GPR Geopolitical risk GJR Glosten, Jagannathan, Runkl MRS Markov regime switching OVX Oil Implied Volatility index PBD Invesco Global Clean Energy PBW Invesco WilderHill Clean Ene PP Phillips-Perron RCM Regime classification measu TAN Invesco Solar Energy ETF VAR Vector autoregressive WTI West Texas Intermediate 524 energy sources which are highly susceptible to geopolitical con- flicts. Such transition may cause a significant increment in the asset prices of new energy firms, further leading to a drop in the levels of volatility. Additionally, the results of GARCH models also reveal an inverse relationship between GPR and the volatility of clean energy asset class. Therefore, when an upsurge is observed in the geopo- litical risk index, the volatility of these clean energy ETFs seems declining. In other words, higher GPR implies lower risk for these green assets. The findings of this paper offer several important policy impli- cations. First, as increasing geopolitical risk is found to exert a positive effect on clean energy asset prices, policymakers are likely to promote the usage of renewable energy sources with the increment of geopolitical conflicts. One could thus expect that ris- ing geopolitical uncertainty encourages the future developments of new energy technologies, which would in turn decrease the dependence on traditional energy sources such as fossil fuels. In addition, the transition to renewable energy also leads to greater energy self-sufficiency and reduces the adverse impacts of climate change. It is, therefore, important to pay more attention to under- stand how the association between geopolitical risk and clean en- ergy markets evolves over time. Second, investors could utilize the information content of geopolitical risk to predict the volatility of clean energy asset class more precisely. Our results are thus important for market participants while finding more efficient ra- tios for portfolio optimization, hedging and risk management. In sum, along with the economic policy uncertainty and crude oil volatility, geopolitical risk can also be used as a supplement for making proper asset-allocation decisions and gaining better port- folio diversification benefits. CRediT authorship contribution statement Anupam Dutta: Conceptualization, Validation, Formal analysis, Writing e original draft. Probal Dutta:Writing e review & editing, Final Revision, Project administration. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. tion - generalized autoregressive conditional heteroskedasticity onditional heteroskedasticity e ETF rgy ETF re A. Dutta and P. Dutta Renewable Energy 196 (2022) 518e525 References [1] W. Ahmad, P. Sadorsky, A. Sharma, Optimal hedge ratios for clean energy equities, Econ. Modell. 72 (2018) 278e295. [2] J. Cunado, R. Gupta, C.K.M. Lau, X. Sheng, Time-varying Impact of Geopolitical Risks on Oil Prices, Defence and Peace Economics, 2019, https://doi.org/ 10.1080/10242694.2018.1563854. [3] R. Demirer, R. Gupta, Q. Ji, A.K. Tiwari, Geopolitical risks and the predictability of regional oil returns and volatility, OPEC Energy Rev 43 (3) (2019), 342-36. [4] A.C. Marques, J.A. Fuinhas, D.A. Pereira, Have fossil fuels been substituted by renewables? An empirical assessment for 10 European countries, Energy Pol. 116 (2018) 257e265. [5] X. Lv, X. Dong, W. Dong, Oil Prices and Stock Prices of Clean Energy: New Evidence from Chinese Subsectoral Data, Emerging Markets Finance and Trade 57 (2021) 1088e1102. [6] K. Yang, Y. Wei, S. Li, J. He, Geopolitical risk and renewable energy stock markets: an insight from multiscale dynamic risk spillover, J. Clean. Prod. 279 (2021), 123429. [7] Y. Song, Q. Ji, Y.J. Du, J.B. Geng, The dynamic dependence of fossil energy, investor sentiment and renewable energy stock markets, Energy Econ. 84 (2019), 104564. [8] A.F. Aysan, E. Demir, G. Gozgor, C.K.M. Lau, Effects of the geopolitical risks on Bitcoin returns and volatility, Res. Int. Bus. Finance 47 (2019) 511e518. [9] J.E. Lee, The world stock markets under geopolitical risks: dependence structure, World Econ. 42 (6) (2019) 1898e1930. [10] C. Park, S. Park, Rare disaster risk and exchange rates: an empirical investi- gation of South Korean exchange rates under tension between the two Koreas, Finance Res. Lett. (2019), 101314. [11] N. Antonakakis, I. Chatziantoniou, G. Filis, Oil shocks and stock markets: dy- namic connectedness under the prism of recent geopolitical and economic unrest, Int. Rev. Financ. Anal. 50 (2017) 1e26. [12] J. Liu, F. Ma, Y. Tang, Y. Zhang, Geopolitical risk and oil volatility: a new insight, Energy Econ. 84 (2019), 104548. [13] I. Henriques, P. Sadorsky, Oil prices and the stock prices of alternative energy companies, Energy Econ. 30 (3) (2008) 998e1010. [14] D.C. Broadstock, H. Cao, D. Zhang, Oil shocks and their impact on energy related stocks in China, Energy Econ. 34 (2012) 1888e1895. [15] J.C. Reboredo, M.A. Rivera-Castro, A. Ugolini, Wavelet-based test of co- movement and causality between oil and renewable energy stock prices, Energy Econ. 61 (2017) 241e252. [16] I. Dawar, A. Dutta, E. Bouri, T. Saeed, Crude oil prices and clean energy stock indices: lagged and asymmetric effects with quantile regression, Renew. En- ergy 163 (2021) 288e299. [17] A. Dutta, Oil price uncertainty and clean energy stock returns: new evidence from crude oil volatility index, J. Clean. Prod. 164 (2017) 1157e1166. [18] L. Pham, Do all clean energy stocks respond homogeneously to oil price? Energy Econ. 81 (2019) 355e379. [19] G.S. Uddin, M.L. Rahman, A. Hedstrom, A. Ahmed, Cross-Quantilogram-based correlation and dependence between renewable energy stock and other asset classes, Energy Econ. 80 (2019) 743e759. [20] A.W. Lo, A.C. MacKinlay, An econometric analysis of nonsynchronous trading, 525 J. Econom. 45 (1e2) (1990) 181e211. [21] T. Krause, Y. Tse, Volatility and return spillovers in Canadian and U.S. industry ETFs, Int. Rev. Econ. Finance 25 (2013) 244e259. [22] X. Zhao, Y. Fan, M. Fang, Z. Hua, Do environmental regulations undermine energy firm performance? An empirical analysis from China's stock market, Energy Res. Social Sci. 40 (2018) 220e231. [23] S. Managi, T. Okimoto, Does the price of oil interact with clean energy prices in the stock market? Jpn. World Econ. 27 (2013) 1e9. [24] R. Bondia, S. Ghosh, K. Kanjilal, International crude oil prices and the stock prices of clean energy and technology companies: evidence from non-linear cointegration tests with unknown structural breaks, Energy 101 (2016) 558e565. [25] T. Saeed, E. Bouri, H. Alsulami, Extreme return connectedness and its de- terminants between clean/green and dirty energy investments, Energy Econ. 96 (2021), 105017. [26] P. Sadorsky, Correlations and volatility spillovers between oil prices and the stock prices of clean energy and technology companies, Energy Econ. 34 (2012) 248e255. [27] J.C. Reboredo, Is there dependence and systemic risk between oil and renewable energy stock prices? Energy Econ. 48 (2015) 32e45. [28] W. Ahmad, On the dynamic dependence and investment performance of crude oil and clean energy stocks, Res. Int. Bus. Finance 42 (2017) 376e389. [29] F.X. Diebold, K. Yilmaz, Better to give than to receive: predictive directional measurement of volatility spillovers, Int. J. Forecast. 28 (2012) 57e66. [30] A. Dutta, E. Bouri, D. Das, D. Roubaud, Assessment and optimization of clean energy equity risks and commodity price volatility indexes: implications for sustainability, J. Clean. Prod. 243 (2020a), 118669. [31] E. Bouri, N. Jalkh, A. Dutta, G.S. Uddin, Gold and crude oil as safe-haven assets for clean energy stock indices: blended copulas approach, Energy 178 (2019) 544e553. [32] T. Xia, Q. Ji, D. Zhang, J. Han, Asymmetric and extreme influence of energy price changes on renewable energy stock performance, J. Clean. Prod. 241 (2019), 118338. [33] M. Yahya, K. Kanjilal, A. Dutta, G.S. Uddin, S. Ghosh, Can Clean Energy Stock Price Rule Oil Price? New Evidences from a Regime-Switching Model at First and Second Moments, Energy Economics 95 (2021), 105116. [34] Dario Caldara, M. Iacoviello, Measuring geopolitical risk, Am. Econ. Rev. 112 (4) (2022) 1194e1225. [35] S.A. Basher, A.A. Haug, P. Sadorsky, The impact of oil-market shocks on stock returns in major oil-exporting countries, J. Int. Money Finance 86 (2018) 264e280. [36] E. Bouri, N. Iqbal, T. Klein, Climate policy uncertainty and the price dynamics of green and brown energy stocks, Finance Res. Lett. 47 (2022), 102740. [37] R. Schaeffer, B. Borba, R. Rathmann, A. Szklo, D.A. Castelo Branco, Dow Jones sustainability index transmission to oil stock market returns: a GARCH approach, Energy 45 (2012) 933e943. [38] A. Dutta, Impact of silver price uncertainty on solar energy firms, J. Clean. Prod. 225 (2019) 1044e1051. [39] A. Dutta, R.K. Jana, D. Das, Do green investments react to oil price shocks? Implications for sustainable development, J. Clean. Prod. Forthcom. (2020b). http://refhub.elsevier.com/S0960-1481(22)01025-4/sref1 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref1 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref1 https://doi.org/10.1080/10242694.2018.1563854 https://doi.org/10.1080/10242694.2018.1563854 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref3 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref3 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref4 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref4 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref4 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref4 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref5 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref5 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref5 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref5 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref6 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref6 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref6 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref7 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref7 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref7 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref8 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref8 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref8 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref9 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref9 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref9 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref10 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref10 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref10 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref11 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref11 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref11 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref11 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref12 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref12 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref13 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref13 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref13 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref14 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref14 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref14 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref15 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref15 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref15 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref15 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref16 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref16 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref16 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref16 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref17 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref17 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref17 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref18 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref18 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref18 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref19 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref19 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref19 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref19 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref20 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref20 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref20 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref20 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref21 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref21 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref21 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref22 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref22 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref22 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref22 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref23 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref23 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref23 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref24 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref24 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref24 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref24 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref24 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref25 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref25 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref25 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref26 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref26 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref26 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref26 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref27 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref27 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref27 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref28 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref28 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref28 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref29 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref29 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref29 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref30 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref30 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref30 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref31 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref31 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref31 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref31 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref32 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref32 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref32 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref33 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref33 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref33 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref34 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref34 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref34 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref35 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref35 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref35 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref35 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref36 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref36 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref37 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref37 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref37 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref37 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref38 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref38 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref38 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref39 http://refhub.elsevier.com/S0960-1481(22)01025-4/sref39 Geopolitical risk and renewable energy asset prices: Implications for sustainable development 1. Introduction 2. Literature review 3. Materials and methods 3.1. Data 3.2. Empirical method 4. Empirical findings 5. Additional tests 6. Discussion 7. Conclusions CRediT authorship contribution statement Declaration of competing interest References